...... ...!

      

     

...I 1 .l n I .
. . III .I ‘I: z .0 ! IV I . V
. 2... 1 V. I'l' 0.1.4,: A. . . . 1|...V‘..¢.Ll.. Cf H“JU"UI....!.¢.3..1MZ.4 ”unhfl.ro.n‘. 1.0.. an . .l;..».fl1.....£‘l t7 . “Conn“. 5.. r.
. . I .... V E . a :3 . V V . V I... o ... .... ... Iv?- !‘ _’..Q’ o .v 9.. . «....
n .V W. 3 *‘l V ‘4 V-h.JVV.qo¢.qwu_V.. a. ”I: 2.1.33.3; unheflHd...KHPh.ol.u‘—Jihh! . ’0!¢...¢’.u.“.!!lm:..ow “Cr"!“an".flfaghumfi.toflwlm“;1l;I...A’L.-v.uou.ni 3).”... 25.0 If...” :1. 03..
L c a ~ i‘23 62.8 o. .I . . . ..3 . .. a. .3} 1 014-! . . V .a‘o...!o.! in]... .. ....3....lo.$oo!o.s.:9a 3'. .93... ......3. . . l4. .. . 0r . . o
. O. o. I. 1 c . . flit?! . out ... .. ...... r V . o it... . ...-x ...
w; ‘ «sawtoofimo‘h v." .501 on.‘"§..!’u ... OsVV Jo,chum.NT.03.JH.’.VOO¢W¢HMOWNQNJ”V¢LV...-r ‘ukoswo‘rrgu‘uofiagibfln; {-lfikltumnmhtiwml ”fin...“ Nu..W>lou~.-.Ofn..3¢l.t 90.....m. gun-L. 53.4... .. 3,:ch . .V... ....v. ...
Q .... . ‘ . . ... . 6V . a Q .| .
a." .."n...6o.. . l... .I. V.. at! ’00....0‘4-4 Ciuomivmsoo1tl3woyi.:0. find” 27"; ...LV V'.“moo‘: " II 3.9” 0'63. at, ...-6. I... .
a... . . .0 V‘ . . ~ . V 31C .
56...“. .vaflud. hath... ...: .1 i . V ... «a. 2V3... 37......V . .0 u. o. o ’9.

 

.. .....O ...

...V .. . .... I ... a . x is?! . ixvflvru... .V . V..
.. on”. .."g .... . . cg “4%.”hu.w ..‘4rlau‘.’ Log. a ..

      

 

 

 

 

    

     

     

        

o . a 0 O
’7’61- .sQ‘ - 4...... ‘. Dc
... V . .. ...... 3... '23.... .... .9. .
. V . V I . £3.06... . p u v .00. oh . a
t. :12... .‘. 39.... .... ... ”Wanna: JvICu. . 1...... . . V..-V... . ... l9. . .. a... ‘...\.os..!t ......ilahuromfluJ’mhnun. o. it“. 9. .1. . r _ I. 3... V
1.....3'33 0 ..V :3. . .....O. ..V 0.... «... . . V. 3.98... .J‘: . .. a. $.61... I... ... .9153}... ......o! ...... .v.... . ..rV .. I ..V ..V . 731..-: . Q
9‘ § ....v ... . o: \ .8 . a o. l . . a “.....vczét’ . .... I!" V . Oil}. 893'... I... #13,}. .1::.... w.$.!..v¢:..o...‘0“o?;l.. 23;.‘91: .oufsoffo. . ....' .1}... "9 V.. o... .b . V..... g
V.“ O‘“u0.a‘vVo ...o. .' C J...9.O“W.Qv ....OVOT’ :- ...-”61 u 27 .00.... "u. ....t 14“. lat-‘0‘ . 4.1.7.“ .... or l."feseou0.c. .....un‘. 9......" 731,0...1.’ .u. .. . 0.6“ ...)“.l .- 6 3m.) anal-.0I’c'1 - v‘oa .v .. é. ‘
3.0.. I... . . a. jazr...‘ ... 21‘... .3. . ..V. . . .M ..V 2 . 6|.C:..I..3...... a v :3... o. 080...... 991231383}: . o I. 291323.06. ... l. ..V-2... -... It. I V V.... 6. V.. 0 . I... V. V. . 50... ....
qwn.n3~.& ..s....:9..duflufl9.. ....r . . ..V .. ... ......‘Juéo. .5. . ...... ...-'29:...331 ....vapt.ttl.!2..y.e£!. . I .VVo? .0. .V [...-.03? ... .17.! .14.... .9 V 0.. .33.... 1...?! 352...... r .15.: ..L . ... :
...}. . c .311...“ o. .53.. ‘1: .....i....? ... ‘97.... 9...... .94 a $2 .1... o. :1. -..... .1390! v lltd. 8...: i9..w..\.~qut 0 wont? VOW..." .0 .....fVV. 11¢). V. a. . Lioi; ... ...n. . .8131? ..- .
.V . ...; ... o c..ll..’ .. 1.50:5}...th ......r . t. ...: ... ......9....il!....iQ-...J . ...! ...: ... . .... ..I. 21......- oi. .37.... 6.1.1.... .... .o 3.} .33.... . I .4 V.- 3v. V. v V... .. s... V.. v1... ..o.).i....... _
. . . 713.53.. I fi”331'€..w|41.v .3...!.~..¢.ollo$§ 1 .... ... a ...}: 537......2‘. .JI....OQ.7.;. ...... .58. V1.1..Po....an . .3; 4.-....13 ..V...,V.IO\...I ....oltotr’.$§.l;. o o. o. ... '3 a. o“ .1"... ....V V V.. .....o‘. 9.390...
I ' ...... 3.... . .033”)! $.10..o'uv.aa0"ou Jun...“ .. o.-...~O’...I It . a... .--. A... .7.- ‘...o.... 03.1....» ‘a.‘ .00. 9 . o «43.1?180-3’ .o .....0. ...?! I... 3! .0..... g 1,. . Ci VV'" 9 I .0 ...1. O‘. 9 .. 38(n 3’.
'0' t o I . a .o . .v V ‘9. 2...! 0“ up. 1’... a v v. .. '-‘.. it. .A .a . n .1. 30.1.1.6. » ‘3“ -
838.329 32.11.... V V. .fI..VvJ”z‘!.I..!§:tianctotaut. I V. . h a .. ..V. t . .2 . .. .r a Q.
of! 0.: 1.1. u ‘ . V V
v0.18. a...».... u

. .0 9“.an v9'r

..3. V‘. cu... V .V . 9 .. no... .. .V . .
...“...w....fi...ummu h...un.fi................ 1...“...st . V. ..V: ..- V.....VVVV. V... ..
My? . .Iv.lh.t.¢.. O. .

90.433... 0;. 9-.133...oI-3 97V. .V. t
V .3... a . V
. ...-2...... ”1.2.1.5... ”fizz-J 42.1.3. . SP.-
. .

- . . VV 0. a}:

....,..P”..za to" ‘ on..’..9 .- ...! . A . 5‘ A:

. L .d'aV’O..."o. 3...... .... . ..- V.. a 2 . V. o . .Qo ...L ...: . .’..70Vo

.0! u .7. a .. 1 a ..3! ..- . ..V...I'o.3..1.pi’ V . . . a . . .... ... a . 592‘ : . _ V ..V. b..‘..o. .. ..

.0 .V IR. u‘v‘ . Q: I .. OJ - . .p 13"... ... .0 . . .... lo ulna...‘ ...,»vzoovaV .3..an o. 2.... .o Ila-.5017... . ...

.19 ...-2J7... ..
F‘tvfioorvoizx .

. v... .- V .
mm.” 3.1.17...

    

      

   

         

  

 

               

          

 

           

 

 

 

      
        

 

 

     

 

       

   

 

      

    

 

 

 

 

 

 

 

  

 

.V . .. .....I’7'.¢J
. ...... .3. :3.” ... ..V 13...... . J r: . ...... .. 1. . . .
.ln V ...-4.3.. .3. e... 21”.. . o.1..0.l".f’s. o. 1.0.. CI V..o‘di. ...?!“IIL. .... V 00111,.h': rwuonfid.
2.0 ... o. '8'... ... ‘93....J‘... 1.. u .- 2'51... .... . «Oi-.Vv... V9V;|V$!. .053»!
outl‘v.’.t V . 9 o. . V1,? .3’v. . .ch. ...-3‘6: .... v 1. ‘04.. ..V...) .6...\..? 3...} 0 . .1. k1,”:
. .. . _ . to. no .7... ...-O... «3.63.... .0! .Va.\1»I’l..0.1r.-lP-o. 3... 139‘... .... '21 r.‘.um....3..
..J-‘tVCo . V . V .. V. w.."rfi.””o.€ ..s‘ovi.£?.¢’..§..u 2.9.13... . .... .. V.....¢..\..13 b, .950 . o .
p. .. a.. . . V V . ... ...?! . o.D?-.o I .1331... o I- ...»...V .3 .s’......o)... .3. . .. .. A1\ .4 .... .
«1.3.3.: V . V . . . _ . V . I .371... ...... 3.69.5.2. 7... . .... V1.3...V79Q 9.; 1.198... ....I V..... .
15...... .... . V V V V V . 0.. . .V. 0...! 1.9990259”... 91......1333 .lv. 1'. .0 .92. 1‘.
... ... 3.. . . V V ... . . . . J 23...... D . i. ....22. ...-3&6: 93... a; IP4.9.V.....V .- ...... .. . ..V
I. .... .1. . . I .....l!..9~f.l::.oc...2 1%.}...‘0‘9’... .5 :9..- . .VV. ... .3‘ ...
.V 514...)...1. 1.....1.‘ .rttv..1$e .. ...... 5.9%.... ..V. ..fi. .3. V .... o
1 .... (...... “.9..uVfinrfl.h..7......o..n....9k.lo.?o. ..6 ..V... (:9. .... . .r . .V
Jvu: on. V . V 0.. V..-... .c .V. ..- .V ‘1‘: on! . 3-.,..\ ..i‘. 00¢... . .V V. . c ... u‘ . o... 39:. ...... .... 2 ‘ Q! Div“... ....V" . a: . Q .
.9. .. . . .. . v o. fi.w.J0.lo.1~ c.3671. n... .J'...o’l\.9..$..do~oil.!oi . ”9......- o. . ......TI’V é!n?“..8\'..q . . V 30...”; ....PJEVUJH... ? V... .V . . 550?...1. o t... V... V 2...! .V V. .........l.. 39.3.1... . ..
....o :0.- a ... .... . .0. . .... 7 9.351. 0' ’9',o.r'.l’.32il'- - .... . . O... .6. .‘vt..-Ouon.o.a o ...: 69. 1.. 58’..- ...-l 0- Q... ... .. .‘.nb. . . a It. .9? ’13-... .V v00.....’ :7. . n .... 4‘; u.- ..o.’ ‘-
1 :0 ...... \v: . V ...-10.1.00 $1.. ... ‘34,". $5.0 to" .V (3.3.3). “o ...... ....QBIJtl-"3V‘.5.7.7..... {3... i ’9. .0 . . V . V V .....rl.’ r. A. 0. P .. ..V n V... .V.. V... a... . V V . V V . r13»
.‘ . :3 . . 0“ .Ja."oo'.)". .1 .33 .....I... v.vPQ3-..n.o .... (0...)...1' .9 4,-9.1: . 3.-....9... ,ou’.....o .ooF... . . . . .0 . 9.16. .. iris-... ..V D . V..?! .. . . . . V . . . . . 0 V up a 3.9.
«a. V. ... ...ou . 9.13. ..V c... 1... 2.9.2.8.. .. .0 . .1. :3 5.3.2.3.... V. . . .. V. v... .. . a. 31-... V... o (I. . V. . . V .. V..)... \.V ..V” .i.-.!2V.. ......Ou... a. .V . 3.196“ .— 3 \d V
:!l¢...d¢nl!..t:s: I. x... J. €241.29. ..... r: . ...... 11.34%. ......Vi... ...... .. ...... Vi...CV1!...2..I _ . . . .. . a ....V a... ....va .... 5:2,... ......I: .V . _ .V a s.
. I ...-up 9 . ....w u- 1... .03....vJo it 3. ...d ...-9‘... .. .V. .c ...vo ~ .1..." ......9; . :9: on! . ..V ...-:9: v.0... . ...: ...‘5. ... . . . V . 0W. . 3"". i0
.n 36.... a» 0'7.8..r 1.: v. [Ibt.7.... 1:3... I ......V... $1.... . . . V . . . V ....61 .. .V.. V. V . V . N «P, ‘0)..1V .I
V . .... TV . V . . V V V . V Its. .v. PM”. V ..
8.3 .3 . . In ...... .....7. “.....nufl. 1. V . . 8L3 0." .“.
...... O... to... ’1 ...-‘7. . V . . _ .. . V. V . gar. I V 3" ,.$.V'
of .
...vv!tV‘.-ot'.ob.t'l.nvb ._ V V . V V . V . . .. V . . V . . . . V V. ,0.-'._445“5 JF‘fi ‘5‘! or
...”...M. .I.. .taVI. ....D..V.OVV.CI. ..f. . V .. V . . . . . V "Mali”... . ... . 0.3“). s. .0
V .l a no _ 3"; JIV. - ...... . V .. V . 1:... .. n. V . .. V V . . .. . ... . . V
I: .6. .0; . ”at... .... 203‘ .to..wi.o.nc. . . .. V . ...... ... . c0: ... v. ... ... ...a‘ . V .V . V . V V . N"
. .40.... ...... ......o. - .7. 1.9.1.8.... .... . l I i .. 1.1.. . . V V . V .. . Q .... 0.. . V. is: V. I. .
.V .93.; .n V.. 0...: .....l .... 06“. 3‘ V ‘0. o 7.. ). 3.3.39! . .. V V . . V . V V . V V V V . ‘3 Q 1!:.V9.V\ ... V.. V..; .
V3 .... .. ...: .1338. I13" .10 a V...- ..V . .V... v... w ..9...v....0 ...-...: 0... c... 1...: V . V V V V . . .V . . V V
VI... .. . . ......6. 1. .- V ... . ......Ouic V..-.... ...... - ..Vi..‘|8...v..1dl..t0.3 V V . . V .
..V. V.. .v as... .‘C. . .1: . v ...V .3591”... .‘....\..;.lo...0.._...: ... . .
J... b-.. a. 5": . ...... ....)3. V

 

.t o. I? .0... . 0.7...
In wr...‘.-D ...! _. a .316...
in. .u ... .. ... . VIII .... ......iof 1. -

 

    

.Is. .0. .037. .93.... ..u.
no“. in. $.23...) .
.... .89..

  

.3. ... 7 ..t
...ooiicoo iv. 1’ .

 

..V... .. ....v Q .Io... V
.u .e.VV.c..o O .-....

           

 

.- ... . o o a .. l9 ..l.:o. 6.0...
w.’ . ...... ...l'.‘ V ... 3a.. ......1. a. .
V ......c . .-.. ..ur . on: 1. (3 A 3...ch! 45 I D..3!£.i.c.u.

V C . . .

 

P
...JV
...... ...}... ...:

   

            

      

 

 

 

o’o.

. ..Q.
. . . ... . . . . 1‘. ,IS..~...’.I.! .... . 9...!
V .. . . . V .V . . .. .. ?Q\on’¢3.V’ V.. A. C ..o. ... .

. V ...f .V.P¢.V ... v3 2‘33; . . .. . V ..6l .91 V. V V V
0 .... 9.. ...-5.... c. .. .. o . .... . $.03... 3.8.0336 coil-...... .V V V V
036...... . .. .... . .0... v 6. .‘lou. .... v.5.l o... ... ..V-5.....015.VD. 1...... i.‘. .V..
V..... .n u. . u .. ... ..V.‘ I. on... V.) ....c? .... ......o... 905.. 35.. ...)... o.
t ' 113,313... . ... ... 1.5... $9.... -

      

.
o . .v“' .00. .1...0'.. >0...“ 45...... ... .
..O: . .(wol? : 1.. 9. ... a ...-......lc. til-...}? .t..t.+. .. 0 V 3. . .....a.
V... . .. ' . . ... .. n ‘ .. ..ov .rJ..-.E.O.,..Jr..out I ...-..V.
’0‘ >9...Pn... t O .30 V‘ ’.'r'.'. ..I
. .V .. V...-

 

             
 

ql’ V .. ...:.6 .10..
V . . V . .. s.
O.‘ .87...." ...; . . . Go", a...

        
         

      

      

V . .. .... 0-391

:t..w..§.41.. .- 359......

’.76 ... ...‘Lo- .v. 00.1». ...f.¢ '0'" 6.. lo. .7.‘v'w. .3553.

3.5.4:... [I ...... :60}. .V. .1. .0. :11... 9.707 .V . . V
... v. I a... V..:.....tno...05-V.“~.3VI‘I’:1- V

 

      

. .. O’Vo’c .0 .l...’.v.. .5- V. V’3.-1~
5......- 0.9....) 2100...... I v V

.... .V .. .6‘?‘ V
.‘s.o..‘._:’§V 03.0on 0 V. A. I. u..- .
. ‘0" V. VV ... ... .3»! .F' 4..

u, ..
{3.1: .0... an. ... .V c v ... . s I V c .-
....V.. “1.1.2... o... V..-IIIu. u...o~..VoV..uO...v.
V ... V . . . .

             

O
‘5 a... .

       

 

   

                 

     
      

                   

   

 

     

      

    
               

   

  
  

     

   

 

 

 

 

 

     

 

 

 

 

  
       

 
 

  
     
   

          

                 

 

 

 

      

   

  

       

 

          

                
 

     
       

    
    

     

                  

      

 

     

 

 

 

 

   

       

 

   
   
     

  
     
        

 

     

 

      

 

 

 

   

 

 

     

     

    

 

      
 
  

       

      

 

   

      
       

   

      

      

                

  

             
  

          

 
         
 

     

       

 

  

   

 
  

     

 

. £3.10... ‘0“.xv‘.r.o..oo .u .p .31...
5-1..V. ...V- . v! . .. . V V .. ... C .V . 0‘. ... ....VV .3 ’-
.. .... h V ‘ 3.. .. V. ....:.VV... ..V.... 2... -......V....H.,V..V..? .V. ......u.... . . .31.)?! .VVMMVV.
v .V . . V . .. . V . . . .. V . V . . . . .. . .. . .
..‘diov 0... V V V . . V. 09‘.... o ......flV .“~.‘.~,-.U.§.. . V 0 V . ...... . V .... O I; 3‘ . . Q . I .. ’1... art...”“a..\£ V v.
{... .13 V. . .90.... O. V... 9.30:. .7... ....t: .....c. .35... ..I. I. . 31?... 8. . .. ...t ... . . .V V); ...
..Q. ...l ...?b d on V. I no. 3"VI V..: c ..QQ‘V .l ..Iil. ...-o vol I I... 0". ...“- . ... 9-3 . . V, V I . . .
V V! . . .. IIV. ...... .....QV... .. . 7?..- .z 0.... :1... . ...93 a Q. ....llnlo yr. V.
. V .... .396. c .35... . ...... £5.03qu . ..V. ..73’... 3.5-1... fi, .. .... .-.-... . o . .
V . .. V.....Js . . ..V. . ... ...:va . ......37 1.3.2.91: . Ito-.... . .q‘ V o -...é........!. . 11-9.. .... in. .... ..oh . 1. ...? ..V . | m. 9...... ... ..V r!
J. N ‘3 .0... V.....V. - ..l. u. a.» .... $6.: anon-1.0.... o. ....laL...r)'1.¢.‘.... V ... .O‘tt‘. . ....V V ..I.....S.D. . .....PH 1%.. . g. V.. .o . ...~ . ... . V V
9 . .01. .. v .0 .1... V.. V .. It... 3... V. ... 5|..‘u ..n . V . «4.0.6.301... o.'p .9: .‘ ... .. .. . ... A. .... .... V..?.01.1. .I . V
I v. .56.. . p ..-... I C. .4 VI. Qi‘.7Vo’-.V.,\l .V'.v...o.o-rlng. . .1... V . ...-fl. .. .s 6 v... .V.... n... 4. . s ..A. . . V V . A. «.0. -
¢.v no. .. a. o. I ..V . o ..3 ......) 3... ..3....’. ....f’... L! .- ....I.A. . . \ .0. (O V. Vr .. o C. Q. .
. - .0330...» o . A ..V “.020... V . (.... ... “nun-V....U-Jlu. 5. 6'... . 1’3 . ...... . . 1. . 32. .V .30.: ... v
V .. .. . . . . V .. ..V. o . V ... I V. V .l. . . . V
. «....o..¢tu.o.o”¢ a V..o\. ... ....v. . . V V . . . . V u..‘.......l..£\ ... .60.}! ...: . V . .... . ...... I. to. . . ... 2...... ! V. V . .c
_. .0 V. . a V. ._ IV . 130.6. . .. V. o.L.o‘-.V..O..... ......VQ u 9.. ......z... y... . V ..V V. v
. I no. ... 2.11.. .v . . ..OV. v . . . . . .1 V..}. ...-:4 . .......¢ .0 V. ‘- ...: \‘V . 5.. . . 1
_ V ... .O: .... .. .V... .D . V V. .. ... . O
. .a .V . . 0. u L . v‘ ...-I. .V V.. a .c 0. L. V ...
...CIA! . . 0.". 0.. .‘I- .4 0|? 0...... .. . V ...
n. K a . II... J a I . ... 0-. ..V-...... _ V . .... ... . .'
..J o I 6‘7 . . . ..6' I.¢.VJ&..&' o.\- u . u! . ... n to.
.O.. 0’ o’V .. ..V .V . v-ovav ...V‘..uVo..OO. In... .v‘. 1 . .. . .. . . ‘6 s ’
I V... f ~ . V I h . . .. . .0 ’32.... .... .0 5.3. c. . 3 ... .. .V . e. ..V .... .I
_ . . ‘3 . I 0 ..| .2. . V . V V VV V V . . . . . .
..l...lo .. n. . ...-I. ..lol I . .. .'. 9 ‘3'}! 0.. .15... . . . .. . . o .0. . iii-13...? V..
.t-u .HC on . ..¢ .. 4.1.341...
.51... .V: . .
_ 6...... . V J I. .0... o 0 . . ... -.. V..Vb. . . ‘2 ..o‘...
o I ‘ I O I . . I .
V.. ..Ji _ V” o ..H .... . I .o i... .0.- 3.1..... . a) c. .
VVJ. .. o . a ..V . .. 6 V. .. 4. 0‘. 5‘7-.. .- ...1.... V . . .
C . . ... O ...u. .VI . . 0 V V.. .‘1V-.o.. O . ..."no...u'c... "u n . .u u. 1 V. o. O.
.....f o 04...... .v‘ . V . .. . V. V . . . . . .. ... ......
_ . .....a . V 4.. _. V o. . V . V V .4 . . . V. V .V, . '6. £33.96 V... . V
.59: . 5.-.- . V .V. V ‘..‘.v..'.‘ 9'. . . V. . . . .. . . . . . A. I
. . .6 n .. n. :2 EV _ 3.0 o. o \6...HV. 3:. p. . . . . . . n f. .
. . V . . ‘1‘ . V V .
.V ..V'. .. ..nv... V.. 1. 4..-... . V V V . . . ... . .... .0 . . .o . .V . . 5%.JSVOY’- Sol...)
—. . . ..0 <u‘ll ... V V. A]. 9.06.,‘Q‘..“ou!0’.
- .0 0.... w......... ..-. 0. ‘4. v. 0.0.-..3VQ- 735......9... a ...“:TI; ...r1.\”..60.u .'
9V .0 .. . .. V . . . . V .7219. . .sf ... a: .. ‘3... 9. 26... . .
_ V. . o c .. . V. . . u”... 1 ...:‘o 9 s
v... .0 J . o r .. .... . . . 2 . o . . V . . . ..’o s . 9.. t ..L a. 16 .3I..9u... \ 1. .. i‘fdx ‘1. . u:
I. a V. ..V . on. V. ..V.. V . V _ V V V V. V . ...V'. v ..V ...‘ak’. I.» 3.1,: ...o. ‘9! 2. .Q
t. V. 6 o... v Oct. ... .2... .....‘ . av..- V .. . . . . . .09... t ... o V‘I V .
II .. .9. _ v .6. ... .1. ...V .v. \o. u. 1‘. . .V . . V
. . v... 3. V . .oV\... 120...... ... .IX. .
c . v3 _ V. 0.... V..\.. ..- . .V .V . 1 n .
.8..... V . ...9. . . .V... ....
V. ... or... 1 o. 4:621}: ..V .v... . ... ....I Q» V. ... . v.0 ... . ...t .. -t V ...... . .0. ....t..... n .. .. V. ...: . .. P331 \
6.. V o a 9-1.6. .. 9. O V . . O ..V: n. a} 5.. ....o?’ a ...: . .. _. .... V ..-.’...~J.o... :3... W... . 4...... ...-I. 5.... . c .. A ... ' ch. . ..V o. . . V... f... .0. .V
V. o ‘ .u .. .1 . V.- .... O- ‘V o.... V..: 3‘... -...Oc; 1:0... 3"! V9! .. V n. V.. V ... .... .. o
.c .. . . V. .q ..Vl ... flu... ..V .. .V.. .35.. A... ... V . ...I .....1.. ..V. . V..
on. _. V. . .o i . .. V .0. .t Z . .. .... . . . .. . ... .OIC.I3.-.ovv \.0 o . . . ... .o . . .- ...
V.. .. 9. V..... .31-... . 4... .v.| . . V . u ... . . ... . . V. .V ..
a .F. ... . V n . V . . 3 . . L... .. . V. V .. oo..olobl o.-V!..V..-. . . .
...... .. V. . ... .0 ... 0.0.0... .9 0. V: ... ...V.b.(.!|t.."| . ...
u. V. v V. V . ..P. ... . .. . . . 5.94.... :
.... P . v . 1......‘2‘ . . .
. . ..‘\...5.. . . o. .‘. . .- .. V.. .... V..
a. . a .v V . .u. I . V V
O. l 1 1! . o o ... . us“:'.00.v.
V. V a. . _. . V. .
C . V. a .10 V V . VV . . . o
.o . u .7 ..v 0 V. o ....l‘l V .. I. . in o . or..- ..V .. .0 .M X.
. .. V ..b....... .0 ... V .. so... ... o . v.3 .... a . u t
«X! . ... V.. 2. . .- Vd. ... . ..V. . o o I. .. Luv-.7 V0903.“ a I... .. «v$..a.- JOAVIIC ‘.
V. o .‘. V ......2u .. a... 0......v0. . .-. .13.. ...olslo u. 9. U V . . . V . .. $ .4 .IVII.
V . V V... ..I... V. . ... .. . . . . II. 0. .0
a . “CHI 9 . . V 5M O ' "05.3:’. ...-.0.
...-.-..I . .. .V . J9......I.A.~.\Ic¢0
... v.0. ... .....-vi .- . . VV. . . .V . . .0‘.-. ... .u .VYV
. 77 v _ ......C.o.. .... V. . V V . . V . .. V
V .t . 4.... .1...o C....;I . .. .V ... V a! .
’0: ...-.....3 V . . .V . . I. . I.
‘ . 4 . ... u. . VV , ... . ...-.1 ‘1... .. . V .V _\.V.V. . O 1...... . V ...9..nl.\vho._ in. iv 93. o CV6. .05 ...-3! II... o"!‘-.‘.“1...w’u
...... . .. CI: . .y.. {V.V... .. ..V . . . .. . . . .. c . . 1......\ 314.. .V' v... ..QJV‘PIo! V to
... 4 ... ..3 1 . .. V .0 ... V . . u t V «0.0» I: _ . 0...... 0... V. ... . .0 ’V.‘ .loV..-V-JVoc no...‘ ,3! t 1‘. . . I
... 2.5.1.. 3.33.... .. v? . ..u .. .. V.. ‘. . . . I! . .J. to .V..... V. ..V ...»..V'qi .
.0 t . s. V.. .. ‘V. . . .. .... V. . . V O. . V . . V . . . . . V
. . o _ ... 0. it .v. .V .30. . .c . V
. . V. .... ..v. ‘o . o 1...' .V .. .V _
.. . a. n .u “.V cm ... ...:«l .o.‘ . .... V
.l n c .. VI. 6. I . .V . 54' ~A.VO
. . V . v . 0.! . . 1 ... . .'l V
J“ . .‘ . . . V . Q '2. . V .V .71! ...... V. ....n .o .. V .... .....f. . . .V .p 1 cl fut V.V. I... . V..;‘I‘hruoilloo tumour’..uihua..l..‘
.0107: ...? . . . A... ..~ 0 .... v.0 V V 0V. -..... .. ..o.... ...... . .... ... . to. .... . V.. 0. at. o ....ch .... .- ... ..I. .GO . IV"V .3 n l -0“ in
. Va... ... . V . 0'. .... 3.... O c. . . 0V _. 01.03... V . .. V .. V ... V . .... ... .. V Ahuo‘ .. I ..o. -.. 1.“ .9.C.'V 4a,- o C. ‘ I .V.‘
..IV . v .- oV . .. .' .. V.. . t 3.1.! . o .. . V. . ...l . V .o .VV . . o o .. .. .. . c . .. ADV. u. .....H. V . .q!‘ .. U . ... lo. a... V. ‘ u. . .. IV. s . ’5 .v t! . t . a I . V. . s V .
. I o a... .! Vt"... ...vq....4.303. V o. .... V ... 1 . . p 3 ‘0. a .1 . V. o I a .IT’.‘ .....l’... i. o. s 9 .V .. :1 1... V... . V . o . - .
. . 20.. V. .3 a .. v 3.. v. ... .. . ...—a' ... V .. . ... .53.... . .. 6 . . . .
l V. I V V v. . . n V L: .V V 1 V . u . . ... O...
. ... . .9 ‘ § 0 O .10. . . V..... .n 1.... I» ... a... ... z . .4 . I . .v 0 A . V.. r. c... . . ... V
. . o . .o . .... O... o I. .t:..! V V. ..V V .. O . Q I v'. . 9...! x.....u‘u...
O. u. a ..V .1 .. « .u' . _ ...... .... u . V V.. .....5: .....JAlqu .- . V . . .
V I. o. d. C. . . .. O o .. u .V a . n. ... (\l;
. .9 V ..V o. 4 ¢ ..v .o . I... . ...V V . ...; . _ . V .. . . .V . .
9 av t . . 9" . v 0... V. .... .. V 1' . .. .... Al‘s... . u . . .V ‘tucfir V.. .V o. - .0. .
. . v . . u... .. ‘ ... . ...: . . .. f ... . V. f. V.... . 1 p a. ._ ..I‘o ..V . ... 3...; .-.-L . ... ~.._ v...
.. . ...! ... . . I ...... 0 V N... . ...-V ....l. .I...’ ............ . '0 n .-V 3...... ._ . .V V . l V V s
‘:i...!- V f .0... o .Av. V....‘o: V ... .csV . 3-0. ... . 3. . ......V. s: . V I I . .... .
VV 5.....10. V .. o .. . its: . z ., . ...I. V . 0
V .9. . ... .. I. . -. . . . .. 01.. . ... u .u o . V .Q . . .30. IV V I’I..A C. V
l 0 c . . . .1 é... o. . . .. s...‘ 0.... .... .V n V? . . V. V ......V. I... Ol V. V ..V V V
V J. .. o co . .- .. I ... . .1... ... . v..- . ~ . . u . V.. 2 . 1”. .03....” .5116. s10...“d\".h\d. i V
‘.’n.. . . . -... I... .0. .... ... .. ... V Vic: .‘ ... ... 9... .Oav.’....0. ‘v99. 30...... V. .13... u .
.. . v. . .. . .t a ’0. .-O. I... ...I. VnV ..IV I V... 0‘. \’....a. . ’6'..6.-a .’0..PV
..V r .... . ... .. ... V. .... ‘9... . . . 1 ...n. .. .- .. .... .. . ‘o 1 o ‘1 V .V’. x . ....V EPIS.
._ A. V . :3... 0...... 0.19.3..- .... .. _ ..V . ...V. . O. .: ... v. I .. 4 2... a... . . . . ..f. .‘au‘ .Vo
. .IVQ. . .. v... I. u . o. I. u. . Y; . ... . .1. . «V . . .... V. . ... o . o .n .. a. I. so u t. u . .
V V .. u.V cos. c. . V... \. .... . .V V 1 . V. ..V. .. ...: . .V . V . ..V . V a 0.. n. V... 9!... ... ... at H. “IOL’V.HM . t!
. .v 0.0.0 .1... IftVzv. I V! . .. I . . . o o .7 o. 0: v V O .0 ... 11 . ..V. . u. ... 53V- O . .V. . . v I. .. .... .V V VI i.l..nw.. -‘ I V.. V. . V
. ~2Av...’ ...c _ v: :3... ... .. u .a: 1.6»: O... . o. . ......or..VVl.. .1 . ‘0. 1' . .c . t .o ... .V. .-. .V...¢V.V.v-va~ ..fi5.l.~. s I.
cl .1. . .0. or. o. . V. a: .‘..l. . .V> I .. V . . . ... .. Io. ... . V . .. . .t V ..V). ... o...l..'fi ...:JqoogoVVVC‘. 2.0-0.5.!
.4.-. 3.. V. — .!. .. A V....V....V......4 .. ...V. V V . . V. V. . V. ..V. ’ ‘.. ~ ...... V...... .....B 921......3: ..V . V ... .a
.... ... .... . I .073...» ......10. 0 V... .... ..\.8...‘ V . . . . V . V V V V. . V V .Vn. 03!... . V... .ch...‘ Q...' . V. . up I....| a.OIo.O.orV
... V..... ‘1...» V. c.- .. I“ o ....... .VV'.| ..... .0. . .. .. . .04. . .‘V. ... Vv .. . . V. . . V. V . . V . V. V. . V . ... a a V v I! .V. p... at. ... ‘1'. l V
v. ...... .. V ....J'.V..\V.4'>:.. I... V -.....t. .I . ..V: .I .. ....V ..V . .00 V . . . . . . .V . . V .. . . . V .l.... o... ... .t..' . ..-. . t a .- ...... 9‘:
Ar :4 v .... t ...; .VV. . V. . . ... C... . vV . 9.... . o... ... ...r... . . . V . .V.. . .. .. :1 so .I g n v. . ... .lv‘IQ\‘O . It. ‘~t.l
i... .0 .4... ... . 5..) 5‘3... .. .-V ... .O a ..V . .0 ....0 .... .....v...-.. 9 . .... 3......31 ... . V. . . . V ....Ic. - o fun-99'1“ .l ‘V3V‘O43I1
..I o . n _ . 3.2... V . V ... V. V V VV .. ...... V .. : .02... . o c..1....t ~.. . . 0......1u: .V .. ..-
v to . .... .... .V . V .... . . . 16.} .0... V O .0. V 2.7.4 .3. I . . . ..I . V .c . . . .Q . ... A... . u -...o ... v s 5. 92“ VI. 3. .
v.4 9.0V. .. . V....... V . V .. . V.. V. J ... ‘3. 3 o 0.! . 0O. . .0 J. . . o I. a . V V V. V V l« !\ < ...V .o .91. \.~.’V.Oo.- ..V. 75.. . . ....W V ts. .. r. . I'd ....
“V. .. -.. -..... V U. I 9V '- .. ... 1:9... 3... f a 6.... . ...-...! . V V V . . .V V . .. V V.. . . o ... . .... ‘ . l.... V It.IA
3.01 0 Y‘o!Vo V..... ._ a V. v: .1! . O. .... .... I! .o v . . V V ... _ .V V ... V V 'I. v: Q . 3.....I .\.I0 n.9,. v ...9 IOUIA -.-:
V.. .o. v. on: . ...-' t.. .V ..V.... V . lint-.1 ......oV. . 0 . . V V .. V . V s . V.. n r I V. fl)... .V... "O O V. n
“- .v -. V? .. ...:- o .13 .... ... $6 .31.. .. .V ..V... V?» . .3 10.... I . ... IO. ...” ...‘v A. 6. a o . V t ..c ...lVa.nx.. .00:I.o0h . .a
. . .. . .0 . I. ...... o... .. 9 ....V . ..D ...! . V . o. .! a .u I. oV . n O u. v .u . to! .I . '. . 0 - v o o . .0 V a . u . , o .90} 1 o J V a . I! a
A ....l I... \...‘.u..6. .13.wa .. a Y .... ...... .. V . L 0.0.... ...-..V; ...? V. . V . .. - ... .. ..o I... V rot. ..I‘. .... V. .. o ‘. .. .... - Vavt . o. ..\. w .p .o 0-. .\.- 0.75 ._ . V.. . ..ut. . . .....4- ..V.” .\
o do .0 . . .... I I o. o O o o 0.0...- .. . .... . t a ' U . o .9 V ... . .. .. I. .3 I .. C v .. on... 05...... ..V! I . . .0. . . 6.5 I O .3. 0.1 o. .n 0. .V .. . . s‘ n . . . .n n . . Q ... up a . b V o ‘ . J c A
C . 3... o .0 . . b V.. 70 ....ovo c a..‘ .- o .o u A o . ‘ .V ..V-lvnun V . .-. 30...... V. v n. . I ... .c I . .00. .0. .. V c \ Q .. :- :6 V . . Q, . Z I. . o. g 0 in o . .1. a! v. o I? .0. . u ..
9 0.. vVV . -.. .. . . ... .. wob 0. V n 0 o $ . ..ou. 1. I. v. .....n. . . .... ......QI‘! - . . 1.9.... Is: . u... .. ... .... .'. o ...... . or. ... v o.“ . a... . .. . . r . D o . v o .l . I‘ a
I .II V o. 6 v. I ‘q . o 9 0| c. ... I 6V 0 V .. U . -a In. I .. .. . . .. . . .- o '. . Jo... .. < -.. I V . _ O . CI. of... a
D notdt V ... u - ..‘. .. . .5 ... ID ..V ... ..V . . o .. V . .1. . .. v n .I Y! ..V. .v .. . . o . . t . .... . . . o I .. “o.“ I. hi—l‘. ..a'L -10 A3...
W... . .5 u .. ' . ‘ a . ..V. .. . ... .10. a . DV. . .3 ... o 0 v I; V a V. . . . d’ k u 0‘ 1 Vacu’ ' .0. . ...Il. . I .
... o. u I ...... ... V6 1.1 .a .. v. .V . . .. - V. .... .v'... ‘. . _. c . In: ... ....OlonSr -. .II ‘,.r a .0..- m. ...; n
. ..t .. o . v A .. .l s . . V. ......c... .o DI1 4‘. I I 0‘ l ..V... u¢ 'V' 0., .-<..“ },....w’:.’: ,0. 90.... V o.
. n\ \o o v- 9 -.. ... O .4 . 2.. . V. V ...... V .. V.. \. b.11l \0 9 .... ..O V 1 m ”If... 24!. . V figw 2.32.3600‘976931. ... ... ... .3
r I .. o .. V. . ..V . . Y . VG . .. . 6.. O t . s 1.. «...... a . . ’v’...«!¢t..l «.... V”. Va. S! .2. .15 gem-aw... 2“... - ..V 2 ... . . . ..V
v o. . Y 2.. .' ¢ . .3 .u .I. . V..... . V. I. 00 4 . 4 . I. . ..‘uu'o‘ . . dillcc. .5? r 352.. \ . v - .... . . .0... :3. 2."... . 2 Q.. V 0.49 c. v. c
r... ... . .1. ..r - . 1L... . .V . V .V l. . ....an M. .~.Aqrn..q.¥.§ .11 ..V. .V aunt ... ...u....uVV...u.... .. - _ V . .
. . V. ‘1 .' .w .. '9: I .0? .5!) o u 7 ...? 05. 9?. -aVTbfl’f. .- ..u’. ...u 3.3.4.... V V..): .5 ... . ..
.9 . .
...6. Vin P Q... t .1... .W sd.”.....1‘t ... ...
V. . II. I. 00.00... . ..1’ J .....3. . ......31... . 3... 0 ..V. |
. .a It . . . . . o. . . a... ...: ...v 03.. . ... .
. ‘t‘vL‘W. a r“ . ‘ ..
<L Jo .71 5 ‘OJ% R. .Jr . ‘tt”. ......
o lo 0 .0. . V V. ... Va... ... ...... 3‘1 . ... .. Q . . I

 

r THESIS l V
3 LIBRARY

7030 Michigan State
University

This is to certify that the
thesis entitled

LCA - A CRITICAL REVIEW AND ESTIMATION OF ITS
UNCERTAINTY IN THE COMPARATIVE EVALUATION OF
PACKAGING SYSTEMS

presented by
DARIO MARTINO

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Packaging

wedge“ ng

Major Professor’s Signature

FJA. M/ Zflflé

 

 

Date

MSU is an Afiinnative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:lC|RC/DateDue.indd-p.1

 

LCA - A CRITICAL REVIEW AND ESTIMATION OF ITS UNCERTAINTY IN
THE COMPARATIVE EVALUATION OF PACKAGING SYSTEMS

By

Dario Martino

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

The School of Packaging

2006

ABSTRACT
LCA —- A CRITICAL REVIEW AND ESTIMATION OF ITS UNCERTAINTY IN
THE COMPARATIVE EVALUATION OF PACKAGING SYSTEMS
By

Dario Martino

Life cycle assessment (LCA), defined as a systematic approach to evaluate
environmental burdens associated with a product, process or activity covering its whole
life cycle (i.e. from raw material acquisition to disposal) has been increasingly used by
firms and government agencies to try to estimate the environmental impact of packaging
options. Among the many LCA studies published, a great number of them are
comparative studies to identify the environmentally better option, but whose results end
up often being challenged by the losing parties.

This dissertation elaborates that conflictive results are a sign of LCA limits, which
in turn can be traced back to the uncertainties inherent to the LCA approach. In order to
fully represent these ideas, this work offers a two-step approach. First a critical overview
of the state of the LCA method, its disparate applications and impact on topics that are
part of the packaging field is offered. Second, the effect of some types of uncertainty (i.e.
inventory data and scenario) in the outcome of a packaging based LCA featuring three
different packaging materials is analyzed for a hypothetical drink product using published
average process values. The three packaging systems are based on PET bottles, PLA
bottles and an aluminum cans, along with their whole set of distribution packaging, and

transportation services and end of life alternatives which were evaluated with regards to

four environmental burdens: energy use. water use. global warming potential (GWP) and
mom depletion potential (ODP).

The stochastic simulation for the simultaneous assessment of uncertainty due to
LCI data. through Monte Carlo simulation. and scenarios relevant to packaging. through
a non-parametric procedure. has demonstrated that. in this particular comparison, LCI
parameter uncertainty seems to have a dominant effect in the outcome of the LCA results
considered. Furthermore. by including the aforementioned noise, or considering temporal
or spatial information or allocation procedures, the overlapping of uncertainty intervals
(at 95% confidence.) indicates that, despite the differences in the primary raw materials,
the systems under study can be similar to one another, at least with regard to the
environmental indicators of this work.

The dissertation also demonstrates that when the knowledge of the systems under
study is limited to average values of the different operations. or to conservative
uncertainty estimations. a clear option is unlikely to surface. On the other hand, the
dissertation shows that when more information is known about how. where and when the
system operates (i.e. location and time of operations, type of technology, raw material
quality) more conclusive results can be obtained within an impact indicator, but it also
shows that this gain cannot be fully exploited unless a priority-based approach across

indicators is used for perforating the impact assessment study.

Copyright by
DARIO MARTINO
2006

DEDICATION

I would like to thank Dr. Susan Selke, my major advisor, for her patience and
endless support. Many times, throughout the duration of this research she encouraged me
to challenge the framework of many ideas that I assumed for granted, and that, I wish to
believe is one of the contributions of this research. I would also like to thank Dr. Satish
Joshi for sharing with me his experiences in LCA research and Dr. Amar Mohanty for
providing me an insight on the tendency of LCA use in the biodegradable material field. I
want to express gratitude also to Dr. Bruce Harte for being a mentor figure not only
during this research but also during my whole stay at Michigan State. Our dialogs will be
among my most cherished experiences of Grads School.

It will be an injustice if I don’t include in these paragraphs, Dr. Seungdo Kim
from the Chemical Engineering Department who allowed me access to the DEAM
inventory database modules that are used in this study. Gratitude also to Dr. Keoleian and
David Spitzley from the Center for Sustainable Systems at University of Michigan, both
of which allowed me to attend seminars in their department that helped me better
understand the nature of LCA. Thanks also to Dr. Huijbregts and Dr. Ragas, from
University of Nijmeden (The Netherlands), for their correspondence with advice in the
use of Crystal Ball, and to Mary Curran (US. EPA) for her advice in evaluating the
literature review chapter of this dissertation.

Last, but not least, I want to express gratitude from the heart to my beloved

Sophia and Karina. who stood by my side throughout the fortunes of the past years.

TABLE OF CONTENTS

Chapter 1 OVERVIEW OF THE RESEARCH .................................................................. 1
I . Introduction ................................................................................................................ 1
II . The need for packaging-based LCA research ........................................................... 2
III . Objectives ................................................................................................................ 7
IV . Hypothesis ............................................................................................................... 8
V . Research methodology .............................................................................................. 9
VI . Scope of the dissertation ........................................................................................ 10
VII . Contributions of the research ............................................................................... 11
VIH . Structure of the Dissertation ............................................................................... 13

Chapter 2 LITERATURE REVIEW ................................................................................. 14
I . Introduction .............................................................................................................. 14
Il . Definition of LCA ................................................................................................... 15
III . Product life cycle ................................................................................................... 16
IV . LCA Stages ............................................................................................................ 18
V . Limitations .............................................................................................................. 21
VI . LCA enhancements ................................................................................................ 28

a) Data-related improvements ............................................................................... 28
b) Streamlined LCA ............................................................ .. ................................. 32
c) Input-Output LCA ............................................................................................. 38
(1) Economic analysis and LCA ............................................................................. 43
VII . LCA and packaging .............................................................................................. 46
a) Stakeholder use of LCA ........................................................................................ 47
b) Third party use of LCA ......................................................................................... 62
VIII Outlook and conclusions ....................................................................................... 70

Chapter 3 COMPARATIVE LCA OF THREE DRINK DELIVERY SYSTEMS .......... 74
I . Introduction .............................................................................................................. 74
Il . Methodology ........................................................................................................... 76

a) Goal and scope definition ................................................................................. 76
Purpose of the study .............................................................................................. 76
Function and functional unit ................................................................................. 77
Product delivery systems (PDS) ........................................................................... 77
System model ........................................................................................................ 79
System boundaries ................................................................................................ 81
Data categories ...................................................................................................... 84
Impact categories .................................................................................................. 85

b) Life cycle inventory .......................................................................................... 86
System ................................................................................................................... 88
Material flow diagrams ....................................................................................... 112
Allocation procedures ......................................................................................... 114
Calculation procedures ........................................................................................ 116

vi

Functional Unit and Losses ................................................................................. 1 17

Assumptions and limitations of this study .......................................................... 122
Input data ............................................................................................................ 124
III . Preliminary conclusions ....................................................................................... 133

Chapter 4 UNCERTAINTY ANALYSIS OF THE COMPARATIVE LIFE CYCLE

ASSESSMENT OF DRINK DELIVERY SYSTEMS ................................................... 134

I . Introduction ............................................................................................................ 134

II . Methodology ......................................................................................................... 137

a) Scenario uncertainty analysis .............................................................................. 137

b) Inventory data uncertainty analysis .................................................................... 139

Presentation of results ......................................................................................... 145

Simulation details, hardware and software ......................................................... 146

III . Results .................................................................................................................. 147

a) Overall comparative LCA results ....................................................................... 147

b) Scenario-specific comparative LCA results ....................................................... 150

IV . Discussion ............................................................................................................ 155

V . Conclusions ........................................................................................................... 168
Chapter 5 OVERALL CONCLUSIONS, RECOMMENDATIONS AND FUTURE

WORK ............................................................................................................................ 169

I . Recommendations for future work ......................................................................... 174

APPENDIX ..................................................................................................................... 176

I Visual basic code used for scenario analysis procedure ........................................... 178

II Results of the quantitative uncertainty analysis ...................................................... 179

111 Details of the sensitivity analysis using Spearman’s Rank correlation method 194

REFERENCES ............................................................................................................... 196

vii

LIST OF TABLES

Table 1. Critical issues at different stages of the LCA process (Based on Huijbregts
(1998a) and BjOrklund (2002)) ................................................................................. 23
Table 2. Summary of LCI databases and managing organizations and their status
(updated from Norris and Notten) (2002) ................................................................. 29
Table 3. Expected variations in LCI data (Finnveden and Lindfors, 1998) ..................... 32
Table 4. Methodology oriented streamlining approaches (based on Todd and Curran
(1999), and Hunt et a] (1998)) .................................................................................. 35
Table 5. Environmentally responsible product assessment matrix (Graedel and Allenby,
2003) ......................................................................................................................... 37
Table 6. Basic structure of an EIO matrix ........................................................................ 39
Table 7. Main hybrid approaches combining process-based analysis and input-output
based analysis (based on Suh et a1 (2004)) ............................................................... 42
Table 8. Differences between LCA and LCC (Norris, 2001b) ......................................... 45
Table 9. Categories of costs according to AlCHE/CWRT (Center for Waste Reduction
Technologies-AIChe, Total cost assessment methodology, 1999) ........................... 46

Table 10. List of selected packaging oriented LCA and LCI involving comparison studies

................................................................................................................................... 52
Table 11. Primary packaging characteristics .................................................................... 78
Table 1.2. Polyolefin resin production system data sources .............................................. 90
Table 13. PET resin material production data source ....................................................... 92
Table 14. PLA resin material production data source ....................................................... 92

viii

Table 15.

Table 16.

Table 17.

Table 18.

Table 19.

Table 20.

Table 21.

Table 22.

Table 23.

Table 24.

Table 25.

Table 26.

Aluminum material production system data source ......................................... 95
Distribution 1 system data sources ................................................................... 99
Blown film extrusion process data source ...................................................... 102
Injection molding process data source ............................................................ 102
Can body and lid making processes data source ............................................. 103
Distribution 2 process data source .................................................................. 106
Distribution 2 process data source - Distribution packaging modules ........... 106
Injection stretch blow molding process data source ....................................... 107
Distribution 3 process data source — Transportation modules ........................ 109
Distribution 3 process data source - Distribution packaging modules ........... 109
End-of-life process data source ....................................................................... 1 11
Different scenarios evaluated in this study ..................................................... 150

Table A. l. Uncertainty factors (k) for categories of parameters used in the screening for

relevant variables .................................................................................................... 177

Table A. 2. Uncertainty factors and distributions of parameters with at least 1%

contribution to the variance of overall LCA outcomes ........................................... 177
Table A. 3. Scenario-specific interval results for the relationship PDS l/PDS2 ............. 179
Table A. 4. Scenario-specific interval results for the relationship PDSl/PDS3 ............. 182
Table A. 5. Scenario-specific interval results for the relationship PDSl/PDS3 ............. 185
Table A. 6. Life cycle results for PDSl .......................................................................... 188
Table A. 7. Life cycle results for PDSZ .......................................................................... 190
Table A. 8. Life cycle results for PDS3 .......................................................................... 192

ix

Table A. 9. Details of sensitivity analysis ....................................................................... 194

LIST OF FIGURES

Figure 1. Typical product life cycle .................................................................................. 17
Figure 2. The stages of LCA and possible applications (ISO 14040, 1997) .................... 19
Figure 3. General representation of main differences between mid-point and end-point
methods for impact assessment (based on Bare et al (1999) and Jolliet et a1 (2004))
................................................................................................................................... 20
Figure 4. Streamlining approaches ................................................................................... 34
Figure 5. Environmental label classification (based on Environmental Labeling Issues,

Policies and Practices Worldwide, United States Environmental Protection Agency

(1998)) ....................................................................................................................... 64
Figure 6. Life cycle phases considered in the study ......................................................... 80
Figure 7. Polyolefin resin material production system ..................................................... 89
Figure 8. PET resin material production system .............................................................. 91
Figure 9. PLA resin material production system .............................................................. 93
Figure 10. Aluminum material production system ........................................................... 96
Figure 11. Distribution 1 system ...................................................................................... 98
Figure 12 Container component manufacture process of PDS l and PDS 3 ................. 100
Figure 13. Container component manufacture processes of PDS 2 ................................ 101
Figure 14. Distribution 2 process .................................................................................... 105
Figure 15. Distribution 3 process ................................................................................... 108
Figure 16. End-of-Iife processes ..................................................................................... 110
Figure 17. Material flow diagram of PDS 1 (PET based) and PDS 3 (PLA based) ....... 113

xi

Figure 18. Material flow diagram of PDS 2 (aluminum based). .................................... 113
Figure 19. Losses in the product delivery system ........................................................... 118
Figure 20. Procedure followed for scenario analysis ...................................................... 138
Figure 21. Inventory data uncertainty analysis procedure (Huijbregts et a1., 2003) ....... 141
Figure 22. Overall uncertainty in the comparison ratios due to the combination of
scenario and LCI data uncertainties (asterisk denotes significant difference from
unity at 95% confidence level) ............................................................................... 149
Figure 23. Scenario-specific results for the relationship PDS l/PDS2 (asterisk denotes
significant difference from unity at 95% confidence level) .................................... 152
Figure 24. Scenario-specific results for the relationship PDSl/PDS3 (asterisk denotes
significant difference from unity at 95% confidence level) .................................... 153
Figure 25. Scenario-specific results for the relationship PDSZ/PDS3 (asterisk denotes
significant difference from unity at 95% confidence level) .................................... 154

Figure 26. Life cycle energy profiles of the three systems (values shown are medians) 157

Figure 27. Life cycle water use profile of the three systems .......................................... 159
Figure 28. Life cycle GWP profile of the three systems ................................................. 162
Figure 29. Life cycle ODP profile of the three systems ................................................. 165

xii

LIST OF ABREVIATIONS

GWP: Global warming potential
ISO: International Organization for Standardization
LCA: Life cycle assessment

LCC: Life cycle costing

LCI: Life cycle inventory

LCIA: Life cycle impact assessment
LDPE: Low density polyethylene
MC: Monte Carlo (simulation)
ODP: Ozone depletion potential
PDS: Product delivery system

PET: Polyethylene terephthalate
PLA: Polylactide

PP: Polypropylene

SETAC: Society of Environmental Toxicology and Chemistry

xiii

CHAPTER 1 OVERVIEW OF THE RESEARCH

I. Introduction

The final goal of this dissertation is to provide a thorough understanding of some
of the reasons behind the complex (and many times conflictive) results of comparative
packaging based LCA. In order to achieve this, two major issues need to be addressed.
First, the scope and limits of the Life Cycle Assessment (LCA) method need to be
known. This dissertation achieves this task by offering a critical overview of the state of
the LCA method, its disparate applications and impact on topics that are part of the
packaging field. Second, as will be seen in later sections of this dissertation, when LCA is
used for comparison purposes — as it is commonly done in packaging — it often reaches its
limits by giving unclear and conflictive results. Many times these limits can be traced
back to the uncertainties inherent to the LCA approach. Thus, the second task of the
dissertation is to analyze how uncertainty in both the LCA method and packaging
operations plays a role in actual comparative packaging-based LCA results.

The remainder of this introduction will provide a roadmap to the subsequent
sections of Chapter One specifically, and the dissertation in general. Chapter One has
several sections that explain in more depth the motivation and the approach taken to
perform this packaging-based LCA study. Section II, in particular, highlights the
relevance of conducting this LCA research. Additionally, this section touches upon some
recent LCA methodology publications that are reviewed in detail in Chapter Two.
Section III enumerates the objectives of the research. Section IV states the hypothesis of

the dissertation. Section V reviews the overall research methods, outlining the steps

followed throughout the dissertation to achieve the stated objectives. Section VI discusses
the scope of the dissertation, while Section VII discusses the contributions of this
research to the field. Lastly, Section VIII details the structure for the rest of the chapters

in the dissertation.

II. The need for packaging-based LCA research

Life cycle assessment (LCA) is considered as one of several strategies known in
the industrial ecology field for environmentally conscious manufacturing, pollution
prevention, and “ecoefficient” or “green” design that is increasingly being used by
stakeholders and third-party organizations. Defined as a process to evaluate
environmental burdens associated with a product, process or activity over its whole life
cycle (i.e. from raw material acquisition to disposal), LCA has been praised as the
approach to assess the environmental footprint of industrial operations.

Almost naturally, packaging situations were one of the earliest applications of
LCA. About three decades ago, public concern over increasing volumes of solid waste
due to the use of plastic in packaging and later concerns about energy consumption
became the major driving forces to study the effects of packaging on the environment,
and life-cycle studies were devised to help assess the environmental footprint of
packaging use. Since then, a great number of packaging-based life cycle assessment
studies have been published. Many of these LCAs were and are comparative studies.

However. due to the uncertainties involved in both the LCA methodology and the
operations that are the subject of the assessment, there is not (and there might not be) a

single, definite one-size-fits-all approach to performing an LCA. This is a major problem

when LCA is used for comparison purposes because this means that any result can be
contested. Nevertheless it is not uncommon to find comparative packaging LCA studies
that portray, deliberately or not, one packaging material as environmentally better than
other.

With this in mind, the motivation of this dissertation is to try to understand some
of the reasons behind the conflicts in comparative packaging based LCA results. In order
to achieve this, it is proposed to address two major issues.

First. there is the need to understand the scope of the current LCA approach in
general and when applied to packaging, and its improvements and the arguments behind
the limitations of the LCA method. That is, there is need to identify major advances
developed by the LCA community as well as current issues of the LCA methodology. For
instance, according to the current understanding of the International Organization for
Standardization (ISO 14040, 1997) and the Society of Environmental Toxicology and
Chemistry (SETAC) (Bamthouse et a1., 1998), LCA comprises four stages which are: (1)
goal and scope definition, in which the purpose, boundary conditions, functional unit and
allocation rules are set; (2) life cycle inventory (LCI), which is the accounting of
emissions, raw materials and energy used; (3) life cycle impact assessment (LCIA),
which links the LCI with identifiable environmental threats; and (4) critical interpretation
of results, which is made throughout the study.

However, due to the uneven pace at which it has been embraced around the world,
several aspects of LCA remain unclear. In fact, though developed more than thirty years
ago in the US, the European willingness to incorporate it as a part of their environmental

regulatory process is often cited as the reason why they lead the way in LCA research

(Curran, 1999). Only since the 1990’s with [80’s release of its 14000 series of standards
on Environmental Management, have many developing countries started to learn about
this concept. This delay in coordination (i.e. internationally and nationally) is one reason
why a common terminology has been slow to be developed and life cycle-based terms
and approaches such as life cycle management or life cycle costing generate confusion.
Further, since many LCA studies still remain unpublished or inaccessible (e.g. due to
language barriers or confidentiality agreements), the assimilation of common
methodologies is even more difficult.

In recent years, though, the private sector in developed countries has appeared
more willing to embrace product life cycle concepts at the product design phase in order
to respond to consumer expectations. As multinational firms operate around the world,
their understanding of LCA concepts has been extended. In turn, efforts on harmonizing
private life-cycle initiatives have started to occur (Hunkeler et al., 2002; Hunkeler et al.,
2001; Sonnemann et al., 2001). Private and governmental agencies of the US. and of
European countries have started the development of frameworks and partnerships with
industry to help with the goal of making LCA a more useful tool for decision-makers.
Current efforts by environmental certification third-party organizations in these countries
follow harmonization steps as well as reaching out to similar bodies in developing
countries. Moreover, a survey of the LCA work in general clearly shows the growing idea
that the life cycle concept has evolved from a tool to help compare products to an
essential part of achieving broader goals such as sustainability.

It is thus of particular importance, and not only to help understand the reasons

behind conflicts in comparative packaging based LCA results. that a critical review of the

literature detailing the state of the LCA method and its implications in packaging be
provided, since utilization of this tool in this field is not restricted to comparative
analysis, but also comprises a broad spectrum that includes product development, product
improvement, development of pollution prevention or waste management policies, and
environmental labeling certifications.

Second, as it will be seen in later sections of this dissertation, when LCA is used
to compare the “environmental performance” of packaging materials, it often reaches its
limits by giving unclear and conflictive results. In this dissertation it is argued that many
times these limits can be traced back to the uncertainties inherent to the LCA approach.
Moreover, it is argued that even within the same comparative life cycle assessment study
of the same packaging options, conflictive results can still be obtained and a supposed
preferability of alternatives can change.

In fact, with regard to the LCA method, Huijbregts (1998a) elaborated that LCA
methodology is the subject of different kinds of uncertainties and variability along its
different stages and because of them its results can be seriously undermined. In industrial
systems, uncertainty (i.e. the lack of sureness about something) and variability (i.e.
inherent variation of measured values) are responsible for many of the problems in an
LCA. Uncertainty is important because when not evaluated, there is more risk that the
impact predicted by the LCA may not match the actual environmental impact. Likewise,
variability is important because it limits the precision of LCA results.

ISO standards (ISO 14040, 1997; ISO 14042.2, 1998; ISO 14041, 1998; ISO
14043, 2000) recognize the limitations that arise from not addressing uncertainty and

variability and recommend the practitioner to do so in the LCA study. Nevertheless, the

standards do not suggest any method to do so. In fact, Ross et a1 (Ross et al., 2002)
reported that out of 30 LCA studies only 10% included a quantitative or qualitative
uncertainty analysis. Thus, he concluded that the standards needed to be revised to ensure
that LCA included at least a qualitative discussion of uncertainty and variability. Among
attempts to provide a systematic analysis, just in the last few years the first general
framework for comprehensive uncertainty and variability evaluation with the definition
of key concepts in the LCA context was proposed (Huijbregts, 1998a; 1998b; 2001;
Huijbregts et al., 2003; Huijbregts et al., 2001). But uncertainty analysis is something
new in LCA (Bamthouse et al., 1998) and thus a number of methodologies for
uncertainty estimation have been proposed. Currently, the most favored approach for
uncertainty and variability analysis in LCA is incorporating it into the life cycle inventory
(LCI) stage. However, though often acknowledged to suffer from uncertainty, the
information in databases required for the analysis (e.g. ranges, standard deviation,
probability distributions) is still very limited. Thus engineering uncertainty estimations
along with computer simulation are becoming the method of choice. A number of recent
uncertainty and variability methodologies have been proposed and incorporated into
LCAs associated with different industry sectors and almost exclusively used Monte Carlo
simulation (Canter et al., 2002; Ciroth et al., 2004; Contadini and Moore, 2003; Guyonnet
et al., 1999; Huijbregts et al., 2003; Maurice et al., 2000; McCleese and LaPuma, 2002;
Sonnemann et al., 2003). Others used fuzzy logic to perform a similar data evaluation
(Guyonnet et al., 1999: Tan et al., 2002). In some of these cases, packaging situations

with very limited scope have been used only as case studies to introduce new and often

intricate uncertainty assessment methodologies (World largest PET LCA, 2004; Canter et
al., 2002; Kennedy et al., 1997: Kennedy et al., 1996).

In turn, uncertainties in many packaging operations can be understood as scenario
uncertainties. For instance, differences in percent recycled content in a packaging
material, packaging end-of-life practices (e.g. landfilling, incineration, recovering), and
product distribution distances all may change the outcome of the LCA. Even within the
same country, the same packaging end-of-life alternative can be viewed both as
environmentally friendly and environmentally unsound. In fact, analyzing several
comparative packaging based LCA results, it is not uncommon to find directly opposite
results that are often the subject of heated debates among industry groups and even
countries (World largest PET LCA, 2004; Hockerts et al., 1999).

Thus, in order to understand the nature behind the conflicts in comparative
packaging based LCA results, it is of key importance to analyze how uncertainties in both
the LCA method and packaging operations affect an actual outcome of a comparative

packaging-based LCA.

III. Objectives
The primary objectives of this study are then:
i. To provide a critical review of the state of the environmental Life Cycle
Assessment (LCA) method and its application to packaging.
ii. To analyze how uncertainties in both the LCA method and packaging operations

play a role in actual comparative packaging-based LCA results.

To help accomplish the second primary objective, it was divided into the two
following secondary objectives:

a. To develop a life cycle assessment (LCA) study to compare the
environmental performance of hypothetical drink delivery systems (PDSs)
featuring three different primary packaging materials.

b. To critically analyze the results of an uncertainty and variability analysis
of the comparative packaging based LCA outcome. Specifically,
uncertainty in packaging operations will be expressed by the analysis of
the comparative LCA outcome of different scenarios relevant to packaging
operations. The uncertainty in the LCA method will be expressed as

inventory parameter uncertainty through stochastic simulation.

IV. Hypothesis

This dissertation investigated the hypothesis that when uncertainties inherent to
the methodology and the nature of industrial processes are considered, the current life.
cycle assessment approach does not provide enough evidence to rule out (or favor) one
alternative versus another when packagingsystems are studied and compared based on
their environmental performance. Indeed, the expression “environmental performance”
by itself is very general and could encompass any set of environmental indicators. More
specifically, it is argued that under the stated uncertainties within a same set of
boundaries and same options of packaging materials, the LCA approach still cannot

provide a clear more environmental option when packaging systems are compared.

V. Research methodology

To achieve the first primary objective, a critical review of the literature with
regards to LCA and its applications to packaging operations was performed and is
presented in Chapter Two. The main bodies of literature reviewed in this chapter involve
actual LCA studies, recent journal publications regarding advances in LCA methodology
and case studies, ISO 14000 series of standards, reports from governmental agencies and
private sector literature.

The research methodology followed to achieve the second primary objective was
the following. First, a comparative LCA study featuring three different packaging
materials was developed for a hypothetical drink product. Following ISO’s LCA steps,
extensive details and information sources used for this comparative LCA are presented in
Chapter Three. The three material/container alternatives involved in this study were a
PET bottle, a PLA bottle and an aluminum can. In order to represent realistic implications
of actual packaging operations, production and use of distribution packaging (i.e.
corrugated trays, stretch wrap, and pallets), as well as transportation steps through several
parts of the packaging life cycle were included. Relevant end-of-life scenarios such as
landfilling, incineration and recycling were also taken into account in this analysis. Each
of the main material alternatives (i.e. PET, PLA, aluminum) along with its whole set of
operations was defined as a product delivery system (PDS) and environmental burdens
associated with each of the three PDSs were included in the study. The actual calculation
procedures used in this comparative LCA were performed using a series of electronic

spreadsheets in Microsoft Excel (Microsoft Excel, 2000).

In Chapter Four the actual uncertainty analysis of the comparative LCA is
performed. Specifically, the uncertainty analysis in this study comprises two aspects.
First, a scenario analysis was performed identifying different domains that may have the
potential to change the outcome of the comparative packaging based LCA. Under each
domain two alternatives are proposed. Then, a procedure is used to select an alternative
from each domain and randomly create different scenarios. The second aspect of the
uncertainty analysis in this study comprises inventory parameter uncertainty. This
analysis is achieved by the use of appropriate published uncertainty estimations along
with stochastic simulation (i.e. Monte Carlo simulation) in order to represent inventory
data uncertainty as probability distributions. By this analysis, relevant inventory
parameters are identified and further analyzed, and final results of the comparative LCA
along with the combined uncertainties from inventory data and scenarios are obtained.
Finally in this chapter, the consequences of the comparative outcomes of the LCA are
extensively discussed.

Last, Chapter Five is where conclusions, future work and recommendations are

presented.

VI. Scope of the dissertation

This dissertation contains information regarding advances and limitations of the
current LCA methodology and its uses and implications for packaging. Regarding the
LCA study presented here, it comprises the stages of a packaging delivery system
including material production, container manufacture, filling, and end-of—life operations.

Transportation and use of distribution packaging to transport packaging components are

10

also included. The life cycle model does not include the usage stage nor the
transportation that is related to this stage because emissions and energy use data about
these operations were not readily available. Furthermore, it was reasoned that the
omission of these operations would not impact or change the characteristics of the LCA
since its contribution was assumed negligible (e.g. when compared to the material
production phase)

The impact assessment phase of the comparative LCA in this dissertation
comprises impact categories such as energy and water use, ozone depletion potential
(ODP), and global warming potential (GWP). In essence, the first two are actually
inventory results for which information is often available in the literature and thus
commonly tracked in LCAs and therefore their selection here. In turn, GWP and ODP are
indeed environmental impact indicators which were selected since by being the only
impact indicators with a global scope (i.e. not dependent on location) and whose
characterization methodologies are among the most accepted in the LCA community,
they offered what were expected to be the most robust environmental impact estimations.

The uncertainty and variability analysis includes the study of the outcomes of a
comparative packaging based LCA result using different scenarios that are relevant to
packaging operations. This study further uses Monte Carlo simulation in order to help

with the parameter uncertainty analysis that comprises the life cycle inventory phase.

VII. Contributions of the research

Primarily, this dissertation is a conceptual analysis and pondering of ideas based

on the packaging-based comparative LCA outcome. The contribution of this dissertation

11

is in providing information and discussion to help understand some of the reasons behind
the conflicts in comparative packaging based LCA results.

This research also satisfies other academic and practitioner interests in life cycle
assessment practices used in the packaging field. For instance, information in the
literature review of the dissertation can serve not only to help understand the current state
of LCA, but also to screen for appropriate resources at the moment of deciding to do a
packaging-based LCA. This section is also of value since it provides an overview of the
LCA use associated with implementing programs with impact on packaging, such as
product take-back initiatives as well as environmental labeling certifications.

With regard to the materials selected for the comparative LCA study, the
consideration of a novel biobased material (i.e. PLA) as an alternative will also help
understand some of the environmental implications of the use of a biobased polymer as a
packaging system associated with a relevant packaging functional unit. As will become
apparent in later chapters of this dissertation, the selection of this novel material also
highlights a critical aspect of LCA, which is the limitation of the environmental
assessment based on engineering estimates, diagrams, and theoretical studies instead of
actual commercial production data.

Insights from the discussion of the effect of uncertainty and variability on the
outcome of a packaging-based LCA outcome will also prove useful to motivate further
analysis and extend it to other packaging systems and/or other scenarios with relevance in

packaging operations.

12

VIII. Structure of the Dissertation

Chapter Two is a comprehensive review of the literature. The main bodies of
literature involve LCA methodology, ISO 14000, publications by governmental agencies
and private sector literature. Chapter Three introduces the details and data sources used
for a comparative LCA study of drink delivery systems featuring three different
packaging materials. Chapter Four presents an uncertainty analysis on scenarios and the
inventory data and discusses the results and their implications. Chapter Five presents
overall conclusions as well as recommendations and future work. Finally there are the
appendices that contain data tables and statistical procedures used for uncertainty

estimation.

13

CHAPTER 2 LITERATURE REVIEW

1. Introduction

Life cycle assessment (LCA) is defined as a cradle-to-grave analysis that can be
used to move towards the ideal of sustainable development (Klbpffer, 2003).
Sustainability, as adopted by the World Commission on Environment and Development
(1987), describes the political goal for the future of mankind. Sustainable development
implies an ideal balanced relationship between natural resources and human activity.
LCA is intended to play an integral part in attaining this balance.

LCA is the result of the evolution of early “waste” and “energy analyses”
performed by a few industries in the past. These waste-energy analyses were aimed at
calculating the embodied energy and generation of solid wastes over different stages of
the product life cycle. Later, due to their identical methodology, these were expanded to
encompass the computation of total life cycle release of pollutants.

LCA is one of a number of environmental impact evaluation techniques and as
such, it is recognized that it may not be appropriate for all situations. In fact, this tool
presents limitations that are intrinsically related to the definition of the scope and
interpretation of the systems that are being assessed, the modeling of emissions, fate and
final impacts of substances released into the environment, and problems due to the data-
intensive nature of the assessment. As a result, LCA is an evolving tool.

The purpose of this chapter is to provide an updated, structured review of LCA
information for packaging-oriented LCA practitioners. To help present the information,

this chapter has several sections. The first section describes the basic nature and

14

characteristics of the LCA methodology, followed by a review of the limitations of LCA
and recent enhancement approaches deveIOped by the LCA community to address some
of the limitations. Next, the use of LCA specifically in packaging applications is
reviewed. A birds-eye description of related private and governmental initiatives with
regard to LCA in packaging is included. The final section discusses the likely future of

LCA applications in packaging.

II. Definition of LCA

In the last fifteen years, several definitions of LCA have been offered with minor
variations from each other. Developed in 1991, the definition by the Society of
Environmental Toxicology and Chemistry (SETAC)(Workshop report: A technical
framework for life-cycle assessment, 1991), was among the first ones and describes LCA
as:
“ . an objective process to evaluate the environmental burdens associated with a
product, process or activity by identifying and quantifying energy and material uses and
releases to the environment, and to evaluate and implement opportunities to aflect
environmental improvements. The assessment includes the entire life cycle of the product,
process or activity, encompassing extracting and processing materials; manufacturing,
transportation and distribution; use, reuse, maintenance; recycling and final disposal

Since 1997, harmonization steps regarding a common LCA understanding have
been made with the appearance of the International Organization for Standardization

(ISO) LCA-related standards (ISO 14040, 1997; ISO 14042.2, 1998; ISO 14041, 1998;

ISO 14024, 1999; ISO 14043, 2000) within the 14000 Series of Environmental

15

Management System (EMS) standards, and their subsequent rapid global adoption. ISO
defines LCA as:
“a compilation and evaluation of inputs and outputs and the potential environmental
impacts of a product system throughout its life cycle

In either of these definitions, the life cycle of the product system, or cradle-to-
grave, starts with the gathering of raw materials from the earth to create the product and
finishes when they are returned back to the earth. LCA then considers the cumulative
environmental impacts that occur due to all stages of the product life cycle.

Worth noticing is the fact that by considering the whole life cycle of a product,
LCA has the potential for avoiding problem shifting. However, understandable or not, the
LCA definition does not require that all possible environmental impacts be accounted for
(i.e. a study that looks only at greenhouse gas emissions on a cradle-to-grave basis can be

called an LCA), allowing room for burden shifting.

III. Product life cycle

LCA is based on the product “life cycle”. Figure 1 shows a simplified diagram of
a typical product life cycle which would start with the raw material acquisition (e.g.
petroleum extraction and refinery for petroleum-based plastic products). After raw
material acquisition, the cycle would include the material manufacture stage. Here, raw
materials would be processed into basic manufactured materials (e.g. manufacture of
plastic resin pellets). These materials would then be moved to the actual product
manufacture stage where they would be made into products (e.g. plastic pellets extruded

and molded into milk jugs). Eventually, they would be used (e.g. by milk distributor and

I6

consumer) and disposed. When disposed, they might go through waste management
programs to be reused, and/or recycled, and/or incinerated, and/or sent to landfills. As
shown in the diagram, all stages are interlinked, and along with the transport required to
move products and materials, require energy and ancillary materials and produce wastes

and emissions.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r - - - - — - - - - - - - - - — — — — — - -
I Raw material |
I acquisition I
Materials recycle I
r ----------------------------------------------------- I
I . .
g : l
I 1 I Material E
I Manufacture : I
Transport : I
I Product remanufacture :
I
.' """""""""""""""""""" 1 I
l I I
I t I Product :
Manufacture Product reuse i I
I Transport r ------------------- :
I I
I : g I
—¥——> Use 5 I
I Transport :
: I
| :
: I
I a Disposal
Transport I
- - - - - - - - - - — - — - - - - - - - - I
Energy . .
Ancrllary materials Wastes and emissions
(Input)

(input) (Output)

Figure 1. Typical product life cycle

17

IV. LCA Stages

LCA is an evolving technique. Currently, ISO and SETAC guidelines divide the

LCA into four stages ([80 14041, 1998) (see Figure 2).
Goal and scope definition: where the purpose of the study, its scope, functional unit and
the procedure for quality assurance of the results are described. This step specifies the
inputs and outputs selected for inventory and selects the functional unit, a common
reference to which the inputs and outputs are related, associated with the function of the
system under study.

Inventory analysis: which is the actual quantitative analysis of inputs of raw
materials and fuels into a system and the outputs of solid, liquid and gaseous wastes from
it. In the Life Cycle Inventory (LCI), data associated with the flows is collected using
literature studies, interviews, measurements, theoretical calculations, data banks and
qualified guesses. In theory, the application of allocation principles (i .e. partitioning the
input or output flows of a unit process to the product system under study) and procedures
should also be explained, and information required in recycle or reuse situations should
also be presented. For transparency, the details of the methods for data collection and
sources of the data should also be provided in this phase.

Impact Assessment: SETAC and ISO define environmental life cycle impact
assessment (LCIA) as the stage whereby the inventory results are linked to the
identifiable environmental problems. This stage is a technical, quantitative and/or
qualitative process to characterize and assess the effects of the environmental emissions

identified in inventory analysis.

l8

 

 

 

 

 

 

 

 

 

 

 

 

LCA framework
G031 and scope fl—b Direct applications:
definition
Product development and
t improvement
Inventory analysis I I Interpretation ' Strategic planning
(LCI) ‘ ’
- Public policy making
t - Marketing
Impact assessment H _ Other
(LCIA)

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2. The stages of LCA and possible applications (ISO 14040, 1997)

There are a number of impact assessment methods (Bare et al., 2000b), and
related concepts and terminology are still being developed, but in general they include
three basic steps: step I, identification of the potential environmental concerns (i.e.
“impact categories”) affected by the LCI component results; step 2 classification, actual
assignment of LCI results under the identified impact categories (one LCI component
may affect more than one impact category); and step 3 characterization, calculation of the
contribution of the effect of LCI components to each identified environmental problem
(i.e. category indicators).

There are two main approaches for estimating category indicators as outlined in
Figure 3. A category indicator can be located at any place between the LCI results and

the environmental “endpoint” (Jolliet et al., 2004).

19

 

 

 

 

 

 

  

 

 

 

   
   
   
 
  

 

 

 

 

   
   
   
        
    
   
 
  

 

 

 

  
 

   

 

 

 
 

 

 

 

 

 

 

 

 

Mid-point End-point
categories categories i\
l Human toxicity
, \ __/
, asualties 'g.\
I\
\\
oise Y‘s-5 Human
\i‘I health
Photochem. A
oxidant \jl
formation
1 . ...“ Biotic &
Ozone depletion 1’7 abiotic
‘31,. natural
, Climate change ‘\ environment Scores
l ‘0
/ Acidification ‘t‘
Inventory 4
N . .
results § Eutrophtcatton ‘
\
\ Ecotoxicity . _
\:~ Biottc &
§~ . .
“Land use impact A \ abtotic
\ ‘ / natural
Species & organis ti resources
dispersal v' I
Abiotic ‘\ Biotic &
. i '
resources i :3 (1)]:de
depletion .
(minerals A envrronment
. energy,
freshwater)
iotic resource
depletion
Legend:
—> Quantitative model End-point process

available

—--->
available ( 1.6. unmeasured emnssrons.

unconsidered types of releases. etc.)

a

No quantitative or only qualitative relationship ‘

path
Mid-point process
path

 

 

Figure 3. General representation of main differences between mid-point and end-point
methods for impact assessment (based on Bare et al (1999) and Jolliet et a1 (2004))

20

One approach conforms to the so-called “mid-point” methods, which link the
inventory results to environmental mid-point categories (e.g. ozone depletion or
acidification or global warming potential). The term “midpoint” indicates that this point
is located on the impact pathway at an intermediate location between the LCI results and
the final environmental damage (i.e. endpoint). The alternate approach constitutes “end-
point” or damage-oriented methods, which link the inventory results all the way to
damages (e. g. damage to human health or animal species endangerment). In doing so, an
additional step may be used to allocate the previous “midpoint categories” into one or
more “damage categories” (Jolliet et al., 2004).

After the characterization step, though not required by ISO standards, additional
procedures are usually followed to better organize the results under some type of rating to
facilitate decision-making. In particular. normalization, valuation methods and/or
weighting procedures of the resulting impact outputs can be used in order to convert
characterization results into “impact scores” with the intention of facilitating the
decision-making process.

Interpretation: where the results from the LCI, alone or combined with those
from the LCIA, are integrated to reach conclusions and recommendations. This stage may
involve the iterative revision of the goals and scope of the LCA, and assessment of the

quality of the data.
V. Limitations

There is almost unanimous opinion about the enormous potential of LCA.

However, the current LCA technique, though gaining international acceptance, is far

21

from perfect. One thought that probably can sum up the skepticism is by M. Densie
(Elkington and Hailes, I993):

“The outcome of the LCA is the result of the inputs. The inputs are the result of the
preferences of those who are paying for the study

Further, since there is not a single, harmonized and standardized approach of
performing an LCA (e.g. due to differences in allocation rules, data collection
procedures, etc.), any result can be challenged. This is critical when LCA is used as a
product comparison tool to determine product preferability. In fact, Finnveden (2000)
argued that as long as no general framework is used in LCA, none of the studies can be
used to show an overall environmental preference for any of the alternatives compared.
The outlook is considered better when LCA is used as a tool for improving a system’s
environmental performance, since under the same approach on the same system, LCA
could give useful information for strategic system improvement (e.g. for selecting the
container size that appears to require the least life-cycle energy after analyzing the effect
of using different container sizes to deliver a product to the consumer).

Regarding the LCA methodology, shortcomings along the different LCA stages
were classified by Huijbregts (1998a) who elaborated that LCA limitations arise mainly
due to different kinds of uncertainties and variability.

In industrial systems, uncertainty (i.e. the lack of sureness about something) and
variability (i.e. the inherent variation of measured values) are responsible for many of the

problems in an LCA (see Table 1).

22

Table 1. Critical issues at different stages of the LCA process (Based on Huijbregts
( 1 998a) and Bjorklund (2002))

 

 

 

 

 

 

 

 

 

 

 

 

Problem LCA phase
Goal and Inventory Impact assessment (LCIA)
scope (LCD Choice of Classifica Characteriz Weighting
impact tion ation
categories
Data Inaccurate or Uncertainty Inaccurate
uncertainty no emission in lifetimes normalizati
measurements of on data
substances
Model Linear instead Impact Contributi Characteriz weighting
Uncertainty of non-linear categories on of ation criteria are
modeling are not impact factors are not
known category not known operational
is not
known
Uncertainty Choice of Choice of Leaving out Using Using
due to choices functional allocation known several several
(i.e. scenarios) unit, methods. impacts characteriz weighting
system technology categories ation methods
boundaries level methods
within one
category
Temporal Differences in Change of Change of
variability yearly temperature social
emission over time preferences
inventories over time
Spatial Regional Regional Regional
variability differences in differences differences
emissions in in distance
inventories environmen to
tal (political)
sensitivity targets
Variability Differences in Differences Differences
between emissions in human in
objects/source between characterist individual
s factories ics preferences
which , when
produce the using the
same product panel
method
Mistakes Any Any Any Any Any Any
Estimation of Estimation of Differences Estimation
uncertainty uncertainty in in human of
inventory characterist uncertainty
parameters ics in potential
impacts

 

 

 

 

 

 

 

 

23

 

Uncertainty is important because when it is not evaluated, there is more risk that
the impact predicted by the LCA will not match the actual environmental impact.
Likewise, variability is important because it limits the precision of LCA results.

In the packaging field several of these limitations have been found to cause
problems. For instance, Oki and Sasaki (2000) describe how sometimes it is difficult to
take into account the packaging function as a basis for comparison. The authors compare
a gas-barrier multilayer container with a monolayer container and claim that the current
state of LCA would exclude the gas barrier function from the assessment. That would
make the monolayer material more desirable because it means less material consumption
and less processing energy, and lower environmental burdens. Though it is true that the
gas-barrier material involves more energy and costs more, this material is winning the
competition in the marketplace and its function reduces the transportation energy and
emissions during distribution by extending the sales period.

Another popular source of uncertainty in LCA occurs when the effect of
different “scenarios” such as for the “energy” used in the inventory stage is not discussed.
It is argued that site-specific energy production data may produce very different
conclusions than average or industrial world energy mixes, which are common scenarios
used when site-specific data is not available (Paine, 2002). For example, results may
differ when coal-generated electricity data is used for operations that occur in regions in
which electricity is produced mainly by hydroelectric power. Coal-generated energy is
considered “non-renewable” while hydroelectric power is produced from “renewable”

resources and thus appears to be more environmentally friendly.

24

But perhaps one of the most uncertain parts of any LCA is the impact assessment
stage. This is because of an array of reasons. For instance, as described earlier, there is no
unified approach for implementation of the impact assessment process. In fact, in ISO
words (ISO 14040, 1997): “There are no generally accepted methodologies for
consistently and accurately associating inventory data with specific potential
environmental impacts”. The two approaches described earlier have their advantages and
disadvantages. For instance, since they can aggregate categories under a common basis
(e.g. DALYs: Disability Adjusted Life Years), endpoint methods may be preferred over
midpoint methods for category weighting, but introduce more subjectivity and
uncertainty (e. g. model, scenario and/or parameter) in the assessment since the closer one
goes to the end-point categories (i.e. towards the right in Figure III), the more the models
are highly dependent on the user’s preferences (Bare et al., 2000a). Thus. though
reconciling efforts are underway within the two main approaches, currently there is no
consensus method (Jolliet et al., 2004) and comparison studies among these two
approaches are often complex to deveIOp and interpret (Dreyer et al., 2003; Pant et al.,
2004).

Further, though according to ISO the LCA goal seems straightforward: “LCA is a
technique for assessing the ‘environmental aspects’ and ‘potential impacts’ associated
with a product throughout its whole life cycle”, a methodology for studying the general
environmental aspects has not even agreed upon. While some consensus has emerged on
assessment methods to evaluate contributions to environmental impacts such as climate

change, stratospheric ozone depletion, photochemical oxidant formation. acidification

25

and eutrophication (Udo de Haes. 2002), the situation is not the same for impact
categories such as resource extraction, land use and human health.

LCIA is further challenged by the fact that impact assessment methods rely on
models, many of which are being developed or updated to account for current changes in
the environment itself. And this stage is further hindered because of the sheer number of
chemicals used today. In fact, it has been estimated that of the around 100,000 substances
presently used in the world, only about 5% of even the 2000 most used substances have
been screened for toxicity and fate (Tukker, 1999).

Lastly, several in the LCA community acknowledge that due to the enormous
amount of uncertainty involved, product LCIA methodologies based exclusively on
mathematical relations representing system flows from and into the environment have
important limitations. An alternative is the use of value-based methods, but these in turn
need to deal with the open issue of LCIA weighting (Bengtsson, 2001; Finnveden, 1997;
Udo de Haes, 2000). Tucker (1998) offered three alternatives to this LCLA conundrum:
use a “reductionalist approach” by reducing the LCIA scope to obtain a truly robust
method; acknowledge the subjectivity of the LCIA method and develop an indicator
system that reflects the views of an authoritative forum; or develop an LCIA method that
includes the views of different sectors of the society and thus yields results in a more
socially acceptable product evaluation.

ISO standards (ISO 14040, 1997; ISO 14042.2, 1998; ISO 14041, 1998; ISO
14043, 2000) recognize the limitations that arise from not effectively addressing
uncertainty and variability effectively and urge the practitioner to do so in the LCA study.

However, the ISO standards do not suggest any procedure to do so. Ross et al (Ross et

26

al., 2002) found that out of 30 LCA studies only 3 (10%) included a quantitative or
qualitative uncertainty analysis and conclude that the standards need to be revised to
ensure that LCA includes at least a qualitative discussion of uncertainty and variability.
Among attempts to provide a systematic analysis, Huijbregts (1998a; 1998b) and
Huijbregts et al (2001) proposed the first general framework for comprehensive
uncertainty and variability evaluation, adding the definition of key concepts in the LCA
context. But uncertainty analysis is something new in LCA (Bamthouse et al., 1998) and
thus a number of methodologies for uncertainty estimation have been proposed.
Currently. the most favored approach for uncertainty and variability analysis in LCA is
incorporating it into the life cycle inventory (LCI) stage. Computer simulation is
becoming the method of choice. A number of recent uncertainty and variability
methodologies have been proposed and incorporated into LCAs associated with different
industry sectors and almost exclusively used Monte Carlo simulation (Canter et al., 2002;
Ciroth et al., 2004: Contadini and Moore, 2003; Guyonnet et al., 1999; Huijbregts et al.,
2003; Maurice et al., 2000; McCleese and LaPuma, 2002; Sonnemann et al., 2003).
Others use fuzzy logic to perform a similar data evaluation (Guyonnet et al., 1999; Tan et
al., 2002). In the packaging sector, despite the increasing use of LCA for evaluating
alternatives, a few Monte Carlo uncertainty estimations have been published but with
limited background information (World largest PET LCA, 2004; Canter et al., 2002;
Kennedy et al., 1997; Kennedy et al., 1996). Moreover, although these studies have
shown that uncertainty and variability can be incorporated in LCA, the exact implications

for decision makers remain unclear.

27

VI. LCA enhancements

Enhancements, in the context of this explanation, refer to attempts and approaches
to overcome some of the previously stated LCA limitations in order to increase or
improve the LCA value, quality and ease of implementation. For the purpose of this
discussion, these approaches are grouped into four categories: a) data-related
improvements, b) streamlined LCA, c) input-output LCA, and d) economic analysis and
LCA. These areas are some to which increasing research has been devoted in recent
years, and it is expected that they will evolve from their current state into more

established methods.

a) Data-related improvements

Data related improvements comprise three major aspects that are not necessarily
separable: a) data collection, b) data availability, and c) data quality. Regarding data
collection, the current status of LCI databases around the world can be used as an
indicator of the situation. Worldwide, government, private organizations and research
institutes are currently either expanding existing inventories or developing new ones (see
Table 2).

Many LCI databases are sector oriented and some already follow ISO 14048
guidelines for data collection. For instance, the US. LCI database which is managed by
the Athena Institute and hosted by the US. National Renewable Energy Laboratory
(NREL) (Norris et al., 2003b) is a project to develop publicly available U.S.-based
inventories originally focused on the building sector but later expanded to other industry
sectors. However, this database has not been peer-reviewed yet and it is not as transparent

as it needs to be for clear utilization. Other LCI databases cover several industry sectors

28

and are usually accessible through commercial LCA-specific software developed by

consulting companies.

Table 2. Summary of LCI databases and managing organizations and their status
(updated from Norris and Notten) (2002)

 

Status

Managed by

Completed?

Planned or under development

 

National and Multi-govemment*

Italy, Switzerland (BUWAL 250),
Switzerland (Ecoinvent). SAEFL

Australia, Canada, Chinese
Taipei. Japan, Korea, Sweden
(SPINE), USA

 

 

Consultants and research Denmark (EDIP), Sweden (CPM), Austria, Denrmrk, France,

institutes** Ecobilan (DEAM). Germany, Sweden, Switzerland.
UK. USA

lndustrial*** [ISL EAA, FEFCO. APME and

PWMI, NiDI

 

Acadenric/Decentralized**** Belgium, China, Chile, Estonia.
Finland, India, Norway, The
Netherlands. Portugal, Poland,
South Africa. Spain. Vietnam,

Argentina. Malaysia. Thailand

 

 

 

 

 

i may be updated

* Coordinated effort to produce nationally representative and accessible database.
Typically involves collaboration between several organizations and some degree of
government funding.

**Inventories produced by research organizations or consultants and made publicly
available in a database, sometimes for a fee (e. g. databases included with LCA software).

***Inventories produced and published under the auspices of a particular industry
organization. Includes cases where data made only partially available (e.g. for a fee, or
only to parties with sufficient motivation for requesting the data). Most often data
compiled by consultants, but includes cases where LCI development is done in-house, or
by academic or other research organizations.

****Includes inventories compiled by academic or other research organizations, made
either partially or fully available on an ad-hoc basis (e.g. through journal publications).
Countries may have some degree of information sharing (e.g. an LCA society), but no
coordinated data gathering effort (i.e. studies are not organized into an accessible
database).

29

Regarding data availability, attempts have been made to improve the situation.
For instance, the SETAC-Europe LCA Working Group on “Data Availability and Data
Quality” was formed in 1998, with the goal to focus on the key features of improving the
efficiency and quality of data collection (Huijbregts et al., 2001). Similarly, NREL in a
joint project with the US. EPA has developed a global LCI inventory matrix to help
centralize multiple-sector LCI databases. The information is obtained by voluntary
submission and is categorized according to industry sector and life cycle stage. However,
many of the specific LCI databases are accessible only by a fee (US. EPA, Global LCI
matrix, 2002).

Data quality has been the subject of many publications in the LCA literature
(Braam et al., 2001; Finnveden and Lindfors, 1998; Huijbregts et al., 200]; Rousseaux et
al., 2001). Quality, often defined as “fitness for use”, has been used to comprise aspects
of the LCA data such as database format, uncertainty, reliability, completeness, age,
geographical location and process technology. Thus, approaches to improve quality
involve standardization in data collection and processing procedures; data validation by
cross-checking of energy and mass balances; use of data quality goals (DQG) and data
quality indicators (DQI): use of parameter estimation techniques; use of higher resolution
models; critical review; and additional measurements (Bjorklund, 2002). Each of these
approaches has its own disadvantages. For instance. since the development of
standardized databases requires consensus among LCA practitioners, it is time and
resource demanding. Likewise, making additional measurements of inventory data or

using higher resolution models to obtain better estimates are time and resource intensive.

30

DQGs and Dle are simpler and more flexible approaches but there is no
consensus about their methodologies. In general, DQG is a qualitative scheme to specify
the data quality requirements before actual data compilation. DQI can be either a
qualitative, quantitative or semi-quantitative technique to assess the quality of already
compiled data.

Alternatively, though limited in scope, sensitivity analysis, which is the study of
the effect of changes in an independent variable on the LCA outcome, gives more
knowledge about the behavior of the model. However, it is time-consuming when not
coupled with some kind of data uncertainty importance analysis. In fact, data uncertainty
importance analysis is a useful screening methodology that can help concentration on
relevant model parameters by establishing some kind of contribution criteria to the
overall LCA uncertainty outcome. However. as stated earlier, uncertainty analysis is still
a new concept in LCA, and the information in databases required for the analysis (e.g.
ranges, standard deviation, probability distributions) is very limited. In fact, publications
considering the evaluation of data ranges are scarce. Among the few, Finnveden
(Finnveden and Lindfors, 1998) analyzed the range of the common inventory data in a
number of databases and his findings are shown in Table 3. The variation in this table

indicates the typical uncertainty of European databases in the mid 19905.

31

Table 3. Expected variations in LCI data (Finnveden and Lindfors, 1998)

 

 

 

Inventory parameter Variation that can be expected
Central. non-substitutable resources Factor of 2.
Less central and substitutable resources A factor of 10 or more if they are completely
substitutable.

 

Outflows that are calculated from inflows (e.g. The same as for the corresponding inflow.
C02)

 

Other energy related air emissions Factor of 10.

 

Other process-specific emissions Factor 10-100 or higher if mistakes or very
different types of technologies can appear.

 

 

Total amount of solid waste. Factor of 10.

Specific types of solid waste The variation can be very high partly due to
different classification systems in different
countries.

 

 

 

 

b) Streamlined LCA

In practical terms, it can be expensive and time consuming to conduct an LCA
according to ISO or SETAC guidelines. Streamlined or abridged LCAs are approaches
that are used to obtain timelier and less expensive results. According to SETAC, there are
two main kinds of streamlining: approaches within the LCA framework and alternative
life-cycle approaches as shown in Figure 4.

Approaches within the LCA framework can be again subdivided into two: process
oriented and methodology oriented. Process oriented streamlining methods deal with the
actual operation of performing the LCA (e.g. making software with embedded ready-to-
use databases, using process templates, etc.). Methodology oriented streamlining methods
deal with simplifying the actual stages of the LCA. By limiting the goal and scope of the

study, the LCI and LCIA steps can be simplified. Table 4 contains a list of the common

32

streamlining decisions that have been used in the past, along with their advantages and
disadvantages. Since all LCAs are streamlined to some extent, the degree of information

“lost” by these techniques cannot be fully accounted for.

33

menacing mamE—Ewebm .v ocswfi

9336.508 .8. 3:333:00
A InH—rm< .l mmOOOHQ \ASUHNHO—r— Gosh—Dx/OHQ COGS—Tum W—utfln—EOO ~05 umha OH meHOHCO>Cm
23% we 22328.. .5 3328.50 .we 33 maze: B we 64

 

 

_ Amos—cocoa mEEcoE 93.38% 32-8-»nt

 

 

 

 

 

 

4
Amowfim coazwfia 55 £wa .33 use 038 325980 FEB BEES 3
205-8: mxm .wov ..sowoamw Havana... wage: .wov awe—ouofioz mains. .wov $805
monomoamm 293-8: @3858? €02,253 <0..— 05 5:55
P _
4

3:08.93 WEE—Emobm

 

0:00: 00:00:.
00 00000000300 05 00... 0000 05

0000000000

02020000 000 0003000 0000

 

00 000 0300 000000000 0003-020000 0000000 00 000 0000002000000 300030 0203 02.000 0050 ”000000000 030003000 000 000000.:

005 0002000 00 000 0000000220000 0003 300—2: .0000 05 .3 00800000: 000000 00000 00 000.00 02003000 0000:: «N .00 0000 00 000000000 00 020000

000000000300 00000000: 00000 000000020000 000000000300 00 0000000 000000 0000080000 00 0000 000 020000 000020m 050000 m000D
0003000 ..000w-00-000w.. 0

.0000000 00 .0000: 00 00 00 00 00000000 000000008 00000000000000 0: 0000 00000000000

000000 30:: 00.000: .0000 .3 00000000 3000000 00 000 0.000 000 000000000 0000000000000 .30 00 :03 0002 00000000300

00 000000000 00 203 00: 00000<

.3000 05 0200: 000000000 000 0000:00w 0000

0002000 00 0020200000 000000000 000.0000 .200

000 -00 $030.00:

 

 

$522000
.0700 00000000 :00: 20.200090 .0E00w000 000000000000 030000000300 0000000.

$0205 0.00 00000000 00000 00 000: 00 000 00—0000 0000000000.: 0000000 0000000500
000 030 000 000 .20me0 000 .203 00 .0000: 203 00: 05 0:23 00000000 02020000 0000 00000000300
00: 05 E 00300 008000;: 000003 0000000000300 0000000000 00000 000.0000 000000000000 3000000 :00: 0000 000000000 =< w00>0E0m

0.2000000 300 0000000 00 0000 0.000 00 000000000

0000000000000 0000 0000000000 00200020 300
025.00— .00300000000 00000008 :00: 30000000 00000000000
.00.. 0000 05 .00 000000000 05 503 0002008 000 00000000:

0—0000 003020 0 2000000
000000020000 00000 000000006

0020000000: 30000000 .00: 00 0200 000000000 :<

w00>00000 3.20000

 

.QSww_ 360 006:?)

000000200000 00003

 

 

 

 

 

00 300 00000000 00 000000000 093200000 000000005 000220000 000 0000000 0000000000000 000 .00: 000000000 000000 0000000800
30000000 300 00 000000000000 05 0000000000 00 002000: 02000000 000000000 000000: 0000 000.000.0000 000200. 600280 000 000000000
0000000000300 000000000: 000 00000000 05 =< .00 020000000 0020 0030000008 00000000 :00: 00 0000 000000000 :< w00>0E0m
00000050
0003000 003000250 003000000 0053:0000 gnu—000005

 

 

2mg 3 _0 00 000$ 000 .800 3 000000 000 000B 00 00008 0000000000 3000—0000000 00000200 0320000002 .v 2000.

5
3

 

08.00.0006

00 008 00000.00 0000000000300 000000000.
5.0.02 00 000000 w0.000w000.0

000 0E0.0> 00 0.00 m0.00000 0m.

00.0 0000000. 00

0>00 000 0000 000 0.0020 00.000 0. 000m.
.0000 00000 0000000 000 000 00000090.
0008 00 00 0.00... 000 0000 00000 00
0000000 000 0800. 00 000500 000 0000.4

.0000 0000 00.0 000 0.00..

000m < 0000.0000 000 00000 5‘. 000 00 00000
00 000. 0000 000. 00.000080 0.000000 30m

0.0000000
300 w0.0.00.‘.

 

0.0000 000

00000 0000000 00 00.00000 .0000000 000
0000000000 0.00 000w00000 0000 000000
0. 000000 00000000 00 00000 0000000006

.0.00..0>000 00 00.300000 0.003
0000 0.00 000 0000.0>00 00 000 00008.00m

0000.000

00.00 00000000 000 0. 5.000000 .00000000 00
40200000 000.0000 00 00000 000000000 00:00.0.
000000000 00.3 0000.000 00 000000000 00000.0m

0000 0000000
000w00000 w0.0D

 

00000000

00000 00 000.000.0000 0. 000 20.00 00..
..000>0 0. 0000000 .0000E000.>00 0000
00 000000006. w0.000000 0. 5.000.005

00000.0000

00 000 ..00000. 000.000 03.003.005.000
000000000000 00 20000000

....0000 000 0000000 000000000300 080m
.205 00.. 000 00 00000 0000 00 00000000
000 00000. .0000E000.>00 .0000000 ..<

50.000000

0000.0 00 00.30050 00 03.00.0000 00 000 0.00
00003 00000 000 00 0000.05 000 000.0> 00000
”0000 000 30.000000 0.000.000 000 000000000000
0.0.03 .000 00.000000 00000000008 0000000

000 00.000000 00000.00 000000000000 00.000000
0020000020. 00.00.0000 20.000000 300. 000000
0000000 0 0.0 0000 0.0003 000.0> 00000000 0.00

0000 00000000 000.
00 000.000.0000 00.03

 

00.000

00.000 00 0005000300 000 00.. 0000

000 00 000 0.000 000.0.000 00000000000
0000 0000.000 00 000 000000000000
.000000000300 000000000. 00000

0000000000 ..0 ..0 00000.00
000.0000 00 0000 02 .0000 000 00 00000000.
020000 00000. 00.0000 00 0000000.

022000
20.00008. 00 0. 0.0000 000 .0030 000 w0.000
00000000000 0. .0000 0000000000 000 0.02.00

20.02.00 008.0000...
00 ..000000003000
.. 0.0.00

 

.0000.
00000 0000000 00 00.00.0000 0000000 00.
0000000000 0.0.: 000m00000 0000 000000

00.000000.
..00. 0003.. .0000000 0.00.0.6 00 .0000: 00

0000.000 000
00.000 000.0 ”00.. 00000000 000 00.0000 .01. :00
00.3 0.020 20.00000 0000 .01. 000 3.0000000

.01. 000000000 0.

 

00 000000 02.00.0000 00 000E 0000w0000m 00E 00.0000 0000003. 000000.000 0000200 000000000 .000.>.00. 000 6000 00.0000 00.000m 00.0000 00.0000 000.03
00000000
000.0000 0000000300. 000000000 00.002.00.00 00.5.2085

 

 

 

 

 

 

.0008. 0000: 0.. 00 0.5: 0.0 .800: 500:0 0.0 0000. so 3030 02008000.. 00000500000 035.00 00282002 .0 2.00

Alternative life-cycle approaches do not involve a complete inventory analysis
and do not follow the inventory/impact/improvement analysis path required by ISO.
Instead, they attempt to evaluate relative differences among alternatives along their life
cycles. One of the best examples of this approach is the one developed in 1993 by
Graedel and Allenby at AT&T (2003) called environmental responsibility product
assessment (ERPA) matrix. The ERPA method divides the product life cycle into 5
stages: pre-manufacture, product manufacture, product delivery, use, and recycling or
disposal; and considers 5 environmental concerns: material choice, energy use, solid,
liquid and gaseous residues. These two dimensions are presented in a matrix format in

Table 5.

Table 5. Environmentally responsible product assessment matrix (Graedel and Allenby,

 

 

 

 

 

 

 

 

2003)
Life cycle stage Environmental concern
Material Energy use Solid Liquid Gaseous
choice residues residues residues

Premanufacture (i.e., (1,1) (1,2) (1,3) (1,4) (1,5)
raw material
extraction/production)
Product manufacture (2,1 ) (2,2) (2,3) (2,4) (2,5)
Product delivery (3,1) (3,2) (3,3) (3,4) (3,5)
Product use (4,1) (4,2) (4,3) (4,4) (4,5)
Recycling, disposal (5,1) (5,2) (5,3) (5,4) (5,5)

 

 

 

 

 

 

 

As indicated in Table 5, each cell of the resulting 5x5 matrix is then assigned a
score ranging from 0 (highest impact of a stage on an environmental concern item) to 4

(lowest impact of a stage on an environmental concern item). By this scoring technique,

37

the method estimates the results of more formal LCI and LCIA, and also takes into
account whether the possibilities of reducing impacts have been utilized or not
(Hochschomer and Finnveden, 2003). The scores are assigned by consideration of
information from actual life cycle studies. checklists, manufacturing surveys and
experience. The ratings in a matrix can be added up (i.e. to sum up to a maximum of 100)
or can be plotted in a circumference target plot for more convenient evaluation (e.g.
center or zero represents highest impact. circumference represents lowest impact). This
technique, which critics say is subjective, was found useful for identifying hot spots and
opportunities for environmental improvement and has been used recently for a number of
different product categories and applications range from re-refined oil in Japan
(Nakaniwa and Graedel, 2002), to evaluating the environmental impact of a residential
refrigeration unit with a proposed maintenance and take back service (Bennet and

Graedel, 2000).

c) Input-Output LCA

In conventional (SETAC/ISO) LCAs, the system boundary is usually chosen with
the assumption that addition of successive upstream production stages has a small effect
on the total inventory. However, truncation errors inherent to conventional LCAs have
been estimated and in cases can be of the order of 50% (Lenzen, 2001). One way to
address the boundary issue in LCA is by using economic input—output methods.
Economic input-output LCA (BIO-LCA) is the result of applying economic input-output
(E10) analysis to help perform a life cycle assessment. The BIG analysis is based on
using the E10 matrices that are regularly estimated for most developed countries and

economies. An [310 matrix is a transaction matrix that shows the relationship between the

38

different sectors that form part of an economy. Table 6 shows the basic structure of such
a matrix in which entries are expressed in dollars. For example, an is the amount of
dollars required to (directly and indirectly) input in sector 1 (e.g. electricity) to obtain $1
worth of sector 2 (e.g. aluminum sheeting) output. For the U.S., the BIO matrix has 519

sectors, so it is a 519x519 matrix.

Table 6. Basic structure of an ElO matrix

 

 

 

 

 

Output
Sector 1 Sector 2 . . . Sector 11
Input
Sector 1 all a12 a1n
Sector 2 a2] a22 a2n
Sector n anl an2 ann

 

 

 

 

 

 

 

 

The basic approach for ElO—LCA then can be compactly summarized using

matrix algebra notation in equations 1 and 2 as described by Lave and colleagues (1998;

199SX
X=(l-D)"F (1)
B=RX a)

In equation 1, X is a vector containing the total output (in dollars) from different
sectors of the economy required to meet a desired final demand, I is a identity matrix (i.e.
to include the output of the aluminum sheeting sector itself), D is the BIO matrix, and F is
a vector representing the desired final demand (e.g. dollars worth of a desired amount of
aluminum sheets). In equation 2, B is the vector containing the economywide
environmental burdens (e.g. toxic emissions or electricity use), and R is a matrix
representing the environmental burden per dollar output of each sector.

By setting the boundary of the LCA on the level of the national economy, EIO-
LCA with the BIO matrix attempts to address the boundary issue including the
interdependence of different processes. However, this analysis has its own problems
including the high level of aggregation (i.e. combination of product and technology
information) in industry or commodity classifications, which limits the level of detail of
EIO-LCA studies. Moreover, there is incompleteness of sector-based environmental
statistics, which in turn limits the accuracy of the EIO-LCA results.

Due to the previous reasons, conventional LCA is often seen as more detail
oriented. However, these analyses are more labor- and time-intensive, and suffer from the
stated truncation error (i.e. due to omission of contributions outside its finite boundary).

In fact, due to the quick and inexpensive nature of the BIO-LCA approach developed by

40

researchers at Carnegie Mellon University, Matthews and Lave (2003) suggested the use
of this system to help in corporate benchmarking efforts to evaluate the environmental
performance of their operations. Moreover, using input-output techniques, Suh (2001) a
researcher at Leiden University has recently developed the MIET, an inventory
estimation tool for missing flows.

“Hybrid” analyses combine process-level data with sector-level input-output
analysis and thus try to get the best of both approaches. According to Suh et a1 (2004),
hybrid approaches can be grouped into three different categories, which are tiered hybrid
analysis, input-output based hybrid analysis, and integrated hybrid analysis. Table 7
summarizes the main aspects of each approach along with their perceived advantages and

disadvantages.

41

 

 

 

 

 

.209 30:20:“ .0028 05 w5mmob v.30: swag:
3&0 9 >03 wave—.50 €009? 889598.035 05 Ba 308-0885 05 85:
03:00 863‘ .0505 05 ~03: 55. .055 D5085 .3 0050852 5 529$ $98.35 503050
03305 0.053 05 .8 58388.: 05 0:53 8008a :80 .8 0:5 5:808 5:: ha 85: 30593 an 352
.850 was .95... .00: 9 5.3.50 x505 E05200 x505 Ame—05.02 a 5 005000.50; 5 50?? 30.3.0885 05. 0353::
.0895 :8: 509.300 xiwfi
5 b.5800 35:0: .5 80505
0050 5.3 854:8 0n Esonm
£8 02.08-35 08 80%
0.5.0065 use 00: c0032 A032
$53000 $3 925 9:58.. .8630 0E003-mu0oo.a .«o _0>0_ 05 3 cows—82 a :08. 50:05
as, 8:2 Baa £52» .0558 .0380 862 3: 85020035255; 0a 9% 58.: 328m 358 p5: Bag
000 0093 0:50.20 5a 05 dong E02800 055 080 5 anEwwwwE 35.3 05 5908 598.255 532 598-255
.05.: 0:0 con—5 .580 .6 2:0. 5 858.950 Ea $0888 .05
5% Bean éafifia 3% E ”2&8 :8 mafia 598
-59: 9.8 0809a 05 803.02 ban—Boo 05 .6 5:553:00
Ha 20:82 .005.“ .5935 535.035 3 8.960
05 35:08:. 35 ..o mctauflsqfl: can 850550 $.53: ..w0v
2:8 5ng8 0:095 050800502 0005 8&2 mugged 5ng 3085 85500
303-3085 5 A0032 $5580.. com: 555596 a 5 35.550 08 0—98 0S .0555 05 00 82050530.. E858:
$3 35: mag :53 $068 9 .300: 0020 030. 50:09: 058 93 32.00-30 28 00:32:35 5250
00:02 wasp—:00 03:50 :53 V0302 09 ~82 .00: o. ism .00: 50.20.5080 ..wuv 050505102 £32238 use 6050 55? 020$.
0m3§>ummfi 0wS§>c< 85058036 500803

 

 

 

 

 

2,303 B 00 saw so U038 533:0 00me 5920-035 98 wax—0:0 802-3085 $559500 m0zomoam0 2.53 502 .b 050%

42

The connection between process-based and input-output-based LCA is a topic
under development (Heijungs and Suh, 2002) and thus much research work still needs to
be done to define this relationship. Nevertheless, recent assessment studies have already
benefited from the information provided by these hybrid analyses. For example, by using
a hybrid approach, Norris et al (2003a) were able to estimate the energy consumption
during the factory-to-mall phase of life cycles that has been universally neglected in
process-style LCAs; and Nakamura and Kondo (Nakamura and Kondo, 2002) have
deveIOped a hybrid approach that expanded the input-output system to include waste

flows and showed that the E10 model is in fact a special case of their model.

(1) Economic analysis and LCA

While LCA can be useful for evaluating environmental attributes of a system, it is
often criticized for not providing monetary information that business managers routinely
need to allocate the often scarce capital resources available to minimize the
environmental footprint of their business operations. Thus, various approaches have been
developed to supplement environmental information with cost information and enhance
the decision-making process. The central challenge is estimating of the “environmental
cost” of business operations and a whole body of concepts and terms has been developed
under the umbrella of “environmental accounting” to address this issue (US. EPA,
Introduction to environmental accounting, 1995; Shapiro, 2001 ).

The scope of the present discussion comprises life cycle based approaches to
estimate environmental costs of products. One of these approaches is life cycle costing
(LCC). LCC is a systematic procedure for identifying environmental consequences along

the life cycle of a product (i.e. product line, process. system or facility), and assigning

measures of monetary value to those consequences using accounting procedures. This
process includes the assessment of material flows (e.g. amount of solid waste generated)
through the product system (i.e. materials accounting, essentially a kind of LCI) as well
as costs (i.e. cost accounting). including environmental costs (e. g. waste disposal).
Despite the apparent compatibility of approaches, LCA and LCC have important
methodological differences. For example, while LCA attempts to evaluate the relative
environmental impact (from a broad societal perspective) of alternative product systems
that perform the same function, LCC intends to estimate the relative cost effectiveness of
alternative investments and business decisions, very often from a private perspective.
Thus. LCA and LCC actually consider life cycles with different spans and flows (i.e.
physical or energy units vs. monetary units) that are not necessarily compatible. In a
succinct table (Table 8), Norris (2001a; 2001b) summarizes the extent of the differences
between life cycle assessment and life cycle costing methodologies. Furthermore, the
LCC outcome is limited when used at the product design stage (e.g. Design for the
Environment programs), since it suffers from greater uncertainty than LCA (Schmidt,
2003). This is because future technological changes have a strong effect on the results,
and because of specific additional factors (e.g. interest rate and market dynamics) that are

not always stable and are independent from technology changes.

Table 8. Differences between LCA and LCC (Norris, 2001b)

 

 

 

supporting usage phase; entire
physical usage

Tools LCA LCC

Items

Objective Compare relative environmental Determine cost effectiveness of
performance of alternative alternative investment and business
product systems for meeting the decisions, from the perspective of
same end—use function. from a an economic decision maker such
broad, societal perspective as manufacturing firm or a

consumer
Scope of life cycle Supply chain of processes Activities directly causing costs or

benefits to the decision maker
during the economic life of the
investment as a result of the
investment

 

Flows considered

Pollutants. resources, and
interprocess flows of materials
and energy

Direct costs and benefits to
decision maker

 

Units for tracking
flows

Physical and energy units

Monetary units (e. g. dollars)

 

Time treatment and
scope

 

 

Timing ignored; all causally
linked flows, and some of their
impacts collapsed in time and
valued equally regardless of
timing

 

Timing is critical; present valuing
(discounting) of costs and benefits;
specific time-horizon sc0pe,
outside of which costs and benefits
are ignored.

 

However, by offering direct opportunities for cost reduction, LCC is perceived to

help to promote life cycle based analysis. In turn, economic analysis with a life-cycle

perspective has the potential of discovering “hidden” costs (see Table 9, cost types 2, 3, 4

and 5) and revenue impacts that are otherwise neglected in conventional economic

analyses. Very often, though, LCC users utilize the pragmatic approach of focusing

exclusively on internal or “private costs” (i.e. type 1 and some type 2). Just recently,

some comprehensive approaches have been developed to bridge the gap between LCA

and LCC and to improve it to allow for easier identification of hidden costs. For instance,

Total Cost Assessment (Total cost assessment methodology. 1999), method developed by

45

 

a joint effort of private companies and the American Institute of Chemical Engineers’
Center for Waste Reduction Technologies, is an approach that facilitates the inclusion of
environmental costs into a capital budgeting analysis by classifying costs into categories

shown in Table 9.

Table 9. Categories of costs according to AICHE/CWRT (Center for Waste Reduction
Technologies-AIChe, Total cost assessment methodology, 1999)

 

Cost Type Description

 

Type 1: direct Direct costs of capital investment, labor, raw material, and waste
disposal. May include both recurring and nonrecum'ng costs. Includes
both capital and operations and maintenance (0&M) costs.

 

 

Type 2: Indirect Indirect costs not allocated to the product or process (overhead). May
include both recurring and nonrecurring costs. Includes both capital and
0&M costs.

Type 3: Contingent Contingent costs such as fines and penalties, costs of forced cleanup.

personal injury liabilities. and property damage liabilities.

 

Type 4: Intangible Difficult to measure costs, including consumer acceptance, customer
loyalty, worker morale. union relations, worker wellness, corporate
image. and community relations.

 

Type 5: External Costs borne by parties other than the company (e.g., society).

 

 

 

 

Several alternative versions of life cycle costing methodologies have also been
developed, mostly by interested companies. and confusion still exists about the concepts,
scope and terminology involved (Total cost assessment methodology, 1999; Sonnemann

et al., 2001).

VII. LCA and packaging
Packaging situations were one of the earliest applications of LCA. Harry E.

Teasley, Jr., manager of the packaging operations at the Coca Cola Company was

46

credited for first devising the analytical scheme of quantification of material. energy and
the environmental burdens of a package over its complete life cycle from raw material to
disposal in 1969 (Hunt and Franklin, 1996). About three decades ago, public concern
over increasing volumes of solid waste due to the use of plastic in packaging and later
concerns about energy consumption became the major driving forces to study the effects
of packaging on the environment (Sonneveld. 2000). A representative study is the
comprehensive energy analysis of production and use of packaging systems published by
Boustead and Hancock (1981). Eventually, these studies evolved into the comprehensive
tool that LCA is today and example studies are those such as that published by the Swiss
Agency for the Environment, Forests and Landscape (SAEFL) (LCI for packaging Vol. I,
1998a; LCI for packagings. Vol. II., 1998b) and the series of “eco-profiles” of plastic
resins and intermediates. conversion processes, and packaging published by the
Association of Plastics Manufacturers in Europe (APME) (2002; 2003).

Applications of LCA in packaging situations can be divided into two main
categories depending on which constituencies use it, namely: a) stakeholder, and b) third-

party organizations.

0) Stakeholder use ofLCA

A stakeholder, is any interested (often private) body that might use LCA, or some
form of LCA, in decision making regarding product design, product improvement,
product comparison. strategic planning. compliance with regulatory policy, marketing.
academic purposes. etc.

Recent examples of the use of LCA for product development and improvement

purposes can be drawn from several industries. Most of them resulted from corporate

47

environmental stewardship programs that generally involve proactive premises such as
design-for—the environment (DfE). DfE intends to integrate health and environmental
considerations into business decisions and it has been applied in Europe as well as in the
US. While in Europe most of these DfE programs are voluntary and often internally
adopted by companies that want to comply with the strict disposal regulations in place, in
the US. DfE is mostly known as a voluntary partnership between the US. Environmental
Protection Agency and industries to attempt to help with pollution prevention (Hart et al.,
1995). DfE principles, along with integrating health risks, aim to use “accepted” results
from LCA studies to create the attractive but unproven concept of an “environmental
preference ranking” or “environmental indices” for the selection of materials to be used
when designing new products (Industrial Designers Society of America. EPA partners
with IDSA. 2005; Toloken, 2004). The Cleaner Technology Substitute Assessment
(CTSA) methodology developed by the EPA, which involves comparative evaluation of
substitute technologies. processes. products or materials, regarding human health,
environmental risk. performance, cost and resource conservation, is another tool (along
with LCA) that is used under the DfE premise (Kinkaid et al., 1996) to try to help with
process selection. Many DfE programs are internally developed by companies that often
claim substantial economic benefits after implementing DfE-recommended
improvements. For instance, Xerox Europe using DfE under its “waste-free initiative”.
reported utility savings by pushing towards the reuse and recycling of equipment
components through appropriate labeling and improving disassembly, as well as the reuse
of packaging components by reducing the number of pallet styles and boxes used for new

equipment. and by switching from conventional single-use corrugated boxes to wooden

48

and steel totes for the collection of used equipment (Maslennikova and Foley, 2000).
Likewise, U.S. examples exist on the use of LCA for product improvement purposes by
studying packaging options. For instance, an LCA conducted to evaluate the
environmental performance of the yogurt product delivery system used by Stonyfield
Farm Inc. (Keoleian, 2001; Keoleian et al., 2004) analyzing different packaging formats
(i.e. 4, 6, 8 and 32 oz polyprOpylene cups and 2 oz linear low-density polyethylene),
estimated that the greatest potential improvements were the redesigning of the primary
packaging and the use of alternative manufacturing techniques for the yogurt cups. The
study indicated that in this case, shifting from injection molding to thermoforrning of 32
oz containers reduced the life cycle energy by 18.6% and solid waste by 19.5%, primarily
due to light-weighting. The authors claimed that elimination of overcaps for 6 oz and 8 oz
containers provided similar advantages. and indicated that the effect of container size was
significant when it was found that delivering yogurt in 32 oz instead of 6 oz containers
could save 14.5% of the life cycle energy and decrease solid waste by 27.2%.

Along with product improvement applications, partnership of industry with
research institutions using packaging related “case studies” attempted to evaluate the
environmental aspects of a number of industry Operations (Bide, 2002; Keoleian, 2001;
Keoleian et al., 2004; Keoleian and Spitzley, 1999; Ross and Evans, 2002) as well as the
LCA methodology itself. For instance, researchers from the University of Melbourne
used a case study involving the utilization of different packaging formats by an
Australian-based maker of refrigerators, to estimate the effects of excluding and
including site-specific data (Ross and Evans, 2002). By limiting their analysis to a single

non-global cumulative impact category such as the presence of significant photochemical

49

precursors in the atmosphere. they reported the ability to assess whether an improvement
in protective packaging produced any noticeable change in this impact category under
two scenarios (i.e. when aggregated into a single global parameter or when spatial and
temporal factors were taken into account).

LCA has been often used for product comparison purposes as well. but due to its
nature, the assessment shows the environmental implications of different choices and the
trade-offs that need to be made, instead of a clear answer. Nevertheless, oftentimes the
complexity in the interpretation of the results is overlooked in many LCA comparison
reports that, deliberately or not, portray one alternative as more environmentally sound
than the other.

A survey of the literature in order to attempt an analysis of the use of LCA in
packaging comparisons was presented by Martino (2005) and is shown in Table 10. The
table presents a list of some packaging oriented LCA and LCI involving comparison
studies which have been released or published in peer-reviewed journals. summary
reports or books, in the last fifteen years. It can be seen from the Table that, regardless of
the packaging formats evaluated in the studies, their scope is relative consistent
comprising processes from raw material extraction to. in most cases. disposal. Likewise.
in most cases, the functional units have been selected comprising the containment of the
product and in some instances including its delivery to the consumer.

Consistency is also noticed with regards to the inventory parameters/impact
assessment indicators since all of them include energy used (though many don’t indicate

whether it is a gross or net value), and some include warming emissions. In some

instances the impact assessment indicators belong to a pre-established set based on a
particular method (e. g. SimaPro).

Not surprisingly, the consistency diminishes when analyzing, the data sources and
the types of scenarios considered, since they are directly related to the systems studied
and the selected functional units. In general, in only one instance an uncertainty analysis
has been included and its results included in the conclusion. It can also be noticed that
when such studies are used for marketing purposes, their conclusions seem less qualified

and more absolute and there is no critical review included.

51

 

3:83

 

 

 

 

 

 

 

 

 

 

 

 

 

539... m. 5.3.050:—
éeaa :55. Ba .96..
8.33.5 3.35 :5. 5:80:65 Exec:
go: £23.35 .2236...“ .0... Rama... ...:::0.: .0... :53?
5.03 .m> 82 ounce...— 333 :5. cm: . 3...: wfinEg
:83 no.8... ...o :5. :5. 02.86 memo. w:_.3o&=:w:. 8.0.398 83-3%
acme—n: Noa— zwbfi. mm... 55.3 .3: 6:3 Enos. ..wEmmoooa 0.3:: 2.: a :5. :83
.w:E=m:oU om: @qom 5:65. :5. mac. :0 .mfitoae 38 083.8895 5.2% an: emo—
:o.oEE=Q Exces— =o .335. 5:33.:3 «\z ..o :26:th wEBxa—z $80.5 0:23.039: 8:5—
636%.:
.:o_mm::o .o: m.
2.8:: 2:8 ...... mfimfioaa
.0 5.38.8 Co $2.33.
on. 5.3 .A:o_.:.o:.o:_
:3. ...: .o 2.2%....» :5. 35%....—
m=o>» ”So. .255 :5. ...: .m:o.mm_:.o .0... 1.8%.:
.wExoozv :8. 82.0.5 .933 :5. 8:
So. .0...va $2.5... :5. .... .3: 6.88.358. :3.
£05.30 ...—«.53.: £030: .583 22.82%: 2.2:: 950.: 3w: 3.6 .38
653502 3:. :8. 3:08 .ana coo. a... 5.. as. has: 8958:: :5. an:
”3.03 255:8. d... 5...... “E: 96 x56 .0: 5353:. 3:. oznéeowan 0:25:30: ....E.
”mg—co: 8% 8.5 mm. 3.35: 33. ..o 35 25% 32 no 5.88.5 3:5 030 380.5 3.. 03:5 .0:
coho—:88
E3385 comm—«9:8 122—Eco
53:558.: 33:32.8 82::— ..£ 8.5% $3.32:
85.9%.. 335: a: £952.... 62023.8 :33sz :e:8=nn< 3.53%. 6:: 3:558
1:: 23> 185.5 Eafifizm Rhea—.3:— _:=e_~o:=.m meta—Sum £39m 5:2 <04 mafia—cum 838.5

 

 

36:; :OmtmmEoo w:_>_o>:_ 54 cc: <04 3.5:: wEwmxoam 8.8.8 ..o 65 .o. 033.

52

 

00:00.0
0. 5:05:55

 

 

 

 

 

Goo. 0.003 000. 00.9.00. 5:
050.0... 05.00.50 0. .0... 3.0.0.:
053.00 0:5....0 0:0 5 0:00.00 0550.50
0.5.0.50 0:50.203 E5005 0.5.0 .50
.0 .009... 000— 50:00... 500.550.: .00:..o-m ..
3:055:50 Pm.— 0:0 $500.00.. 50:55.0
0:... moo. 55:55.0 b00500. 0:0
5.0.500: 00 $55 0. 05:03:00 50:00.: 3
550.50 :02: 00 0:0 .6050... 000% .5055
0:00... 5.. 50% .03 0.003 0:8 .0 m5=c 0:0 -0: Cm...
500.0800. 0:: :0... 0:23.55 5500.50 3.550.305 0.205.505.
.05002 0... 35:0 000. 0:585 0:0 .3 0000.55: $500000... 00000.00 05.2.0.5:
5. 00.0.0000... 00 00:50:00 0:35.03 5:0 .50 .5 0.0.5.0... 30. 0.0350000: ”0550.50 0.5.0
...—.50.". ...mE .359“.— 0:o=0w coo. .5 5.50.05 00000..— 0005005 com
0050.050
0.5.00.0... 50.59.50 00.09.50
53035.5 509.: 5.. 00.500 05.5008
00:05.0. 3030.. 0:35—0:00 0.00.5.5. 3.32:»... 53083.7. 3095:». 0:0 0.08.5.
0:0 50> 30th 5 5085.5 .m..3:0>:— 5:030:00— 0.50m <04 0:50:00.— 00:00...—

 

 

 

 

 

 

 

 

A0..:00v 00.00... 50:00:50 w:_>_o>:. 51. 0:0 <04 0050.5 06500.00: 00.00.00 5 .05 .o. 050...

 

 

82m?

2: 60¢
88th

as 85
8:88 05
8 8882
can team—ab

 

 

 

 

 

 

 

 

 

 

 

 

88 8:89
38 E
828888
cam 9.888%
8.82
8E3»:
@8085:
can wczmaa
..8 .385
892868 2
8:85 “8%:th Ea
82m .33 28 88:9 8:: 88.883
52: Q88 82 82:38 bfito. new
300— 85380 82w :88? ©8808 @8285
8288800 28522 8V 8 mm 8 8:: 85 8 88288 EEw—omv 8:89
53V 32 mm 88 888 8:8... 5 85 828:5 .wcéc 8893mm mmflw
£3 98 :23 .8538 8 Soc. 8 .22 2 8:89 8 833.5 28532
828800 teams“: begun can 3:982 :88“: 98 808g 98
§> oZ 8 82283. wacm wags—8m 82w $2 8838*. 83m 880 8805 8.3 25 Sum
8282989 @8888
83385 818.2:
Etna—88m 8.3:.— .835988 8.53. 98
92—988.. 3982 9882289 8.. $9325. 8.85889 mini—w:— =833=n=< 5.589: 3358
98 80> 833.5 8 ran—55m R889»:— 8:czo:=m mom—«sow 2.8m :82 <04 wiwaxuam 828.:—

 

 

APEOQV 883% 88388 wE>8>E ~01— Ea <04 882.8 wcmwaxomm B828 8 7,5 A: 2an

 

8888

 

 

 

 

 

 

 

 

 

 

 

 

E8: 93%

828230:

Ea Ema:

3:85:88:

88359200

.35: BE .6800

:8 2% 885988:

38588 88:8 808::

dosh—83:00 08:: mSE< 888m

883 082 588.88 “0888.. 80.: A95:

8 .888. E 8 :08 8:880 083 8:25.830:

PM: .3 3:0 83858.: «.8830: 8.8::

88:93 8058 3088:0038: :wE AAEV o

2 <1?— 8: 38:8 :23 88 8273 8:30:50:

:8: 85 85 883 :8 8:588 .<:E 80 8:092:38 .Amed

80888 .83 :88 888008 82885:: :30 2835830:

E50 82 88888 :0 E8: 8808:: 588:

80200:: :8 w:::83 5:88:50: :8 88:38 mSE< 30— ADA:

38m :8 8 38:8 88w Eco—w : :8 38:2: 28% 88888:: :8 o: 8080:8030:

085... moon 8. 88 .388 8:50: _ 888:8 .3028. 88m b88888: «>FS88: .o :032
.8 8 083 02 E8: <4: 80:0 8 w! — <4: 0: 88:80 8.805 8.88:: .08 :032 32

8829.00
mug—50:2: Esau
83558.: 8:9:— :0m_._anfi00 8.50m 3388.:

8:308.— 3989. 2.082280 :8 £8.58: «882800 8:3in :0=3=AE< 5.5893 6:: 5:88.—

98 88> 18.—3.5 8 Esafiam R8282: 8:030:3— 8:5:on 2.09m :82 3: mamas—9:.— 838.5

 

 

20.808 88:: 8.88980 w:_>_0>:_ 54 :8 <04 888:0 mfiwmxomm 8828 8 83 d— 835.

55

 

8::

::5 8:3:

0:: 508550
0m:0::00:w
<4: 023850
2880005
:0: 800

<4: 82%:
5083100

0:: 5:00:— 50::
508250 8w
0m:0::00:w ::
0:: <4: .08:
m: :0:::0::>0
80::0 05:00

(A: :0

5520:8002:

_E0::_ 02
50525500

 

 

:0:3 :0:8 :0 0:: 8:00:00 00:03::
:_w::5 55:3 2850:: :0: m:
m: 00:08::5 5:3 0::505 85 8:00::
$8300: 0:0: 0: <4: :0 :0 @5300”:
000:: :0: 00m: :0::0::w000_: A5555:—
m_ :0_:::0::>0 5:00:1— 8: 88:20 8:0
80::0 085 5002—020 .8: .w 0058:0585
m: w:0_ :8 3:0 0:0: 050:: 5:205:85 5::
Goon :: ::5 38:0 85:58 30:0: :0: 85:58 092880 0:238:30:
.::5_:0m: 80: 05:58 :3 00:22:: 5:50:88 88885 08 0:: 24::
voom 8:0 <4: :80» 80::0 2:5 3:: 35:08:: 0:000:30:
.::5::0m 0058:0585; :0 m: 83 0:: 03:00 :0 8:85;”: 3000:: Now
00:25:00
5:88:85 50:89:00 00.500 0.5885
00:08.8: 3080: 0:33—0:00 :0: £02.25. 00808—80 8:332: 5:00:93: 3.5893 0:: 30:50:
0:: 80> 805:0 :0 5035.5 8:020:5— 008::00m 0:00m <04 ”5938::

 

 

 

 

 

 

 

 

 

 

20.500: 855: :0m::::500 w:_>_0>:_ an: 0:: <0:— 005000 w:_w::0:: 02008: :0 85 d: 28:.

56

 

 

 

6:08:33 :0
8:22:38
:5
838.355
£22338 3::
ho 5:855
.93 £838.:

 

 

88:8:
wonaoxm
.ufizaoN 3oz
5 2:06?
wfiwfioa 03.
6052: 05 :0 38:2:
:3qu E: @2982
ooom mam on. .8: €2.38.
5 3:25:00 .wESSfigcafi .802.
803:: 3:8 wEmaxoma @xmaotém
05 $80: .852: $6003 Ema—:5“ A89 .9: 829?
8:8 5:83 .3300 a :5 82L. 5 BEBE .328: .252 wfiwfiumq
of 55 oooN 328% @xwaobém 958$ .wdv wEthE 033:8 £3 809::
353 95% 5:03 : biota dd 53$? wEEwBB basin 9%85 -oEooam
ix :5 35:50 Oct. §:uE:o:_>:m_ wEwuxoma Son: 25 223$: :5: ESE : US$213
3‘: voom @fiaotém 05 E 32:05 cums—a 523:0?“ >3: h.o ,m2m<Bomaa :owooa £05380
5x :5 03 of. 303:: 2.: c8 :5 < b::~:oo:: 228332 3:3 3:0 $805 3:233:00 tzwo>
69335.3
$53365 cave—358 68388
53:83: 8:9:— ..8 8.58 mafia?
3:282 33>?— Efimfiu—Sa $932.:— 33238 map—win:— =czuo=an< San—Bab es: 353:8
6:: 39> .3320 u: anus—Sam >335;— .::cuo:=h movie—cum 2.8m 5:2 <0: wfiwuxoam 828.:—

 

 

 

 

 

 

 

 

 

 

 

 

8L:03 363m :oflhwmfioo mE>E>£ ~04 Ea <0: BEBE wEwmxomm 8328 mo 35 .2 2pm.:

 

.0380:
0. 0000.01.06 0...

 

 

 

 

 

 

 

 

 

 

 

 

0.0.0.? 000000 ...000.000

00.0.20 00.050. 00:00:30..

0.0 000.00 00059.0

km“. :23 2.. 0%:

€0.30 00.00.00. 00800.0

Avoom {Ou— 00:...0O .0: m. .0...
Ema. .w0w..0_ 0... 5.0.3 8:00... 00
00026 00.030. 0.0 imam $50.33.

08.0. 0000.6 .000 00.3. >030 .0": 80.9? .2505: 000205
$50.83. 05 00 .0...0 0.30.. $500.00. 000 00.0 0.0
050.000 ..0 Ema... 5.0.3 00 00.. .onm 00000 0000.050... 000.0 .050

km; 0... 0.5 000.00 30% 0000020830 ...00000 0:0 0.: 000.00 00.2w
.0.. 30:50 _ pd 0302.0. 000500 $55.09 30000.0 032000.
.w.0£00_0: 0:. 0:0 000.00 000 000.0000. 0.0.0.? .00.. .w: _ 5.0
... 5mm: ....mE $3-000. £00005060 @5050. 00.0035... 00.8530 0.... Chat .0203
20.0000”. _ m. 0... 2003.00 0050.00. 300.. 000 00:02:00 00.500 0.205.320. 00.05:.
S..0E=0..>..m 0000000006 £005.80 00:00:00 £000.08 ... 00.0.0000 000350300 000
0:0 300...”.— .009... wag? 03.0.60 00.3.00. 30. .0 00200000 .33 00.5.0
.00 0.300.: .3:0E:0..>:m .0005 00 _. coo. 008.00 00000.06 000000.002 00000.5 -000 _ 2 com
00.00.0000
000000.05 000.0388 00.00500
5.00808... 0009.: 00.. 00.500 03.00008
00:280.. 3030.. 0:33—0:00 £00.50... 000029.00 madam—aw... 0033.304. 3.509: 0:0 $0800.—
05. 03> 3030.0 ..0 b.2555 R..0...0>:— 10:050.... .— 8taa8m 0000m £02 <04 wing—00.— 00:00...—

 

 

A0....00v 00.030 000000800 w.._>_0>:_ 54 0:0 <04 00.00.00 mammaxomm 00.00.00 00 .05 .o. 030...

 

£5580 mam

 

 

 

 

 

 

 

 

 

 

 

 

umaoacoem
Ea
8mm; 22:
8883
as 3 >80: 82:8
Avoom ma 8:: c2982
.muoowtow moxom .388 go 282
SEO 8888a 95 .68:
-5435 :52 95E 85.8. 85:38 620288
.8 .5895 $8 85 803 83 3:8 93
hop. £580 838: 8E: £08038 3:0:an8 mam—080m
wEwaxoam Se 55 9 £8: wEwmxoa .Rmommfi
.8 @982: 82: 203 808 com .82 85038 2
293 833 28%? mo wows—8a .88 .8an 823.893: damn—a
83V voom .25 .2983 A888 wag 688888 .onE .223 .88
Jim— .m.D 2: ..om $2865 90 EwBB 6:925 war—88a ho 889:8
was 5:80 2388:8160 a 93 -EwB: 3.538 08 .520 9 $3
388coh>cm wcmocoscfi 8308688 5:: EOE 882888“ $835 28
mo ~888qu 58& 30:8 .833 “ES - ..m.m 82:2: .Bzfiflscme owaccsu no 808
88.0 of 39: 2: .waowmflEu x ...S x :2.— mEwaw 9.3688 89: got? c8
.8 82883 2 wags—8a 833 28 EN 6.: 8:53 wcmwaxoan .88.“be warms». 832 5:5 3x8 .88
225:”. mm; ho Ems? .8: 38%. Good. @625 REES 33”— wEwacoSn— $8on 8.82.80 -5434
8839.8
9.38%.: 839:8
852885 88:: 88.858 8:8 23.53:.
88.3%... 33>?— 83888 ..8 33.5%:- 8.5288 3:333.— :c_a8=na< 32:85 6:: 83:8
.23 58> 185.5 we mus—5.5m R5353:— 1288:: m 8.8:an 2.8m £32 <04 mafia—8m 8:8...—

 

 

szcoov 86:: 88888 mE>_o>E ~04 was <01— BEoto 9:888 8828 .6 a: .o_ 2an

 

.8:OQm

 

 

a: 8808:: .0
.8m :08: 8888:: 880:8:
E 8:800: @9388: 8:83
8:228:8— ::0 0:03 :08: :8 m2:<
9 88:8 8 5:28:80. 88:5
80:80:85 8:828:80: 28855 08:0 0:5
8 00:0: 8:00:88: 5 802:: 8:02:08:
8:3: .0: 8:05:80 8:30:85 23:08:: 8:83:80
:o:8_:::-8m_ 8:83 :8 888.8: 8:
3 8:02:80 Boo—w: 8:58:0— 8 88:08 :8 088:5
080:8 0:: 8:08 08:50 .0: 8:80:88: 8:05.: .
08:0: 008 8050:: 0.: .8: 88:00: :5: 08 AWE—23m:
8528 :2: 8308:708— 0 E 30:80.: 0:78-80 . 8:38:88: 88:80: 0:08:
00:80. :08: 0.082580 :0::80: 820:5 $38008 m:m_ 8 0:0 8:08:08:
Amoom .85. 88:8 ow. :o_m:0> 0::: 0 888: 3:903 85:30:88.0: 8:88:32 8:85:80
:5: 8.5 0:08: <0... 2:065 8 8:05: 8:08.: 002 :08: :00: :8
moom .85: 888:0: E 8:205 0>:008:: 8:0: 800:: .8888 :8: Sam—8008 08:30.20: :0::80:
:8 8:. 50m 0:00:8— _0::0::_ 08% 8 088:0: 3:03.: 3:: 080 0008:: 8883 0:0,:
800288 88:58
08808:: 08:08.:
8:88.88: «0::—E 588980 :38 0088 :8
00:08.8: 3030: 888—080 .8: £888: 8.8288 38:2»:— 0=::< «890:5 888
:8 80> 803:0 .8 8888.6 R8802: 88:88: 8.8:00m 0:8@ 82 <04 9.85—8.— 8:8:

 

 

 

 

 

 

 

 

 

 

 

 

6.803 8830 808980 w:_>8>:_ :04 :8 <0: 880:: 8808:: 880.0: 8 :05 .0: 038:.

60

Another popular use of LCA in packaging is when LCA has been used from a
waste management perspective to attempt to identify environmental burdens of a certain
waste management operation, or to determine what is the environmentally better waste
treatment system for packaging materials (Arena et al., 2003; James et al., 2002;
Neumayer, 2000; Wollny et al., 2002). Within the waste management programs often
studied by LCA, often included are take-back programs, which are taking root in a
number of industries (i.e. automobiles, computers and other electronic devices, adhesives
and garments) for dealing with the disposal of their products. Product take-back programs
and “extended producer responsibility” (EPR), have become popular in many countries
(Scarlett, 1999; Schifflcr, 2002a; b). The take-back systems are founded on the idea of
product “recovery” by the manufacturer or “reverse logistics” and by involving
appropriate planning, managing, and optimizing of the forward as well as reverse
distribution streams of both new and used products. They have to be not only cost-
effective, but also to reduce the environmental footprint of an operation, but not
necessarily by packaging reuse. While a number of these programs have developed as the
result of legislation, some of these programs are voluntary, set by industries. For instance,
the electronic components industry, due to its high volume of bulk shipments to a
relatively small number of globally distributed locations and the near pristine condition of
the packaging used after shipment, seems to be especially suitable for instituting
packaging take-back systems (Matthews and Axelrod, 2004; Matthews, 2004).

Furthermore, these take-back systems may be part of broader initiatives such as DfE.

61

b) Third party use of LCA
A third party, in this case, is any independent (governmental or private) body that
might use LCA, or some form of LCA, for either of two purposes: 1) help with
environmental labeling programs, or 2) help with policy-making.
Environmental labeling programs
By a third party, LCA can be used to help obtain information required by some
environmental labeling schemes. For most developed (and some developing) countries,
environmental labeling in general comprises more or less the types of environmental
labels listed in Figure 5.
In recent years, steps for harmonizing and standardizing environmental labeling
programs worldwide have been taken by ISO, with the release of a set of standards in
which it recognizes three types of voluntary environmental labels: Type I (Environmental
labels and declarations), Type II (Self-declared environmental claims) and. Type III
(Technical report-environmental labels and declarations). ISO Type II labels, also called
“green claims” or “green symbols” are issued by the interested party which itself creates
the label, applies the label, and establishes controls to ensure that its product meets the
claims on the label (Weber Marin and Tobler, 2003). In some cases, these claims may be
certified under single-attribute third-party certification programs. Green claims are the
most widely used environmental labels (Lavallée and Plouffe, 2004) and are not based on
product life-cycle concepts. Instead, they are general statements about whether a product
is recyclable, contains recyclable material, is degradable/biodegradable/photodegradable,
or comPostable or source reduced, or refillable, or ozone safe, etc. In countries with

regulations that allow the use of this type of labels, claims are required to be accurate and

62

not misleading in order to comply with national legislation and trade regulations. For
example, in the U.S., federal (Federal Trade Commission) as well as state bodies have

developed guidelines to regulate such claims (Levy, 2000).

63

2mg: 8:0w< 8808:: 8:058:35,“— 008%
85:3 6:36.83 80:00:: :8 0.0.0.8: .0080. 8:081. .0::0E:0:.>:m 8 800:: 8:00:80? .08. 8:058:35. .m 08%.:

50:0. :08:-<U\.

 

 

 

 

 

 

 

>
\ )
20:0. .>:0 A .
.3: 00. .32 >5 . .
.3 3.23:: 2;... 0: ..e E; 3.2.: w e 222 Es a
:80 ::0:0~. 8:05.228 0:: :. m 0: mt: :
:0 .00m 0:=:.::.:-0.w:.w : : 0
20:0. 3.: .:0:: 00:05:88: €33 ... 0: :0:0:0:A0.::.0»_00: .w.0:
8880:: 0:50:80 0.550.280 :0 33:0: 28:95 0 0000:: .w.0: A0 0 .>:0
8:00.380 .02 <:m $0: 50:85.00: .w.0:§:0..::0:.>:0 8::500 .w.0v : . : .
. . m::00.::E : 0:»... Om.
.3:0E:0:.>:m. 088.00.: w:_:83 0:08:80 w::00.::E .
. . . 88.8-0080 :0: 05.0.0
:08588. :0 :88: 8 8:059: 8:52-008“: _ ;
x:§:=.0> 80:88.). 88.208.59.00 :0::.0:-:0::0::

T : . _
_ 4

05:38:: w:..0::. .:::0E:0:.>:m 3880:: m:._0:0. .3:0E:0:.>:m
b:::-::.::r b:::-:m:.n.

_ _
A

w:._0::. 23:05:05.8”:

64

Though the purpose of these claims is to provide the consumer with accurate
information about positive environmental attributes of products and to help with
international trade, a couple of issues arise with their use. First, due to economic
limitations it is very difficult for an average consumer to challenge or question these
statements since tests are often expensive and time consuming. Second, these claims
often focus on one stage of the product life cycle, disregarding other stages that
potentially may be more harmful to the environment, and thus providing misleading
information. In fact, Lavallée and Plouffe (Lavallée and Plouffe, 2004) argue that the
widespread use of such labels has hurt the development of LCA-based labels.

Type I and Type III labels are issued by a third party and involve LCA—based
analysis for the certification. Type I labels, also known as seal-of-approval, are the
outcome of what are usually called eco-labeling programs in which a product, process, or
management system is certified to meet specific environmental criteria as established by
a third party organization. Oftentimes, manufacturers make prior use of LCA under
programs such as DfE that help environmental stewardship, to “self—certify” their
processes before third-party validation. The third party can be the owner or administrator
of the label program, which usually has three basic steps: (a) selection of product
category (e.g. by similar function, and/or similar environmental impacts, and/or
importance of product in marketplace); (b) development of requirements to be met (e.g.
by using some form of LCA along with peer-review process, selection of the system’s
most relevant contributions to environmental impacts and guidelines for their reduction
are set); and (c) certification and licensing (i.e. compliance verification and testing,

applicant licensing and monitoring). Worldwide, coo-label programs vary on how they

65

are run or sponsored. They can be administered by governments, private companies (for
profit and non-profit), non-govemmental organizations, or some combination of the
above.

Though both ISO Type I and Type III environmental labels involve LCA
concepts, they differ with respect to the way they convey the information. By assuming
that the information from LCA results is too complex and too extensive to present on a
label, Type I label programs first decide which stages of the LCA are the most significant
for the determination and weighting of the certification criteria, and finally evaluate and
issue the seal-of-approval for qualifying products. Alternatively, ISO Type III labels aim
at presenting to the consumer much more detailed environmental information, including
items such as energy use and environmental impacts in a report-card format, and assume
that consumers can themselves prioritize across environmental burden categories and thus
themselves do the judgment.

Several issues stir debate and complicate the use and implementation of these
environmental labeling programs. For example, while ISO standards require an LCA
compliant with ISO 14040, in-depth LCAs are seldom used for awarding these labels
because they are cost- and time-intensive. Instead. these programs end up considering
only certain stages of the life cycle (Davis et al., 1998), usually by extrapolating from
environmental performance results of similar products offered on the market.
Furthermore. regarding type III labels, since the selection of labeling criteria is not based
on the same LCA methodology, product comparisons cannot be made, thus confusing the

consumer at the moment of judging the preference.

66

Issues may also occur due to the complexities added by global economic trends,
trade agreements and logistical practices when considering imported goods and the
environmental assessment of their life cycle (Appleton, 1997). Further, due to the nature
of the ecolabeling programs and their LCA-based approach, often the programs end up
awarding preferability seals to products made with state-of-the-art technologies that are
difficult to obtain in less developed regions or countries, thus creating animosity towards
the results of these studies (World Trade Organization, Environment: Trade and
environment news bulletin, 1996).

Standardization efforts, though very costly and/or technically challenging in some
cases, have been made in order to achieve harmonization and/or mutual recognition
among programs. In fact in 1994, national and multinational ecolabel licensing programs
founded the Global Ecolabeling Network (GEN) with the objective to “improve, promote
and develop the ecolabeling of products” (Global Ecolabeling Network, Introduction to
ecolabeling, 2004). Currently GEN has twenty-six members with programs such as the
well-established Green Seal (U.S.), Blue Angel (Germany), TerraChoice (Canada),
European eco—label (EU) and Eco Mark (Japan).

Policvmaking

The use of LCA for policy making is a practice that sometimes faces strong
opposition from trade and industry organizations and even from countries. So far, the
governments within the European Union have the most experience in using LCA
concepts for developing policies.

The Integrated Product Policy (lPP) developed in 1999 by the European

Commission, is a product-oriented approach to government policy that attempts to reduce

67

environmental degradation by addressing all phases of a product's life cycle. The IPP
approach uses a number of instruments, such as economic assessments, product
stewardship programs, substance bans, voluntary agreements, environmental labeling and
product design guidelines, to address the system life cycle impacts of products and
processes.

Not surprisingly, though, actual LCA-based policy making, as for any other type
of environmental policy, has many critics, and may have economically sensitive
consequences for many industries when policymakers consider taxing or restricting what
are found to be environmentally unsound products (Europen, Use of LCA as a policy
tool, 1999; Europen, Economic instruments in packaging, 2000). For example, many
European countries have used some form of LCA to develop federal packaging mandates
that require manufacturers to take back packaging discards or pay for their recycling.
Germany requires companies that do not participate in its Green Dot program to take
back their packaging and pay the cost of recycling it themselves, with no exceptions for
foreign companies. This measure has broad implications since the take-back burden is far
greater for those companies that ship their products long distances to Germany. Thus
several manufacturers exporting to Germany from within the EU and beyond argue that.
due to its nature, the Green Dot label program places imported goods at a market
disadvantage. Moreover, industry and trade organizations within Europe argue that the
degree of diversity between countries and even regions within the same country is so
large that the preferred waste management method in one area may not be appropriate for
other areas. Thus. these constituencies claim that waste management decisions should be

made on a case-by-case basis (Europen, Use of LCA as a policy tool, 1999).

68

There are no federal packaging mandates of a similar kind in the United States
(US. EPA, Environmentally Preferable Purchasing, 2005). However, there are a number
of federal and state initiatives that involve the use of LCA based tools. For instance, since
1997, the US. EPA has been promoting the concept of extended product responsibility
(EPR) which is a product-oriented instead of a facility-oriented approach to pollution
prevention by using product life cycle concepts (Davis et al., 1997). Within this principle,
programs such as Environmentally Preferable Purchasing (EPP) promote the use of LCA-
based tools. In fact, originating from executive Order 13101 on "Greening the
Government through Waste Prevention, Recycling, and Federal Acquisition", EPP is a
nationwide program that uses the leveraging strength of federal buying power as an
incentive for industry to develop environmentally preferable products. Guidelines for
EPP are developed by the US. EPA for use by other federal agencies; however, the
program encourages state and local government and the private sector to incorporate
environmental considerations into their purchasing processes as well.

Another kind of initiative is the US. EPA Design for the Environment Program
(US. EPA-DfE) (Environmentally Preferable Purchasing, 2005) which is a voluntary
govemment-industry partnership that seeks to incorporate environmental considerations
into the design and redesign of products, processes, and technical and management

systems.

69

VIII Outlook and conclusions

The future of LCA for packaging, as for any other product category (e.g. energy,
automobiles. appliances), is necessarily linked to the future of LCA and its maturation
into a more reliable tool.

Thus, with regard to LCA in general, challenges remain due to the uneven pace at
which it has been embraced around the world. In fact, though developed more than thirty
years ago in the U.S., the European willingness to incorporate it as a part of their
environmental regulatory process is often cited as the reason why Europe leads the way
in LCA research (Curran, 1999). On the other end, it is only since the 1990’s with [80’s
release of its 14000 series of standards on Environmental Management that many
developing countries have started to learn about this concept. The delay in coordination
(i.e. internationally and nationally) is one reason why a common terminology has been
slow to develop. and terms and approaches such as life cycle management or life cycle
costing generate confusion. Further, since many LCA studies still remain unpublished or
inaccessible, the assimilation of common methodologies is even more difficult. On the
other hand, the increasing tendency of the private sector to look at product life cycle
concepts and embrace them at the product design phase in order to respond to consumer
expectations may be a sign of what is next. As multinational firms extend their operations
around the world, along are spread their philosophies and their understanding of LCA
concepts. This is why efforts on harmonizing private life-cycle initiatives have started to
occur. For example, under the UNEP/SETAC Life Cycle Initiative various workgroups
on inventory, impact, and Life Cycle Management (LCM) are trying to achieve this

international coordination and discussions have been proposed to adopt LCM as the

70

platform from which to build and execute private environmental stewardship programs
(Hunkeler et al., 2001), with an “LCM toolbox” with LCA and LCC as components
(Hunkeler et al., 2002; Sonnemann et al., 2001). The open partnership of the private
sector with environmental research institutes and regulatory bodies has often been cited
to as one of the reasons why many European countries have a healthy LCA activity
(Hansen, 1999), and this is a reason why international harmonization measures in the
private sector are important. Thus, coordination efforts reflected by the numerous
guidelines from SETAC and ISO, and working groups and workshops will need to
continue to catalyze the harmonization process and to address current limitations.
Furthermore, private and governmental agencies of the US. and of European countries
will need to continue their development of frameworks and partnerships with industry to
help with the goal of making LCA a more useful tool for decision—makers. Current efforts
by environmental certification third party organizations in these countries will need to
focus on continuing their homologation steps as well as reaching out to similar bodies in
developing countries. Likewise, work remains to be done on controversial matters such as
data quality and harmonization of inventory procedures and impact assessment methods
as well. LCLA improvement will also depend on the success of efforts on modeling the
fate of chemicals released into the environment and development of weighting
procedures. But perhaps, due to the nature and subjectivity of many of the components of
an LCA, several issues will still remain unresolved.

For LCA for packaging, future challenges exist. For example, as packaging
remains a necessary item in the market and one of the preferred fields to which LCA is

often applied, world population growth and a higher overall quality of living are

71

indicators that packaging waste management Options (e.g. reuse, incineration, recycling)
will remain issues for further LCAs. Furthermore, optimization in packaging design with
regard to the environment will necessarily use life cycle-based methods as expressed by
emerging industry groups such as the Sustainable Packaging Council (Johnson, 2005).
The consequences of trade agreements and a “global economy” with regard to production
and movement of products and their packaging from one region to another and the
consequent implications for resource (c. g. material and energy) use and emissions release
will also have to be investigated with a “life cycle thinking” philosophy. Lastly, as ever
more complex packaging concepts are designed in order to meet increasingly higher
consumer expectations (Ver-Bruggen, 2004), complexities in their LCA will require
further research. Improvements in radio frequency identification (RFID) technology for
product tracking, new dynamic promotion/information capabilities through active labels.
product quality sensors, nanoscale optimization of material properties, development of
fully biodegradable as well as biodegradable composite materials are some examples of
scientific breakthroughs that are starting to be used (Biodegradable Plastic Society.
Inventory of GreenPla products, 2004; Lingle, 2004) or are being investigated to build
packaging materials and packaging components (Anyadike, 2004; Harrop, 2002) in the
future. In turn, inventory databases will need to be developed not only for “conventional”
packaging materials and components but for these new packaging concepts as well.
Furthermore, since with the use of these new capabilities, new functions would be added
to packaging (e. g. degradability, traceability, sensing) besides the traditional containment,

the definition of the functional unit for an LCA will be a subject for future debate. All

72

these challenges are indications that LCA research for packaging options should remain a

high priority in the future.

CHAPTER 3 COMPARATIVE LCA OF THREE DRINK DELIVERY SYSTEMS

I. Introduction

This chapter presents information regarding a comparative life cycle assessment
(LCA) based on a hypothetical drink product for which we evaluate three main container
material alternatives: (1) a polyethylene terephthalate (PET) bottle; (2) an aluminum can;
and (3) a polylactide (PLA) bottle.

Following ISO’s LCA steps, details and information sources used for this
comparative LCA are presented. In order to represent realistic implications of the use of
the three main material/container alternatives and their actual packaging operations,
production and use of distribution packaging (i.e. corrugated trays, stretch wrap, and
pallets), as well as transportation steps through several parts of the packaging life cycle
were included. Relevant end-of—life scenarios such as landfilling, incineration and
recycling were also taken into account in this analysis. Each of the main material
alternatives (i.e. PET, PLA, aluminum") along with its whole set of operations was
defined as a product delivery system (PDS) and environmental burdens associated with
each of the three PDSs were included in the study.

The calculation model of the present study was based on previous comparative
packaging based LCAs published elsewhere (Keoleian, 2001; Keoleian et al., 2004;
Sellers and Sellers, 1989). The data used in this study come from a number of publicly
available sources and can be divided into background and foreground categories.
Background data comprise the life cycle inventory data associated with all the activities

and processes involved in the drink delivery systems. This type of information was

74

obtained from the Data for Environmental Analysis and Management - DEAMTM -
database developed by the Ecobilan Consulting Group (Paris, France) (DEAM module
databases and manuals. 1999). The DEAMTM database represents industry activities by
“modules” whose inventory data in turn were gathered from a number of different
sources (e.g. official reports, surveys, engineering estimates, etc.). Each module is, in
fact, the result of a “cradle to gate” inventory analysis. With this convention, then, a
complete product life cycle (i.e. cradle to grave) inventory analysis can be modeled by
appropriately linking several of these modules. The linking of these modules requires the
knowledge of foreground data. Foreground data comprise data specific to a certain
product and its requirements (electricity requirement for PET injection stretch blow
molding, etc.) Foreground data were obtained from a number of different sources, and
for organization purposes their details are provided in later sections of this chapter.

Regarding this study’s impact assessment component, energy, water use and two
characterized impacts are included: global warming potential (GWP) and ozone depletion
potential (ODP). Other impacts were not included since we restricted the focus of our
analysis of the comparative LCA results to packaging-related causes and effects of
uncertainty in the inventory data. Thus, as explained in Chapter Two and aware of further
uncertainty considerations in the impact assessment methodologies of other impact
indicators (e.g. natural resource use), we opted for using GWP and ODP as the two most
widely accepted, least uncertain impact indicators, according to the LCA community.

For organization purposes, this whole chapter is devoted to presenting the
information used in the comparative LCA study as transparently and openly as possible

and clearly giving specifics on the LCA scope, data origin, and the systems considered.

Actual results of the LCA are in the next chapter, Chapter Four, along with the actual
uncertainty estimation and the discussion of its results.

The following sections of this chapter explain the methods and present the details
of the LCA. In particular, Section II discusses the methodology for the study, including
the goal and scope of the analysis. Details on the definition of the functional unit and
product delivery systems are provided. Specifics about the system model and its
boundaries, data categories and impact categories are also included. Later in this section,
information regarding the life cycle inventory analysis itself is presented, including
specifics on material allocations, system diagrams, data sources, calculation procedures.

assumptions and limitations of the analysis.

11. Methodology

a) Goal and scope definition

In this section, the goal and scope of the analysis is presented. Details on the
purpose of the study, definition of the functional unit, and definition of the product
delivery systems are provided. An overview of the basic PDS model and its boundaries,

accounted data categories and impact categories is also included.

Purpose of the study

The primary purpose of this study is to provide an LCA model to evaluate
robustness of the results when subject to scenario and parameter uncertainty. This study

also provides information to increase the knowledge of some of the potential

76

environmental impacts associated with the life cycle of a hypothetical product delivery
system considering three packaging material alternatives.

The alternative product delivery systems modeled in this study are:

. PDS l: utilizing PET bottles and PP caps as one choice for drink delivery.

. PDS 2: utilizing aluminum cans as another choice for drink delivery.

. PDS 3: utilizing PLA bottles and PP caps to investigate the use of renewable material
feedstocks.

The findings in this study are intended for research purposes. In particular, the
information in this study is expected to give insights on the behavior of a packaging—
based LCA outcome. Though the findings of this work may be useful, due to the
streamlined nature of this study in particular, and the intrinsic limitations of LCA in
general, it should not be considered as the only source of information regarding the
environmental performance of the materials investigated or the product delivery system

considered.

Function and functional unit

The function of the system was the delivery of a hypothetical drink to the market.
In this study, market is defined as the first destination after the product filling stage. The

functional unit was 1000 liters of drink delivered to the market (distributor or retailer).

Product delivery systems (PDS)

The PDSs were defined as the systems employed for the distribution of the

hypothetical product to customers. They included the primary and distribution packaging,

77

as well as all manufacturing, and transportation associated with the function of the
system.
. Primary packaging: The hypothetical container formats, materials and sizes
considered in this study are listed in Table 11. The plastic bottles under
consideration were assumed to be injection stretch blow molded and the

aluminum can was assumed to be a draw and iron two-piece container.

Table 11. Primary packaging characteristics

 

 

 

 

 

 

 

 

 

Format Materials Individual Number of
size, fl. Oz. containers
per tray
Bottle/sleeve/cap/carrier ring PET/LDPE/PP/LDPE 16 24
Bottle/sleeve/cap/carrier ring PLA/PLA/PP/LDPE 16 24
Two-piece can/carrier ring Aluminum/LDPE 12 24

 

As is a common practice, all formats use LDPE ring carriers and paperboard
trays. Inks used in primary packaging or related printing processes were not included in
the study.

Distribution packaging: This is defined as the components of a PDS used for the
delivery of materials for the manufacture of containers, and the materials used to protect
the primary packaging during the transportation to and from the drink filling facility.
Specifically, the distribution packaging comprised corrugated boxes, stretch wrap and
pallets.

Transportation: The shipping mode considered in the PDSs is truck.

Location and time considerations: The operations involved in the product delivery

process are assumed to be current and to happen in the US.

78

S Vstem model

The system model of the LCA divides the PDS into phases (see Figure 6). Each
PDS was represented as a nine-phase system with phases labeled as follows: material
production, distribution 1, container component manufacturing, distribution 2, filling,
distribution 3, wholesale retailer, product consumption and end-of-life.

Though identified in the PDS model. the wholesale retailer and product
consumption phases were not included in the life cycle analysis (i.e. their associated
environmental burdens were not compiled) because emissions and energy use data about
these operations were not readily available. Furthermore, it was reasoned that the
omission of these operations would not impact or change the characteristics of the LCA
since its contribution was assumed negligible (e.g. when compared to the material

production phase).

79

 

 

3: .3 Ucm

 

 

33.0. of E wobEmcoo momma £98 £5 .0 oeswE

 

Bzfiom "
>532:me

 

95E

 

$220835”:
$5350

 

 

 

 

 

5620.605
.3832

 

 

 

 

 

 

 

 

m 8:555

 

 

L

 

 

 

 

 

 

223955 11 Zoeseeea A.

 

 

 

 

 

 

80

System boundaries

The boundaries of the system are presented next, following the order of the phases

of the PDS according to Figure 6.

Material Production: The Material Production phase includes the extraction of the
raw materials, the materials manufacture (i.e. the processing of the raw materials
into intermediate materials), and the corresponding transportation of raw materials
to a location where they are processed into intermediate materials. Material
Production includes only production of materials used in the primary packaging.
Material production of distribution packaging is included in the Distribution
phases. Additives (e.g. plasticizers, antioxidants) and color concentrates (i.e.
pigments and carrier resin) that the PET. PP, LDPE and PLA materials used for
the primary packaging may contain usually represent a very small percentage of
the plastic mass and thus their associated production burdens were excluded from
the system model. As a result, the materials were modeled as 100% plastic resin.
The aluminum for the aluminum can was also assumed 100% pure.

Distribution 1: Distribution 1 includes the transportation of primary packaging
input materials from the location where materials are produced to the container
component manufacturing location. Material production and manufacturing of the
distribution packaging used to ship materials between the Material Production
phase and the Manufacturing phase were considered to be outside the boundaries
of this study. Likewise, since most of the shipments in this phase utilize bulk

trucks, distribution packaging weight was not included.

81

Container component manufacture: this phase includes only the manufacture of
the primary packaging. Manufacturing of distribution packaging was included in
the Distribution 2 and 3 phases. Burdens associated with the production of the
manufacturing equipment are excluded from the system model. These burdens
were expected to be small relative to the burdens associated with the processes
that were inventoried. The one exception was that road transport in many of the
published data modules contained the burdens associated with the construction of
the vehicles and replacement parts such as tires and batteries. Burdens associated
with the construction of the manufacturing facilities are excluded from the system
model. Burdens associated with human activities, including driving to and from
work, are excluded from the system model. Printing inks are excluded from the
system model due to lack of data.

Distribution 2: this phase includes the transportation of primary packaging
materials from the manufacturing location to the filling facility. The
environmental burdens associated with material production and manufacture of
the distribution packaging used during this transportation were included. The
burdens associated with the transportation of the distribution packaging from the
supplier to the manufacturing facility were also included. Materials required for
the shipment of the distribution packaging were considered to be outside the
boundaries of this study because emissions and energy use data about these
operations were not readily available. Furthermore, it was reasoned that the

omission of these operations would not impact or change the characteristics of the

82

LCA since its contribution was assumed negligible (e.g. when compared to the
material production phase)

Filling: Energy and emissions burdens associated with the filling process were
excluded from the system model. It was assumed that the facility in which the
filling occurs is also used for the drink production and therefore burdens
associated with the PDS are difficult to distinguish from the burdens associated
with drink production. For this reason, Filling was determined to be outside the
scope of the study. The only exceptions were two: a) the inclusion of the injection
stretch blow molding energy (i.e. in the form electricity) prior to the filling in the
case of bottles; and b) the solid waste generated from the losses of primary
packaging during the filling/packing processes, which were directly related to the
PDS and therefore included.

Distribution 3: This phase includes the transportation of the drink and its primary
and distribution packaging from the filling location to the distributors/retailers.
The environmental burdens associated with material production and
manufacturing of the distribution packaging used during this transportation are
included, as well as the burdens associated with the transportation of the
distribution packaging from the supplier to the filling facility. Materials required
for the shipment of the distribution packaging were considered to be outside the
boundaries of this study. Distribution 3 includes only the burdens associated with
transportation to the first destination.

Wholesale retailer: Environmental burdens associated with activities in the

Distributor/Retailer phase were excluded from the system model. Disposal of

distribution packaging used in the Distribution 3 phase is counted in the
Distribution 3 phase.

Drink consumption: Environmental burdens associated with the transportation of
the product from the retailer to the consumer and consumption of the drink were
excluded from the system model.

End-of—life: Environmental burdens associated with the End-of-Life phase include
burdens of landfilling, incineration and recycling operations associated with the

main material alternatives (i.e. PET, PLA and aluminum).

Data categories

Four data categories were used to classify the items inventoried throughout the

study. The data categories were intended to reflect the emissions or resource use for each

area of interest. The data categories and their components can help make overview

statements pertaining to the environmental impacts of the PDS.

Energy: The study tracks total primary energy consumed at each life cycle phase.
The DEAM1M database further decomposes primary energy into Renewable, Non-
Renewable, and Feedstock energy. Renewable and Non-Renewable refer to
energy generated from renewable fuels (hydroelectricity, wood and biomass) and
non-renewable fuels (coal, lignite, oil, natural gas or uranium) respectively.
Feedstock energy is the part of total primary energy that is embedded within used
material such as combustible fuel material. This level of detail was maintained in
this study and will be referred to in the discussion of results section.

Water use: The data category for total water use referred to the tracking and

aggregation of water used at each phase of the life cycle, measured by volume.

84

Impact categories

Two characterized impact categories were analyzed:

. Global warming potential: Global Warming Potential is the category indicator
measuring possible contribution by the PDS to the “greenhouse effect.” The
greenhouse effect refers to the ability of some atmospheric gases to retain heat
that is radiating from the earth. The measurement of GWP employed in this report
is based on the model compiled by The Intergovernmental Panel on Climate
Change (IPCC) (1995). The GWP index is defined as the cumulative radiative
effect between the present and a chosen time horizon (this study arbitrarily chose
20 years) caused by a unit mass of emitted gas, expressed relative to that for some
reference gas (IPCC and this study use C02). A single indicator is produced for
the greenhouse effect. in which:

E = Z GWP, * mi

6",,
l

where, for a greenhouse gas , m; is the mass of the gas released (in grams), and
GWP, is its Global Warming Potential, expressed in C02 equivalents. Although
the GWP calculation is one of the most accepted of all life cycle impact
assessment index methodologies (Udo de Haes, 2002), it is limited since GWP
assumes an even distribution of the gases being tracked, and indirect GWP
effects, such as the emission of one greenhouse gas leading to the formation of
another greenhouse gas, are not considered.

. Ozone depletion potential: Deterioration of the ozone layer allows more radiation

to reach the Earth’s surface, potentially destabilizing ecosystems as well as

causing adverse effects on agricultural productivity, human health and climate.

85

ODP is an impact category measuring possible contribution by the PDS to
deterioration of the stratospheric ozone layer. The most accepted ODP model,
developed by the World Meteorological Organization (Scientific assessment of
ozone depletion, 1998), is employed in this study. The impact category is
evaluated as follows:

Ozone Depletion = Z ODP, * m.
where, for an ozone depleting gas “i”, m is the mass of the gas released (in

milligrams), and ODP, is its Ozone Depletion Potential. Ozone depletion is

expressed in milligrams of CFC—l l.

[2) Life cycle inventory

In this section infomration and data sources associated with the life cycle
inventory analysis of the three drink delivery systems is presented including detailed
information on the systems components, material allocations, systems diagrams,
calculation procedures and assumptions and limitations. The data used in this study come
from a number of publicly available sources and can be divided into background and
foreground categories. Background data comprise the life cycle inventory data associated
with all the activities and processes involved in the drink delivery systems. This type of
information was obtained from the DEAMTM database developed by the Ecobilan
Consulting Group (Paris. France)(DEAM module databases and manuals, 1999). The
DEAM1M database represents industry activities by “modules” whose inventory data in
turn were gathered from a number of different sources (e.g. official reports, surveys,
engineering estimates, etc.). Each module is in fact the result of a “cradle to gate”

inventory analysis and includes inputs (e.g. bauxite from ground, water used, etc.) and

86

outputs (e.g. air, water and soil emissions) associated to the provision of a certain
intermediate product (e.g. PET resin production, aluminum ingot production, etc.) or
service (e.g. energy provision by the US electricity grid, transportation service of a 40-
ton truck, etc.). A product life cycle (i.e. cradle to grave) inventory analysis was modeled
by appropriately linking several of these modules. The linking of these modules was
made by using foreground data. Foreground data comprise data specific to a certain
product and its requirements (e.g. weight of a PET bottle, electricity requirement for PET
injection stretch blow molding, materials required for corn farming in PLA production,
etc.). Foreground data were obtained from a number of different sources and are detailed
in every case when they were used. Foreground data in this study were used provided
they were readily available (e.g. literature estimates. previous LCAs, etc.) and from a
reference source.

Additionally, besides foreground and background data, there are input data. Input
data were measured (e.g. weight of a PET and PLA bottles, aluminum can), estimated
using software and/or literature data for average operations (e.g. distribution packaging
formats and weights), or arbitrarily set (for all PDSs equal) with the spirit to represent a
relatively average operation involving drink products (e. g. transport distances).

DEAMTM modules are designed to be viewed with a spreadsheet like interface. In this
study all calculations and system representations were done by appropriately linking and

formatting Microsoft Excel spreadsheets (Microsoft Excel, 2000) containing the modules.

87

System
Material Production: The Material Production phase represents the material
production activities for primary packaging. The primary packaging materials were
detailed in Table 1.
. Polyolefins (PP, LDPE): The life cycle processes of the production of polyolefin
resins are illustrated in Figure 7 below. The production of these materials includes
Oil and Natural Gas Extraction, Petroleum Refining, Cracking of naphtha, and
Polymerization. The difference between PP production and LDPE production is
the intermediate produced in the cracking process. The interim monomer ethylene
(C3H4) is polymerized into polyethylene and propylene C3H6) is polymerized into
polypropylene.
The detailed system model of the material production process depends on the data
modules obtained from secondary data sources. The following Table 12 contains

the sources and brief explanation about the data modules used in this study.

88

£05.? 35355 BEBE Emu.— ccoBbom .5 2:3”—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mmoooa seventeen Ema ace—030m

man:
Al.
mm
.1. .............................................................. _
A .3838 “co—oxoom m :outhoEx—onm m
" mi $39: 3:832
A as; 2.8 m w m
" 52080 A] WA
“ ENDHW .
T mcoafiEo " m Quoumeoom + .355
. . u . $883 .Aweocm
ESE—om by. u u
M fence new mi
A mcofimmEo " 5308?». :0 m 53>?
ESE—om .533 u m

89

Table I2. Polyolefin resin production system data sources

 

Module Source

 

PP resin Background data:
production DEAMTM module name: Polypropylene: Production.l

 

 

LDPE Background data:
resin DEAMTM module name: Low Density Polyethylene (LDPE):
production Production. 1

 

 

 

Polyethylene terephthalate (PET): As shown in Figure 8, the production
subsystem of PET includes an oxidization and hydration processes and a poly-
condensation process, instead of the polymerization process shown in the PP and
PE production process. Table 13 contains the sources and explanation about the

data modules used in this study.

90

8293 539255 353E Ewen Ema .w ousmfi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

became can .533be :0 L855

223:5
ESE—om 53>?

 

 

 

$8.8m couozeoa Ewe Ema

HmE . .................................................................. .
A m :oummcoecoobom m
A 3538 83.80% m w m
m cove—ozficouanmxo MA 3:9: Eran—z 9
A u "
8%; nzom " u
" was—85 m
_ .+
. :8on .
A mcoamgo “ m Eoogmnofl + $55
ESE—om :< m m 3888 352m

me’K

Table [3. PET resin material production data source

 

Module Source

PET resin Background data:

production DEAMTM module name: Polyethylene Terephthalate (PET, Resin):
Production]

 

 

 

 

 

. Polylactide (PLA): As shown in Figure 9, the production system of PLA includes
corn production, wet milling, dextrose conversion into lactic acid, purification and
polymerization processes. Table 14 contains the sources and explanation about

the data used for this module.

Table 14. PLA resin material production data source

 

 

 

Module Source

PLA resin Foreground data:

production Corn farming and corn wet milling: (Gemgross, 1999; Shapouri et al.,
2002)

Polymerization: (Gemgross, 2000; Vink et al., 2003).

Background data:
Corn farming and corn wet milling:
DEAMTM module names
. Limestone (US, C3CO3)Z Quarrying]
. Nitrogen (US, N2): Production.2
. Potash (KCI): Production]
. Superphosphate (Normal): Production]
. Natural Gas (US): Production.2
. Natural Gas (US, Industrial Boiler): Combustion.2
. Petrochemical Feedstocks (US): Production.2
. Electricity (US, average): Production.2
9. Diesel Oil (US): Production.2
10. Diesel Oil (US, tractor): Engine combustion.2

OOQQ'JI-DWNF‘

 

 

 

92

 

E009? :28:on 3.023: Ema «in .o 05w?—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘J
<15
m cosztBEbom A1—
A 3.509: 020.30% m
m cougctsm 0:0
i r " 568250 200 0:03
0083 Bow m
A manual coume0c0w
BEE—om :< m 200 0:004
A 0:052:00 " wEEE
. . . 003 500
805:3 00:03 “
W as

mm000E 5:0:on Ewe <13

-
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
l
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
l
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I.

A

 

A

8:9: 3:032

 

A

A

0.005000% + $05

$0093 300cm

 

00:03

93

Aluminum: As shown in Figure 10, a standard practice to represent (in prepared
databases) the production system of aluminum cans is by “mixing” in a linear
fashion virgin and recycled-based aluminum production systems (Norris, 2003).
For the virgin system, it was assumed that the bauxite mining and alumina
production (which often happen near the mines) was done outside the US. since
in the US. almost one half of aluminum products are made from imported
alumina (Australia and Jamaica account for most of the imported alumina). Under
these assumptions, then, the electricity requirements of the DEAMTM alumina
production module was linked to the corresponding burdens related to the
Australian electricity grid (Ecobilan Group Inc., DEAM module databases and
manuals, 1999; Norris, 2003). However, transportation of alumina from Australia
was considered outside the scope of this study. The smelting of the alumina to
produce aluminum metal was assumed to happen in the US, so the DEAMTM
U.S.—based aluminum ingot production module was used. To represent the
burdens of the production based on recycled aluminum the DEAMTM U.S.—based
recycled ingot production module was used. The aluminum ingots are next
transformed into sheet material for later container manufacture process. Table 15
contains the sources and brief explanation about the data used for this module.
The inventory data for ingot casting in the US. was calculated by subtracting the
burdens from alumina production and the electricity requirement from the US.
electricity grid from the DEAMTM U.S. ingot-production module. The burden data

for aluminum for the manufacture of aluminum sheets was obtained by

94

subtracting the DEAMTM aluminum ingot production module from the DEAMTM

aluminum sheet production module.

Table 15. Aluminum material production system data source

 

 

 

Module Source

Aluminum Foreground data:

roll Alumina requirements for aluminum production (Life cycle

production assessment of aluminum, LCA of aluminum, 2003; Norris, 2003)
Background data:

DEAMTM module name
1. Aluminum (Al, ingot) production
2. Aluminum (Al, sheet) production
3. Aluminum (US. 100% primary, ingot) production
4. Aluminum (US, 100% recycled, ingot) production

 

 

 

95

2::
8:58.02

‘1

E000? 5302005 3:038 E=EEE< d— ouswfi

 

02350

+

90020000 835820 :3 mag

 

 

 

 

 

 

 

 

82 32:8. 008:
k cocoa—000m How”: 62:83;

 

 

 

 

 

 

mezzo E=EEB<

 

 

 

 

 

‘

3:036 00—0z00m

 

0:03 ozom

8:03.255 .08:
50:02:03 How—z E=EEE<

J

 

 

 

 

 

223:5
ESE—om 02

3:03.05 mO~_<

 

 

 

 

 

3:32»:

 

223:5
0:03:00 00:03

 

 

_ 5302008
»~_0_:00_m

 

 

 

$0008 cozosooa 2.8 E=EEE<

 

3:9: 23.—0:02

 

AA

--------------------.A.--"----_----------

039808 + $0020

$0083 300cm

 

00:03

96

Distribution 1: the Distribution 1 phase includes environmental burdens
associated with transportation of materials used for the manufacturing of the primary
packaging. The extent of Distribution 1 was the transportation between material
producers and the respective container component manufacturers. The burdens of fuel
production and fuel use were taken into consideration. The transport load is the sum of
the weight of materials for manufacturing of primary packaging (drink container) and the
weight of distribution packaging used for the transportation of these materials.
Distribution 1 does not include the production of the distribution packaging, since it is a
common practice to ship the materials for manufacturing in bulk and thus the use of non-
reusable distribution packaging is avoided. Each material was assumed to be transported
from the producer to the container manufacturer by truck. The same arbitrary distance
was used for all PDSs and the value are presented in the Assumptions section of this
Chapter. Any waste associated with the losses of carried materials during transportation
was assumed to be landfilled at a rate of 100%. The inventory model for truck
transportation includes both the production of fuel and the use of fuel for transportation.
Table 16 includes descriptions of the truck model used for the distribution 1 phase in this

study.

97

 

 

wctsfiflngE

E28800
00:30:90

 

E000? _. cocsnEma .2 00=wE

 

550$ 3230005506 80:09:00

 

 

A

3002.2:
00—050”—

 

A

00003
ozom

0058.80 8 3503 cogsnoa
Sec 3.0qu 00 cozmtommceb

 

 

 

 

‘

A

2235.0
ESE—om
h<

 

 

 

cougcoa Ban

 

 

 

2.060080
ESE—om
00003

mm00oE _ 5:2:me

.—
l
I
I
I
I
I
I
I
I
I
I
I
l
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
l
I
I
I
h

 

A

$29:
3:302

 

4

30030000
+ 300:0
$0083 300cm

 

00:03

98

Table 16. Distribution 1 system data sources

 

Module Source

 

 

Truck Background data:

transportation DEAMTM module names

1. Road Transport (Truck 40 t, Diesel Oil, kg.km).1
2. Diesel Oil (US): Production.2

 

 

 

Manufacturing: Energy requirements for the different manufacturing processes
were taken into account. The energy requirements were met by the US electricity grid.
Four models for the Manufacturing phase were created. The first three models are for the
production of the plastic container components: injection molding, which is used to
manufacture caps, injection stretch blow molding, which is used to manufacture bottles
and extrusion/lamination which is used to manufacture sleeves and carrier rings. The
other model is for the manufacture of the aluminum can body and end. Figure 12 and
Figure 13 show the manufacturing operations involved in the different PDS. Following

the illustrations, there is a detailed description of each model.

99

m mam ::0 _ mam :0 $0005 0:300:50: 30:09:00 550300 Q 05E":

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A m m
22:03:: 00—053: " H
Al 0:03 2.0m m mA 3::5 3:032
" :0_m::x0 mansi— A n .
A 0:060:80 m MA
ESE—0: :< M m 300530.: + >305
m mic—0:: :0m:00.a:_ mm A m 3000:: >w00:m
A 0:060:80 " m
ESE—0: :03? m mi 00:35
W wEE0E :0_:00_.:_ Em: A M
Alli” .m
$0005 " 3000:: m:c::0&::wE 3:0:0QE00 $50200 " 3000::
m 8:20:35 _ .......................................................... . _ 5:30:55

 

 

 

 

 

lOO

N mm: :0 83000:: 0:::0::::::: ::0:0:E00 $5850 .3 8:3:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A

m::::: 3:302

 

m 3020000: + >30:0

3.08:3 >30:m

 

:08?

 

 

 

. m m
5:088 :0_0>00m " u

m ill W

A 805:5 m n
2326: :2 m _

m 0:303:52: _ m

A 2050.5 m 2: as: :88 52:52.: "
203:0: :03? u m

m 3000:: 3:308:52: m

3000:: AITI 3:0:0:::00 :0:::::0U m

m 8:23:35 .. ....................................................... .

 

 

 

3000::
: 8:32:55

 

 

 

101

. Film Extrusion: as shown in Table 17, the process requirements for blown film
extrusion process were obtained from the SAEFL study (LCI for packagings. Vol.
11., 1998b). The published energy requirements (i.e. electricity) for Swiss plants

was converted into US. environmental burdens using US. specific DEAMTM data

 

 

 

modules.
Table 17. Blown film extrusion process data source
Module Source
Blown film Foreground data:
extrusion SAEFL report: Blown film extrusion energy requirements (LCI for

packagings. Vol. II., 1998b)

Background data:
DEAMTM module name: Electricity (US, average)

 

 

 

. Injection molding: As shown in Table 18, process requirements for the film
extrusion process were obtained from the SAEFL study (LCI for packagings. Vol.
11., 1998b). The published energy requirements (i.e. electricity) for Swiss plants
were converted into US. environmental burdens using US. specific DEAMTM

data modules.

Table 18. Injection molding process data source

 

 

 

Module Source
Injection Foreground data:
molding SAEFL report: Injection molding energy requirement (LCI for

packagings. Vol. 11., 1998b)

Background data:
DEAMTM module name: Electricity (US. average)

 

 

 

. Can body and lid manufacture: process requirements were obtained from a
Mitsubishi Materials Corporation LCA study. The published energy requirements
(i.e. electricity) for a Japanese aluminum container plant (Mitsubishi Materials
Co., LCA of aluminum can, 2003) were converted into US. environmental
burdens using US. specific DEAMTM data modules. More details are contained in

Table 19.

Table 19. Can body and lid making processes data source

 

Module Source

 

 

Can body and Foreground data:

lid making Mitsubishi Materials Corporation - Registration number: S-EP00022
— l to 12, Energy requirements for can making (no. 8) (LCA of
aluminum can, 2003)

Background data:
DEAMTM module name: Electricity (US, average)

 

 

 

Distribution 2: the Distribution 2 model includes environmental burdens
associated with transportation of manufactured container components to the filling
facility. The burdens of fuel production and fuel use were taken into consideration. The
transport load is the sum of the weight of components used in the filling process and the
weight of distribution packaging used for the transportation of these materials.
Distribution 2 and 3 include the production of the distribution packaging. Each
component was assumed to be transported from the producer to the container
manufacturer by truck. Any waste associated with the losses of carried materials during
transportation was assumed to be landfilled at a rate of 100%. The inventory model for

truck transportation includes both the production of fuel and the use of fuel for

103

transportation. Table 20 and Table 21 include descriptions of the truck and distribution

packaging model used for the distribution 1 phase in this study.

104

 

3000::
0:00:

 

 

 

 

3000:: N 0003535 .3 08me

 

 

 

 

 

_0._:0:08 00.983:

 

0303 20m

 

3:23:80
808:0: ::<

 

3:23:80
808:0: :80?

———-----——--—-—-—---——--—---—--—--r-n-nnnu

 

3000:: w0_=_:

0: 52:00: w0_:800:::08
80:: 3:000:88

:00 03000 :0 0000203005.

 

8000:800 :080800

8080080008

 

 

 

I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I

 

 

 

 

3500: 808008008
8000:800 :00 :0800 0: 52:00:
w0._:800::008 80:: w8w0x00:

02805320 :0 0000:0300;

 

 

 

 

 

00:00:08:
30.:

 

 

00:00:08: w8w0x00:

 

803020

 

 

3000:: m 00:80:35

..
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I_

A

a

A

 

08:8 3:302

 

+

3003000: + 3:000
3008:: 3:05—

 

:0003

105

Table 20. Distribution 2 process data source

 

Module Source

 

 

Truck Background data:

transportation DEAMTM module names

1. Road Transport (Truck 40 t, Diesel Oil, kg.km).1
2. Diesel Oil (US): Production.2

 

 

 

Table 21. Distribution 2 process data source - Distribution packaging modules

 

 

Module Source
Corrugated Background data:
board DEAMTM module name: Corrugated Cardboard (Recycled Fibers):

Production. 1

 

Stretch wrap Background data:
DEAMTM module name: Low Density Polyethylene (LDPE,
Film): Production]

 

 

Wood Background data:
LCI of kiln-dried lumber (WW PA, LCI of Western Wood Prod.
Assoc., 1995)

 

 

 

Filling: Injection stretch blow molding: process requirements for injection stretch
molding process were obtained from the SAEFL study (LCI for packagings. Vol. 11.,
1998b). The published energy requirement (i.e. electricity) for Swiss plants was
converted into US. environmental burdens using US. specific DEAMTM data modules.

More details are in Table 22.

106

Table 22. Injection stretch blow molding process data source

 

 

 

Module Source
Injection Foreground data:
stretch blow SAEFL report: Injection stretch blow molding energy requirement (LCI
molding for packagings. Vol. II., 1998b)
Background data:
DEAMTM module name: Electricity (US. average)

 

 

Distribution 3: Distribution 3 model includes environmental burdens associated
with transportation of the drink container from the filling facility to the
distribution/retailer. The burdens of fuel production and fuel use were taken into
consideration. The transport load is the sum of the weight of components used in the
filling process and the weight of distribution packaging used for the transportation of
these materials. Distribution 3 includes the production of the distribution packaging. Each
component was assumed to be transported from the producer to the container
manufacturer by truck. The same arbitrary distance was used for all PDSs and the value
are presented in the Assumptions section of this Chapter. Any waste associated with the
losses of carried materials during transportation was assumed to be 100% landfilled. The
inventory model for truck transportation includes both the production of fuel and the use
of fuel for transportation. Table 23 and Table 24 include descriptions of the truck and

distribution packaging model used for the distribution 1 phase in this study.

107

 

 

5030000
000000

 

 

 

 

3000:: m 00:00:35 .m: 80%:

 

 

 

 

 

A

00:80:: 0000080:

 

0303 000m

 

30003000
:00:0:0: :2

 

308380
:00:0=0: :90?

-------_-----—-—-- -----------—-----J*-----‘--

 

:0:00::3_0\:0:0:0:

9 0000:

00:00 ES: :00: .000 +0
8000:: :0 00:80:80;

 

 

 

0000: 05:0 0:080 9 00:00:

m0::0:00:0000: w0mw0x00:
00:00:30 0:8: w0:w0:_00:
00:00:03.6 :0 00 :0::0:m00: H

 

 

 

 

 

00:03.8:
wEm0v—00: 00:00:35

 

 

 

 

 

00:0000::
_00n_

 

 

3000:: m 00:00:35

A

 

0500
00000::

 

 

b

 

A

30:0: 3:802

 

A

0003000: + 3:000

30083 3:05:

 

a

:0003

108

Table 23. Distribution 3 process data source - Transportation modules

 

 

 

Module Source
Truck Background data:
transportation DEAMTM module names

1. Road Transport (Truck 40 t, Diesel Oil, kg.km).1
2. Diesel Oil (US): Production.2

 

 

 

Table 24. Distribution 3 process data source - Distribution packaging modules

 

 

Module Source
Corrugated Background data:
board DEAMTM module name: Corrugated Cardboard (Recycled Fibers):

Production. 1

 

Stretch wrap Background data:
DEAMTM module name: Low Density Polyethylene (LDPE,
Film): Production.1

 

 

Wood Background data:
LCI of kiln—dried lumber (WWPA, LCI of Western Wood Prod.
Assoc., 1995)

 

 

 

End-of-life: End-of-Life process modeling accounted for the environmental
burdens that stemmed from waste treatment of used containers. It was assumed that all of
drink products produced in the drink Filling phase were consumed at market and all the
containers were sent to a recycling site and/or a waste treatment site for incineration
and/or landfilling. Table 25 contains details of the modules used for the end-of-life

processes.

109

83000:: 0.::.:0-:0m .0: 0:03:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m 00:85:30; m
m @0300: m
A J" AE0EEE0 :0 u
00208 80:80 " :00: 0090: m A
u % " 80:0: 3:202
A 0:020: m
0303 2.0m “ w0=0>00m A m
m 2092002 "A
A u m 0:003:00: + 3:000
300:3:00 “ 3:050: " 3000::v 3:00m
502.2: :2 m u
m +|| r
A 3:23:50 " wE—EfiEMA W 5:03
203:0: :80? m m
m 00:0:0500: A . M
m m :00:0:000
m m 0m0:0>00 :033

110

Table 25. End-of-life process data source

 

 

 

 

 

 

Module Source
Incineration Background data:
(plastic) DEAMTM module name

Plastic (US, with energy recovery): Incineration.2
Incineration Background data:
(aluminum) DEAM1M module name:

Aluminum (US): Incineration.
Landfilling Background data:

DEAMTM module name:

Landfilling without energy recovery (US, MSW average).2

Recycling Background data:

 

DEAMTM module names
1. Commingled recyclables with plastic recovery
2. Commingled recyclables with aluminum recovery
3. Road Transport (Truck 16 t, Diesel Oil, kg.km).1
4. Diesel Oil (US): Production.2

 

lll

 

 

00:35:20;

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00:00:00:
3:300: 00m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00:00—08:
E0500: 0:305 .. _0_:0:00:
:05 030:5
>580:
..--.. @0030: 1 00:00:08:
. .
m 392002 00:80:: 5:00:
_
w0:0»00m j 00:00:00:
1 mime—00:
00:20:60: A 00:00:35
I 00:35:30.0;
0023 4
4
0:::O :0m ll.— m 0o:0£:3_Q
:0E03000 m
\:0:0:0m 0::: 200m
00:00:35

 

 

 

 

 

00:00:08:
00:80:: xom

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00:00:08:
T 3:200:
:83 00:0:w
00:00:08:
1 3:300: 00:0:
¢
00:00:08
1 w0mw0xo0:
00:00:35
00:80.0 £000 :0:
1 0000000300; 0000000300;
N0030£E£Q A. :300_:0£bm:n_ Al]
Al 00558030000 300:0000::
J w0:\0>003 3:302
\:00\0::0m m0_:\0>00_3
08:00:00

 

 

 

 

 

 

 

 

 

 

 

 

 

 

005%:50 20%:01083

112

 

 

00:85:30;

 

 

 

E0500: E0035?

 

 

.0032 555520: N 30: :o 50320 32: .0030: .0: 2.00::

 

00:8:
m0:0>00m

 

 

W _0::0:00E

 

 

w0:0>00m A;

 

 

 

 

00:0:0505

 

 

 

 

 

 

0023 4

 

<

 

‘

 

00:00:0::
3:800: 00m

 

 

00:00:0::

.. 3:800:

:00,» 0000::m

 

 

 

00:00:0::
0:30.: :0:0:

 

 

 

 

 

00:00:0::
w0.:w0000:
00:00:35

 

 

 

00:38:30;

 

 

 

 

 

 

 

 

 

 

00:0 20 AI—

0 80:50:00

 

 

 

 

 

 

 

 

50:03:00
02000
0502500

 

 

0r

 

¢

 

¢

 

 

00:00:0::
0:205 00m

 

 

 

 

00:00:0::
3:200:
:00: 0000::m

 

 

 

 

00:00:0::
_0:0:00: 0:0:

 

 

 

 

00:00:0::
w0:w00_00:
00:00:35

 

 

 

 

00:0::o:300::.

 

 

 

 

 

0 80:00:00

 

00:000.: 0000 :0:
00:00:09,300;

1L

 

 

 

 

 

 

 

 

 

 

 

05:0 :00

 

 

 

 

 

_ 22:50:00 I|_

 

 

300:00:0::
:0:0:0E

 

 

w0:0:00.:000E
0:00.00

0:0 0:00
0:050:20

 

 

%:0:00003
:00 0:00: ::

 

 

 

00300 5% 0 3:: 0:0 630:: 00:: : 30: :0 0.2000 3o: 00222 .2 2:00

113

Allocation procedures

Allocation procedures are used to partition inputs and/or outputs within a product
system. An allocation procedure is required when a unit process within a system shares a
common pollution treatment infrastructure or where multiple products or co-products are
produced in a common unit process. The system on which this research focused was a
product delivery system so it did not produce any co-product, except for the electricity
generated at municipal solid waste management sites. However, several allocation rules
were adopted in each phase: the following sections discuss them phase by phase.

. Material Production:

Co-Product Allocation: Since all the input/output data in the Material Production
phase were obtained from secondary data sources, the allocation rules were defined for
these data modules in the sections above.

Recycling Allocation: As described above. recycling allocation rules also depended on
the data modules from secondary data sources whose information was unavailable and
out of the scope of this study

. Distribution 1:

Co-Product Allocation: There was no co—product from the Distribution 1 phase.
Recycling Allocation: There was no recycling allocation in the Distribution 1 phase.

. Manufacturing:

Co-Product Allocation: Since all the input/output data in the Manufacturing phase

were obtained from secondary data sources, the allocation mles were defined for these

data modules in the sections above.

114

Recycling Allocation: As described above, recycling allocation rules also

depended on the data modules from secondary data sources.
. Distribution 2:

Co-Product Allocation: There was no eo-product from the Distribution 2 phase.

Recycling Allocation: 100% of the corrugated cardboard boxes used in
Distribution 2 was assumed to be sent to a recycling facility to be reused. However, since
such boxes are reused in an open loop process (i.e. outside the PDS), the mass of the used
boxes was classified in separately as recycled material. Pallets instead were assumed to
be reused and some damaged. Expert estimations (White, 2005) assume that wooden
pallets are used for about 16 roundtrips on average. with a reuse rate of about 94%.
Similar operations (Keoleian, 2001) make the assumption that the remaining 6% of
pallets are damaged. Then, it was assumed that these damaged 6% of total pallet were
replaced with new pallets in each delivery cycle, and this is what finally was counted as
pallet material (wood) input to the distribution phase under study.

. Filling:

Co-Product Allocation: Since all the input/output data in the Filling phase were
obtained from secondary data sources, the allocation rules were defined for these data
modules in the sections above.

Recycling Allocation: As described above, recycling allocation rules also
depended on the data modules from secondary data sources.

. Distribution 3:

Co-Product Allocation: There is no co-product from the Distribution 3 phase.

115

Recycling Allocation: Allocation rules within the Distribution 3 phase were

exactly the same as within the Distribution 2 phase.
. End of life:

Co-Product Allocation: Some incinerators have energy recovery devices that
generate electricity by incinerating wastes. Electricity generated in this way at the End-
of-Life phase was treated as a co-product, and the environmental burdens that would
otherwise occur with normal electricity production were subtracted from the total
environmental burden of the End-of-Life phase.

Recycling Allocation: There are several materials that were converted from solid
wastes to reusable materials. In this study, it was assumed that some of the PET, PLA and
aluminum is sent for recycling. For sake of simplicity, the rest of the materials were

assumed to have a zero percent recycle rate.

Calculation procedures

This section explains the procedures by which inputs and outputs generated from
each unit process were calculated. All the calculations of inputs and outputs from each
unit process were based on the mass of product output from each unit process. For this
reason, it was important to determine how much product output each unit process
generates. The following section describes (1) how the product output from each unit
process was determined and (2) how the inputs and outputs (burdens) from each unit

process associated with the product output was calculated.

“.6

Functional Unit and Losses

The functional unit of this system is 1,000 L of drink delivered to wholesale
retailers. Figure 14 illustrates how the product output from each process unit was
calculated considering various losses that occurred in the downstream unit processes. To
facilitate calculations, loss in this study is expressed as the fraction of desired process

unit output that is lost in the process.

117

0:05? 303.0: 00300:: 00: 0: 8000.0 .2 0:0mE

  

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00:80 A 00 0 >
:0 _ 89 0m 000— 000005 A 00 0 00 0
000— 000080
002 00000::
n ............. _ =0 0+0: 0: $000+:
" 0000800000 " A03. 9+CAO+S
0:: :0 :50 Ar“ .002 H w0:=E Kl, m0:0:00.:0008 , 00:00:00.5
u. ............. N 00.00000 0:802 0m AN+CAO+ C
$000+:
Q+§0+C
0 803.020 5| 080350 22:30:00 :0 0+0
0020+: 6+ :0? v
50+: 0003+: _ , :
A0+CQ+C
000 00% 00::
000. 00:00:00.5 « E 000— 00:00:05 ‘08— 00:00:05

118

For example, if a distribution loss fraction at Distribution 3 was defined as 2, the
drink Filling unit process has to produce 2 additional functional units of drink as its
product output, because 2 out of 1,000 L is to be lost before it reaches retailers. The
product output from each process unit was calculated by multiplying each of the down
stream loss fractions plus unity by the functional unit, as shown in Figure 19.

The Material Production, Manufacturing and Filling phases included losses. The
loss fraction “a” for Material Production is embedded in the individual material
production DEAMTM modules. The loss fractions “b” and “c” (in Figure 19) were defined
as single fractions for the container manufacturing and filling unit processes respectively
and details of the values are in the secondary data sources used (SAEFL, LCI for
packagings. Vol. 11., 1998b). In all cases, “x”, “y” and “2’ fractions were assumed zero.
Once the product output was determined by the formulas described above, total.
environmental burdens from each unit process were calculated based on the mass of
product output. The following sections describe calculation procedures for all phases in
order from downstream to upstream.

To obtain the mass of drink containers sent to municipal solid waste (MSW)
management site, it was assumed that the entire functional unit of drink products
delivered to wholesale retailers was all carried to the MSW facility. In other words, a
zero loss fraction after Distribution 3 phase was assumed. The total mass of waste was
classified into three categories; MSW sent for incineration, MSW sent for landfilling, and
material that is sent to a municipal recycling facility. The total inputs and outputs were

calculated by multiplying the mass of each category (i.e. mass for incineration) with mass

based environmental burden data, corresponding to each category.

119

In Distribution 3, three categories were taken into consideration to aggregate the
total environmental burdens of this phase. They are: Transportation, Material production
of distribution packaging, and Solid waste

Transportation: In order to calculate the burdens of transportation, total km-kg

was calculated by aggregating the following pairs of weights and distances.

 

 

 

 

 

 

 

Weight Distance

Drink product Filling facility to retailers/distributors
Drink containers Filling facility to retailers/distributors
Distribution packaging Filling facility to retailers/distributors
Distribution packaging Supplier to Filling facility

 

 

The weight of drink product consisted of the weight of a functional unit of drink
and the weight of the drink containers. Distribution packaging included
corrugated boxes, pallets, and stretch wrap.

Material Production of Distribution Packaging: The burdens of the material
production were calculated by multiplying the weight of materials used for the
distribution packaging and each material’s burden data obtained from secondary
data sources. The materials included corrugated board for corrugated box, wood
for pallet, and LLDPE film for stretch wrap.

Solid Waste: The masses of solid waste from Distribution 3 were calculated in the
two categories described below.

Loss from transportation: A transportation loss fraction was allowed and
aggregated by multiplying it with the weight of drink carried by truck.
Distribution packaging: Some of the distribution packaging was assumed to be

recycled in an open loop and so counted as a Recycled Material output. The rest

non-recycled distribution packaging was landfilled. Total burden was calculated

by multiplying the mass of waste by the DEAMTM MSW Landfill module data.

In distribution 2, three different categories again were included. They are:

Transportation, Material Production and Solid waste.

Transportation: In order to calculate the burdens of transportation, total kg-km

was calculated by aggregating the products of the following weights and

distances.

 

Weight

Distance

 

Container components

Container comonent manufacturers to filling facility

 

Distribution packaging

Container component manufacturers to filling facility

 

 

Distribution packaging

 

Supplier to Container manufacturers

 

The distribution packaging consisted of corrugated boxes, sleeves, separator

cardboard, pallets and the stretch wrap.

Material Production of Distribution Packaging: Calculation procedure was the

same as for Distribution 3

Solid Waste: Calculation procedure was the same as for Distribution 3

The mass of all materials required for the functional unit was calculated based on

the weight of product output and composition data of each container component.

Total environmental burdens were calculated by multiplying the mass of each

material and corresponding data modules.

Loss from transportation: Calculation was the same as for Distribution 3.

121

 

Assumptions and limitations of this study

General

Burdens associated with distribution phases were based on the assumptions and
calculations of the DEAMTM Road Transport module (Truck, 40 ton capacity,
Diesel, kg/km, 100% efficiency).

Wooden pallets were assumed to be reused until worn or damaged, at which time
they were typically rernanufactured. The re-manufacturing process entailed
replacing only the worn or damaged components of the pallets. Therefore,
burdens associated with the production of pallets were calculated using only the

quantity of wood consumed per functional unit.

Material Production

The environmental burdens associated with material production of PP, LLDPE,
PET, and PE were based on European production data. Material production
burdens for Europe were assumed to be similar to those in North America.

No additives or fillers were accounted for in the production phase.

The environmental burdens associated with PLA polymerization were based on

estimates by the manufacturer (Vink et al., 2003).

Distribution '1

All materials were assumed to be shipped in bulk trucks and therefore did not use
distribution packaging.
40-ton trucks were assumed to be, on average, full on the delivery leg of the trip

and empty on the return leg of the trip.

122

Manufacturing

The PLA injection stretch blow molding requirements were assumed to be the
same as for PET per kg basis, but the weight of the PLA bottle was assumed to be
6% lighter than that of PET.

The loss fractions for blown film extrusion, injection molding and injection
stretch blow molding were based on estimates from SAEFL (LCI for packagings.
Vol. 11., 1998b).

The loss fraction for the can making process was assumed to be 0.01.

Distribution 2

Filling

All distribution packaging was assumed to be shipped by truck from the supplier
to the manufacturer.

Trucks were assumed to be full on the delivery leg of the trip and empty on the
return leg of the trip.

It was assumed that damaged pallets were replaced with new pallets in each
delivery cycle, and this is what finally was counted as pallet material (wood) input
to the distribution phase under study.

Distribution packaging formats for PLA containers was assumed to be identical to

the distribution packaging currently being used for PET containers.

The PLA and PET containers were assumed to be similar in design, and therefore
existing filling. sealing and palletizing equipment can be used. Requirements for
Injection stretch blow molding, which was included in this phase, were assumed

identical for both PLA and PET.

Distribution 3
. Trucks were assumed to be, on average. full on the delivery leg of the trip and
empty on the return leg of the trip.
. The same distribution packaging was used for PDS 1 and PDS 3.
End-of-Life
. All wastes produced during the Material Production, Manufacturing, Filling and
Distribution phases (all except End-of-Life) was assumed to be sent to landfill.
. One hundred percent (100%) of the containers shipped to the
distributor/retailer/consumer unit process in Distribution 3 are assumed to reach
the End-of-Life phase. In other words, none were assumed to be disposed of

improperly or maintained by the consumer to be reused for other purposes.

Input data

Besides foreground and background data, there are input data. Input data were
measured (e.g. weight of a PET bottle, aluminum can), estimated using software and/or
literature data for average operations (e. g. distribution packaging formats and weights), or
arbitrarily set (for all PDSs equal) with the spirit of representing a relatively average
operation involving drink products (e. g. transport distances).

Input data used in this study is included in this section. The details about the
variable input data (i.e. data were varied during the uncertainty analysis) is postponed for

discussion the next chapter.

0 Material production phase inputs

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PDS For component Material Mass per package, Kg Recycled content, %
resin

1 Bottle PET Variable 0
Cap PP 0.0025 0
Sleeve LDPE 0.0005 0
Ring LDPE 0.002 0

2 Can Aluminum Variable Variable
Ring LDPE 0.002 O
3 Bottle PLA Variable 0
Cap PP 0.0025 0
Sleeve LDPE 0.0005 0
Ring LDPE 0.002 0

0 Distribution 1 phase inputs
PDS For component Material Transportation distance, Loss fraction
resin Km

1 Bottle PET 400 0
Cap . PP 400 0
Sleeve LDPE 400 O
Ring LDPE 400 0
2 Can Aluminum 400 O
Ring LDPE 400 0
3 Bottle PLA Variable 0
Cap PP 400 O
Sleeve LDPE 400 O
Ring LDPE 400 O

 

 

 

 

 

125

 

 

0 Container manufacturing phase inputs

 

 

 

 

 

 

 

 

 

 

 

 

PDS For Material Process Electricity Loss Information
component resin requirement, fraction source
MJ/Kj
1 Bottle PET Injection Variable 0.0049 (SAEFL, LCI
molding for packagings.
Vol. II., 1998b)
Cap PP Injection 3.51 0.01 (SAEFL, LCI
molding for packagings.
Vol. 11., 1998b)
Sleeve LDPE Conversion 1.29 0.04 (SAEFL, LCI
for packagings.
Vol. 11., 1998b)
Ring LDPE Conversion 1.29 0.04 (SAEFL, LCI
for packagings.
Vol. II., 1998b)
2 Can Aluminum Can 12.67 0.01 (Mitsubishi
making Materials, LCA
of aluminum
can, 2003)
Ring LDPE Conversion 1.29 0.04 (SAEFL, LCI
for packagings.
. Vol. 11., 1998b)
3 Bottle PLA Injection Variable 0.0049 (SAEFL, LCI
molding for packagings.
Vol. H., 1998b)
Cap PP Injection 3.51 0.01 (SAEFL, LCI
molding for packagings.
Vol. II., 1998b)
Sleeve LDPE Injection 1.29 0.04 (SAEFL, LCI
molding for packagings.
Vol. II., 19989
Ring LDPE Injection 1.29 0.04 (SAEFL, LCI
molding ' for packagings.

 

 

 

 

 

 

Vol. II., 1998b)

 

126

 

0 Distribution 2 phase inputs

Inputs for primary packaging transportation

 

 

 

 

 

 

 

 

 

 

 

 

PDS For component Material Transportation distance, Loss fraction
resin Km (assumed) (assumed)

1 Bottle PET 400 0
Cap PP 400 0

Sleeve LDPE 400 0

Ring LDPE 400 0

2 Can Aluminum 400 0
Ring LDPE 400 0

3 Bottle PLA 400 0
Cap PP 400 0

Sleeve LDPE 400 0

Ring LDPE 400 0

 

 

 

 

 

127

 

Inputs for distribution packaging transportation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PDS For Distribution Component Weight Transporta— Loss Infor-
eomponent packaging units per of unit, tion fraction mation
dist. Kg distance, (assumed) source
packaging Km
unit (assumed)

I, 3 Bottle Corrugated 12000 3 100 0 (Giles.
preform box 1999)
Stretch wrap 12000 0.32 100 0 (White.

2005)

Wood pallet 12000 23.7 100 0 (Giles.

1999)

Cap Corrugated 20000 0.8 100 0 (Giles,
box 1999)
Stretch wrap 20000 0.32 100 0 (White,

2005)

Wood pallet 20000 23.7 100 0 (Giles,

1999)

Sleeve Corrugated 38500 0.8 100 0 (Giles,
box 1999)

Stretch wrap 385000 0.32 100 0 (White,

2005)

Wood pallet 385000 23.7 100 0 (Giles.

1999)

Ring Corrugated 38500 0.8 100 0 (Giles,
box 1999)
Stretch wrap 385000 0.32 100 0 (White,

2005)

Wood pallet 385000 23.7 100 0 (Giles,

1999)

2 Aluminum Pads 36] 0.5 100 0 (Giles,
can body 1999)
Stretch wrap 7942 0.32 100 0 (White,

2005)

Wood pallet 7942 23.7 100 0 (Giles,

1999)

Aluminum Sleeves 35 0.25 100 0 (Giles,
can lid 1999)
Stretch wrap 3500 0.32 100 0 (White,

2005)

Wood pallet 3500 23.7 100 0 (Giles,

1999)

Ring Corrugated 38500 0.8 100 0 (Giles,
box 1999)
Stretch wrap 385000 0.32 100 0 (White,

2005)

Wood pallet 385000 23.7 100 0 (Giles,

1999)

 

 

 

 

 

 

 

 

128

 

0 Filling inputs

 

 

 

 

 

 

 

 

 

 

 

 

PDS For Material Process Electricity Loss Information
component resin requirement. fraction source
MJ/Kg
1 Bottle PET Injection 8.40 0.001 (SAEFL,
stretch LCI for
blow packagings.
molding Vol. II.,
1 998b)
Cgp PP N/A N/A N/A
Sleeve LDPE N/A N/A N/A
Ring LDPE N/A N/ A N/A
2 Can Aluminu N/A N/A 0.001 Assumed
m
Ring LDPE N/A N/A 0
3 Bottle PLA Injection 8.40 0.001 Assumed
stretch
blow
molding
Cap PP N/A N/A N/A
Sleeve LDPE N/A N/A N/A
Ring LDPE N/A N/A N/A

 

 

 

 

 

 

 

129

 

0 Distribution 3 phase inputs

 

 

 

 

 

 

 

 

PDS Distribution Product Weight Transp. Loss Information
packaging units per of unit, distance, fraction source
dist. Kg Km (assumed)
packaging (assumed)
unit
1, 3 24 0.26 100 0 (CAPE, CAPE
Corrugated Systems:
board tray palletization and
package design
software, 2000)
1296 0.32 100 0 (White, 2005)
Stretch wrap
1296 23.7 100 0 (CAPE, CAPE
Wood pallet Systems:
palletization and
package design
software, 2000)
2 24 0.26 100 0 (CAPE, CAPE
Corrugated Systems:
board tray palletization and
package design
software, 2000)
1296 0.32 100 0 (White, 2005)
Stretch wrap
1296 23.7 100 0 (CAPE. CAPE
Wood pallet Systems:
palletization and
package design

 

 

 

 

 

 

software, 2000)

 

130

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SOON 000$: :00 lw0:__.::001_
0%: H0.: 2.: s as; 2.8 002550: 0.0: 0 <2 E
Doom 00:0)”: :00 00:0:0500:
as: ”.00 05 5 as; 0:8 092.5250 0.0: 0 <2 :0
28.0. 00:03: :00 E0000:
08: ”0.0 20 E 233 0:8 aegiszfim .000 0 <2 0 050
A SON 00:09: :00 w0_=:.::00x:
08: 00.0 20 s as; 0:8 Recaszfim .000 0 <2 0:
doom 005%.: :00 00:0:0500:
08: “0.0 2:: 5 as; 0:8 0005.250 .000 0 <2 40
A Son 00:09: :00 @0003:
as: ”0.0 20 5 233 0:8 0000:3200 .00: 0 <2 0 0520
Soon 005%.: :00 w0_=_.::00.._
300.: 0WD 0:: 0: 00003 :28 39200:)??5 .md: o <\Z on
A SON 02%: :00 00:0:0060:
800.: 0WD 0.: 0: 0:003 ::0m EQ:0_00E<mm .m.Dv o <\Z 0N
A :80 .83»: Ba 0380
08: ”0.0 20 a: as; 0:8 09250200 0.00 0 <2 0 0“0
3:00.03
0 «<2 0.00:0>
Zoom 080%.: :00 00:0:0500:
28.: U0.0 2: 5 as; 2.8 022524.00 .000 0 <2 .02 _
50003:
0 O2 0.00::0> 0.00m m .—
A:00:0mm0v 3000003
00:00: 00:00:: EV: 000002:
00:00:80: 000.: 8:003:00; $000080; 0:00.00 40m 00009000 mam

 

 

 

 

 

 

 

 

02:5 0:: .6 20 o

 

 

 

 

 

 

 

 

:80 02:00 25 050023
0000: 00.: 05 0: 0:00; :08 00920020005 0.2: 0 $2 0:.
A Sea 00:03: :00 0000:0060:
000: 0.0.3 05 E 0003 :08 :00_0_002<mm 0.3: 0 $7: vm
Coom 00:09,: :00 20300:
000 H0.0 2: s 0003 0:8 0060:0200 0.2 0 <2 0 050
05:00:00
0 $2 03000:»
zoom 00:00.: :00 0000:0500:
0000: ”.05 0.0 E 0:00)» 0:00 BEBEZSE 0.3: 0 <2 00:
20.500: 0:0: 000
o 8. 030:; 05.0052 N
9000:0000: 3080000:
00:08 00:00:: EV: 000000:
00:08:80: 003 0200:0300; 0w00000:0n: E0000 40m: 000009000 man:

 

 

 

 

 

 

 

 

132

III. Preliminary conclusions

This chapter has presented detailed information regarding a comparative life cycle
assessment (LCA) based on a hypothetical drink product for which we evaluate three
main container material alternatives: (1) a polyethylene terephthalate (PET) bottle; (2) an
aluminum can; and (3) a polylactide (PLA) bottle.

It has followed ISO’s LCA steps, and details and information sources used for this
comparative LCA were presented. Representative realistic implications of the use of the
three main material/container alternatives and their actual packaging operations,
production and use of distribution packaging (i.e. corrugated trays, stretch wrap, and
pallets), as well as transportation steps through several parts of the packaging life cycle
were included. Relevant end-of-life scenarios such as landfilling, incineration and
recycling were also taken into account in this analysis. Each of the main material
alternatives (i.e. PET, PLA, aluminum) along with its whole set of operations was
defined as a product delivery system (PDS) and environmental burdens associated with
each of the three PDSs were included in the study.

As shown, the procedure for performing an actual LCA according to [80’s
recommendations is both time consuming, and greatly data intensive and heavily relies
on gathering of quality data. More significant is the fact that, in spite of the effort made in
this study to identify quality sources of data and represent production systems in as
standard a form as possible, this comparative LCA is still a streamlined version and thus,
its results and conclusions need to be handled acknowledging such limitation.

The actual results and discussions of the LCA presented here are in the next

chapter along with the actual uncertainty estimation.

133

CHAPTER 4 UNCERTAINTY ANALYSIS OF THE COMPARATIVE LIFE CYCLE

ASSESSMENT OF DRINK DELIVERY SYSTEMS

I. Introduction

Uncertainty analysis is something still new in LCA, as discussed in Chapter Two.
Just in the last years a number of methodologies for uncertainty estimation have been
proposed. Currently, the most favored approach for uncertainty and variability analysis in
LCA is incorporating it into the life cycle inventory (LCI) stage. In fact, the study of
uncertainty in the inventory data is a natural inclination since LCI is the actual
quantitative analysis of an LCA and relies heavily on the quality of the information
regarding inputs of raw materials and fuels into a system and the outputs of solid, liquid
and gaseous wastes from it. At this stage. data associated with the flows is collected using
literature studies, interviews, measurements, theoretical calculations, data banks and
qualified guesses. In theory, the application of allocation principles and procedures
should also be explained, and information required in recycle or reuse situations should
also be presented. However, this is not the reality in most cases, and some LCA experts
even argue that since there are no standards for co-product or recycling allocations of
environmental emissions, it is impossible, for instance, to look at an off-the-shelf
computer LCA program and understand how the numbers were generated. This is a
serious problem for the credibility of the LCA method since, with the idea of user
convenience and to speed the process of performing an LCA, it is not uncommon to find
private groups developing computer programs using prepared databases which are, in

turn, compilations of other external data sources. Moreover, leaving the uncertainty due

to variation in allocation procedures aside (which can be up to 50%), Finnveden and
Lindfors (1998) found that apparent mistakes in the reporting of inventory data can be
real and more detrimental. In fact, mistakes in LCI data are not uncommon and have been
found to explain some very large variations in emission data (sometimes orders of
magnitude different). On the other hand, data variations (large or small) can be real since
they may simply reflect natural variations in emission data or differences in types of
technologies.

But, as explained earlier as well, uncertainties in LCA are not restricted to LCI.
Very commonly and with great consequences in packaging, uncertainties due to the
analysis of different scenarios may be the reason for conflicts in LCA results. For
instance, differences in percent recycled content in a packaging material. packaging end-
of-life practices (e.g. landfilling, incineration, recovering), and product distribution
distances all can greatly change the outcome of the LCA. Even within the same country,
the same packaging end-of—life alternative can be viewed both as environmentally
friendly and environmentally unsound. In fact, analyzing several comparative packaging
based LCA results, it is not uncommon to find directly opposite results that are often the
subject of heated debates among industry groups and even countries (IFEU, World
largest PET LCA, 2004; Hockerts et al., 1999; Hocking, I991; Saphire, 1994; Wells et
al., 1991).

In this dissertation we argue that conflicts in the results of comparative packaging
based LCAs may arise mainly due to uncertainties in both the LCA method and the actual
packaging operations under consideration. Thus, in order to understand the nature behind

the conflicts in LCA results, this chapter presents an analysis of how uncertainties

135

actually affect the outcome of the comparative packaging-based LCA developed in
Chapter Three.

Specifically, the uncertainty analysis in this study comprises two aspects. First,
we perform a scenario analysis by identifying different domains that may have the
potential to change the outcome of the comparative packaging based LCA. Under each
domain, two alternatives are proposed. Then, a technique is used to select an alternative
from each domain and randomly create different scenarios.

The second aspect of the uncertainty analysis in this. study comprises inventory
parameter uncertainty. The study on inventory parameter uncertainty is achieved by the
use of apprOpriate published uncertainty estimations along with stochastic simulation (i.e.
Monte Carlo simulation) in order to represent inventory data uncertainty as probability
distributions. By this analysis, relevant inventory parameters are identified and further
analyzed, and final results of the comparative LCA along with the combined uncertainties
from inventory data and scenarios are obtained.

The next section discusses with more detail the methodology for the study,
including details of the areas selected for the scenario analysis and the procedure
followed. Later in that section, information regarding the parameter uncertainty analysis
of the inventory data itself is presented, including specifics about the uncertainty
distributions based on engineering estimates and computer simulation procedures. Lastly,
we present the results of the uncertainty evaluation based on the comparative LCA results
along with an extensive discussion of the reasons for the differences and their

COHSCQUCHCCS.

Chapter Five summarizes the overall conclusions of the dissertation along with several

recommendations for future work.

11. Methodology

(1) Scenario uncertainty analysis

Figure 20 outlines the procedure followed for the scenario analysis in this study. First, for
the PDSs explained in Chapter Three, domains associated with packaging situations with
perceived effect on the LCA outcome were identified. In this study, we used three
domains: material recycled. content percent, breakdown of end-of-life streams, and
Distribution 1 distance in the case of PLA. Material recycled content percent was
perceived as relevant since, for instance. in the case of aluminum, it has been reported to
greatly decrease the energy consumption of the material production operations.
Differences in the end-of-life streams were also assumed relevant, since for instance in
the U.S., deposit refund systems, which may or may not be in place or not, have been
found to greatly affect the percentage of material sent for recycling. The last domain,
Distribution 1, which specifically targets the distance traveled to supply PLA resins to a
component manufacturer, was assumed relevant, since so far in the whole us. there is
only one facility commercially producing this plastic, and thus the effect of transportation

for large distances may be significant.

0:0»:000 0::00000 :0: :030::0: 0:0:000:n: .8 2000::

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l .
N 9:00:82 : 03:00:00? N 03:00:00.0. 00”: m0=0>00m .MM:
”m £0.08 ”N 0:0E00 H: 0:0E0Q 0:53.:
0 0::0000m
o
o
o
o
N “imam“? - : 03:00:00.0. : 03:00:82
Hm 0:00:00 , ”N 000an 0: 0:00:00 0:05:
0 oE=000 0:50 =<
_ - a 05:80 00
: 05:08
00:00:00: b0>0
0:0E0: :00 9.008000 000 :0 00:00:00 E0:00~:
000: #:m <15 000 ”my: <15 000 ”in: 000 H.010:
EV: comm .EV: 8m 00mm 0%: 0:00:00? 000: "mm E0EEE< 000: HE0EE02< 000m HE00:E0:<
N: 00.000505 <40: N: 00000505 5: 000m Hmy: :m: 0.00: ”my: .5: 000 Fm: 000 “:mn:
”N 0>00E0=< 0: 9:00:00? ”N 03:00:00? H : 9:00:00? um 03:00:00? H: 03:00:82
: 00:00:05 ”m 0:008 0:: :0 :0m: ”m 50an 0000000 8:933: H_ 0:0E0n:

 

 

 

 

138

The scenario analysis procedure continues with the identification of two
alternatives for each domain. In this study, though alternatives for each domain were
arbitrarily selected, they in many cases can still represent actual situations. After the
identification of the alternatives, an iterative procedure was put in place in order to, by
random selection of one alternative per domain at a time, create a scenario and calculate
its LCA outcome. In this study, this calculation was performed 10,000 times, and
statistics (i.e. median, confidence intervals) were calculated afterwards. The resulting
output distribution reflects the uncertainty of the comparative LCA outcome, taking into
account the alternatives in the domains under consideration. The Visual Basic code
developed to perform this scenario analysis is in the Appendix.

In this analysis, to determine the potential impact of the identified alternatives on
the comparative study outcome, all alternatives per domain were given equal probability.
To illustrate the consequences of ignoring or reducing scenario uncertainty, the results of
this simulation will be compared with the outcomes that would have resulted if one

specific alternative had been chosen.

1)) Inventory data uncertainty analysis

Inventory uncertainty analysis was performed using Monte Carlo (MC)
simulation. MC, in general, is a numerical technique to randomly generate or sample
data to obtain approximate solutions to complex mathematical and statistical problems. In
fact, MC is a technique to randomly generate a number under a defined probability
distribution and can be used to represent uncertainty and variability of inventory
parameters of an LCA (Heijungs and Sub, 2002; Huijbregts, 1998a; 1998b; 2001;

Huijbregts et al., 2003). To actually perform MC simulation in LCA, first each uncertain

input parameter is specified as an uncertainty distribution. Next, the method randomly
selects one number from the specified uncertainty distribution for each parameter. All
selected numbers are then used for the calculation of the desired output. Iterative
calculations end up producing a distribution of the predicted output values, reflecting the
combined parameter uncertainties.

Since this LCA study involves an enormous amount of parameters, it was not
feasible to specify the uncertainty distributions for all these parameters in detail. This
problem was overcome with a stratified procedure developed by Huijbrets et a1 (2003). A

diagram of the method is presented in Figure 2].

I40

 

Start uncertainty analysis

 

 

T

 

Specify LCI model

 

{r

 

Specify rough uncertainty
distribution
of input parameters

 

V

 

 

First sensitivity analysis

 

 

T

     

 

   
   

 

 

 

 

 

No Individual input
parameter potentially
important?
Use rough uncertainty distribution Estimate detailed distribution H—

 

 

 

 

 

Second sensitivity analysis

 

 

 
 
   
   

   

No

 

Negligible
contribution of rough input
parameters?

 

 

Monte Carlo simulation

 

r

 

 

Model outcome

 

 

Figure 21. Inventory data uncertainty analysis procedure (Huijbregts et al., 2003)

141

First, each uncertain input parameter was assigned a distribution in order to reflect
a conservative estimation of its uncertainty based on experience or engineering
estimations. In this study, all inventory data was assigned the same log-normal
distribution defined using upper and lower ends covering 95% probability in the
distribution. The upper and lower ends were estimated by the approach used by Slob

(1994) and expressed by the following formula:

 

P(MI((X)< X <kxM(X))=O.95 (l)

where M(X) is the median value of each parameter, and k is the uncertainty factor
assigned to that parameter. It was assumed that all the inventory values reported in all
DEAMTM modules were medians. Their initial uncertainty factors (k) were obtained
using expert recommendations published in the literature (Huijbregts et al., 2003) and are
listed in Table l of the Appendix.

Next, a first sensitivity or screening analysis was performed to identify the
parameters that had a strong correlation (i.e. were “important”) with the LCA results.
This analysis was performed by running a first MC simulation followed by a Spearman’s
rank correlation procedure. This rank correlation method was done by first ranking each
of the simulated values of each LCI parameter and their correspondent LCA result in
ascending order. Next, for each simulated LCI parameter-LCA result pair a correlation
coefficient (i.e. Spearman’s coefficient) was calculated between them using their ranking
values instead of their actual values (since a correlation coefficient measures the strength
of the linear relationship between two variables, it is unlikely to find any linear

relationship between variables that have different distributions as in our case: thus, using

142

their rank instead of their actual values, that limitation was avoided) (Decisioneering Inc.,
Crystal ball - User manual, 2004). Afterwards, for each LCI parameter-LCA result its
average correlation coefficient was calculated based on the total number of simulated
values. Finally, the contribution to the overall variance of each LCI parameter was
estimated by squaring each averaged rank correlation coefficient and normalizing them to
100% (i.e. adding all squared coefficients, dividing each squared coefficient by this sum
and multiplying it by 100%) (Technical note from Decisioneering, How does Crystal Ball
calculate sensitivity?, 1998). Parameters that contributed at least 1% to the overall
variance of each LCA result were considered “important”. More details about this
calculation and Spearman’s formula are included in the Appendix.

When more complete uncertainty information was available, those important
parameters were specified in more detail, and that information is presented in Table 2 of
the Appendix. The importance of those parameters was confirmed in a second sensitivity
analysis. This second analysis was necessary because defining parameters in more detail
may have affected their uncertainty importance. When important inventory parameters
were specified in detail, and keeping the “unimportant" parameters with their rough
uncertainty distributions, a final MC simulation was performed to quantify the output
uncertainty.

Some notes of caution about the actual accuracy of this uncertainty analysis are
required at this point. First, as stated earlier, the uncertainty analysis is something new in
LCA, and the use of MC simulation as well as the appearance of other approaches and
methodologies for uncertainty estimation in the last few years only reinforces the idea

that the quantification of uncertainty is itself uncertain. Moreover. within the same

uncertainty estimation approach, uncertainties in the uncertainty analysis itself may not
be avoided. For instance, a case can be made for the use of different distributions (i.e.
triangular, uniform, gamma, etc) to represent parameter uncertainty. In this study, for the
sake of simplicity, a log-normal distribution was chosen to represent uncertainty (i.e. lack
of knowledge) of most of the parameters. The few exceptions are parameters such as
packaging material weight and product weight for which a normal distribution was used,
representing variability (i.e. natural variation) rather than uncertainty (see Table 2 of the
Appendix). Worth noting is the fact that the use of a logmormal distribution appears to be
a standard practice by risk analysis experts to represent uncertainty since it avoids
negative values, it captures a large value range, and the uncertainty in many processes
and parameters follows a skewed distribution (Slob, 1994).

A final word of caution is regarding the application of this analysis to prepared
data such as those contained in the DEAMTM database. Since this database is a
compilation of data from a number of other sources, and Ecobilan itself does not
guarantee that the original studies have delivered high quality results (Ecobilan Group
Inc., DEAM module databases and manuals, 1999), the basis for the application of this
uncertainty analysis is to represent the uncertainty contained in those sources (in fact
Ecobilan in its LCA specific software called TEAMTM, which in turn works with its
DEAMTM database. allows for uncertainty analysis using MC simulation). Moreover,
since many details about the procedures followed by Ecobilan to adapt those databases
(e.g. to convert them to the single format that it uses) are not easily known or validated

and outside the scope of this analysis, the study also represents that uncertainty. Thus,

144

results of this uncertainty analysis should be interpreted to be a conservative estimation

of that of the overall results of the comparative LCA under consideration.

Presentation of results

All results of this uncertainty analysis refer to and use the same identification
notation of the three product delivery systems (PDS) explained in Chapter Three. As
explained there, each PDS was defined as the system employed for the distribution of a
hypothetical drink to customers and included the primary and distribution packaging, as
well as manufacturing, and transportation associated with the function of the system. The
alternative product delivery systems considered are:

. PDS 1: utilizing PET bottles and PP caps as one choice for drink delivery.

. PDS 2: utilizing aluminum cans as the other choice for drink delivery.

. PDS 3: utilizing PLA bottles and PP caps to investigate the use of renewable
material feedstocks.

To facilitate the comparison of the LCA results (i.e. the characterized and not
characterized environmental impact indicators as explained in Chapter Three) for each of
the three PDSs. they were related using the following relationship (Huijbregts et al.,

2003f

 

CR = (2)

where CRu is the comparison ratio for impact indicator u (dimensionless) and ru,n and

ru,m are the environmental impacts related to PDS n and m, respectively (e.g. expressed

in M] for energy use). The environmental impact indicators of the product systems
compared can be considered to be significantly different, for instance, if 95% of the

iterations gave results above 1.

Simulation details, hardware and software

In this study, simulations for both inventory data uncertainty analysis and scenario
uncertainty analysis were simultaneously carried out. The number of iterations used for
all simulations was 10,000, as recommended in the literature to achieve a representative
distribution of cumulative results (Huijbregts et al., 2001). MC simulation and scenario
analysis were performed using codes and tools available in Crystal Ball 2000.v5 (Denver,
CO) (Crystal ball - Forecasting and risk analysis for spreadsheet users, 2004), and Visual
Basic programming and spreadsheets of Microsoft Excel (Microsoft Corp., Microsoft

Excel, 2000).

146

III. Results

This section will be divided into two main parts. First, overall comparative LCA
results will be presented with the inclusion of both scenario and LCI data uncertainty.
Next, outcomes that would have resulted if one specific scenario had been chosen are

presented.

a ) Overall comparative LCA results

Figure 22 shows the overall uncertainty in the comparison ratios of impact
indicators due to the combination of scenario and LCI data uncertainty. For a given pair
of PDSs, if the ratio of the indicator is higher than 1, the PDS in the numerator has a
higher environmental impact indicator than that of the PDS in the denominator, and vice
versa for a comparison ratio value lower than 1. In Figure 23 to Figure 25, each
comparison ratio is represented as an interval in which the top end represents the 95th
percentile, the mark is the median and the bottom end is the 5‘“ percentile. When the
comparison ratio value is significantly different than 1 (with one sided, 95% confidence
level), an asterisk (*) is placed by the interval. Details of the data on which Figure 23 to
Figure 25 are based are presented in Tables 3 to 5 in the Appendix.

A first look at Figure 22 shows that when comparing PDS 1 with PDS 2 (left
side), it appears that the system using aluminum cans has significantly higher ozone
depletion potential (ODP) value than that of the system using PET bottles. When we
compare PDS 1 with PDS 3 (middle part), the global warming potential (GWP) value of
the system using PLA appears to be significantly lower than that using PET. Lastly, when
we compare PDS 2 with PDS 3 (right side), the system using aluminum cans appears to

have significantly higher ODP and GWP values than the system using PLA bottles. All

147

other comparison ratios do not differ significantly. Figure 22 also shows that the widest

uncertainty ranges correspond to water use and ODP ratios.

148

 

 

 

:32 8:03:80 $3 an E25 Eob 8:20.:6 “:85:me
mecocon var—2mg 355388: 38 ~04 new 05.:on mo 2052588 05 9 26 8:8 :cmtmmEoo 2: E 35850:: :Eo>O .NN Eswfi

 

 

 

 

 

 

 

395QO 8938“. again:
cm: em: 8:
ADC 950 83>? zmcocm n50 E50 83>? 389m. A50 950 .53? 38cm
l _ ll — L _ _ _ g k h H — 0
AV
4 e .M H
.. H W W H w e _
1-1.1:; llllt I -lli 11* it I it} 11. i- l...lli IIII i rN
*
.. m
l ililll l i ll ll rill ill 1 llllllll 1 IV
m

 

 

 

 

 

 

ODP-H

149

b) Scenario-specific comparative LCA results

In order to evaluate whether any comparison ratio differs significantly when
scenario uncertainty is not included, or vice versa, the results of the specific scenarios are
needed. All 8 scenarios were evaluated separately; Table 26 indicates their characteristics
along with the notation code used in Figure 23 to Figure 25. In every evaluation of each

scenario, LCI data uncertainty was included.

Table 26. Different scenarios evaluated in this study

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Identification code Domain: Recycled Domain: End of life: Domain:
of the scenario content % Recycling Rate Distribution 1
(RR)

A PET: 0% PET RR: 10% PLA Distribution 1:
Aluminum: 30% Aluminum RR: 19% 200 Km.
PLA: 0% PLA RR: 0%

B PET: 0% PET RR: 10% PLA Distribution 1:
Aluminum: 30% Aluminum RR: 19% 2500 Km.
PLA: 0% PLA RR: 0%

C PET: 0% PET RR: 39% PLA Distribution 1:
Aluminum: 30% Aluminum RR: 75% 200 Km.
PLA: 0% PLA RR: 10%

D PET: 0% PET RR: 39% PLA Distribution 1:
Aluminum: 30% Aluminum RR: 75% 2500 Km.
PLA: 0% PLA RR: 10%

E PET: 0% PET RR: 10% PLA Distribution 1:
Aluminum: 60% Aluminum RR: 19% 200 Km.
PLA: 0% PLA RR: 0%

F PET: 0% PET RR: 10% PLA Distribution 1:
Aluminum: 60% Aluminum RR: 19% 2500 Km.
PLA: 0% PLA RR: 0%

G PET: 0% PET RR: 39% PLA Distribution 1:
Aluminum: 60% Aluminum RR: 75% 200 Km.
PLA: 0% PLA RR: 10%

H PET: 0% PET RR: 39% PLA Distribution 1:
Aluminum: 60% Aluminum RR: 75% 2500 Km.
PLA: 0% PLA RR: 10%

150

 

Figure 23 to Figure 25 show the intervals of the comparison indicators that would
have resulted if each specific scenario had been chosen. For reference purposes, the
intervals for the combination of scenario and data uncertainty are shown as “All” in each
Figure as well.

Inspection of Figure 23 to Figure 25 shows that the only cases in which an
indicator appears not to agree with the overall uncertainty results (at the same confidence
level at least) are the ODP ratios between the system using aluminum cans and that using
PLA bottles (Figure 6d, scenarios G and H) in which aluminum recycled content is 60%
and the end of life involves 75% and 10% recovery rates for Aluminum and PLA
respectively. Nevertheless, the confidence level for the agreement in the ODP values
between these systems in these scenario-specific results is still very high (94.5% and

94.7% for scenarios G and H respectively).

151

 

95>“: 50:50:50
$3 5 3:: 50¢ 855:6 EdanwB .5855 56553 $95QO 9:25:32 55 58 8:52 oEoommétmcoom .mm oSmE

oraceom 2.5:on

I U n— m D U m < :< I 0

 

 

 

md

 

 

 

 

 

 

l l-—--O—l
l I-—-O-—l
011133
a}
l
l
l
l
l
l— l
l
'l— l
l
F-H
l
l—r-O—‘l
l
l
*— l
7
In '— V3
—- c
01193

 

a a a
W.» . i i .w * ... *l H l l* l l.* 2”." . M r, N

152

8558 220:6 525 one “50 SE: mom a 8:203 225:6 525 0:2 526 895 mm: 3

 

 

 

25:on 05:on
I O m m G U m < :< I O m m— Q U m < =<
» _ r L r L a L _ p _ p _ m . O
u H H H w a w H - no

 

one};

 

 

 

 

 

 

 

 

I—y—O-l
l
.—-—
l
f
|-——
l
[——
l
H—
l
}—a——
l
|
|
|
}—.'__
|
l
l
l
M N '— 0
open
liar
I—H
I—H
l—H
F-H

 

m.—

 

 

8:958
255:6 595 0:8 8: .553 mmflns man— S 85:58 “55:6 .525 05E $.55 meASQO 3

2532 50:25:00
gem: 5 2:: Est 855:6 Emowtcwmm 58:5: €558 mmn—nSQO 25:25.5: 2: :2 33.5: 3:53-25:on .vm EsmE

 

 

 

 

ougoom
I U ”I I Q U I < =<
_ . r _ H F L _ O
md
__l m law lilm-ll Hill. will Mil-H11, “limit I
m m;

 

22558 .:o.5.t6 56:: 25.. IQO meQ Imam— 6

 

 

 

 

 

 

 

 

 

 

 

 

 

25:on
I O m m G U m < :<
vrm_fl_.._¢_¢LHr¢_v
8.558

555$: 52:. 25.. mm: 553 mmDASQO 3

Nv—O

outta

C":

ones

0

2.5:on

n— m D U m < :<

 

 

P

WWW w #41.“ imlmm

*

 

 

8:258 228% 52: one :30 5258: o

 

 

 

255%
I O m m— D U m < :<
_ bl l— » I p L I O
..-.l-| ll l i i ll - i - i i -- I i 2 - .-.; wd
:- -I - d : +1 _
:1- -, I. ll Z. 3..- 2
N

 

225:on “55:6 5::: 25: 35:5 mmflmhmmm A:

09821

open

153

955. 55:52.58
553 5 ~95: E2: 55:55:56 555555 .555: 57.553 mmmfimmmm 25:255: 55 :8 5:55: 5£55%-25:55m .mm 55w?—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OT—acoom DEN—boom
I O m m G U m < :< I O n— m G U m < :<
m r L F L w L I b m o _ L L L L _ t. L o
q:.+ididiigt-titi+i.. :.
.1 .r .... .... .... ._- w .V W; N
H .. .. . l_ 5 o m
l l lll ll 1 i i l I i l“. w ll-ll l l i i l-l I i: ll 1 i ..V
8588 5:25.25 5:5 25 5:0 55:55:: 6 8553 225.56 55:: one :35 89:55:: a
Otmcoom otdcoom
I O ”I I Q U m < :< I O m m. G U m < =<
r _ I _f L . I — To L LI L p _ p _ _ o
... .. w I H H
l ... ll 1 I N
m

 

 

85:55.5
555:5 55:: 2.5: 55: 553 meENQO 3 525.555 522:: 5.5:: 25.. $55 mmnfimmam 5

open

onvu

154

IV. Discussion

For organization purposes details of the data in which the figures included in this
section are based are located in Tables 6 to 9 in the Appendix.

The present discussion is based on the results presented in the previous section

and on the systems considered and assumptions made in Chapter Three of this
dissertation. However, the aim of this section is not only to present a discussion based on
the quantitative uncertainty assessment, but also to highlight other potential factors not
included in the quantitative estimation of uncertainties in the comparative LCA.
Looking at the results presented in the previous section, we can identify two important
considerations. First is that LCI parameter uncertainty seems to have a dominant effect in
the intervals of the LCA outcomes. In fact, most of the times the uncertainty intervals of
scenario-specific results appear to be of similar magnitude to those in which scenario
uncertainty is included. Nevertheless, in many cases scenario-specific results do seem to
affect the spread and location of the median of the comparison ratio value, and this is
further discussed in later paragraphs.

The second consideration is that most of the times the uncertainty intervals
contain one, which in fact indicates that the systems under study are similar to one
another, at least with regard to the environmental indicators included here. In fact,
looking at the life cycle energy profile of the three systems (with all scenarios
considered) shown in Figure 26, their similarities in median values and 95% confidence
intervals are more noticeable. For instance, in the material production phase, the system
using PLA bottles consumes per functional unit a median of 4080 MI (68% of the total

energy use), the system using PET bottles consumes per functional unit a median of 3978

155

MI (66.8% of the total energy use), and the system using aluminum consumes per
functional unit a median of 4858 MJ (69.8% of the total energy use) with a somewhat
greater interval since scenarios containing 30% and 60% recycled content affected this
phase of the system based on aluminum cans.

This last consideration is important since it ultimately means that the perceived
preferability of one system over the other, at least in this case, is strongly dependent on
the quality of the data (which defines the interval).

Further considerations can also be made by analyzing the scenario-specific results
of Figure 23 to Figure 25, which reveal that several uncertainty ranges show systematic
differences between the scenarios chosen. For instance, when comparing systems using
aluminum cans, the effect of the use of recycled content is always noticeable since it
generally lowers the energy use, and the emissions of global warming and ozone
depleting substances due to material production and thus it helps move the median of the

ratio closer to one (Figure 23a, c, d to Figure 25a, c, d,).

156

5:562: 55 :32: 55:50 5:55.? 55:: 55 .5 5552: 35:5 5.98 5::— .0m 5::wE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ovSoom o _ .m I m- M m _ .mmo 0“wa _ mcwm wmémm mndm omdwov mmflm I
Sumac 3&2- _. $.de 0N0- __ #062 mmdw: #0.: £0me mmflm§
:ummow modem- . bmgmo voéwom _ 00.00 $.03 mwdm afiwnom EASE.
_ W 5:350:55: :2 5: 0:
58h. 5:— .:o Em— “ :05: m: m. wEEHI _ :2 :N: 2 .0205 :2 : I: m. 5.”.5W a
m. . n. a .Q _ .5 2.: .9 5:55:00 .5 0.: .Q _ . 5 I2
_. 00.000m-
FM“— 00.0
l1 00.000m
4. 00.000:V
t .. 00.0000
... | 00.000w
. 00.0000—

 

 

 

 

 

00.000m _

warm "1 000mm ‘Kfiwua

157

Water use, in contrast, (Figure 23b to Figure 25b), does not seem to show this
systematic variation. Instead, these graphs show large intervals in spite of the consistent
appearance of a median value below one. This situation highlights the effect of the actual
uncertainty analysis method (and the assumptions) on the LCA result. In fact, the large
intervals for water use seem to occur because the uncertainty factor used for this
parameter in this study was a conservative value of 10, at least 5 times bigger than that
used for energy values and values for global warming and ozone depleting substances.
More details of the effect of the magnitude of the uncertainty factor can be noticeable
when we look at the life cycle water use profile shown in Figure 27.

As clearly shown, the largest uncertainty intervals belong to the material
production phase, in which the system using PLA bottles has the greatest. This peculiar
situation seems to occur by the concurrence of two factors: the large uncertainty factor
used for this impact and the larger number of inventory “cradle to gate” modules used to
model the corn growing and corn wet milling operations included in the material
production phase of this system. That is, the more “cradle to gate” modules a system has,
the larger the resultant uncertainty spread. Obviously this situation does not help in
drawing a conclusion, in spite that, since it has a higher median, it may be perceived that

the system using PLA uses more water.

158

585?? 55:: 55 .00 5.02: 55: :58; 298 503 SN 5.5mm”—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r.--

wdmtm 00.7 % gamma .: 3N N53 St.— mmdm Kéomm memI
wit: 3; _1 mvwwm H 0 M26: Nw.m flaw p.29 mmmm§
wmdom— NV..- _ :dmm Sago. 50m :0; mad— oo._mn and—U
:33. 5. o : “l m _ wE _ m EzmwwmmE _ 5:52:20
0.: U m ”:ESLEEQM .:.m :ocsgcaa $58550 ”22:55.55 3:532
l t _. . 8.82-
llllllllall l l I ll l l i l 4 00.0

00.000m

- t 00.000.“q

.r 00.0000

I - - - - 00.000w

- - - - - - i r 00.0000—

- -r 00.0002

- . 0. 00.0003

fi|mrl l 00.0000—

00.0005

)IUIJP '1 0001/1 3919/“

159

On the other hand, the effect of end of life recycling rate in the scenarios
considered is highlighted in Figure 24d. In fact, following estimations from the
manufacturer of PLA (Vink et al., 2003), PLA incineration was modeled in this study
using the same environmental burdens as for paper incineration (with energy recovery).
Since the incineration energy is harnessed. the corresponding DEAMTM module credits
(i.e. subtracting to account for the avoided burdens from generating electricity by burning
fossil fuels) several of the emissions of the inventory data of the paper incineration.
DEAMTM also applies this same reasoning to the module for plastic incineration with
energy recovery used in this study to model the incineration of PET. However, the credit
given to methyl bromide in the plastic incineration is almost three times higher than the
credit given to paper incineration. Since methyl bromide is an ozone depleting substance,
the effect of the previous difference can be noticeable when we analyze the ODP
comparison indicators. In fact, by using the high recycling rate alternative (i.e. Aluminum
RR: 75%, PET RR: 39% and PLA RR: 10%), there is less mass of PLA and PET sent for
incineration. This is further affected by the fact that, since PLA is less dense than PET
(Leaversuch, 2003) and following performance estimations from the PLA manufacturer
(Vink et al., 2003) this study assumed that the weight of a PLA bottle is 6% less than that
of a PET bottle. Less mass of any of these materials means less methyl bromide credits,
but this effect is more important for PLA since quantity of this inventory value is three
times smaller than that of PET. The overall effect would be then a comparison ratio
moving away from one, which is exactly what happens in scenarios C, D, G and H in

Figure 24d.

160

Though in this case a significant difference in the ODP ratio of PDS l and PDS 3
was not found, the appearance of substances with a high ODP contribution such as
methyl bromide and halon 1301 in the incineration inventory data of a material such as
PET needs to be further discussed, since this situation actually highlights the uncertainty
in the allocation method used for emission data. In fact, it may be argued that there is no
“chemical correlation” between PET incineration and halon 1301 emission for example,
and thus its inclusion in the incineration inventory is not required, at least from a
chemical standpoint. However, if for this incineration process, a mass allocation applied
to the incineration facility is in fact used, the situation may well change with potentially
great consequences for the calculated ODP of PET.

The significant differences regarding the GWP value between PDS 1 and PDS 3,
and PDS 2 and PDS 3 indicated in Figure 22 also require a closer look. Inspecting Figure
28 which shows the life cycle GWP profile for the systems, it appears that the main
reason for the higher GWP of the systems using the aluminum cans and PET bottles is
because in their production phases the emissions of global warming substances are much
higher than in the system using PLA.

In fact, in the material production phase, the comparative study shows that the
system using PLA bottles emits, per functional unit, a median of 243 kg C02 eq (34.5%
of its total). the system using PET bottles emits per functional unit a median of about 639
kg C02 eq (58.3% of its total) and the system using aluminum emits per functional unit a
median of 824 kg C02 eq (69.3% of its total), with again a somewhat greater interval
since scenarios containing 30% and 60% recycled content affected this phase of the

system based on aluminum cans.

'l6l

555?? 55.5 55 .00 5520 030 293 5.05 .wm 5.53m

 

 

 

, mo 35 as? 54 am 2 m: M 55.2 g is 3.2 _. mag an: I
m «woo: ~52 . :3 8.: H 54.2 £52 3.5 m 3.4% $55 It
r Ill) 1. l t-.. it. ...t tr l il .-t I. ..l 1.- tltl. - ... l ll-1 l 7. ill] [1 : - -- l I l ._ l. i l: l, lit 1 ill
W 2.32 :5». 3% 55.3: m 8.3 2.8. 5:: 55.58 55.5
._ “ Ohaommzfiwe _ C0503 .—

% Egon; Om: MO mun—m CO— SMC m— MHZ—Zn m :0_ SN—bm- .QEOU G05: "5m— m m—hOuMO Aw

W .5 n. s .5 w .3 a. .5 5:585 . a. .3 _ . 2

2- t t r l l l l l t l t t l l l t l I l l l l l . Litl Illl bl l .l- l: l . 8.8m:

 

 

 

 

 

 

 

00.0
x x . \\
\ \ \ \
\
\\ . \

t 00.00m

k .

 

 

 

00.000—

 

 

 

00.00m 0

 

 

 

 

00.000N

162

auup "I 0001 /b9 zoo JO 8a 'dMD

Further analysis of the material production phase of these systems shows that the
main GWP contributors in this phase are carbon dioxide and nitrogen oxides that are
emitted in the extraction, processing of intermediates and any related transportation of the
raw materials as they were modeled in the DEAMTM databases. It is important to
emphasize now that the differences found in Figure 3 with regards to GWP correspond to
the eight scenarios previously considered, and an alternative scenario may actually give
different results. Thus, for instance, if an aluminum can with 100% recycled content is
used in the comparison, the GWP of the material production phase of this system drops
dramatically to 244 kg C02 eq, almost the same value as that of the system using PLA
bottles.

The importance of the use (and its inclusion in the comparative LCA study) of
packaging components is also highlighted in this analysis as well. For instance, the effect
of not including the LDPE sleeve in the system using PET bottles roughly reduces its
energy use by 89 MJ, while not including the PLA sleeve in the system using PLA bottles
reduces its energy use by about 90 M]. This difference becomes critical when systems are
similar, and quickly shows us why, even though aluminum is generally more energy
intensive, the fact that it includes less components may allow it to be comparable to
systems which, though less energy intensive when considered individually, require more
components overall.

The significant differences regarding the ODP value between PDS 2 and PDS 3
indicated in Figure 22 require a closer look as well. The contributing substances to the
ozone depletion potential identified in this study are: carbon tetrachloride, halon 1301,

methyl bromide and methyl chloride. Inspecting Figure 29 which shows the life cycle

163

ODP profile of the systems, it appears that the main reason for the higher ODP of the
system using the aluminum can is because in the production phase of this system the
emissions of ozone depleting substances are much higher than in other systems.

Further analysis of the material production phase of the system using aluminum
indicates that the main ODP contributors in this phase are halon 1301 and methyl
bromide, which actually account for 51% (at 60% recycled content) to 56% (at 30%
recycled content) of the total ODP of this system. However, it should be noted that halon
130] and methyl bromide are no longer produced in the U.S. and their use is heavily
restricted (U.S. EPA, EPA guidance on Halon emission, 2001; U.S. EPA, Ozone
depletion rules & regulations, 2005). It would be expected, then, that updates in data
would reflect significant reductions in emissions of ozone depleting substances, and this
in turn suggests that the ODP values of these three systems may actually appear similar to
one another. Moreover, this situation underlies another aspect of the LCA result, which is
its dependence on the location of the operation. For instance, as stated earlier in Chapter
3, alumina production for aluminum manufacture is mostly done outside the U.S. in
countries such as Australia, Jamaica or Brazil, in some of which restrictions in the use of
these ozone depleting substances may not yet be in place. Then again, updates in data
with regard to origin of the operations may change back the standing of the system using

aluminum cans.

164

555?? 55:5 55 00 552: 000 296 5.2..— .om 55mm”—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gimp—00m
2.5 3: 8.5 .5 88 88 8 who 8::-
85 8.: 03. W as 3 o N; 8555
as Ed 85 H 8.: 8.0 8.0 8.8 GED
ma . 5:355ng :28: o:
:85 5.2.8 25 m 3 55:5 . N :58 _ . e 5
8:32:23 ” 8:355: 55585 8:25:55 322...:
v . cod-
.. 8.: m
. 1 - - - - 85. M
-.3- 8.2 m.
“a 8.2 Q
i m 8.8 3
r . _ .m
t r CO.WN .ln
- L. - 8.8 m
- - .4 8.2 m
I - I ..| - 8.8 w.
.... V4.

00.?

165

In Figure 29 the consequences of the uncertainty analysis itself requires a
discussion as well. The large uncertainty spread in the material production phase for the
system using aluminum seems to occur because of the effect of the uncertainty factors of
the “contributor” parameters (i.e. halon 1301, methyl chloride, etc) of the ODP value.
Ultimately, this is clearly a problem for characterized impact indicators since they are
usually compilations (e.g. weighted sums) of several contributor items, which in turn
have their own inherent uncertainty. and so the more inventory parameters compiled. the
larger the resultant uncertainty spread.

In fact, we can further ponder on the implications of the uncertainty analysis
results shown here. As was evident from the previous quantitative assessment of the LCA
uncertainty, uncertainties have a propagating effect and do affect the LCA outcome. This
situation also implies the somewhat counterintuitive idea that the more the number of
inventory parameters considered for a system in a study, the more (not less) uncertain the
result may be. On the other hand, the other extreme is no better. For instance, for
practical reasons the compiling of just major inventory parameters (i.e. based on emitted
weight, feasibility of the measurement) is not unusual. However, if traces of substances
that have a large contribution on a characterized impact (e.g. ODP, or GWP) are
neglected, the conclusions can potentially change. This, in fact, may be a problem when
“newer” technologies, for which only limited and proprietary information exist, are
compared with heavily studied ones. For instance, with regard to the information
gathered for PLA in the comparative LCA presented in Chapter Three, polymerization
data was obtained from the manufacturer of this resin, which in turn used engineering

estimations to account for just selected inventory parameters while other substances were

166

neglected. Therefore, the results for ODP or GWP or any other characterized indicator
should be interpreted acknowledging this situation.

Lastly, it needs to be clarified at this point that values of the energy analysis
presented here are total energy values, this is, the summation of both renewable as well as
non-renewable energy. This note is important since the use of annually renewable
biomass (e. g. corn), as opposed to petrochemicals (i.e. oil or natural gas) as the feedstock
for the production of polymers is one of the core reasons for the desire to develop PLA or
any other biobased polymer. In fact, based on the assumptions of this study, about 26%
of the total PLA energy consumption came from renewable resources, while the other
systems had smaller renewable energy consumption percentages (about 11% and 0.3%
for aluminum and PET, respectively). Nevertheless the percentage of renewable
resources consumed in the systems based on aluminum and PET can still greatly change
if we use, for instance, location-specific energy production data which may produce very
different conclusions than the average electricity generation values used here. Clearly,
results may differ when coal-generated electricity data is used for operations that occur in
regions in which electricity is produced mainly by hydroelectric power. Coal-generated
energy is considered “non-renewable” while hydroelectric power is produced from

“renewable” resources and thus appears to be more environmentally friendly.

167

V. Conclusions

The present study has shown the use of a simultaneous assessment of uncertainty

analysis due to LCI data and scenarios relevant to packaging.
The results presented in this study have shown the important consideration that LCI
parameter uncertainty seems to have a dominant effect in the outcome of the LCA results
considered. Furthermore, in most cases the uncertainty intervals contained unity, which in
fact indicated that the systems under study are similar to one another, at least with regard
to the environmental indicators included here.

It was also shown that location and age of data and allocation procedures, along
with specific packaging scenarios, all affect the impact assessment method and confuse
the outcome for results that originally indicated significant differences among the
systems compared.

The implications of these considerations are important since it shows that though
the systems compared are based on the use of different feedstock materials, uncertainty in
the LCI data and the scenarios interfere with determining conclusive differences.

Further improvements in the application of the methodology are needed, including the
development of a life-cycle inventory database with uncertainty information, and the
further development of location specific and time specific impact assessment models for

a more systematic analysis of scenario and model uncertainty.

168

CHAPTER 5 OVERALL CONCLUSIONS, RECOMNIENDATIONS AND FUTURE

WORK

In this dissertation, a thorough understanding of the complexities of comparative
packaging based life cycle assessment (LCA) outcomes was sought. To achieve this, the
research was divided into two major parts. The first part involved a critical review of the
literature with regards to LCA and its applications to packaging operations, which was
presented in Chapter Two. The main bodies of literature reviewed in this chapter
involved actual LCA studies. recent journal publications regarding advances in LCA
methodology and case studies, ISO 14000 series of standards, reports from governmental
agencies and private sector literature.

As shown in Chapter Two, the LCA method covers a great variety of aspects of
packaging systems and has been used to provide information for purposes such as
strategic planning, marketing, academia, environmental labeling, etc. However,
uncertainties play a major role on the LCA method throughout the phases of the study.
Uncertainties in the quality (e.g. age, source, completeness) of the data used in the study,
uncertainty in the scope and boundaries of the processes included and uncertainty in the
scope of the impact assessment stage are some of the most critical ones.

The second part of the research comprised the actual quantitative study of the
effect of some types of uncertainty in the outcome of a packaging based LCA. For this, a
two-step process was followed. First, a comparative LCA study featuring three different
packaging materials was developed for a hypothetical drink product. Following ISO’s

LCA steps, a comparative LCA study was developed and its extensive details and

169

information sources were presented in Chapter Three. The three material/container
alternatives were a PET bottle, a PLA bottle and an aluminum can, which along with
their whole set of distribution packaging (i.e. corrugated trays, stretch wrap, and pallets),
transportation services and end of life alternatives were evaluated with regards to four
environmental burdens: energy use, water use, global warming potential (GWP) and
ozone depletion potential (ODP).

The second step was the actual quantitative uncertainty analysis of the
comparative LCA presented in Chapter Three. The uncertainty analysis comprised two
aspects. First, a scenario analysis was proposed by identifying different domains with
effects on packaging situations. Under each domain two alternatives were proposed and a
procedure was used to select an alternative from each domain and randomly create
different scenarios. The second aspect of the uncertainty analysis comprised the inventory
parameter uncertainty. This analysis was achieved by the use of appropriate published
uncertainty estimations along with stochastic simulation (i.e. Monte Carlo simulation) in
order to represent inventory data uncertainty as probability distributions. Relevant
inventory parameters were identified and further analyzed, and the final results of the
comparative LCA along with the combined uncertainties from inventory data and
scenarios were obtained. A discussion of those results was presented in Chapter Four.

Several conclusions can be drawn from these last two chapters. For instance, the
procedure for performing the LCA according to ISO’s recommendations was both time
consuming, and data intensive and heavily relied on gathering of quality data. However,
this comparative LCA is still a streamlined version and thus, its results and conclusions

need to be handled acknowledging such limitation. Nevertheless, its average results are in

170

 

 

line with a similar comparison study performed by Franklin Associates (The
environmental impact of softdrink delivery systems, 1995). This is, the average energy
values of PET and Aluminum packaging systems are within the confidence interval of the
present work (e.g. PET system uses 6140 MJ/ 1000 L drink at 36% recycling rate; soft
drink; 6698 MJ/ 1000 L drink at 62% recycling rate). This situation suggests a point in
favor towards the accuracy of the LCA result of comparative studies, since the data
sources used for this and the aforementioned study can well be considered different.

However, as explained in Chapter One, this dissertation investigated the
hypothesis that when some uncertainties inherent to the methodology (i.e. data used) and
the nature of industrial processes (i.e. scenarios) are considered, the current life cycle
assessment approach does not provide enough evidence to rule out (or favor) one
alternative versus another when packaging systems are studied and compared based on
their environmental performance.

As shown in Chapter Four, the simultaneous assessment of uncertainty due to LCI
data and scenarios relevant to packaging has demonstrated that, in this particular
comparison, LCI parameter uncertainty seems to have a dominant effect in the outcome
of the LCA results considered. Furthermore, in most cases the uncertainty intervals
indicated that, despite the differences in the primary raw materials, the systems under
study can be similar to one another, at least with regard to the environmental indicators
included here, by including the aforementioned noise, or considering temporal or spatial
information or allocation procedures.

Based on these results, it can be seen that the hypothesis of this study holds true

especially when the knowledge of the systems under study is limited to average values of

I71

the different operations considered. This also confirms that when more information is
known about how. where and when the system operates (i.e. location and time of
operations, type of technology, raw material quality) more conclusive results and
conclusions can be obtained. For instance, it was seen that if the alumina extraction
operations were in fact performed in the U.S., and with more updated information
regarding emissions that affect ozone depletion (i.e. carbon tetrachloride), the ODP value
of aluminum would have been much lower and comparable to those of PLA or PET.

Not having a priority—based approach for performing the impact assessment study
could be also detrimental when interpreting the results. In fact, the inclusion of different
environmental indicators might as well affect the overall outcome, since there is no
agreed basis for the expression “environmental performance” which in fact could
encompass any set of environmental indicators. The survey presented in Table 10 of
Chapter Two of this dissertation has shown that comparison studies usually limit their
scope to listing inventory parameters such as energy use and some air emissions. That
approach, naturally followed because that kind of data is often more readily available and
thus saves time, indirectly avoids the added complexity of estimating environmental
impact indicator scores. Nevertheless, the problem still remains as to what inventory
value or impact indicator is more important.

For instance in this dissertation, considering the ODP and GWP values of PET
and PLA, we see that their median values presented in Chapter Four are very much the
same for ODP values, but when GWP is considered, PET has a higher emission profile. If
the median water use profile is included, this study showed that the PLA system would

consume more water than PET. In such a case then, a logical systematic approach that

172

can be followed is the use of a priority-based system that can help assign weights to the
different impacts. In turn, the development of such priority-based methods is the subject
of several publications. and by allowing user decisions has the perceived disadvantage of

subjective assessments.

1. Recommendations for future work

A fundamental understanding of the interrelationships governing the results of
comparative LCAs was investigated in this research. Critical considerations affecting the
LCA results were identified. A methodology was used in order to help structure the
uncertainty analysis, aggregating the different types of uncertainty, and identifying the
most important sources of uncertainty. However, there are many areas that need to be
investigated further.

The following recommendations can be made to improve the LCA method and to
further the understanding of situations with regard to their environmental footprint:

1. Improvements in both the reduction of uncertainty in the LCI information, and the
development of LCI databases with uncertainty information, are needed. This is
especially important for new data and new measurements. Though qualitative in
nature, the use of data quality goals (DQG) can help provide a means for such
uncertainty reduction. Further development of location specific and time specific
databases and impact assessment models for a more systematic analysis of
scenario and model uncertainty are needed.

2. Application of similar uncertainty studies to other packaging situations involving
different scenarios, products and packaging formats will be required. Moreover,
uncertainty considerations in other aspects of the LCA methodology (e.g.
functional unit selection) and its effect on conventional as well as future
packaging concepts and functions (e.g. degradability, traceability, sensing), need

to be investigated.

174

3.

Inventory databases will need to be developed not only for “conventional”
packaging materials and components but for new packaging materials and
concepts as well. Details and more thorough technical data from new technologies
(e.g. PLA polymerization) will be needed to provide the clearest environmental
picture possible. Failure to do so will reward the least known technology with a

supposed environmental preferability.

175

APPENDIX

176

Table A. 1. Uncertainty factors (k) for categories of parameters used in the screening for

relevant variables

 

 

 

 

 

 

 

 

 

 

 

 

Item k factor* Distribution
Resources used
Central resources 2 Log-normal
Non-central resources 10 Log-normal
Emissions
C02, SO; (calculated from corresponding inflows) 2 Log-normal
Other energy related air emissions 10 Log-normal
Other process specific air emissions 10 Log-normal

 

 

* values obtained from Finnveden (Finnveden and Lindfors, 1998) and Huijbregts et a1

(Huijbregts et al., 2003)

Table A. 2. Uncertainty factors and distributions of parameters with at least 1%
contribution to the variance of overall LCA outcomes

 

 

 

 

 

 

 

 

 

 

 

 

 

Item Characteristics Distribution
PET bottle weight, Kg k factor: 1.01* Log-normal
Drink volume (bottle), Kg (16, 16.16)** Normal
Drink volume (can), Kg (12, 12.12)** Normal
Injection molding energy k factor: 3.02*** Log-normal
Can making energy k factor: 1.01 *** Log-normal
Halon 1301 in extraction and energy generation processes
(i.e. A1203 production, electricity generation) k factor: 2.3**** Log-normal
NOx in energy generation processes
(i.e. electricity, natural gas extraction) R factor: 3**** Log-normal
NOx due to road transport processes
(i.e. service by 16 ton and 40 ton truck) k factor: 2.9**** Log-normal
Methane air emission as leakage
(i.e. PET and PLA polymerization. aluminum production) k factor: 1.1**** Log-normal

 

* Own estimations based on review of commercial literature

** Own estimation assuming a threshold of 3 standard deviations. Upper and lower 3 sigma

values shown.

*** value obtained from SAEFL report (LCI for packagingS- Vol- 11., 19981?)
new»: value obtained from Huijbregts et a] (Huijbregts et al., 2003)

177

 

I Visual basic code used for scenario analysis procedure

Sub Macrol()

'Macrol Macro

' Macro recorded 7/17/2005 by Dario Martino

Dim A

Dim B

Dim C

Randomize [Timer]

A = Rnd

If A >= 0.5 Then
Windows("PDS31.xls").Activate
Sheets("Distribution 1").Range("C8") = "2500"
Else
Windows("PDS31.xls").Activate
Sheets("Distribution 1").Range("C8") = "200"
End If

B = Rnd

If 3 >= 0.5 Then
Windows("PD531.xls").Activate
Sheets("Material production").Range("E6") = "0"
Windows( "PDSZ l .xls" ).Activate
Sheets("Material production").Range("E8") = "30"
Windows("PDSl 1.xls").Activate
Sheets("Material production").Range("E6") = "0"
Else
Windows("PDS31.xls").Activate
Sheets("Material production").Range("E6") = "0"
Windows("PDSZ1.xls").Activate
Sheets("Material production").Range("E8") = "60"
Windows( "PDSI 1.xls" ).Activate
Sheets("Material production").Range("E6") = "0"
End If

C = Rnd

If C >= 0.5 Then
Windows("PDSl 1.xls").Activate
Sheets("End of life").Range("D3") = "10"
Windows( "PD821.xls").Activate
Sheets("End of life").Range("D3") = " 19"
Windows("PDS31.xls").Activate
Sheets("End of life").Range("D3") = "0"
Else
Windows("PDSl 1.xls").Activate
Sheets("End of life").Range("D3") = "39"
Windows("PDS2.1 .xls").Activate
Sheets("End of life").Range("D3") = "75"
Windows("PDS31.xls").Activate
Sheets("End of life").Range("D3") = "10"
End If

End Sub

178

 

IIResults of the quantitative uncertainty analysis

Table A. 3. Scenario-specific interval results for the relationship PDSl/PDSZ

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A11
95th tile- Median- Significantly
PDS 1 vs Median = 5th tile = different
PDS2 Median 95th tile Upper end 5th tile Lower end than 1?
Energy 0.85 1.26 0.41 0.55 0.3 0
Water use 0.69 2.78 2.0 0.17 0.52110
GWP 0.9266 1.4532 0.5266 0.5528 037381110
ODP 0.4319 0.9087 0.4768 0.144 0.2s7zlves
Scenario A
05th tile- lvledian- Significantly
PDS 1 vs Median = 5th tile = different
PDS2 Median 95th tile Upper end 5th tile Lower end than 1?
Energy 0.741 1.09 0.35 0.49 0.25110
Water use 0.6 2.63 1.97 0.15 0.51 no
GWP 0.8063 1.2781 0.4718 049811 0.3079no
ODP 0.4131 0.8697 0.4566 0.13m 0.282 es
Scenario B
95th tile- Median- Significantly
PDS 1 vs Median = 5th tile = different
PD82 Median 95th tile Upper end 5th tile Lower end than 1?
Energy 0.75 1.09 0.34977 0.49 0.256567no
Water use 0.67 2.56 1.894987 0.15 052034200
GWP 0.8080 1.2691 0.461101 0.4877 03202351110
ODP 0.4095 0.8705 0.461019 0.1286 0.280883[yes
Scenario C
95th tile- Median- Significantly
PDS 1 vs median = 5th tile = different
PDSZ Medran 95th tile pper end 5th tile Lower end than 1?
Energy 0.7 1.10 0.351927 0.50 02413261110
Water use 0.65 2.52 1.869611 0.15 0501137110
GWP 0.80741 1.2783 0.4709341 0.4853 0322081110
ODP 0.39551 0.8900 0.494449! 0.1203 0.275226 es

 

 

 

 

 

 

 

179

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Scenario D
95th tile- Median— Significantly
PDS 1 vs Median = 5th tile = different
PDS2 edian 95th tile Upper end 5th tile Lower end than 1?
Energy 074 1.08 0.3 0. 0.24m
Water use 0.66 2.64 1.98 0.15 0.5 1 no
GWP 0.8073 1.2793 0.472 0.4916 0.3157n0
ODP 0.3929 0.8681 0.4752 0.1999 0.193 es
Scenario E
95th tile- Median- Significantly
PDS 1 vs Median = 51h tile = different
PDS2 Median 95th tile Upper end 5111 tile Lower end than 1?
Energy 0.95 1.34 0.383011 0.67 0.27801no
Water use 0.70 2.73 2.034543 0.19 05078881110
GWP 1.0264] 1.5573 0.530945 0.6831 0343271110
ODP 0.4758T 0.9274 0.451577 0.1823 0.293524|yes
Scenario F
95th tile- Median- Significantly
PDS 1 vs Median = 5th tile = different
PDSZ Median 95th tile Upper end 5th tile Lower end than 1?
Energy 0.95 1.35 0.403741 0.67 0.28043 0
Water use 0.70 2.72 2.021116 0.19 0512914an
0.695243
GWP 1.03139169 1.5733 0.541882 182 0.336149Ino
0.178472
ODP 0.4765881] 0.9324 0.455764 687 0.298115 es
Scenario G
95th 010— Median- Significantly
PDS 1 vs Median = 5th tile = different
PDS2 Median 951h tile Upper end 51h tile Lower end than 1?
0.674190
Energy 0.95 132584727 0.377787 503 0.27387no
0.187439
Water use 0.70 2.82054096 2.1236 779 0.509501no
0.680428
GWP 1.0350 1.57705584 0.542031 21 0.354596n0
0.167696
bDP 0.4570 092586769 0.46891 09 0.289261 es

 

 

 

 

 

 

 

 

 

180

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Scenario H
951111116- Median- Significantly
PDS 1 vs Median: 5th tile: different
PDS2 Median 95th tile Upper end thtile Lower end than 1?
Energy 0.95 1.32 0.370207 0.68 0.267389 0
Wateruse 0.69 2.67 1.986353 0.19 0.50148 0
0.682816
GWP 1.03315513 1.58352526 0.55037 785 03503381110
0.169693
ODP 0457036541 0.9345447 0.477508 439 0.287343 es

 

 

181

 

 

Table A. 4. Scenario-specific interval results for the relationship PDSl/PDS3

Scenario A11

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

95111tile- Median-5th Significantly
1305 Median = tile = Lower different
l/PDS3 Median 951h tile Upper end 5th tile end than 1?
Energy 0.99 1.43 0.44 0.68 0.31 No
Water use 0. 2.3059 1.9059 0.05 0.35 No
GWP 1.5124 1.96 0.4476 1.0998 0.4126Yes
ODP 0.9181 1.0244 0.1063 0.7165 0.2016 N0
Scenario A
95th tile- Median-51h Significantly
PDS Median = tile = Lower different
l/PDS3 Median 951h tile Upper end 5th tile end than 1?
Energy 0.99 1.43 0.441 0.68 0.3 1 no
Water use 0.41 2.06 1.65 0.06 0.35-‘10
GWP 1.5752 2.3719 0.7967 1.17141 0.4038 es
013? 0.9433 1.0331 0.0898 0.75981 0.1835jio
Scenaiio B
95th tile- Median-51h Significantly
PDS Median = itile = Lower different
l/PDS3 Median 9Sth tile Upper end 5th tile end than 1?
Energy 0.97 1.41 0. 0.67 0.30110
Water use 0.39 1.85 1.461654 0.05 0332677110
GWP 1.464 2.2147 0.750673 1.0812 0.382823 es
ODP 0.9421 1.0313 0.08923 0.755% 0.186735 0
Scenario C
95111 tile- Median-5th Significantly
PDS Median = tile = Lower different
l/PDS3 Median 95th tile Upper end 5th tile end than 1‘?
Energy 1.00 1.46 0.46 0.69 0.31n0
Water use 0.41 1.90 1.496007 0.05 0.35463 no
GWP 1.5625 2.3793 0.816795 1.1496 0.412949 es
ODP 0.8848 1.01 15 0.12671 0.7006 0.184135p0

 

 

 

 

 

 

 

 

182

 

 

 

 

Scenario D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9Sth tile- Median-51h Significantly
PDS Median = tile = Lower different
1/PDS3 Median 95m tile Upper end 51h tile end than 1?
Energy 0.99 1.41 0.42 0.69 0.31110
ater use 0.39 1.92 1.53 0.06 0.3 o
GWP 1.4453 2.1899 0.7446 1.0678 0.377 es
DP 0.8849 1.0127 0.1278 0.6995 0. 18541110
Scenario E
95th tile- Median-5th Significantly
PDS Median = tile = Lower different
1/PDS3 Median 85th tile Upper end 5th tile nd than 1?
“nergy 0.99 1. 0.45 0.68 0.31 no
Water use 0.41 1.91 1.4988141 0.05 0.35529 0
GWP 1.5723 2.3642 0.791854 1.1591 0.413205 es
ODP 0.9427 1.0340 0.0913124 0.7583 0.18432no
Scenario F
95111 tile- Median-5th Significantly
PDS Median = tile = Lower different
l/PDS3 Median 95111 tile Upper end 5th tile end than 1?
Energy 0.9 1.40 0.427719 0.67 0.301291 no
Water use 0.39 2.82 2.425947 0.06 0.337271 no
1.0782631
GWP 1.471083934 2.1989 0.7278 07 0.392821 es
0.7477748
ODP 0942525948 1.0323 0.08973 1 0.194751 no
Scenario G
95th tile- Median-5th Significantly
PDS Median = 'le = Lower different
1/PDS3 Median 5th tile Upper end 5th tile end than 1?
0.6893836
Energy 1.00 1.458332093 0.457097 01 no
0.0564527
Water use 0.42 1.976645174 1.554277 43 0.365916no
1.14232
GWP 1.57102.376320845 0.805328 43 0.428669 es
0.6951944
ODP 0.8855 1011939197 0.126399 22

 

 

 

 

 

 

 

 

 

183

 

 

 

 

 

Scenario H

 

 

 

 

 

 

 

 

 

 

 

 

 

 

951h tile- Median-5th Significantly
PDS Median = tile = Lower different
1/PDS3 Median 9Sth tile Upper end 5th tile end than 1?
Energy 0.99 1.43 0.45 0.69 0.301110
Water use 0.39 1.87 1.483024 0.05 0336426110
1.0657115
WP 1.4517 2203100239 0.751395 3 0.385993 es
0.6907850
ODP 0.8837 1.016803825 0.1331041 5 01929151110

 

 

184

 

Table A. 5. Scenario-specific interval results for the relationship PDSl/PDS3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

All
95th tile- Median- Significantly
Median = 5111 tile = different
PDS 2 vs PDS3 Median 95111 tile Upper end 111 tile Lower end than 1?
Energy 1.16 1.84 0.68 0.78 0.3 0
Water use 0.57 2.82 2.25 0.08 0.49no
GWP 1.6518 2.724 1.0722 1.1233 0.5285Yes
ODP 2.0762 5.5667 3 .4905 1.0563 1.0199Yes
Scenario A
951h tile- Median- Significantly
[Median = 51h tile = different
PDS 2 vs PDS3 Median 95th tile Upper end 51h tile Lower end than 1?
Energy 1.34 2.02 0.68 0.88 0.46110
ater use 0.65 3.36 2.71 0.09 0.5 o
GWP 1.9765 3.1086 1.1321 1.3619 0.614 es
()DP 2.2285 6.3081 4.0796 1.1419 1.086 es
Scenario B
9Sth tile- Median- Significantly
Median = 51h tile = different
PDS 2 vs PDS3 Median 951h tile Upper end 51h tile Lower end than 1?
Energy 1.30 1.96 0.652798 0.88 0.419601no
Water use 0.61 3.00 2.393281 0.09 0.52219 0
WP 1.8249 2.8995 1.074626 1.2725 0.55241 es
ODP 2.2565 6.4841 4.22756 1.1348 1.121762 es
Scenario (3
95th tile- Median- Significantly
Median = 5th tile = hifferent
PDS 2 vs PDSB Median 95th tile Upper end 5th tile Lower end than 1?
lnergy 1.35 2.02 0.670103 0.91 0.440561n0
Water use 0.63 3.17 2.544903 0.09 0541021116
WP 1.957 3.1275 1.169853 1.3399 0.617703 es
ODP 2.2095 6.4191 4.209585 1.035 1.174521 es

 

 

 

 

 

 

 

 

 

185

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Scenario D
95111 tile- Median- Significantly
Median = 5th tile = different
PDS 2 vs PDS3 Median 95m tile bpper end 51h tile Lower end an 1?
Energy 1.33 1.9 0.65 0.9 0.43110
Water use 0.61 3.05 2.44: 0.09 0.52116
GWP 1.8101 2.8776 1.0675 1.24541 0.5647 'es
DP 2.2196 6.3531 4.1335 1.05891 1.1607 es
Scenario E
95thtile- Median- Significantly
Median: 5th tile: different
PDS 2 vs PDS3 Median 951h tile Upper end 5th tile ower end than 1?
Energy 1.04 1.48 0.442204 0.73 0.305302no
Water use 0.58 2.59 2.008596 0.0 04972391110
‘WP 1.5440 2.2596 0.715639 1.1306 0.413409 es
DP 1.9532 4.5243 2.571071 1.0730 0.880167 es
Scenario F
951h tile— Median- Significantly
Median: Sthti1e= different
PDS 2vsPDS3 Median 95m tile Upper end thtile Lower end inan 1?
Energy 1.02 1.4 0.418628 0.72 0.3062461no
ater use 0.56 2.44 1.881807 0.08 0.47498 0
GWP 1.43354428 2.069756 0.636212 1.044521 0.389023 es
1.0665454
ODP 1.9487205 4.515195 2.566475 53 0.882175 es
Scenario G
95th tile- Median- Significantly
Median = 5111 tile: different
PDS 2 vs PDS3 Median 95th tile Upper end 5th tile Lower end than 1?
; 0.7474856
Energy 1.06 1.502824 0.44587 35 0309461110
0.0806949
Wateruse 0.60 2.774471 2.178857 11 0514919110
1.0990384
GWP 1.5235 2.227834 0.704369 25 0.424429 es
0.9968852
ODP 1.9110 4.612746 2.701768 75 0.914093an

 

 

 

 

 

 

 

 

 

 

186

T—

 

 

 

 

 

 

 

 

Scenario H
9Sthtile- Median- Significantly
Median = 5th tile = different
PDS 2 vs PDS3 edian 951h tile Upper end 5th tile Lower end than 1?
Energy 1.04 1.4 0.420986 0.7 03047391110
Water use 0.56 2.4 1.837566 0.08 0478341110
1.0287657
GWP 1.40781838 2.073229 0.665411 53 0.379053 es
0.9834693
ODP 1.9112555] 4.4528 2.5415641 8 0.927786po

 

 

 

 

 

 

 

187

 

Material Production

Table A. 6. Life cycle results for PDS]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

95111 tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 3978.79 5853.6 1874.85 2740.74: 1238.05
Water use 751.99 3666.83 2914.841 194.23 557.76
GWP 638.86 1134.71 495.85 407.72 231.1
ODP 0.00 0.00 0.00 0.00 0.00
Distribution 1
951h tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 22.83 43.97 21.141 1 1.941 10.89
Water use 12.99 94.49 81.50 1.80 11.19
GWP 10.78 22.63 11.85 5.73 5.05
ODP 0. 0.01 0.01 0. 0.
Component manufacturing
95th tilc- Median-5th
Median = tile =
Median 95m tile Upper end 51h tile Lower end
Energy 556.34 1476.91 920.57 225.92 330.42
Water use 1.9 15.57 13.63 0.23 1.71
GWP 100.12 276.65 176.5- 40.42 59.70
DP 0.93 6.00 5.0 0.19 0.141
Distribution 2
95th tile- Median-51h
Median = tile =
Median 951h tile Upper end 5th tile Lower end
Energy 60.06 83.95 23.89 45.00 15.06
Water use 26.81 119.71 92.90 7.69 19.12
GWP 14.03 26.60 12.57 8.21 5.82
ODP 0.08 0.16 0.08 0.041 0.04

 

 

 

 

 

 

 

188

Filling

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

95th tile- Median-5th
Median = tile =
Median 95th tile Upper end Bth tile Lower end
Energy 1084.041 1915.77 831.73 357.62 726.42
Water use 3.441 25.72 22.28 0.34 3.10
GWP 191.88 378.10 186.22 62.33 129.55
ODP 1.641 9.79 8.15 0.241 1.40
Distribution 3
95th tile- Median-5th
Median = tile =
Median 95th tile Upper end 6th tile Lower end
Energy 651.37 976.3 324.97 443.6 207.70
Water use 229.77 1160.3 930.53 61.5 168.20
GWP 59.91 127.88 67.97 35.85 24.06
oop 2.62 5.30 2.68 1.31 1.31
End of life
95th tile- Median-5th
Median = tile =
edian bSth tile Upper end Sth tile , Lower end
Energy -564.02 -380.04 183.98 -838.86 274.84
Water use -1.42 7 .15“ 8.57 -21.94I 20.52
WP 30.11 94.57 64.46 -26. 10 56.21
DP 0.7 7.141 6.37 -072 1.49
Total
95th tile- Median-51h
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 5955.71 8086.07 2130.36 4419.80 1535.91
Water use 1200.38 4387.42 3187.041 417.06 783.32
GWP 1094.52 1655.33 560.81 787.5 306.98
ODP 7.35 22.60 15.25 3.11 4.2

 

 

 

 

 

 

 

189

Material Production

Table A. 7. Life cycle results for PDS2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

95th tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 4858.18 8088.68 3230.50 3229.52 1628.66
Water use 1217.77 4973.07 3755.30 368.62 849.15
GWP 824.99 1494.67 669.68 51 1.92 313.07
ODP 9.12 30.2 21.1 4.03 5.09
Distribution 1
95th tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 14.94 28.62 13.68 7.70 7.241
Water use 8.45 62.18 53.73 1.15 7.30
GWP 6.97 14.45 7.48 3.73 3.241
ODP o. 0.00 0.00 0.00 0.011
Component manufacturing
95m tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 1180.53 1871.33 690.80 744.69 435.841
ater use 3.82 26.83 23.01 0.54 3.28
GWP 207.86 372.04 164.18 127.33 80.53
ODP 1.90 10.75 8.85 0.41 1.49
Distribution 2
951h tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 135.64 193.18 57.5 97.28 38.36
Water use 49.18 211.17 161.99 13.53 35.65
GWP 15.49 29.93 14.441 9.54 5.95
ODP 0.45 0.91 0.461 0.23 0.22

 

 

 

 

 

 

 

190

Filling

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

95th tile- Median-51h
Median = tile =
Median 951h tile Upper end 5th tile Lower end
Energy -0.29 -0. 18 0.1 1 -0.46 0.17
Water use 0.00 0.00 0.00 0.00 0.00
GWP 0.03 0.09 0.06 0.02 0.01
ODP 0.00 0.01 0.01 0.00 0.00
Distribution 3
95th tile- Median-51h
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 849.78 1283.38 433.60 570.93 278.85
Water use 288.42 1396.53 1108.11 75.041 213.38
GWP 70.17 156.88 86.71 42.47 27.70
ODP 3.49 7.01 3.52 1.73 1.76
End of life
95th tile- Median-5th
Median = tile =
Median 95th tile , Upger end 5th tile Lower end
Energy 472.92 -68.19 104.78 -393.75 220.78
Water use 1.99 13.90 1 1.91 0.3 1.69
GWP 25.32 65.05 39.73 13.0 12.23
ODP 0.65 4.71 4.0 0.1 1 0&1
Total
95th tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 6954.1 10205.44 3251.34 5150.63 1803.47
Water use 1777.48 5924.91 4147.43 663. 1113.641
GWP 1190.32 1881.50 691.18 840.16 350.16
ODP 17.99 42.69 24.70 10.17 7.82

 

 

 

 

 

 

 

 

191

Material Production

Table A. 8. Life cycle results for PDS3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9Sth tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 4080.39 6128.49 2048.10 2807.50 1272.89
Water use 2596.71 16160.74 13564.03 489.64! 2107.07
WP 243.25 447.17 203.92 177.99 65.26
DP 0.52 0.99 0.47 0.28 0.241
Distribution 1
95111 tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 29.73 189.50 159.77 7.59 22.141
Water use 20.95 315.87 294.92 1.39 19.56
GWP 16.14 93.43 77.29 3.58 12.56
ODP 0.00 0.02 0.02 0.00 0.00
Component manufacturing
95th tile- Median-51h
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 524.28 1387.48 863.20 214.19 310.09
Water use 1.75 14.23 12.48 0.22 1.53
GWP 94.70 260.91 166.21 37.83 56.87
DP 0.87 5.82 4.95 0.17 0.70
Distribution 2
95th tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
Energy 58.63 80.66 22.03 44.09 14.5
Water use 25.72 112.11 86.39 7.33 18.39
GWP 13.22 24.98 11.76 7.841 5.38
ODP 0.08 0. 16 0.08 0.04 0.041

 

 

 

 

 

 

 

192

Filling

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

85th tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile ower end
Energy 987.26 1755.48 768.22 313.85 673.41
Water use 2.94 22.70 19.76 0.29 2.65
GWP 172.16 345.68 173.52 53.98 118.18
ODP 1.4 8.66 7.20 0.21 1.25
Distribution 3
95th tile- Median-5th
Median = tile =
Median bSth tile Upper end 5th tile Lower end
Energy 655.15 975.85 320.70 442.67 212.48
Water use 232.45 1102.4 869.95 60.37 172.08
GWP 59.42 127.88 68.46 36.27 23.15
DP 2.64 5.19 2.55 1.31 1.33
End of life
9Sth tile- Median-5th
Median = tile =
Median 95th tile Upper end 5th tile Lower end
EnergL ~515.1 -357.20 157.99 -758.20 243.01
Water use -1.69 2.19 3.88 -9.54 7.85
GWP 43.49 127.72 84.23 14.80 28.69
ODP 1.14 8.61 7.47 -0.07 1.21
Total
9Sth tile- Median-5th
Median = tile =
Median 95th tile Upper end th tile Lower end
Energy 5997.49 8172.05 2174.56 4461.40 1536.09
Water use 3132.80 16700.86 13568.06 828.81 2303.99
GWP 704.03 1050.56 346.53 526.841 177.19
ODP 8.19 22.95 14.76 3.97 4.22

 

 

 

 

 

 

 

 

193

IIIDetails of the sensitivity analysis using Spearman’s Rank correlation method
Technical note from:
http://www.decisioneering.com/support/simulation/cbl_gen__004A.html
How does Crystal Ball calculate sensitivity?
Sensitivity Analysis is an option in the Run Preferences dialog. When it is on, Crystal
Ball calculates how important each assumption is to each forecast while the simulation is
running.
Crystal Ball calculates sensitivity this way:

1. During a simulation, Crystal Ball saves all the calculated assumption and forecast

values.

2. After a sample size (another Run Preferences option), N , number of trials, Crystal
Ball generates a list of ranks. The smallest number is 1, and the largest number is

 

 

 

 

 

the N.
Table A. 9. Details of sensitivity analysis
Assumption Assumption Assumption Assumption Forecast 1 Forecast 1
1 values 1 ranks 2 values 2 ranks values ranks
12.43 3 —2.44 3 23.01 2
4.82 1 -5.90 2 47.22 3
6.01 2 -12.53 1 7.70 l

 

 

 

 

 

 

 

3. Crystal Ball pairs up each assumption rank list with each corresponding forecast
rank list and then calculates a Pearson correlation coefficient for each pair.

Numerical example

XY
12
25
36

194

 

zxv :11112)+121151+(31161= 30
2X : 1 + 2 + 3 = 6
z><1’=11’+21’+32 = 14

2? = 2 + 5 + 6 = 13

2V2: 22 + 52 + 62 : 65

N:3

zxv—zxzvm = 30 ~1611131/3 = 4
2x2 - (2102/11 = 14 - 62/3 .-. 2
2112— 1'2:an = 65 432/3 = 6.667
r : 4 x11 121166667 = 4/4.16333

: .9608

4. If the model has more than one forecast, Crystal Ball also pairs up all the
corresponding forecast rank lists and then calculates a Pearson correlation
coefficient for each pair of forecasts.

5. Crystal Ball stores the running sums of the Pearson correlation coefficients for
each pairing in separate matrices.

6. After the simulation, Crystal Ball averages the computed correlation coefficients.

If the sensitivity is measured by Rank Correlation (1n the Sensitivity Preferences dialog),
these average values appear in the Sensitivity chart.

1f the sensitivity is measured by Contribution To Variance (in the Sensitivity Preferences
dialog), Crystal Ball normalizes the square of the sensitivity coefficient. Crystal Ball:

Squares the sensitivity coefficient.

Adds all the squares.

Divides the squares by the sum of all the squares to ensure a sum of 1.
Converts each normalized value to a percentage.

P1919.“

These normalized percentage values appear in the Sensitivity chart.

For the most accurate analysis. set the Sample Size option to the number of trials for the
simulation.

195

REFERENCES

Anyadike, N. Atalink Ltd for Pira lntemational. 2004. Nanotechnology in packaging.
London, UK.

Appleton, A.E. Kluwer Law lntemational. 1997. Environmental labelling programmes :
international trade law implications. Kluwer Law International, London ; Boston.

Arena, U.. M.L. Mastellone, and F. Perugini. 2003. Life cycle assessment of a plastic
packaging recycling system. Int. J. LCA 8:92-98.

Bare, J.C., P. Hofstetter, D.W. Pennington, and HA. Udo de Haes. 2000a. Life cycle
impact assessment workshop summary midpoints versus endpoints: The sacrifices
and benefits. Int. J. LCA 52319-326.

Bare, J.C., P. Hofstetter, D.W. Pennington, and HA. Udo de Haes. 2000b. Life Cycle
Impact Assessment Workshop Summary
Midpoints versus Endpoints: The Sacrifices and Benefits. Int. J. LCA 5:319-326.

Bare, J .C., D.W. Pennington, and H.A. Udo de Haes. 1999. Life cycle impact assessment
sophistication. International workshop. Int. J. LCA 42299-306.

Bamthouse. L.. J.A. Fava. K. Humphreys, R.B. l-lunt, L. Laibson, S. Noesen, G. Norris. J.
Owens, J.A. Todd, B. Vigon, K. Weitz, and J. Young. Society of Environmental
Toxicology and Chemistry (SETAC). 1998. Life-cycle impact assessment: the
state-of-the-art. Pensacola, FL.

Ben gtsson, M. 2001. Weighting in practice. Implications for the use of life-cycle
assessment in decision making. Journal of Industrial Ecology 4:47-60.

Bennet, E.B., and TE. Graedel. 2000. “Conditioned air”: evaluating an environmentally
preferable service. Environ. Sci. Technol. 34:541-545.

Bj0rklund, A. 2002. Survey of approaches to improve reliability in LCA. Int. J. LCA
7:64-72.

Bohlman, GM. 2004. Biodegradable packaging life cycle assessment. Environmental
Progress 23:342-346.

196

Boustead, 1. Association of Plastics Manufacturers in EurOpe (APME). Ecoprofile of
polyethylene (HD) bottles. Brussels.

Boustead. 1. Association of Plastics Manufacturers in Europe (APME). Ecoprofile of
polyethylene (HD) resin. Brussels.

Boustead. 1. Association of Plastics Manufacturers in Europe (APME). Ecoprofile of
polyethylene (LD) film. Brussels.

Boustead, 1. Association of Plastics Manufacturers in Europe (APME). Ecoprofile of
polyethylene (LD) resin. Brussels.

Boustead, 1. Association of Plastics Manufacturers in Europe (APME). Ecoprofile of
polyethylene terephthalate (bottle grade). Brussels.

Boustead, 1. Association of Plastics Manufacturers in Europe (APME). 2002. Ecoprofile
of polyethylene terephthalate conversion processes. Brussels.

Boustead, 1. Association of Plastics Manufacturers in Europe (APME). Ecoprofile of
polyethylene terephthalate film. Brussels.

Boustead, 1. Association of Plastics Manufacturers in Europe (APME). Ecoprofile of
polypropylene injection moulding. Brussels.

Boustead, 1. Association of Plastics Manufacturers in Europe (APME). Ecoprofile of
polypropylene resin. Brussels.

Boustead, 1. Association of Plastic Manufacturers of Europe (APME). 2003. Eco-profiles
of the European plastics industry: Methodology. Brussels.

Boustead, 1., and GP. Hancock. Ellis Horwood Ltd. 1981. Energy and packaging. Ellis
Horwood Ltd., West Sussex, England.

Braam, 1., TM. Tanner, C. Askham, N. Hendriks, B. Maurice. H. Mfilkki, M. Vold, H.
Wessman, and A.S.H. de Beaufort. 2001. Energy, transport and waste models.

Availability and quality of energy, transport and waste models and data. Int. J.
LCA 6:135-139.

197

Canter, K.G., D.J. Kennedy, D.C. Montgomery, J.B. Keats, and W.M. Carlyle. 2002.
Screening stochastic life cycle assessment inventory models. Int. J. LCA 7.

CAPE Systems: palletization and package design software. Release 2000.2000. CAPE
Systems Inc., Allen. TX.

Ciroth, A.. G. Fleischer, and J. Steinbach. 2004. Uncertainty calculation in life cycle

assessments: A combined model of simulation and approximation. Int. J. LCA
9:216-226.

Committee draft: Environmental management - Life cycle assessment - Part 3: Life cycle
impact assessment.[ntcrnational Organization for Standardization. 1998.
lntemational Organization for Standardization, Geneva. Switzerland.

Contadini, F., and RM. Moore. 2003. Results of a life cycle assessment using uncertainty
analysis of fuel cell vehicles within the south coast air basin of California in 2010.
Gate to EHS:l-15.

Crystal ball - Forecasting and risk analysis for spreadsheet users. Release Professional
academic edition v. 5.5.2004. Decisioneering, Inc., Denver.

Crystal ball - User manualDecisioneering. Inc. 2004. Denver.
Curran, M.A. 1999. The status of LCA in the USA. Int. J. LCA 4: 123-124.

Data for Environmental Analysis and Management (DEAM) module databases and
manuals. Version 3.0.Ecobilan Group, Inc. 1999. Paris, France.

Davis, G.A., United States Environmental Protection Agency Prevention Pesticides and
Toxic Substances, and Abt Associates. U.S. Environmental Protection Agency
Office of Prevention Pesticides and Toxic Substances. 1998. Environmental
labeling issues, policies. and practices worldwide. Washington, DC.

Davis, G.A., C.A. Wilt, P.S. Dillon, B.K. Fishbein, United States Environmental
Protection Agency Office of Solid Waste, U.o. Tennessee, T. University, and 1.
Inform. U.S. Environmental Protection Agency Office of Solid Waste. 1997.
Extended product responsibility: a principle for product oriented pollution
prevention. Washington, DC.

198

Dreyer, L.C., A.L. Niemann, and M2. Hauschild. 2003. Comparison of three different
LCIA methods: EDIP97. CML2001 and Eco-indicator 99. Int. J. LCA 8:191-200.

Economic instruments in packaging and packaging waste policy.The European
Organization for Packaging adn the Environment (EUROPEN). 2000. Brussels,
Belgium.

Eide, M.H. 2002. Life cycle assessment (LCA) of industrial milk production. Int. J. LCA
7:1 15- 126.

Elkington, J ., and J. Hailes. 1993. The LCA sourcebook: a European business guide to
life-cycle assessment. London.

Environment: Trade and environment news bulletin.World Trade Organization. 1996.
TE/013 - 27 September 1996 Geneva, Switzerland.

The environmental impact of softdrink delivery systems.National Association for Plastic
Container Recovery (NAPCOR). 1995.

Environmental labels and declarations - Type [environmental labelling - Principles and
procedureslntemational Organization for Standardization. 1999. ISO
14024: 1 999(E) Geneva, Switzerland.

Environmental management - Life cycle assessment - Goal and scope definition and
inventory analysis.1ntemational Organization for Standardization. 1998.
lntemational Organization for Standardization, Geneva, Switzerland.

Environmental management - Life cycle assessment - Life cycle
interpretation.1ntemational Organization for Standardization. 2000. ISO
l4043:2000(E) Geneva, Switzerland.

Environmental management - Life cycle assessment - Principles and
frameworklnternational Organization for Standardization. 1997. ISO
14040: 1997(E) Geneva, Switzerland.

Environmentally Preferable Purchasing US Environmental Protection Agency. 2005.
[Online]. Available from US Environmental Protection Agency
httpz/lwwwcpagov/opptintr/cpp/ (posted December 8, 2004; verified March 14,
2005).

199

EPA partners with IDSA Industrial Designers Society of America. 2005. [Online].
Available from Industrial Designers Society of America
http://www.idsa.org/webmodules/articles/anmviewer.asp?a=885 (verified

February 9, 2005).

 

Finnveden, G. 2000. On the limitations of life cycle assessment and environmental
systems analysis tools in general. Int. J. LCA 5:229-238.

Finnveden, G. 1997. Valuation methods within LCA: where are the values? Int. J. LCA
21163-169.

Finnveden, G., and L. Lindfors. 1998. Data quality of life cycle inventory data — Rules of
thumb. Int. J. LCA 3:65-66.

Gemgross, T.U. 1999. Can biotechnology move us toward a sustainable society? Nature
Biotechnology 17:541-544.

Gemgross, T.U. 2000. How green are green plastics? Scientific AmericanAugust 2000).

Giles, G.A. Sheffield Academic. 1999. Handbook of beverage packaging. Sheffield
Academic.

Global LCI matrix US Environmental Protection Agency. 2002. [Online]. Available from
US Environmental Protection Agency
httpz/loaspubepagov/pls/lcaccess/LCA.MATRIX (posted November 19, 2002;
verified March, 2005).

 

Graedel, TE, and B.R. Allenby. Prentice Hall. 2003. Industrial ecology. 2nd ed. Prentice
Hall, Upper Saddle River, NJ.

Guidance for the EPA Halon emission reduction rule (40 CFR Part 82, Subpart H).U.S.
Environmental Protection Agency. 2001. EPA430-B-01-001 Washington, DC.

Guyonnet, D., B. Come, P. Perrochet, and A. Parriaux. 1999. Comparing two methods for
addressing uncertainty in risk assessments. Journal of Environmental
Engineeringz660-666.

Hansen. DJ. 1999. Status of LCA activities in the Nordic region. Int. J. LCA 4:315-320.

Harrop, P.J. Pira lntemational Ltd. 2002. RFID in packaging. Pira International Ltd..
Surrey, UK.

Hart, K.M.. D.L. Boger, and M.A. Kerr. EPA. 1995. Design for the environment:
partnership for a cleaner future. EPA/744/F-96/018A.

Heijungs, R., and S. Suh. Kluwer Academic Publishers. 2002. The computational
structure of life cycle assessment. Kluwer Academic Publishers, Dordrecht.

Hendrickson, C., A. Horvath, S. J oshi, and L. Lave. 1998. Economic Input-Output
models for environmental life cycle assessment. Environmental Science and
Technology 32: 184A-191A.

Hochschomer, E., and G. Finnveden. 2003. Evaluation of two simplified life cycle
assessment methods. Int. J. LCA 8:119-128.

Hockerts, K., T. Rotherham, and D. Esty. Centre for the Management of Environmental
and Social Responsibility (CMER). 1999. Bottle wars: using LCA in a trade
dispute [Online]. Available from Centre for the Management of Environmental
and Social Responsibility (CMER)
http://www.insead.edu/CMER/research/ccon/bottlcwars.htm (verified July, 2005).

Hocking, MB. 1991. Paper vs. Polystyrene: A complex choice. Science 251(4993):504-
505.

Houghton, J .T. Cambridge University Press. 1995. Climate change: the science of climate
change. Cambridge University Press, Cambridge, NY.

Huijbregts, M.A. 1998a. Application of uncertainty and variability in LCA. Part I: A
general framework for the analysis of uncertainty and variability in LCA. Int. J.
LCA. 32273-280.

Huijbregts, M.A. 1998b. Application of uncertainty and variability in LCA. Part 11. Int. J.
LCA 3:343-351.

Huijbregts, M .A. 2001. Uncertainty and variability in environmental life—cycle.
assessment, University of Amsterdam, Amsterdam.

20]

Huijbregts, M.A., W. Gilijamse, A.M.J. Ragas. and L. Reijnders. 2003. Evaluating
uncertainty in the life cycle assessment. A case study comparing two insulation
options for a Dutch one-family dwelling. Environ. Sci. Technol. 37:2600-2608.

Huijbregts, M.A., G. Norris, R. Bretz, A. Ciroth, B. Maurice, B. von Bahr, B. Weidema,
and A.S.H. de Beaufort. 2001. SETAC-Europe LCA working group 'data
availability and data quality'. Int. J. LCA 62127-132.

Hunkeler. D., G. Rebitzer, and A.A. Jensen. 2002. The UNEP/SETAC Life cycle
initiative: Impressions from the LCM Workshop in Copenhagen, August 30.
2001. Gate to EHS: 1-3.

Hunkeler, D., G. Rebitzer, A.A. Jensen, and M. Margni. 2001. Life Cycle Management:
bridging the gap between science and application. Int. J. LCA 6.

Hunt, R.B., T.K. Boguski, K. Weitz, and A. Shanna. 1998. Case studies examining LCA
streamlining techniques. Int. J .LCA 3:36-42.

Hunt, R.B., and WE. Franklin. 1996. LCA - How it came about. Int. J. LCA 1:4-7.

Introduction to ecolabeling.Global Ecolabeling Network Secretariat. 2004. Global
Ecolabeling Network Secretariat, Otawa, Ontario, Canada.

An introduction to environmental accounting as a business management tool: key
concepts and terms.U.S. Environmental Protection Agency. Office of pollution
prevention and toxics. 1995. Washington, DC.

Inventory of GreenPla products Biodegradable Plastics Society. 2004. [Online].
Available from Biodegradable Plastics Society
http://www.bpsweb.net/02 english/ (verified March, 2005).

James. K., T. Grant, and K. Sonneveld. 2002. Stakeholder involvement in Australian
paper and packaging waste management LCA study. Int. J. LCA 7: 151-157.

Johnson, A. Sustainable Packaging Coalition - Green Blue. 2005. 49 Organizations
gather to discuss sustainable packaging solutions [Online]. Available from
Sustainable Packaging Coalition - Green Blue www.sustainablepackagingorg
(verified May 1, 2005).

202

J olliet, O., R. Mtiller-Wenk, J. Bare. A. Brent, M. Goedkoop, R. Heijungs, N. Itsubo, C.
‘ Pena, D. Pennington, J. Potting, G. Rebitzer. M. Stewart, H. Udo de Haes, and B.
Weidema. 2004. The LCIA midpoint-damage framework of the UNEP/SETAC
life cycle initiative. Int. J. LCA 9:394—404.

Kennedy, D.J., D. Montgomery. D.A. Rollier, and J .B. Keats. 1997. Assessing input data
uncertainty in life cycle assessment Inventory models. Int. J. LCA 22229-239.

Kennedy, D.J., D.C. Montgomery, and B.H. Quay. 1996. A probabilistic approach to
incorporating variable input data quality. Int. J. LCA 1:199-207.

Keoleian, G.A. University of Michigan. 2001. Life cycle assessment of the Stonyfield
Farm product delivery system. Ann Arbor, MI.

Keoleian, G.A., A.W. Phipps, T. Dritz, and D. Brachfeld. 2004. Life cycle environmental
performance and improvement of a yogurt product delivery system. Packaging
Technol. and Sci. 17285-103.

Keoleian, G.A., and D. Spitzley. 1999. Guidance for improving life-cycle design and
management of milk packaging. Journal of Industrial Ecology 3:111-126.

Kinkaid, L., G. Davis, and J. Meline. U.S. Environmental Protection Agency. 1996.

Cleaner technology substitute assessment: A methodology and resource guide.
EPA/744-R-95-002 Washington, DC.

Kléipffer, W. 2003. Life cycle based methods for sustainable development. Int. J. LCA
8:157-159.

Lavallée, S., and S. Plouffe. 2004. The Ecolabel and sustainable development. Int. J.
LCA 9:349-354.

Lave, L.B.. E. Cobas-Flores, C.T. Hendrickson, and EC. McMichael. 1995. Using input-
output analysis to estimate economy-wide discharges. Environmental Science and
Technology 29:420-426.

LCA results of aluminum can. Product: Aluminum beverage cans made from used
beverage cans (UBCs). PSR DraftMitsubishi Materials Co. 2003. Registration
no: S-EP00022-1 to 12.

203

Leaversuch, R. 2003. Materials: Renewable PLA polymer gets green light for packaging
uses. Plastics Technology.

Lee, 8.6., and X. Xu. 2004. A simplified life cycle assessment of re-usable and single-
use bulk transit packaging. Packaging Technol. and Sci. 17:67-83.

Lenzen, M. 2001. Errors in conventional and input-output—based life-cycle inventories.
Journal of Industrial Ecology 4:127-148.

Levy, G. Aspen Publishers, Inc. 2000. Packaging policy and the environment. Aspen
Publishers. Inc., Gaithersburg.

Life cycle assessment of aluminum: inventory data for the worldwide primary aluminum
industry.Five Winds Europe GmbH for the lntemational Aluminum Institute.
2003. Donzdorf, Germany.

Life cycle inventories for packagings. Vol. I.Swiss Agency for the Environment, Forests
and Landscape (SAEFL). 1998a. Berne.

Life cycle inventories for packagings. Vol. lI.Swiss Agency for the Environment, Forests
and Landscape (SAEFL). 1998b. Berne.

Life cycle inventory of packaging options for shipment of retail mail-in-order
softgoods.Franklin Associates, Inc. 2004. Prairie, KS.

Life Cycle Inventory of Western Wood Products Association (WWPA).Scientific
Certification Systems. Inc. for the WWPA. 1995.

Lingle, R. 2004. PLA makes a splash in bottled water. Packaging world
magazineDecember, 2004):20.

Martino, D. 2005. LCA of softdrink delivery systems, Fall Meeting, Symposium G -
Materials Research Society. Materials Research Society, Boston, MA.

Maslennikova, 1., and D. Foley. 2000. Xerox's approach to sustainability. Interfaces
30:226-233.

204

Matthews, D., and S. Axelrod. 2004. Whole life considerations in IT procurement. Int. J.
LCA 92344—348.

Matthews, HS. 2004. Thinking outside 'the box': Designing a packaging take-back
system. California Management Review 46: 105-1 19.

Matthews, HS, and L. Lave. 2003. Using input-output analysis for corporate
benchmarking. Benchmarking 10:152-167.

Maurice, B., R. Frischknecht, V. Coelho-Schwirtz, and K. Hungerbuhler. 2000.
Uncertainty analysis in life cycle inventory. Applications to the production of
electricity with French coal power plants. Journal of Cleaner Production 8295-108.

McCleese, D., and RT. LaPuma. 2002. Using Monte Carlo simulation in life cycle
assessment for electric and internal combustion vehicles. Int. J. LCA 72230-236.

Microsoft Excel. Release 2000.2000. Microsoft Corporation, Seattle, WA.

Municipal solid waste in the U.S.: facts and figures U.S. Environmental Protection
Agency (EPA). 2001. [Online]. Available from U.S. Environmental Protection
Agency (EPA) http://www.epa.gov/epaoswer/non-hw/muncpl/msw99.htm.

 

N akamura, S.. and Y. Kondo. 2002. Input-output analysis of waste management. Journal
of Industrial Ecology 6:39-63.

Nakaniwa, C., and TE. Graedel. 2002. Life cycle and matrix analyses for re-refined oil in
Japan. Int. J. LCA 7295-102.

Neumayer, E. 2000. German packaging waste management: a succesful voluntary
agreement with less succesful environmental effects. Eur. Env. 102152-163.

Norris, G. Pre Consultants. 2003. Database manual. The Franklin U.S. LCI library. Pre
Consultants.

Norris. G. 2001a. Integrating economic analysis into LCA. Environmental Quality
Management259-64.

Norris. G. 2001b. Integrating life cycle cost analysis and LCA. Int. J. LCA 6:118-120.

205

Norris, G.. F. Della Croce, and O. J olliet. 2003a. Energy burdens of conventional
wholesale and retail portions of product life cycles. Journal of Industrial Ecology
6:59-69.

Norris, G., and P. Notten. 2002. Current availability of LCI databases around the world.
Working draft 2a.

Norris, G., W. Trusty, and G. Meil. American Center for Life Cycle Assessment
(ALCA). 2003b. How to use the US LCI database. Seattle, WA.

Oki, Y., and H. Sasaki. 2000. Social and environmental of packaging (LCA an
assessment of packaging functions). Packaging Technol. and Sci. 13:45-53.

Ozone depletion rules & regulations U.S. Environmental Protection Agency. 2005.
[Online]. Available from U.S. Environmental Protection Agency
www.cpa.gov/ozone/titlc6/phaseout/index.html (posted March, 2005; verified
July, 2005).

Paine, F. 2002. Packaging reminiscences: some thoughts on controversial matters.
Packaging Technol. and Sci. 15:167-179.

Pant, R., G. Van Hoof, D. Schowanek, T.C.J. Feijtel, A. de Koning, M. Hauschild, D.W.
Pennington, S.I. Olsen, and R. Rosenbaum. 2004. Comparison between Three

Different LCIA Methods for Aquatic Ecotoxicity and a Product Environmental
Risk Assessment. Int. J. LCA 92295-306.

Plastic stretch wrap vs. Lock'n pop unitization system: A comparison of waste,
environmental and energy impacts for unitizing pallets.Dumbleton Consulting.
1992. Green Bay, WI.

Ross, 8.. and D. Evans. 2002. Excluding site-specific data from the LCA inventory. Int. J.
LCA 72141-150.

Ross, 8., D. Evans, and M. Webber. 2002. How LCA studies deal with uncertainty. Int. J.
LCA 7247-52.

Rousseaux, P., E. Labouze, Y. Suh, l. Blanc, V. Gaveglia, and A. Navarro. 2001. An
overall assessment of life cycle inventory quality. Int. J. LCA 62299-306.

206

Saphire, D. Inform, Inc. 1994. Case reopened: reassessing refillable bottles. Inform, Inc.,
New York.

Scarlett, L. Center for the study of American Business. 1999. Product take-back systems:
mandates reconsidered. Policy study no. 153 St. Louis. Montana, US.

Schiffler, H. Duales System Deutschland AG. 2002a. The dual system in Germany - An
assessment of sustainability and prospects for the future. Status quo and
development scenarios. A study by the Prognos AG for Duales System
Deutschland AG. Cologne. Germany.

Schiffler, H. Duales System Deutschland AG. 2002b. The green dot and its benefit for the
environment. Life cycle Analysis of scenarios for the future. A study by the Oko-
Institut e.V. on commission for Duales System Deutschland AG. Cologne,
Germany.

Schmidt, W. 2003. Life cycle costing as part of design for the environment. Int. J. LCA
82167-174.

Scientific assessment of ozone depletion.World Meterological Organization (WMO).
1998. Washington, DC, U.S.A.

Sellers, V.R., and J .D. Sellers. Franklin Associates, Ltd. 1989. Comparative energy and
environmental impacts for softdrink delivery systems. Prairie Village, KS.

Shapiro. KG. 2001. Incorporating costs in LCA. Int. J. LCA 62121-123.

Shapouri, H., J.A. Duffield, and M. Wang. Office of the Chief Economist, Office of
Energy Policy and News, United States Department of Agriculture (USDA).
2002. The energy balance of corn ethanol: an update. 814 Washington, DC.

Slob, W. 1994. Uncertainty analysis in multiplicative models. Risk analysis 14:571-576.

Sonnemann, G., M. Schuchmacher. and F. Castells. 2003. Uncertainty assessment by a
Monte Carlo simulation in a life cycle inventory of electricity produced by a
waste incinerator. Journal of Cleaner Production 11:279—292.

207

31' __

 

Sonnemann, G., A. Solgaard, K. Saur, H.A. Udo de Haes, K. Christiansen, and A.A.
Jensen. 2001. Life cycle managment: UNEP-Workshop: sharing experiences on
LCM. Int. J. LCA 62325-333.

Sonneveld. K. 2000. The role of Life Cycle Assessment as a decision tool for packaging.
Packaging Technol. and Sci. 13:55-61.

Suh, S. CML, Leiden University. 2001. MIET 2.0: An inventory estimation tool for
missing flows using input-output techniques. Leiden, The Netherlands.

Suh, S., M. Lenzen, G.J. Treloar, H. Hondo, A. Horvath, G. Huppes, O. Jolliet, U. Klann,
W. Krewitt, Y. Moriguchi, J. Munksgaard, and G. Norris. 2004. System boundary
selection in life—cycle inventories using hybrid approaches. Environmental
Science and Technology 38:657-664.

Tan. R., A.B. Culaba, and MR. Purvis. 2002. Application of possibility theory in the life
cycle inventory assessment of biofuels. Int. J. Energy Res. 26:737-745.

Tan, R.B.H., and H.H. Khoo. 2005. Life cycle assessment of EPS and CPB inserts:
design considerations and end of life scenarios. Journal of Environmental
Management 74: 195 -205.

Technical Note cbl-gen-004A: How does Crystal Ball calculate sensitivity?
Decisioneering, Inc. 1998. [Online]. Available from Decisioneering, Inc.
http://www.decisioneeringcom/support/simulation/cblggen 004A.html.

Todd, J.A., and M.A. Curran, (eds) 1999. Streamlined life cycle assessment: a final
report from the SETAC North America streamlined LCA workgroup. Society of
environmental toxicology and chemistry (SETAC) and SETAC Foundation for
Environmental Education.

Toloken, S. 2004. Groups strive to simplify life-cycle rating. Plastics news: 12.

Total cost assessment methodology.Center for Waste Reduction Technologies, American
Institute of Chemical Engineers. 1999. New York. NY.

Tukker. A. 1999. Best available practice regarding impact categories and category
indicators in life cycle impact assessment. Gate to EHS:1-2.

208

Tukker, A. 1998. Uncertainty in Life Cycle Impact Assessment of Toxic Releases. Int. J.
LCA 32246-258.

Udo de Haes, H. 2000. Weighing in life cycle assessment: is there a coherent
perspective? Journal of Industrial Ecology 3:3-7.

Udo de Haes, H.A. Society of Environment Toxicology and Chemistry. 2002. Life-cycle
impact assessment : striving towards best practice. Society of Environment
Toxicology and Chemistry, Pensacola, FL.

Use of life cycle assessment as a policy tool in the field of sustainable packaging waste
management.The European Organization for Packaging adn the Environment
(EUROPEN). 1999. Brussels.

Van Doorsselaer. K.a.L., F. 1999. Estimation of the energy needs in life cycle analysis of
one-way and returnable glass packaging. Packaging Technol. and Sci. 12:235-
239.

Ver—Bruggen, S. Atalink Ltd for Pira International. 2004. Diagnostic packaging. London.
UK.

Vink, E.T.H._. K.R. Rabago, D.A. Glassner, and RR. Gruber. 2003. Applications of life
cycle assessment to NatureWorks (TM) polylactide (PLA) production. Polymer
degradation and stability 80:403-419.

Weber Marin, A.. and M. Tobler. 2003. The purpose of LCA in environmental labels and
concepts of products. Int. J. LCA 82115-1 16.

Wells, J .H.A., N. McCubbin, R. Cavaney, B. Camo, and MB. Hocking. 1991. Paper vs.
Polystyrene: environmental impact. Science 252: 1361-1363.

White. M. 2005. E-mail correspondence: Inquiry on average stetch wrap material per
pallet, In D. Martino, (ed.). Center for Unit Load Design. Virginia Polytechnic
Institute and State University.

Wollny, V., G. Dehoust, U.R. Fritsche, and P. Weinem. 2002. Comparison of plastic
packaging waste

management options feedstock recycling versus energy recovery in Germany. Journal of
Industrial Ecology 5249-63.

209

Workshop report: A technical framework for life-cycle assessment.Society of
Environmental Toxicology and Chemistry (SETAC) and SETAC Foundation for
Environmental Education Inc. 1991. Society of Environmental Toxicology and
Chemistry (SETAC) and SETAC Foundation for Environmental Education Inc.,
Washington, DC.

World Commission on Environment and Development. Oxford University Press. 1987.
Our common future. Oxford University Press, Oxford ; New York.

World largest PET Life Cycle Assessment - One-way PET levels with refillable glass -
Summary report.IFEU in Heidelberg, Germany, for the German Market - PET
container recycling Europe (PETCORE). 2004. Heidelberg, Germany.

210

 

......

||I1111111111111111111111111111111111

 

31293 02845 1221