. 13.3.77.
kxkfinfl . ..
7.77.5.r7nfl77777 _
.. «7... . . . . . 7 ..
I 7 . .7 . 7 x 7 .. 7 0 V‘. I‘l‘7Il .I
-57... 7
. . .7 BMW: 77?”.

   

7 7. 7. 7 7773?.777 7
.. 7 . . . . 7 .7fi7nu...4
7 7 7. t. Inuit! .7 .. AW“!

.7 3 t I. . v. . 7 7 . 7

. . 77 71w. 77n7£€h7x7u¢7777777u777ww77 . .

7 W“. 7.70%.”; .0 Klanuyf If 7‘

. 7 . . m.» 77.7“... .n. .7 . . ”3.7 r7...
7 7 nll‘ '
. . . “chm Erurvw... . .7 72771.7: ,7.. . . . . 7
.. . . . .77....777...» 37 . ,7 , q _ . . . _ 7 .
c {.7 NHL» . 7 7 7 7 . 7 7 7 . 7 . 7 . . 7 7 ..77I7.I 7 7
.77.“. flab?» 7 .. . . . . . . 7 . . 7 . 7. . . ...7... 7 73:77..
7 . . . 7 7 7 .7 7 I'ltprvalotloI-JI;
. . 7 . . 74.77177777177fi777.)
. 7 7 7 v 7 l'171t 7”.
4 I 1 vl‘ '7‘
.n 4.7.77‘1‘ 7'
.0 I'VDtID-VMWDNDMh—IV

fining”... 7
. 7 L a: 7.7.7. 5.7. .

 

 

 

.7

. L77. .7.
77 747.77%..7 ,. - 7.7.7 . 7 7...
. {NJ-773+“ t. 7 7 7 . f. 7 :51 I finswfi .
low E7 07.7 70.7% . . ..7nll. ‘7 L. a 71.77 7.
3775.}. . 7 .2 7 . . lhlwkfiwfi.
77..» ...w...7m_.77. 7. 77.97.77. . “77.7.7.7...
L17 .7 7mm». 77.. 77.77%. . . 77.77....
. In ... 7. 7.7 7 .. 77v...
.7 ”77¢..."W7;¢777.u77.7mr777: 7. . . . .
a: 7%. 7... 7.7 7 ~77...“ LwV Armada. _ 7 . 7 . ...77 7 . .

«£7... Par. .7 777.777.1777? .
.7. fl‘flu. “A7. . 701% 57.770. 7
u. 777.. a... théxuufi
«

,7..

. ....-...r.7.1...n77. 7 7777 7777.7 7

,7.7.7 .7 7&7.“me 4.7%.». . 7 ..

.7. ,.77 7.9.4.7 7... 77.79... ,777 .HE

7\.

 
 

 

. draw. 7

. . 7. 7 7n. ”a IMI: .

7 \I . . 7 7 .

3 .7? . . . .47“ hw7wwmv7....2¥ . .

7 I 17 b7. 7.. 75.7 77. 1.71 . 7 7 7 7

.7337? 77.77.77r...7a.o.77... . . 7 , .7 . . . .872.
' 9.7.37h777577; .7. 7 7 777.7777. 70717.7

. . .u.
«777 . 7.7%. 7
. 7. o v M7 . 7 . . 7 I 77!
3.77M!- . . H7377 77 x
U 777.. I 1 l .\
7 a. . 7.5.:

7.775 7.77%".
777.7% . 7.
.fiN7N7977 73¢! . $727... . 7... .r. J 7.0.... 7 . .
.77. urr...777u... 7.77.77.72.77 7 . . .. . .7....7a...777..7§7. ..7 “7.37.37. . . .
77.7.7.7...» fivfifififi.» 7m. .. fihmufifiw we... . , . . . .. . . . . .
7 7.... 77.77.“. 7.77777; .. . . .. . . 7 . .. 7 .. .7.
,. 7 7 7 Ill.‘ 7 7 n 77. r . v tfl"! I
. .. . 37.133.777.77.“

 
 

 

7. 7.7. 7
7.
.7... «v.1... Mania... 7
7 7 gap ”77.777.77.717. 77. 77.777.77.317. 7.. ag‘utx . . 7|
77.777.57.311..- 7.77.77". 7177.777.A»7.777..777.« . . 7 .. 7 7
77.77.73- 7.7777777“7.7777.77.7.77%77w.7777777777 7.77777 .777 . . 7.727.. 77.. .. 7 .. .. 7...
.‘ lo|7tvlcV7¥P7 Ilulfl't'i7‘77‘7l 7. 7 3.74 7 . . €7.47 . 7 7. sIIt~977I
7 7 777 7.7. .77777777 .7 . 7 7 7.7.--.
7 17.77 7 737 7Hr777ev I 7 .7 $157,777,742”! 77 . . .. [.7 Pfu...oo..7.7777777
V 7 Y 7 . 7. '0‘! I
7 7.777.777 . 7777....
7 ‘ IIA 7A E

 

I

‘4

 

I

1177.9!0 7 .
[477.777.777.737 . 7 . .
7. 77177 7.77 .. 7 . . 7
77777.71...) £17 . . .
v 7777 77.77.79.777 . . .77 77”.”? 777 .
.7 7.7.3.7.! 7 77.77 .7 . .7 777 . . , .
7 7 7 . 79,77 7.7.7.7741. 74.777. , . .7 7
7 71.7747 7 7 7.. .. fififi‘hmfl
777‘! II? . .7\a. 7.77.7
V ‘7' - 7 7. ‘0‘ ‘ 'IJV -
07.77. 717.5 it . x7. hvhhbn7§nv777§ . 7 .
nhulfliuthv . 7 8.7377 .7: 16777“... 7 7. . . . .
7.1 7 . .7.7.m.r7u7~...w777 7 .7 .. 7 . 79|7\unlv|bo7a .X‘I?lrru7
7. .. 79.. 77177 .77 . . . . . 7-777.777“ 1
. 7 . i «1. 2‘1}?! 7 A V I‘LI‘ I :‘I\.
.7 . 7 II?“ 7 7 I 7 ’7!’
..
77%... 775'. 7

 

77‘7757 77
7. .07twn177. \7)\h
.7'7 77' .7777. 77

"$1!
«‘2‘:

I770.
Iva:
on-

J "-z-hgl.
.17
7227?:
- .1353:
1:753?"
‘ $735}!
.7». :7
N "

a
a 2

iv} ‘
' cl'
if.
‘:
{0

7! I MHWWVJV .
7 a; 757539...

 

7477797... 77.77
7 7.777777 7.777 ..
77’17777777777v777777uu
77.77.. .77 7.7.. 70777. 577.777.7777
7'7 .77. 79.173 .77»
.777

7... .

777.

7 .‘11". ‘7 7“
7 777.7377!
7‘7777777

7727.77 7 7

77 .77....7 7‘. . .
. 7777777 7.77.7771 7R.-- 7 .1
.77 7.7 777.7 77 77 777. 7 7.
7 2. ”77777777. 7 7. 77 .77. 77.77777; r... 7.7 it. 77.77.77.177. 777.7 7..»
7.7.7773. 7.. 2.77. 77.777777 7 . 7. 72777977777777... 777.777 77.777737777717777? 77..l77 75.
Ol‘ -7 )‘I‘Vu 7 277‘ 7" n 7 '7 ’V7O‘ICI7’.‘|\AIC7O\V>I7o7
7 777.777.757.777} . . . . . 7 $7.77 7 77771777733731».
'5‘ '- -J 7 7 ‘DIQ'LI|¢ )1 V i 7 VMII.|I DLHUI ‘J'I|17|
777.7. 71. 7.......7777 77.. 7777 .373. 7 7.
u 7' v“\l7 .

 

 

o \vlylo7 . .
.7VLI7VN77709‘ .,. 7 7
777‘ ‘757‘71 7 7 .
7I777l?!‘lvv\l7~.7s 7. 7 . HHI‘7 7 ‘7 77 7 7 17
7777.7. . «.llvd717 7 77.777II7771h («£7
7: 77.777711377471777hfl

.77 71.77.77. '
l....|..l77771t 77777777175777
.7... . . 77777 “77727777777777. .
ha». 777 7.7... 7.777! 777.777 .
I7 7777777 777777... .
hum. . 7....7..77..777~7Av77r7.77....
777 .. . 77777177107L37777777| 7 . “777.777.7077: 7. . ..
7 77 7 7 l07\‘7~IV7777|0|7.¢Ip7\l77777“57 7 7 7
7 I U . A|7I.O‘|l7il"‘

 

   

      
  

7 .7. 7.7.. 71.77!
.774]. 7.777. 7V.7.u.7177..m. 77777. 7 7777 77:7
.7: .7 777. .7 :7 7 . n74.777.77777577777.377377,.!7
. .7377: .7
7. 7 .7_LI li“.."lv" Owfl‘iogiu.‘
. 77777777 777.77.71.77 7777.77.73 .
4 . 7 7 7 7| t i I It: 7II V‘
. , 777.77....777..77.777;7.71..1...r.
. 1.....-t..77..77 7.7.
7 I.... ‘V1 '1-7“l' V-..» . 7 7‘7 I‘ ’0‘ I
|I‘ I 7 . 7 7. It .77 . . .
III .7 7. 777.... . .7. 1777......7 7717...}...32 7777.
I‘ . 1.777 7. . 7
III 7.7 v
I‘ ll 7 . 7| I I
7 .7 7..
lll
7 ‘It

ms

\‘3

                                                                                                                          

31293 01558 4968

This is to certify that the

dissertation entitled

The Ecological Impact of Austrian Pine
(Pinus nigra) on the Sand Dunes of
Lake Michigan: an Introduced Species
Becomes and Invader

 

presented by

Lissa Maria Leege

has been accepted towards fulfilltnent
of the requirements for

Ph.D. degree in Botany & Plant Pathology

Quit Min

Major professor

Date 11 August 1997

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

LIBRARY

Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove We checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

i” 1’ e.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU le An Atflnnetlve Action/Equal Opportunity Inetltuion
W

m1

 

THE ECOLOGICAL IMPACT OF AUSTRIAN PINE
(PINUS NIGRA) ON THE SAND DUNES or LAKE MICHIGAN:
AN INTRODUCED SPECIES BECOMES AN INVADER

By

Lissa Maria Leege

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

DOCTOR OF PHILOSOPHY

Department of Botany and Plant Pathology
1997

ABSTRACT
THE ECOLOGICAL IMPACT OF AUSTRIAN PINE (PINUS NIGRA)
ON THE SAND DUNES OF LAKE MICHIGAN:
AN INTRODUCED SPECIES BECOMES AN INVADER
By

Lissa Maria Leege

The sand dunes of the Great Lakes axe exposed to continuous natural disturbance
and may be vulnerable to invasion by exotic species. In this study, the effects of an
introduction of more than 25,000 Austrian pines (Pinus nigra), planted in 1956-1972 as a
stabilization measure, were investigated on the Lake Michigan sand dunes in Allegan
County, Michigan. The post-planting establishment success of viable trees, their biotic and
abiotic effects on the dune system, and their reproductive capacity, and recruitment success
were examined in four seral stages: foredunes, forest edges, wetpannes (ephemeral dune

ponds), and inland blowouts.

Measures of establishment success (height, dbh, and stem volume) indicate that P.
nigra grows as well or better on the dunes than in its native range. Tree size differs more
within seral Stages than among them, indicating that the coarse-scale differences in
environmental factors among stages, such as soil moistuxe and wind exposure, account for
only some of the variation in growth. The broad ecological tolerance of P. nigra appears to

be a major attribute in its successful establishment in a variety of dune conditions.

Pinus nigra inhibits the growth of native dune vegetation in all four seral stages,
and depresses species richness by up to 80% relative to sites lacking P. nigra in all stages
but foredunes. It provides colonization sites for forest species absent from the adjacent
dune communities, however, and a higher density of woody seedlings and saplings such as
Sassafras albidum occur under pine canopy at the edge of native forest than in comparable

sites lacking P. nigra. Additionally, wetpannes surrounded by P. nigra were drier than the

native Pinus banksiana wetpannes and were characterized by a higher density of shrubs.

Pinus nigra foredune seed production per tree is 10-40 times that of other seral
stages, but survivorship of seedlings is five times higher in blowouts and wetpannes.
Light availability is positively correlated with survivorship in all stages (Spearman Rho =
0.52) and is postulated to regulate recruitment of seedlings after initial germination.
Demographic projection models indicate that population growth rates are highest in
wetpannes (lambda = 0.998 to 1.201), despite the higher understory cover and species
richness of this seral stage. These findings suggest that P. nigra has made the transition

from introduction to invasion on the sand dunes of Lake Michigan.

ACKNOWLEDGEMENTS

I would like to thank my advisor, Dr. Peter Murphy, for his boundless enthusiasm
and unfailing support. Pete has a way of transforming every chore into an exhilarating
challenge, and I hope to emulate his positive spirit as I continue in the academic world. I
could not have asked for a better mentor. I also owe many thanks to my committee, Drs.
H. Rique Campa, Kay Gross, and Frank Telewski, who have all provided helpful insights
and excellent feedback throughout the process.

Funding for this research came from the Michigan DNR Parks and Recreation
Division and from the Michigan DNR Natural Heritage Program - Nongame Wildlife Fund.
I am especially grateful to Ray Fahlsing, of the Parks and Recreation Division, and Kim
Herman of the Natural Heritage Program for their invaluable support throughout the
project. I thank Dave Thibault, park manager, for research permits and access to field
equipment each summer. Thanks also go to the Michigan Civilian Conservation Crew,
who braved the dunes with chainsaws to girdle more than 200 trees. I am also appreciative
of the supplemental funding provided by the Paul Taylor Fund of the MSU Department of
Botany and Plant Pathology and by the Ecology and Evolutionary Biology Program.

I owe much to my field and lab assistants and undergrad researchers, without
whom this project could not have been completed. Daria Maksimowich, Chris Wichterrnan
and Jules Staser are to be commended for their tolerance of ticks and tedium, and I thank
them for being hard workers and great company. Thanks also go to Graham Hunting,
Mark Peterson, and Susie Wells for their painstaking lab and computer work. Special
thanks to Jeremy Emmi, who completed much of the 1994 reproductive output work for
his undergraduate research, and to Dave Kovach, Brad Chapman, Rebecca Cifaldi, and

iv

Torn Park for the further insights their work brought to this project.

This work would simply not have been possible without the help of more than 60
student and faculty volunteers who spent a Saturday or a weekend or many weekends at the
dunes with me, collecting cones, coring trees, or counting Pitcher’s thistle. Special thanks
to lab-mate, Ian Rarnjohn, who sacrificed many a fall Saturday to join me in the field, and
to Raj Lal, another repeat field volunteer and a great source of support.

I thank Anita Davelos, Alice Murphy, Kristi Sherfinski, Miles Silman, Kurt
Stanley, and John Wilterding for their careful reviews of earlier drafts of this dissertation
and for their moral support throughout the process. They, and many others have made
graduate school a pleasant experience, and much of what I know, I have learned from
them.

I owe much to my dear friend, Ethel Burns, for tolerating three summers of sand in
her freezer (and carpet, and washing machine, and...), and for treating me like family. I
also wish to thank my grandmother, Esther Schad who has taught me the value of
education, and my brothers, David and Kurt, who have inspired me with their intelligence
and determination. Finally, much love to my parents, Pat and Dave, who have always

believed in me and have cheered me on all the way to the finish line.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. viii
LIST OF FIGURES ............................................................................... xiii
CHAPTER 1
GENERAL INTRODUCTION ..................................................................... 1
CHAPTERZ

GROWTH AND PRIMARY ESTABLISHMENT OF THE INTRODUCED
PINUS NIGRA ON THE SAND DUNES OF LAKE MICHIGAN

Introduction ................................................................................... 5
Invasion theory ...................................................................... 5
Non-native species invasions on the sand dunes of the Great Lakes ......... 7
Distribution and characteristics of P. nigra ...................................... 8
Forestry practices and cultivation of P. nigra ................................... 9
Research questions ................................................................ 10

Methods ................................................................................ 11
Study site ........................................................................... 11
Sampling sites ..................................................................... 11
Characteristics of individual trees .............................................. 12
Stand characteristics at SDSP ................................................ 14
Tree growth ........................................................................ 15

Results .................................................................................. 17
Pinus nigra survivorship ......................................................... 17
Tree characteristics and stand density at SDSP ............................... 17
Tree growth ........................................................................ 21

Discussion ................................................................................... 23
Overview ........................................................................... 23
Primary establishment of Pinus nigra at SDSP ........................ 24
Pinus nigra establishment in even-aged plantations ........................... 28
Growth analysis: Pinus nigra establishment native habitat vs. SDSP ....... 29
Conclusions ........................................................................ 32

CHAPTER 3

BIOTIC AND ABIOTIC ALTERATIONS OF FOUR SAND DUNE HABITATS BY THE
INTRODUCED PINUS NIGRA

Introduction ................................................................................. 46
Methods ............................................................................... 50
Study site ........................................................................... 50
Biotic effects of P. nigra .......................................................... 50
Disturbance regime: sand movement ............................................ 52
Abiotic effects of P. nigra ........................................................ 54
Analysis ..................................................................................... 55

vi

Biotic effects ....................................................................... 55

Disturbance regime: sand movement ............................................ 58
Abiotic effects ...................................................................... 58
Results ....................................................................................... 59
Biotic effects of P. nigra ......................................................... 59
Disturbance regime: sand movement ............................................ 76
Abiotic effects of P. nigra ........................................................ 76
Discussion ................................................................................. 85
Geomorphological effects: Disturbance regime ................................ 85
Population and community level effects ........................................ 86
Assessment of Austrian pine effects ............................................ 91
Conclusions .................................................................... 93

CHAPTER 4

REPRODUCTION AND RECRUITMENT OF THE INTRODUCED PINUS NIGRA
ACROSS DUNE SUCCESSION: DEMOGRAPHIC AND COMMUNITY
CONSIDERATIONS

Introduction ................................................................................ l 19
. Invasion ecology ................................................................. 120
Population projection models ................................................... 121
Research objectives .............................................................. 122
Study species ..................................................................... 123
Methods ................................................................................... 124
Study site ......................................................................... 124
Reproductive output ............................................................. 124
Demography of naturally regenerating Pinus nigra at SDSP ............... 126
Microsite variables .............................................................. 126
Analysis ................................................................................. 128
Reproductive output ............................................................. 128
Demography... 1.28
Microsite variables ............................................................... 132
Results ................................................................................ 133
Reproductive output: seed production and laboratory gerrninability ....... 133
Demography ................................................................... 134
Microsite variability ............................................................. 146
Discussion ................................................................................. 150
Reproductive effort .............................................................. 150
Demography ................................................................ 152
Microsite variables .............................................................. 156
Conclusion ........................................................................ 158
CHAPTER 5
GENERAL CONCLUSIONS .................................................................... 169
LIST OF REFERENCES ......................................................................... 175

LIST OF TABLES

Table 2-1. Average 35-year survivorship of P. nigra from three different
seed sources in a plantation at Kellogg Forest, Augusta MI and at SDSP,
Allegan Co., MI ..................................................................................... 17

Table 2-2. AN OVA for DBH, height, and stem volume, where sites were

nested within seral stage (foredune, forest edge, wetpanne, inland blowout);

and Kruskal Wallis test, where data were pooled within seral stage.

Measurements were taken in the growing seasons of 1994 and 1995, SDSP,

Allegan Co., MI ..................................................................................... 18

 

Table 2-3. Mean size and stand characteristics for P. nigra in four seral

stages at SDSP, Allegan Co., MI (data from 1994 and 1995), and for three

seed lots in a 35-year old P. nigra plantation at Kellogg Experimental Forest,

Augusta, MI (1996 data) ........................................................................... 19

Table 2-4. Spearman’s rank conelations between site slope and several
measures of tree size in P. nigra at SDSP, Allegan Co., MI (n=20, size data
taken in 1994 and 1995) ............................................................................ 20

Table 2-5. Linear regression of height against tree age (total number of
whorls) in four seral stages at SDSP. Data taken in 1996, Allegan Co., MI ................ 22

Table 2-6. Spearman’s rank correlation between June precipitation (current
and previous) and tree-ring width in four seral stages. Tree cores collected
in Fall 1994, Allegan Co., MI .................................................................... 22

Table 2-7. Comparison of height, dbh, stem volume, and stand density for

even-aged P. nigra plantations in Secrest Arboretum, Wooster, OH (Burns

and Honkala 1990), and southwest France (V annier 1984), and for uneven

-aged native stands in Turkey (Ata 1989). Yield classes“ are given only for

France ................................................................................................. 30

Table 2-8. A comparison of measures of P. nigra establishment success
among four seral stages at SDSP. Postulated growth regulators are also listed ............. 33

Table 3-1. Vegetation types present in each portion of the transects in sites

lacking Austrian pine. P indicates the portion of the transect analogous to

that characterized by the presence of Austrian pine canopy in Austrian pine

sites. N is the adjacent vegetation, similar in sites with and without Austrian

pine .................................................................................................... 52

viii

Table 3-2a. Average diversity indices across transects in foredune and forest

edge seral stages. 8: species richness/2m2, D= Simpson’s index, and

H’= Shannon-Weaver index. In foredunes, wetpannes and blowouts,

intervals commence under pine or cottonwood canopy and progress out

towards the bordering vegetation so that interval one is located at the edge of
pine/cottonwood stand. In forest edge sites, intervals commence under pine

canopy or in blowout vegetation in sites lacking pines and progress out

towards the bordering deciduous vegetation. Interval one is the final meter

before the forest vegetation ......................................................................... 62

Table 3-2b. Average diversity indices across transects in wetpanne and

inland blowout seral stages. 8: species richness/2m2, D= Simpson’s index,

and H’= Shannon-Weaver index. Interval one commences just beyond pine

canopy and the transect progress out ten meters further into the bordering

vegetation ............................................................................................. 63

Table 3-3. Top five important species in foredune sites along 20-m transect.

Each number represents a two-meter interval, starting under cottonwood or

pine canopy and ending 10m from the edge of the canopy in the bordering

native vegetation ..................................................................................... 64

Table 3-4. Top five important species in forest edge sites along 20-rn transect.
Each number represents a two-meter interval, starting under pine canopy or in
blowout vegetation and ending 10m into native deciduous canopy ........................... 65

Table 3-5. Top five important species in wetpanne sites along 20-m transect.
Each number represents a 2-m interval, starting under pine canopy and ending
10m into native wetpanne vegetation ............................................................. 68

Table 3-6. Top five important species in inland blowout sites along 20-rn
transect. Each number represents a 2-m interval, starting under pine canopy
and ending 10m into blowout vegetation ......................................................... 70

Table 37. Mean stem density (# stems/m2) 18E (n=10) of all woody species
21cm DBH across four Sm-intervals in forest edge and wetpanne seral stages.
Data from 1994-1995 ............................................................................... 73

Table 3-8. Size structure of two tree Species in S-m intervals across transects
in forest edge sites with and without Austrian pine. Values represent number

of individuals in each DBH class. Data from 1994 and1996 .................................. 74

Table 3-9. Size structure of Pinus banksiana across transects in Austrian

and Jack pine wetpannes. Values represent number of individuals in each
DBH class. Data from 1995 ....................................................................... 75

Table 3-10. Sand movement around pine stands in three seral stages during
the period of 1994-1995 ............................................................................ 77

Table 3-11. Solum thickness in four seral stages under canopy and in

bordering vegetation. Significance of main effects and interactions in two-

way AN OVA for each within-seral stage comparison is shown. Data from

1 995 - 1996 ............................................................................................ 7 9

Table 3-12. Minimum, median, and maximum pH in four seral stages.
Data from 1995 ...................................................................................... 80

Table 3-13. Repeated measure MANOVA for foredune soil moisture across
the 1995 growing season ......................................................................... 82

Table 3-14. Repeated measure MAN OVA for wetpanne soil moisture across
the 1995 growing season ........................................................................... 83

Table 3-15. Repeated measure MANOVA for soil moisture in inland blowout
sites across the 1995 growing season ............................................................. 84

Table 3-16. Correlations between light intensity and understory cover, species
richness/2m2, and Simpson’s diversity index in four seral stages. Values are
Spearman’s rho. Data from 1994-1996 .......................................................... 84

Table 3-17. Actual outcome of P. nigra introduction and projected future trajectory
of native dune populations in four seral stages at SDSP ....................................... 92

Table 4-1. Laboratory germinability for Austrian pine seed collected from

four seral stages in 1994 and 1995. Germination trials performed with seeds

from individual cones: n is the number of cones for which trials were run in

each seral stage. Germinability per cone was significantly different across

years, but also among seral stages in 1995 ..................................................... 134

Table 4—2. AN OVA for 1995 arcsine transformed germination data.
Germination values were determined for individual cones collected from trees
in foredune, wetpanne, forest edge, and inland blowout seral stages ....................... 134

Table 4-3. Repeated measure MANOVA for density of all P. nigra seedlings
in three sites within each of three seral stages and at each of two light levels
across the growing seasons of 1994-1996 ...................................................... 135

Table 4-4. Repeated measure MANOVA for density of P. nigra seedlings
greater than two years of age within each of three seral stages and at each of
two light levels across the growing seasons of 1994-1996 .................................. 136

Table 4—5. ANOVA Table for log height of seedlings in foredune, wetpanne
and inland blowout seral stages and in canopy and gap plots in established
quadrats in 1996. 181 seedlings were measured ............................................... 136

Table 4-6. Seedling survivorship across two growing seasons in three seral

stages and in canopy and gap plots. N represents the number of quadrats

which contained seedlings for which survivorship was determined across the

time period. Contrasts between arcsine transformed seedling survivorship in

canopy and gap pairs within each seral stage were performed within each time

period to determine significant differences. Means marked with an asterisk

were significantly different among light treatments within seral stage (p<0.05) ........... 137

Table 4-7. AN OVA Table for 1994-1995 arcsine transformed survivorship

of seedlings in foredune, wetpanne and inland blowout seral stages and in

canopy and plots in established quadrats. Survivorship was calculated for

188 quadrats ....................................................................................... 138

Table 4-8. AN OVA Table for 1995-1996 arcsine transformed survivorship

of seedlings in foredune, wetpanne and inland blowout seral stages and in

canopy and gap plots in established quadrats. Survivorship was calculated

for 123 quadrats ................................................................................... 138

Table 4-9. Population size of regenerating P. nigra in three years in three
seral stages in gap and canopy plots. No adults were included ............................. 139

Table 4-10. 1994-1995 transition matrices for regenerating Pinus nigra
populations in three seral stages in canopy and gap plots .................................... 141

Table 4-11. 1995-1996 transition matrices for regenerating Pinus nigra
populations in three seral stages in canopy and gap plots ..................................... 142

Table 4-12. P. nigra population growth rates in gap plots in three seral

stages as calculated from projection matrices from 1994-1995 and 1995-1996
transitions. Inland blowout population growth rates were inestimable for the
second time period because none of the seeds transitioned to juveniles during

this time ............................................................................................. 143
Table 4-13. Projected population size in three seral stages calculated from
transition matrices for 1994-1995 and 1995-1996, where P,,=0.99 ........................ 143

Table 4-14. Elasticities” for 1994-1995 and 1995-1996 transition matrices
for regenerating pinus nigra populations in three seral stages in gap plots,
where Paa=0.99 .................................................................................... 144

Table 4-15. Elasticities" for 1994-1995 and 1995-1996 transition matrices
for regenerating Pinus nigra populations in three seral stages in gap plots,
where Paa=0.95 .................................................................................... 145

Table 4-16. Average % needle and deciduous litter cover, and % bare soil in

canopy and gap plots in three seral stages. N refers to the number of m2

quadrats sampled. Significant ANOVA results for Stage, Site(Stage), Light

and Light*Stage interactions are listed as w=stage effect, x=site(stage) effect,

y=light effect, and z=light*stage effect .......................................................... 146

Table 4-17. Average soil moisture and soil pH at two depths (0-5cm and 10-

15cm) in canopy and gap plots in foredune and wetpanne seral Stages. N

refers to the number of m2 quadrats sampled. Significant ANOVA results

for Stage, Site(Stage), Light and Light*Stage interactions are listed as w:

stage effect, x=site(stage) effect, y=light effect, and z=light*stage effect .................. 147

Table 4-18. Average light intensity and canopy openness in canopy and gap

plots in three seral stages. N refers to the number of m2 quadrats sampled.

Significant AN OVA results for Stage, Site(Stage), Light and Light“Stage

interactions are listed as w=stage effect, x=site(stage) effect, y=light effect,

and z=light*stage effect ........................................................................... 148

 

Table 4-19. Average species richness/m2 and understory cover in canopy

and gap plots in three seral stages. N refers to the number of m2 quadrats

sampled. Significant ANOVA results for Stage, Site(Stage), Light and Light

*Stage interactions are listed as w=stage effect, x=site(stage) effect, y=light

effect, and z=light*stage effect ................................................................... 149

Table 4—20. N onpararnetric Spearman rank correlations between Austrian

pine seedling survivorship and abiotic and biotic variables. Soil moisture

and pH values were taken for a subsample of quadrats in only foredune and

inland blowout sites, therefore sample sizes are smaller for these correlations.

It indicates the number of quadrats used for the correlation analysis ........................ 149

LIST OF FIGURES

Figure 1-1. Cross section through a typical dune system. Inland blowouts
(not shown) are formed when vegetation is removed as a response to natural

or human-caused disturbance ..................................................................

Figure 2-1. Native distribution of P. nigra in Europe and Asia (adapted from
Vidakovic 1991; based on a map from Svoboda, 1953). Shaded areas indicate

occurrence of P. nigra ..........................................................................

Figure 2-2. Location of study Site (Saugatuck Dunes State Park) on eastern

shore of Lake Michigan, Allegan, County, MI (adapted from Cowles 1899) ..........

Figure 2-3. Aerial photograph of the study site at Saugatuck Dunes Natural
Area, and on adjoining private property, in Allegan County, Michigan (MI
DNR 1988). Pinus nigra plantings are evident as dark lines in blowout areas
and as scattered patches in foredunes. Sampling sites were located throughout

four seral stages in areas with and without Austrian pine. Boundary of study

area is demarcated with bold lines. One cm a 120m...

Figure 2-4. Mean diameter at breast height 18E of planted P. nigra in five

sites in each of four seral stages at SDSP, Allegan Co., MI (n=5).* .................

Figure 2-5. Mean height iSE of planted P. nigra in five sites in each of

four seral stages at SDSP, Allegan Co., MI (n=5).* ......................................

Figure 2-6. Mean stem volume 18E of planted P. nigra in five sites in each

of four seral stages at SDSP, Allegan Co., MI (n=5).* ...................................

Figure 2-7. Scattergram of relationship between dbh and height in Austrian
pines planted in four seral stages, where x=foredune, y=forest edge,
z=wetpanne, and o=inland blowout. Measurements taken June 1994 and

1995 at SDSP, Allegan Co., MI ..............................................................

Figure 2-8. Mean stand density 18E of P. nigra in four seral stages at SDSP,

...... 3

....34

..... 35

...36

..... 37

..... 38

..... 39

..... 40

Allegan Co., MI (n=5).* ........................................................................... 41

Figure 2-9. Mean tree height iSE across time in four seral stages at SDSP,

Allegan Co., MI.* ...............................................................................

Figure 2-10. Mean height growth rate iSE in P.nigra in four seral stages in
five-year increments. Number of whorls was used as an estimate of tree age. NO
SE. is shown for the last value of inland blowout sites because it is based

on a sample of one ..............................................................................

....42

....43

Figure 2-11. Mean radial growth iSE in Pinus nigra in four seral stages.

Mean annual increment calculated for five-year intervals. Data taken from

10 cores from each of three sites in forest edge, wetpanne, and inland

blowout seral stages. Only three cores were taken from each of three sites

in foredunes because of the limited number of trees available in these sites.

Number of tree rings was used as an estimate of tree age ...................................... 44

Figure 2-12. Diagrammatic representation of patterns of BAI (basal area
increment) of P. nigra across time at SDSP, Allegan Co., MI ................................ 45

Figure 3-1. Diagrammatic representation of 20-m belt transect at the interface
between pine and native dune vegetation. All measurements were taken at
l-m intervals. Multiple transects were located in each seral stage ............................ 51

Figure 3-2. Measurement of sand movement patterns around P. nigra and

Popular: deltoides stands. Sand movement in both windward and leeward

directions was determined. Each x represents a stake, used to measure sand

movement ............................................................................................ 53

Figure 3-3. Understory cover iSE in a) foredune sites with Austrian pine or
Cottonwood and b) forest edge sites with and without Austrian pine. Data
from 1994 and 1996 ................................................................................ 95

Figure 3-4. Understory cover iSE in a) wetpanne sites with Jack or
Austrian pine (1995 data) and in b) inland blowout sites with Austrian
pine (1994 data) ..................................................................................... 96

Figure 35. Mean species richness/2m2 iSE-across 20-m transects in

foredune sites with cottonwoods or Austrian pines. Data from 1994
and 1996 .............................................................................................. 97

Figure 3-6. Mean species richness/2m2 18E across 20-m transects in forest-
edge sites with and without Austrian pine. Data from 1994 and 1996 ....................... 98

Figure 3-7. Mean species richness/2m2 18E across 20-m transects in
wetpanne sites with Jack or Austrian pines Data from 1995 ................................... 99

Figure 3-8. Mean species richness/2m2 18E across 20-rn transects in inland-
blowout sites planted with Austrian pine. Data from 1994 ................................... 100

Figure 3-9. Life form composition of understory cover in a) cottonwood
and b) Austrian pine sites in foredunes. Data from 1994 and 1996 ......................... 101

Figure 3-10. Life form composition of understory cover in forest-edge sites
lacking Austrian pine. Data from 1994 and 1996 ............................................. 102

Figure 3-11. Life form composition of understory cover forest-edge sites
with Austrian pine. Data from 1994 and 1996 ................................................. 103

Figure 3-12. Life form composition of understory cover in Jack pine
wetpanne sites. Data from 1995 ................................................................. 104

xiv

Figure 3-13. Life form composition of understory cover in Austrian pine
wetpanne sites. Data from 1995 ................................................................. 105

Figure 3-14. Life form composition of understory cover in Austrian pine
inland-blowout sites. Data from 1994 ........................................................ 106

Figure 3-15. Similarity in community composition and cover across

a) Cottonwood and b) Austrian pine foredune transects. Solid lines indicate

similarity of the community to that of interval one, in the core of the

Cottonwood/Austrian pine community, and dotted lines indicate similarity

of the community to that of interval 10, in the core of foredune vegetation.

Data from 1994 and 1996 ......................................................................... 107

Figure 3-16. Similarity in community composition and cover across

a) forest edge sites lacking pine and b) Austrian pine forest edge transects.

Solid lines indicate similarity of the community to that of interval one, in the

core of the Blowout/Austrian pine community, and dotted lines indicate

similarity of the community to that of interval 10, in the core of deciduous

forest vegetation. Data from 1994 and 1996 ................................................... 108

Figure 3-17. Similarity in community composition and cover across

a) Austrian pine and b) Jack pine wetpanne transects. Solid lines indicate

similarity of the community to that of interval one, in the core of the Austrian/

Jack pine community, and dotted lines indicate similarity of the community to

that of interval 10, in the core of wetpanne vegetation. Data from 1995 ................... 109

Figure 3-18. Similarity in community composition and cover across Austrian

pine inland blowout transects. Solid lines indicate similarity of the community

to that of interval one, in the core of the Austrian pine community, and dotted

lines indicate Similarity of the community to that of interval 10, in the core of

blowout vegetation. Data from 1994 ............................................................ 1 10

Figure 3-19. Density of trees (21cm dbh) across Austrian pine forest edge

transects. PINI = Pinus nigra, COST = Camus stalamfera, PRSE = Prunus

seratina, PRVI = Prunus virginiana, SAAL = Sassafras albidum, TIAM = Tilia
americana, ACSP = Acer sp., FAGR = F agus grandifalia, FRAM = F raxinus
americana, PIBA = Pinus banksiana. Data from 1994 and 1996 ........................... 11 1

Figure 3-20. Density of trees ( 2 1cm dbh) along transects in forest edges

lacking Austrian pine. Qusp = Quercus sp., Cost = Camus stalanifera, Saal =

Sassafras albidum, Pode =Papulus deltar'des, Tiam = Tilia americana, Prvi =

Prunus virginiana, Acsp= Acer sp., Frarn = F raxinus americana, Prse =

Prunus seratina. Data from 1994 and 1996 .................................................... 1 12

Figure 3-21. Density of woody species ( z 1cm dbh) along transects in Jack
pine wetpannes. Cost = Camus stalamfera, Piba = Pinus banksiana, Pode =
Papulus deltaides, Sasp = Salix sp., Besp = Betula sp. Data from 1995 .................. 113

Figure 3-22. Density of woody species (2 1cm dbh) along transects in

Austrian pine wetpannes. Cost = Camus stalam'fera, Piba = Pinus
banksiana, Pini =Pinus nigra, Sasp = Salix sp. Data from 1995 ............................ 114

XV

Figure 3-23. Light intensity iSE across a) foredune cottonwood and
Austrian pine transects and in b) forest edge sites with and without Austrian
pine. Data from 1994 ............................................................................. l 15

Figure 3-24. Light intensity 18E across in a) Jack and Austrian pine
wetpanne sites (1995) and in b) inland blowout sites with Austrian pine (1994) .......... 116

Figure 3-25. Mean monthly soil moisture in a) foredunes and b) wetpannes
across the 1995 growing season. Note difference in scales ................................. 117

Figure 3-26. Monthly soil moisture in inland blowout sites across the 1995
growing season .................................................................................... 1 18

Figure 4-1. Life cycle for Pinus nigra where terms indicating the stages of

the life cycle are defined as follows: 3 = seed, g = gerrninant (first season

seedling), jl = 0-10crn seedling, j2 = 10-25 cm seedling, j3 = >25 cm

seedling, a = reproductive tree ................................................................... 160

Figure 4-2. Matrix model for P. nigra. Life cycle terms are defined in

Figure 4—1. P refers to the probability of transition from the life cycle stage

at time t to the life cycle stage at time t +1. F is a fertility coefficient that

describes the adult contribution of offspring to time t + 1 during time t ..................... 160

Figure 4-3. Mean 1994 and 1995 seed production per tree iSE in Austrian pine

stands in four seral stages. Means were calculated from cone counts from

up to five reproductive trees, and seed counts in up to five cones per

reproductive tree in five sites per seral stage, 1995. Data taken from 10 trees

in each of three sites per seral stage in 1994. Bars demarcated with different

letters denote statistically significant differences in seed production among

stages and across years (p=.05, Tukey all pairs comparison HSD) ........................ 161

Figure 4-4. Mean 1994 and 1995 seed production per unit area iSE in Austrian

pine stands in four seral stages. Means were calculated from cone counts

from up to five reproductive trees, and seed counts in up to five cones per

reproductive tree in five sites per seral stage, 1995. Data taken from ten trees

in each of three Sites per seral stage in 1994. Significance testing was done

only for 1995 data Bars demarcated with different letters denote statistically

significant differences in 1995 seed production among stages (p=.05, Tukey

all pairs comparison HSD) ...................................................................... 162

Figure 4-5. Seedling density iSE in three seral stages in canopy and gap plots.
Data are averages of densities of seedlings in three sites in each light level and
in each seral stage in 1994-1996 ................................................................. 163

Figure 46. a) Seedling density across three years in three seral stages:

foredune (D), wetpanne (M) and inland blowout (0). Least square mean

seedling densities are plotted against year

b) Seedling density across three years in two light conditions: canopy (C)

and gap (G). Least square mean seedling densities are plotted against year ............... 164

xvi

Figure 4-7 a) Density of >two-year—old seedlings across three years in three

seral stages: foredune (D), wetpanne (M) and inland blowout (0). Least

square mean seedling densities are plotted against year.

b) Density of >two-year-old seedlings across three years in two light

conditions: canopy (C) and gap (G). Least square mean seedling densities

are plotted against year ............................................................................ 165

Figure 4-8. Mean height iSE of juvenile Austrian pine seedlings (21y) in
foredune, wetpanne and inland blowout seral stages in canopy and gap plots.
Seedlings measured in three sites in each seral stage in 1996 ................................ 166

Figure 4-9. Seedling survivorship 1SE in three seral stages in canopy and
'gap plots over two growing seasons: 1994-1995 and 1995-1996 ........................... 167

Figure 4— 10. a)-c) Size structure in regenerating Austrian pine populations

in three seral stages in gap and canopy plots. Data represent prOportions of

all individuals in four life history stages in each of three growing seasons

(1994-1996). Life history stages are as follows: germinant (g), juvenile 1-3

(il-j3) and where the germinant stage (g) was represented only by newly

germinated individuals in their first growing season, and the juvenile stages

were determined by height ( j 1<10cm, 10cm5j2<25cm, 25¢m<j3). None

of the juveniles were reproducing ................................................................ 168

xvii

CHAPTER 1
GENERAL INTRODUCTION

Biological invasions are among the most pressing ecological issues of our day. The
rate of non-native species introductions has dramatically accelerated with the improvement
of human transportation systems, and the biota of the world is rapidly being homogenized
(Lodge 1992). Sometimes termed “biological pollution,” invaders present a more
persistent and pervasive threat than other types of pollution, because they are alive and
reproducing and do not disappear when the inputs cease (Crank and Fuller 1995).

Defined from a biogeographical perspective (Rejmanek 1995), biological invaders
are species which occur outside their native range and are capable of forming self-
sustaining populations and persisting in a novel, receptor community. It is well known that
not all non-native species become problematic invaders; in fact, only a small subset of
introduced species goes on to reproduce and persist, transitioning from introduction to
invasion (Loope 1989, Williamson 1996). Williamson and Fitter (1996) quantified this
pattern of invasion success in describing the “tens rule.” Of all introduced species, only
about 10% escape and are found in the wild; of escaped species, only approximately one in
ten form self-sustaining populations and persist or become naturalized, and only about 10%
of naturalized species go on to become economically damaging “pests.” Though this rule
represents a rough estimate, it holds true for both angiosperms and Pinaceae in Britain, and

for insects, fishes, mollusks and plant pathogens in the US. (Williamson 1996).

Of plant species that do become invaders, most have negligible effects upon the
systems into which they have been introduced (Loope 1992, Williamson 1996). Others,
however, can dramatically alter their receptor communities by depressing growth rates or

causing local extirpation of native plant papulations (Walker and Vitousek 1991).

2

Biological invasions may significantly alter species composition (Woods 1993), and may
even have consequences at the ecosystem level (V itousek 1990). Tarnarisk (Tamarix
ramasissima) lowers water table in river systems of the American West (Brook 1994); the
fire tree (Myricafaya) fixes nitrogen in volcanic soils of Hawaii, altering the trajectory of
succession in these nitrogen poor communities (V itousek and Walker 1989); and ice bush
(Mesembryanthenum crystallinum) concentrates salt which inhibits germination of native
species in coastal California (Williamson 1996). The impact of an invader is not simply a
function of its attributes, however, but is mediated by the characteristics of the receptor

community.

Plant communities tend to differ in their susceptibility to invasion (Fox and Fox
1986, Williamson 1996). Ecosystems that experience some sort of regular disturbance
(e.g. grazing or fire) seem to be particularly vulnerable to exotic species invasions (Ewel
1986, Fox and Fox 1986, Orians 1986, Hobbs 1989, Hobbs and Huenneke 1992). The
sand dunes of the Great Lakes are an ideal system for the study of invasions because they
are continually disturbed by the natural process of sand movement which may open up
space for an invader gain a foothold. In addition, sand dune systems comprise an array of
habitats within a small space, and simultaneous invasions in more than one dune
community type can be compared to gain insight into the differential invasibility of

communities.

Great Lakes sand dune systems are partitioned into distinct community types which
are termed “dune habitats” or “seral Stages” (Figure 1-1, Olson 1958). Foredunes are
defined as the youngest dunes just behind the beach, usually vegetated by grasses, forbs
and small shrubs, and regulated almost entirely by geomorphological processes (Peterson
and Dersch 1981). Wetpannes, or interdunal swales (ephemeral dune ponds), are low-
lying slack areas located behind the foredunes, which are inundated with standing water

3

during a portion of the growing season. They include wetland vegetation, and represent an
unusual assemblage of species in this ecosystem otherwise characterized by water deficit
(Ranwell 1972). Inland blowouts are areas of Shifting sand which form as a result of
natural or human-related disturbance and are located inland from foredunes and wetpannes.
A diverse community of bunch grasses and forbs provide sparse vegetation cover in this
seral stage (Peterson and Dersch 1981). Forest edges, defined as areas of open dunes
adjacent to mature oak-beech-maple forest vegetation, normally occur inland from the
wetpannes and along the edges of large blowouts. They are vegetated by the blowout

grasses as well as by small trees and Shrubs.

 

f oredune wetpanne
nearshore dune forest

Figure 1-1. Cross section through a typical dune system. Inland blowouts (not shown) are formed when
vegetation is removed as a response to natural or human-caused disturbance.

A great deal of research has been directed towards aquatic invaders of the Great
Lakes; for example, the dynamics and effects of zebra mussels (Dreissena sp.), alewives
(Alasa pseudaharengus), and lamprey (Petramyzan marinas) are well known (Mills et al.
1993), but very little work has focussed on the invasion threats to the terrestrial component
of the Great Lakes Basin. The sand dunes of the Great Lakes represent the most extensive

4

freshwater dune system in the world (EPA 1995), and are under tremendous pressure from
development and destructive recreational interests. These pressures translate into an even
more critical need to determine other potentially threatening impacts to the dune system.
This dissertation begins to fill the gap in our understanding of the impact of biological

invasions on the dune system of the Great Lakes.

This study focusses on the impact of Pinus nigra Arnold (Austrian pine), a non-
native pine introduced into a dune system of Lake Michigan, at what is now Saugatuck
Dunes State Park, Allegan County, MI. Between 1956 and 1972, more than 25 ,000 P.
nigra individuals were planted into the dunes as a sand stabilization measure. Pinus nigra
is a long-lived pine which reaches reproductive maturity within 15-20 years. Because the
trees planted in the dunes had just begun to reproduce within the previous 15 years, this
investigation presented a unique opportunity to focus on the earliest stages of invasion
while they were occurring. Chapter 2 documents the primary establishment phase (i.e.
post-planting survivorship and growth) of the pine on the dunes and investigates the
differential success of P. nigra among four seral stages, as evidenced by growth in height,
dbh, and stem volume. Chapter 3 examines the effects of the pine on community
composition and structure in four seral stages and relates these biotic effects to alterations in
abiotic characteristics brought about by the invader. The effects of P. nigra stands on
community structure are compared with those of native cottonwood and jack pine stands.
To determine P. nigra’s potential to persist in the dune system, Chapter 4 examines
reproduction and recruitment of P. nigra in foredune, wetpanne, and inland blowout seral
stages. A demographic projection matrix model is developed to predict population growth
rates, and both invader and community attributes are examined to explain invasion success.
Overall conclusions of this study and suggestions for future research are summarized in
Chapter 5. The terms Austrian pine and P. nigra will be used interchangeably throughout

this dissertation.

CHAPTER 2
GROWTH AND PRIMARY ESTABLISHMENT OF THE INTRODUCED
PINUS NIGRA ON THE SAND DUNES OF LAKE MICHIGAN

INTRODUCTION

Invasion theory

Among the most serious threats to natural ecosystem conservation is the
introduction of exogenous species capable of forming self-sustaining populations and
dispersing among sites. Species introductions can affect all levels of ecological
organization, from the dynamics of native populations (Walker and Vitousek 1991,
Equihua and Usher 1993, Paton 1993) and community interactions (W itkowski 1991 ,
Tyser and Worley 1992, Woods 1993, Pysek and Pysek 1995) to the alteration of
ecosystem function (Singer et a1. 1984, Ramakrishnan and Vitousek 1989, Vitousek 1990,
Musil and Midgley 1990, D’Antonio and Vitousek 1992). Though attributes of successful
invaders and characteristics of invasible communities have been described repeatedly (e. g.
Bazzaz 1986, Orians 1986, Noble 1989), only recently has there been some success in
predicting the potential for invasions by exotic species (Rejmanek 1995, Pysek and
Srnilauer 1995, Rejmanek and Richardson 1996). The present study seeks to determine the
invasive potential of Pinus nigra (Austrian pine) on the sand dunes of Lake Michigan,

where it was introduced for dune stabilization purposes.

Vermeij (1996) separates the invasion process into three phases: arrival,
establishment, and integration. The anival phase is primarily concerned with dispersal and
transport routes, while establishment involves the persistence of the imnrigrant population

through local reproduction and recruitment. Integration is the final phase in which the

6

exotic species, once established, influences the dynamics of the recipient community and in
turn is influenced by the community of which it has become a part. This outline of the
invasion process may be appropriate over long time-scales or for species with rapid life
cycles, but it de-emphasizes the critical survival and growth phase that must occur after
arrival and prior to reproduction if the invader is to be successful in its new environment. I
suggest a five-part framework more appropriate for the study of long-lived woody species
in which the establishment phase is split into three separate phases: primary establishment,
reproduction, and recruitment. Primary establishment is defined here as the period of
survival and growth that must occur following the introduction and is treated separately
from regeneration. Primary establishment is followed by a reproduction phase, and finally
a recruitment (or secondary establishment) phase, during which the population must

become self-sustaining prior to integration.

Knowledge of potential limitations within the arrival, primary establishment,
reproduction, and recruitment phases of the invasion process will aid in making predictions
regarding the potential success of the invader. The arrival phase is predominately
controlled by human activities (Mack 1990), but primary establishment is regulated by the
interaction between the invader and its environment; therefore, both genetic and
environmental factors may limit this phase. The reproduction phase may be stimulated or
inhibited by the environment, but is primarily under genetic control. Indeed, from an
extensive list of pine species attributes, Rejmanek and Richardson (1996) found intrinsic
reproductive characteristics to be of highest value in predicting their potential for invasion
success. Recruitment represents the establishment phase of successive generations, and
like the primary establishment phase, is regulated by interactions between invader and

environment.

While some attention has been given to anival (Mack 1990, Williamson 1996),

7

reproduction (Rejmanek 1995, Rejmanek and Richardson 1996), and recruitment
(D’Antonio 1993) phases of the invasion process, few ecologists have considered the
primary establishment phase in detail. This is a critical step in the invasion process which
may be of particular interest to managers assessing the invasion potential of long-lived

woody species and is the focus of this paper.

Methods for assessing the primary establishment success of a potential invader in
its novel habitat must include an analysis of survivorship as well as some measure of
growth or productivity. A significant difference in survivorship or in Size characteristics
such as height, dbh (diameter at breast height), or stem volume among trees of similar age
introduced into different habitats would suggest that extrinsic factors limit successful
primary establishment. Information about species survivorship and growth in the home
range might provide a baseline comparison with primary establishment in the novel habitat
(Sukopp and Starfinger 1995). It is also useful to directly compare environmental
characteristics such as latitude and climate in native and exogenous ranges in order to
assess the similarity of environmental factors which may regulate the establishment of an

invader.

Nan-native species invasions an the sand dunes of the Great Lakes

The sand dunes of the Great Lakes comprise a dynamic, yet fragile ecological
system maintained by continuous endogenous disturbance in the form of sand movement.
As a result of their unstable substrates, available space (i.e. uncolonized ground), and small
population sizes, dune systems may be particularly vulnerable to invasion by exotic species
and may be severely impacted by new species which threaten to alter the natural disturbance
regime. Despite the potential vulnerability of dunes to invasion, however, exotic tree

species have been introduced repeatedly as potential dune stabilizers in the Great Lakes

dune system (Lehotsky 1972).

For example, in one Lake Michigan sand dune system, private land owners planted
more than 30,000 conifer seedlings, the majority of which were P. nigra, in an effort to
Stabilize the shifting sand and maintain property values. During a 16-year period beginning
in 1956, the pines were introduced into four distinct seral stages: foredunes, forest edges,
wetpannes, and inland blowouts. The present study concerns this group of over 25,000 P.
nigra individuals in the dunes at Saugatuck Dunes State Park (SDSP), Michigan. Stands of
P. nigra have thrived in all of these seral stages since planting 25-40 years prior to the
study, and the species is becoming an increasingly prominent component of the dune

vegetation at SDSP.

Distribution and characteristics afPinus nigra

Pinus nigra Arnold is native to Southern Europe, Northern Africa, and Asia Minor
and is distributed from 5° W to 40° E longitude and from 35° N to 48° N latitude (Figure

2-1, Burns and Honkala 1990). Its distribution is highly discontinuous within its native
range, and it is quite genetically variable, hence, there has been some disagreement as to the
most appropriate taxonomic treatment of the species (V idakovic 1974). It is now generally
recognized, however, that the species is composed of a number of interfertile subspecies,
many of which were formerly of species rank (V idakovic 1974). Pinus nigra ssp. austriaca
(Hoss) Bid., the subspecies present at SDSP, occurs in Austria, Slovenia and western
Croatia where it inhabits areas from 250m to 1000m above sea level (V idakovic 1991).

In its native range, P. nigra often occurs in pure stands where it attains heights of
30-40m and may live 300 years (Vergos 1985, Vidakovic 1991). First reproduction

typically occurs between 15 and 20 years of age, but under harsh conditions cones may not

9

be produced until the pines are more than 40 years old (Vidakovic 1991, Burns and

Honkala 1990). The species is known to be frost tolerant in the northern reaches of its

native range; northern varieties can withstand temperatures as low as -30° C, and southern

varieties, -7° C. Annual precipitation is between 610mm and 1020mm in its native range.

Forestry practices and cultivation of P. nigra

In North America, P. nigra is successfully planted from Nova Scotia to northern
Missouri (Burns and Honkala 1990). Cultivated in the US. as early as 1759, P. nigra was
grown both in National Forests for timber, and as a landscape and windbreak species in the
Great Plains states by the early 19005 (Burns and Honkala 1990). It tolerates a wide
variety of soil conditions and is planted extensively along highways because of its high
tolerance to road salt runoff (Burns and Honkala 1990, Wheeler et a1. 1976). It has
become naturalized in parts of New England and the Great Lakes states, but is known only

as a localized escape elsewhere in North America (Burns and Honkala 1990).

Insight into the success of P. nigra in North America may be gained by examining
the rationale for its appeal as a reforestation Species in Greece. Matziris (1989) lists four
reasons for its extensive use. First, it tolerates a wide range of environmental conditions
and has been shown to have three ecotypes in Greece. Second, it can grow in highly
disturbed environments and has few special requirements. Third, it competes well with
other plants at a young age and can establish with minimal management, and fourth, it
grows rapidly and produces high quality wood. These characteristics make P. nigra an
excellent reforestation species in its native habitat and attractive to managers. Its tolerance
of a wide range of environmental conditions may also make it an excellent invader in novel

environments.

10

The vigor of P. nigra and its ability to colonize many types of sites and soils
(I-Iepting 1971) give it the potential to alter a variety of communities, including sand dunes.
The dense canopy of P. nigra stands creates deep shade in which an understory is usually
absent. Its tendency to stabilize shifting sand may also be detrimental to some native dune

species which rely on open sand for colonization (McEachem 1992).

Research questions

This paper addresses the following questions:
1) which dune habitats are most conducive to primary establishment of P. nigra , and hence
may be most impacted by its presence?
2) are weather patterns and Site conditions such as slope and wind stress correlated with P.
nigra growth in the dune ecosystem at SDSP, and
3) how does the ecology of P. nigra at SDSP compare with that of P. nigra growing in its

native European range?

To answer these questions, growth and size measures of P. nigra trees and stands
were made in foredune, forest edge, wetpanne, and inland blowout seral stages during the

growing seasons of 1994 to 1996.

METHODS

Study site

This investigation took place between May 1994 and August 1996 in the state-
designated Natural Area of SDSP, and on adjoining private property, in Allegan County,
Michigan (42 41 ’N, 86 12’W, Figure 2-2 and 2-3). Interviews of previous land owners
conducted by former MSU student, Debra Reynolds, revealed that approximately 26,000
Austrian pines were planted throughout the 120-ha site while the land was still in private
ownership. The plantings, for which the seed source is not known, took place between
1956 and 1972 and were initiated to stabilize the dunes. Study plots were distributed

throughout foredune, wetpanne, forest edge, and inland blowout seral stages.

Weather data from a Holland, Michigan weather station approximately 13 km north
of SDSP, indicated that this site falls within the natural climatic variation for P. nigra in its
native range. All weather data used in this Study were derived from the Holland weather

station. Average annual precipitation for 1955-1996 in Holland, MI was 911mm and mean

growing season (May - August) temperature, 19°C. In February 1978, six years after P.
nigra plantings had ceased, a low temperature of -15° C was recorded, 15° higher than

-30°C temperatures in the northern native range of P. nigra .
Sampling sites
One goal of this study was to establish baseline data for long-term study of
dispersal and invasion of P. nigra on the dunes. Consequently, 20 permanent sampling

sites were randomly selected from a larger set of potential sites on a vegetation map of

11

12

SDSP, and five sampling sites were established in each of the four dune habitats where P.
nigra had been planted. Pinus nigra occurred in small stands of 3-11 trees in foredunes,
in large stands of hundreds of trees in wetpannes and at forest edges, and in stands of
variable size (25 - hundreds of trees) in inland blowouts. Sampling sites selected for study
in the large stands included only a portion of the trees, but included all the trees in small

and moderately sized stands (up to 50 individuals).

Sampling sites were established in areas of P. nigra occurrence as follows: five in

foredune sites, defined as small, discrete stands of trees covering areas of 150m2 - 420m2;
five at the forest edge, all portions of larger Stands (576m2 - 1890m2); five in wetpanne
sites, all portions of larger stands (550m2 - 1225m2); and five in inland blowouts, two of
which were discrete stands and three of which represented portions of larger stands (312m2

- 916m2). Sampling sites extended only as far as did the canopy of P. nigra in these self-

contained stands and individuals that were more than 15m from pines at the edge of discrete
stands were not included. Because stand area differed among sites, all comparisons among
sampling Sites were made on a percent or unit-area basis. All sampling Sites were
established in May and June 1994, except for two wetpanne sites which were established
in May 1995. Seral stages were defined by the type of characteristic native vegetation

bordering the planted pines.

Characteristics of individual trees

Size measurements.--Ten mature, planted P. nigra individuals, hereafter termed
“focal trees,” were randomly selected for size measurements in each wetpanne, forest-edge
and inland-blowout sampling site. Due to a limited number of trees in foredune sites, only

three trees were sampled in each of these sites. Height was measured with a Haga altimeter

13

and diameter at breast height (dbh) was measured at 1.25m above the soil surface for each
sample tree in 18 sites in May and June 1994, and in the two remaining wetpanne sites in

May 1995.

Because formulae specific to P. nigra were unavailable, cubic volume of the main
stem was calculated for each focal tree using the formula for a parabaloid:

y3 = l/2Abh;

where Ab=1tr2 = basal area, 1: 1/2 tree dbh at 1.25m above the soil surface, and

h=height.
Actual stemwood volume for conifers typically falls between that of a cone and a

parabaloid (Husch 1982), so this formula may slightly overestimate actual stem volume.
Height, dbh, and cubic volume of Austrian pines were compared within and among seral
stages using a nested ANOVA in which stage was treated as a fixed effect and Site(Stage) as
random. Transformations did not improve non-normal distributions of height, so Site data
were pooled by stage and compared among stages using a non-parametric Kruskal Wallis
AN OVA.

Size carrelations.--To investigate specific environmental factors that might be
correlated with tree size, slope measurements were taken with a clinometer from the edge of
the Austrian pine canopy to 10m into the pine stand, and “wind exposure” category was
determined for each site. Category 1 included sampling sites which were completely
isolated from any other trees which might buffer them from the wind, previously described
as “discrete stands.” Category 2 included Sites which were bordered on one side by a stand
of trees, and Category 3 sampling sites were surrounded by trees on all sides. Mean
height, dbh and stem volume for each site were analyzed relative to percent slope and
degree of slope using a non-parametric Spearman’s rho. Mean size variables for each site

were also analyzed by wind exposure category with a Kruskal Wallis ANOVA.

l4

Kellogg Forest comparison.» To compare P. nigra growth in the dunes with that
of an established plantation, height and dbh were measured on a P. nigra provenance trial
stand (plantation 61), established in 1961 by J. Wright of Michigan State University. The
site is described as “rolling with slopes of 0-25%, with Oshtemo loamy sand,” and is
located at Kellogg Experimental Forest in Augusta, MI, approximately 75km east of SDSP
(Wheeler et a1. 1976). To demonstrate the range of potential P. nigra growth in this site,
trees from Corsican (413), Austrian (423) and Greek (425) seed lots were selected for
sampling from a pool of 29 types. In 1976, at 15 years of age, these seed lots represented
the minimum, median, and maximum heights of trees in the plantation (Kellogg Forest,
unpublished data). Height, dbh, and parabaloid volume were obtained as previously
described for all surviving trees of each seed lot (Corsican: n=6, Austrian: n=19, Greek:
n=15). Stand density was calculated by dividing the total number of surviving trees by the

area occupied by the original plantings.

Stand characteristics at SDSP

Stand density.--Stand density was determined by counting every planted Austrian
pine in the 20 sampling sites at SDSP and dividing by the appropriate area Naturally
regenerated individuals in these stands (<10 years old) were smaller than 1cm in dbh and
1m in height and were not included in calculations of stand density. Stand density of pines
within and among seral stages was compared using a One-Way ANOVA with Tukey-
Kramer HSD multiple comparisons.

To evaluate the association among tree and stand characteristics, Spearman’s rank

correlations were determined for height, dbh, parabaloid stem volume and stand density.

15

Tree growth

Height growth -- Annual height growth of P. nigra on the dunes was determined by
measuring the distance between branch whorls (representing one year of growth) with a
meter tape for the first 2m of the tree and with an altimeter for the portion of the tree above
2m. Growth in years 1-10 may be under-represented in the whorls present. Annual height
increment of 10 trees was measured in foredune sites, and 11 trees were measured in each
of the other seral stages. To further examine the growth pattern, cumulative height was

regressed against tree age (total number of whorls).

A five—year growth rate was calculated by summing height growth across five-year
increments and plotting the value against age. Transformation to normality was not
possible, so separate nonparametric Kruskal Wallis One-Way ANOVA was used to test for
differences in height growth rate across seral stage and in growth rate across five-year
intervals. Spearman’s rank correlations were used to determine the relationship between

five-year growth interval and height growth rate.

Radial growth and age estimation. --To estimate tree age and to determine radial
growth rate, P. nigra trees were cored with an increment borer in the fall of 1994. Focal
trees (see Methods: characteristics of individuals) were cored in three of the five
established sampling sites in all seral stages. Rings were counted and measured on
polished tree cores using a Henson measuring machine. The paucity of signature years in
the relatively young trees (20 - 35 rings) made cross dating of the cores impossible (Stokes
and Smiley 1968), but two factors indicated that ring counting was an appropriate approach
to age estimation and dendroclimatological procedures (Telewski and Lynch 1991). First,
the number of rings present corresponded well with the range of planting dates. Second,

the presence of relatively wide annual increments for most years and two very narrow rings

16

from the drought years of 1987 and 1988 gave good reason to believe that the cores had no
missing rings. Average annual increment for five-year intervals was calculated for each
tree core and averaged within sites and within seral stages. Data were square-root
transformed to near-normality. Because the profile of annual radial growth across the life
of the tree did not differ within seral stages, sites were pooled and seral stages were

compared with a two-way AN OVA, with stage and growth interval as fixed effects.

To more directly compare growth across time, basal area increment (BAD was also
estimated for each year of growth from each core. Cumulative radial annual increment of
the tree radius was used to calculate the basal area of the tree at each ring; the basal area of
the tree in the previous year was subtracted from that of the year in question to determine
the BA] of the tree for that ring alone. This method assumes that double the cumulative
radius represents the true diameter of the tree. This may not always be an accurate
assumption, however, if wind stress to the tree varies from one side to another, and results
in unequal growth around the circumference of the tree. Patterns of growth, as measured
by BAI, were compared within and among seral stages. The association between age and

size (height and dbh) was determined using Spearman’s rank correlations.

Tree growth and climate.--To determine the relationship between tree growth and
annual climatic variation, Spearman’s rank correlations were calculated between mean
annual ring width within each seral stage and a variety of precipitation variables, including
annual precipitation, growing-season precipitation (April-August), June, July, and August
precipitation, as well as precipitation from the previous year in all the prior categories.
Several monthly precipitation values were missing (August 1979, February 1985, and
August 1987) and these years were eliminated from the analysis when the missing data
affected the outcome of the correlation. The average ring widths for each seral stage used

in the correlation analysis were determined by averaging ring widths of the same year for

17

all trees within each site, and then averaging yearly site averages together for each seral

stage.

All statistical analyses were performed using JMP software (SAS 1995).

RESULTS

Pinus nigra survivorship

Of the approximately 26,000 pines planted at SDSP, an estimated 21,366 were still
present on the dunes in 1988: an 81% survival rate for trees which were 16-32 years of age
(Reynolds, unpublished), although no data specific to seral stages were available.
Survivorship on the dunes compared favorably with 35-year survivorship of the Austrian
seed lot at Kellogg Forest plantations (79.2%, Table 2-1) but was much higher than that of
the Greek seed lot (37.5%).

Table 2- 1. Average 35-year survivorship of P. nigra from three different seed sources in a plantation at
Kellogg Forest, Augusta MI, and at SDSP, Allegan Co., MI.

     

Average Standard

 

Seed Lot Survivorship Error
(96) (96)
Greek 37.5 7.2
Austrian 79.2 1 1.9
Corsican 62.5 10.7
SDSP"I 81

 

*Reynolds, unpublished data for survivorship from time of planting (1956-1972) to 1988.

Tree characteristics and stand density at SDSP

Tree age. -- Tree age, as measured by number of tree rings, was only moderately

18

correlated with height of P. nigra planted on the dunes (rho=0.265; p<.05, n=86) and not
at all correlated with dbh or stem volume (rho=0.007, p=0.95, n=86; and rho=0.051,
p=0.64, n=86). Because of these negligible relationships between age and size within the
age class of the trees planted on the dunes (25-40 years), it was not possible to directly

compare height, dbh and volume of individuals trees based on age.

Tree size.» Tree height, dbh, and Stem volume were more variable within seral
stage than among stages at SDSP (Table 2-2, Figures 2-4 to 2-6). When sites were pooled
within stage, however, tree height was found to vary Significantly among stages. Mean
DBH was nearly different among stages (though this result was not statistically significant;
p = 0.055), and mean stern volume did not differ among stages.

Table 2-2. ANOVA for DBH, height, and stem volume, where sites were nested within seral stage

(foredune, forest edge, wetpanne, inland blowout); and Kruskal Wallis test, where data were pooled within
seral stage. Measurements were taken in the growing seasons of 1994 and 1995, SDSP, Allegan Co., MI.

 

Source of f

 

 

 

Tree characteristic Variation (1‘ ms ( ANOV A) p
DBH Stage 3 139.392 0.709 0.560
(model r2=0.448) Site(Stage) 16 204.711 6.616 <0.0001

Error 147 30.941
Height Stage 3 68.881 2.427 0.103
(model r2=0.509) Site(Stage) 16 29.554 6.658 <0.0001
Error 147 4.439
Stem volume Stage 3 0.014 0.298 0.827
(model r1=0.380) Site(Stage) 16 0.048 5.345 <0.0001
Error 147 0.009

«grimly ”(Km
DBH Stage 3 7.602 0.055
Height Stage 3 28.505 <0.0001

Stern volume Stage 3 6.075 0.108

 

19
Mean dbh of P. nigra was lowest in forest edge Sites (14.8cm 1 1.0) and highest in

foredune Sites (20.3cm i 1.8; Table 2-3, Figure 24). Within each seral stage, mean dbh
of the Site with the highest mean dbh was more than two times that of the site with the

lowest. Mean height was lowest in foredunes (6.9 * 0.5m) and highest in forest edges (9.6

* 0.3m, Table 2-3, Figure 2-5). Foredune and inland blowout stages were most variable:

mean tree height of the site with highest mean height exceeded that of the lowest by more
than two times. Wetpannes were least variable: the mean height of the site with lowest
mean height was 80% that of the site with highest mean height.

Table 2-3. Mean size and stand characteristics iS.E. for P. nigra in four seral stages at SDSP. Allegan

Co., MI (data from 1994 and 1995) and for three seed lots in a 35-year old P. nigra plantation at Kellogg
Experimental Forest, Augusta, MI (1996 data).

 

 

 

 

 

Seral Stage or Height (111) DBH (cm) Stern Volume Stand Density 11
Sad Lot (m3) (trees per ha)

SDSP # sites
Foredune 6.9 1 0.5 20.3 1 1.8 0.138 1 0.03 274 1 34 5
Forest Edge 8.9 10.5 14.8 1 1.0 0.110 1 0.02 1176 1 124 5
Wetpanne 9.6 1 0.3 17.2 1 0.8 0.137 1 0.02 1173 1 156 5
Inland Blowout 7.1 1 0.4 15.6 1 1.0 0.100 1 0.01 693 1 88 5
Kellogg Forest # trees
Corsican 14.7 1 0.6 14.9 1 2.9 0.152 1 0.05 1051 6
Austrian 15.4 10.5 20.7 1 0.9 0.277 1 0.03 1051 19
Greek 16.6 1 1.2 22.2 1 2.4 0.378 1 0.07 1051 15

 

Mean stern volume did not differ across seral stage. Variability of height and dbh
was amplified in the calculation of stem volume, however. In foredunes, stem volume of
the highest Sites was more than seven times that of the lowest (Figure 2-5). Wetpannes

were the least variable with just 2.5 times difference between the highest and lowest Sites.

Stem volume varied from a low of 0.019 j-0.008m3 in inland blowout Site five to a high of

0.273 1 0.079m3 in foredune site four (Figure 2-5).

20

Allametric relationships.» Despite Similarities in mean stem volume among stages,
tree morphology, as described by the combination of dbh and height, differed across seral
stages. Foredune trees were shorter, but had larger dbh than wetpanne and forest-edge
trees which were taller, and had lower dbh (Table 2-3, Figure 2-7). Inland-blowout pines
were intermediate in height and dbh. Nonparametric Spearman’s correlations showed a

strong positive relationship between height and dbh across all seral stages (rho=0.6247,
p<0.0001).

Environmental carrelates.--Plant size was independent of site characteristics. None
of the correlations between mean P. nigra size variables and slope of the site were

statistically significant at SDSP (Table 24). Mean tree height per site differed with

exposure to wind (X2=6.21, df=2, p<0.05, n=20), but mean dbh and stem volume within

sites did not (x2: 2.10, and 0.37 respectively, df=2, p>0.05, n=20).

Table 2-4. Spearman’s rank conelations between site slope and several measures of tree Size in P. nigra at
SDSP, Allegan Co., MI (n=20, size data taken in 1994 and 1995).

Size Variable Rho p

DBH -0.235 0.3 18
Height -0.216 0.3 16
Stem Volume 0.163 0.493

 

 

Stand density.--Stand density of planted P. nigra was lowest in the foredunes at an
average of 274 i 34 trees/ha (though no stands exceeded 0.04 ha in size), intermediate in
inland blowouts (693 i 88 trees/ha), and highest in wetpannes and forest edges (1173 :l:
156 and 1176 i 124 trees/ha, respectively; Table 2-3, Figure 2-8). Foredune and inland
blowout stand densities were significantly different from those of forest edges and
wetpannes (Tukey-Kramer HSD, q=2.86, p=0.05).

21

Relationships between tree size and stand density.» Tree height and stand density
were positively correlated, but only 28% of the variation in height was explained by density
(rho: 0.28, p=0.0002, n=20). Dbh and stand density Showed a weaker negative
association (rho=-0.19, p=0.0137, n=20) and explained less of the variation. No
statistically significant relationship was noted between stem volume and stand density (rho
= -0.07, NS, n=20).

Comparisons with Kellogg F arest.--DBH for the 25-40 year old planted P. nigra
on the dunes was comparable to that of the 35-year—old trees in Kellogg Experimental
Forest, where it averaged between 14.9 :1; 2.9cm and 22.2 1; 2.40m depending on seed
source (Table 2-3). Trees at Kellogg Forest were nearly double in height, however,
ranging from 14.7 1; 0.6m to 16.6 1; 1.2m. Stern volume at SDSP was comparable to the
Corsican seed lot at Kellogg Forest, but was less than half that of the Austrian seed lot.
Stand density in foredunes and inland blowouts at SDSP was only 0.25 and 0.66 that of

Kellogg Forest, but in wetpannes and at forest edges was 1.1 times that of Kellogg Forest.

Tree growth

Height growth.» Tree height followed a linear trajectory across time in all seral
stages (Figure 2-9, Table 2-5). The rate of height growth did not vary across seral stage,
but it did with tree age (Figure 2-10). Growth rate was positively correlated with tree age
(rho=0.6155; p<0.0001).

Radial growth. » Radial growth rate differed among seral stage; foredune and inland
blowout had the highest, and forest edge and wetpanne, the lowest rates of radial growth
(Figure 2-11). In addition, average annual increment generally decreased with increasing

age (Spearman rho= -0.4461; p<0.0001). In contrast, height growth increased with age.

22

Table 2-5. Linear regression of height against tree age (total number of whorls) in four seral stages at
SDSP. Data taken in 1996, Allegan Co., MI.

 

 

Seral Stage R2 p
Foredune .80 <.0001

Forest Edge .89 <.0001
Wetpanne .86 <.0001

Inland Blowout .82 <.0001
Pooled Stages .85 .0000

 

Table 2-6. Spearman’s rank correlation between June precipitation (current and previous) and tree-ring
width in four seral stages. Tree cores collected in Fall 1994, Allegan Co., MI.

 

 

Month of
Precipitation
Seral stage Rho June p :rhegious Junep
Faedune -0.0879 NS -0.0121 NS
Forest Edge 0.4136 0.0150 0.3198 0.0652
Wetpanne 0.4350 0.0184 0.391 1 0.0359
Inland Blowout 0.5480 0.0038 0.4154 0.0348

 

Tree growth and climate.--Annual increment was positively correlated with June
precipitation of the year in which growth occurred in all but the foredune seral stage (Table
2-6). In inland blowout and wetpanne sites, a weaker positive correlation existed between
ring width and precipitation during the previous June. None of the other correlations

examined between annual increment and climatic variables were significant.

DISCUSSION

Overview

In order to examine the primary establishment success of P. nigra in a novel
environment, its differential success was examined among seral stages in the dunes to
determine if one habitat restricts P. nigra growth more than others, or if the introduced
species is equally successful in all habitats. Environmental correlates with growth were
also explored. Comparisons were then drawn with trees of similar age in North American
P. nigra plantations growing under conditions less harsh than those in the dunes.
Plantations are expected to represent the upper limits for measures of pine establishment
since trees are managed as crops for which the highest possible productivity is desired.
Primary establishment success at SDSP was also compared with that of a European
plantation located in close proximity to the native distribution of the species. This was
thought to give a more realistic picture of what constitutes successful establishment in its

native distribution.

Finally, comparisons were made with the native distribution. Climatic comparisons
were drawn to determine if the conditions at SDSP are similar to those within its native
range. The size and stand—density characteristics of trees at SDSP were then compared
with those in a natural uneven-aged stand within the native range. If primary establishment
of trees at SDSP is comparable to that of a region in which the trees are known to persist
naturally, it is likely that they have the potential to become successful invaders in the novel

dune system.

23

24

Primary establishment of Pinus nigra at SDSP

Survivorship.» Despite the depauperate soils and droughty conditions of the sand
dunes, P. nigra survivorship has been high (estimated at approximately 80%) since
planting at SDSP. In a study of primary establishment success of landscape plants derived
from Yugoslavia and grown in a broad range of environments in the North Central United
States, Widrlechner et a1. (1992) found lO—year survivorship of P. nigra from two
Slovenian seed sources to be 64% and 19%. The majority of mortality occurred during the
first year after planting, and lO-year survivorship, calculated only for the individuals that
survived the first year, was found to be approximately 75% (W idrlechner et a1. 1992).
Estimates of 16- to 32—year survivorship of P. nigra at SDSP exceed these values;
therefore, early mortality after planting does not appear to be a major banier to P. nigra

invasion potential in the dune system.

Size.» Tree Size, as measured by dbh, height, and stem volume, differed more
within seral stage than among seral stages at SDSP (Table 2-2, Figures 2-4 to 2-6). This
indicates that the coarse-scale differences in environmental factors among stages account
for only some of the variation in tree size. Some of the variation within seral stages may be
due to genotype, because the trees were planted over a 15-year period and likely came from
different seed sources during that time. In addition, tree height-growth may have been

reduced by exposure to wind, as determined by the absence of trees (either native or exotic)

surrounding the sampling sites (Xzzfl = 6.21, p<0.05), but because categories of “wind

exposure” were confounded with seral stage, this environmental factor does not help to
explain within-stage height variation. For example, all of the foredune sampling Sites were
in the “most exposed” category; therefore, variation in foredune height could not be

attributed to differing levels of wind stress. Additionally, initial planting densities may

25

have influenced eventual tree dimensions, because stand density was positively correlated
with height (rho=0.28) and negatively correlated with dbh (rho=-0. 19). Further study is

needed to determine the causes of variation in P. nigra size within seral stage, however.

Despite the intra-stage variation in size variables, pooled analyses showed
differences in height across stages (Table 2-2). Wetpanne and forest edge trees were, on
average, significantly taller than foredune and inland blowout trees. If dominant tree height
is taken as an indication of site quality in even-aged stands (Oliver and Larson 1990), these
data would suggest that wetpannes and forest edges were of higher site quality for primary
establishment of P. nigra than either foredune or inland blowout seral stages (Table 2-3).

Because stem volume included measurements of both height and DBH, and was
independent of stand density in this study (rho=-0.07, NS), establishment success may be
better evaluated by this measure than either height or dbh. Contrary to tree height, stem
volume did not vary across seral stage in pooled analyses (Table 2-2 and 2—3), suggesting
that P. nigra was equally tolerant of very different conditions among seral stages. Stem
volume estimates did not include branch volume, however, and therefore could not be
taken as a complete evaluation of establishment success. Branches appeared to be more
dense and numerous on foredune trees than on wetpanne or forest edge trees, possibly as a

response to shading of the lower branches in the denser stands.

Stand density.--Much of the difference in P. nigra stand density among seral stages
was likely attributable to differences in planting practices. In foredune sites, a single row
or two of trees served as a windbreak, whereas in the sandy, destabilized inland blowouts
and on the steep slopes of the forest edge, many rows of pines were introduced to reduce
sand movement. In addition, fire occurred on the foredunes in 1972 and again in 1985

(Reynolds, unpublished data), and this may have reduced stand densities in foredune

26

sampling sites. Fire scarring is still present on many of the standing foredune trees and
charred, fallen trees are evident as well. Differential early mortality of the planted seedlings
among seral stages may also have been responsible for some of the differences in stand

density, but no records of this exist.

Pinus nigra growth.» Tree height, when plotted as a function of age, generally
follows a sigmoid curve (Oliver and Larson 1990). Stressful conditions can alter this
general pattern, by reducing the rapid juvenile phase of growth and lengthening the decline
phase, however (Fritts 1976). Height-growth patterns at SDSP suggest that P. nigra was
still in a rapid phase of growth in all seral stages (Figure 2-9). Rates of height growth were
still increasing in inland blowout and wetpanne sites, but had begun to level off and even
decrease in foredune and forest edge sites (Figure 2-10), however. This suggests that the
trees in these seral stages were approaching a Size at which height becomes limiting (Oliver
and Larson 1990). A decline in height growth rate occurred five years earlier in foredune
than forest-edge trees, possibly indicating that the foredune seral Stage was of poorest

quality for primary establishment.

Despite their taller stature, wetpanne and forest-edge trees exhibited slower radial
growth (Figure 2-11). Because woody plants preferentially allocate energy to height
growth, however (Lanner 1985), radial growth data are inferior indicators of site quality
(Oliver and Larson 1990). The larger annual increments of foredune and inland blowout
trees may be related to increased wind stress in these seral stages (Telewski 1995).
Foredune trees in particular were located in exposed areas, and because stand densities

were also lowest in this seral stage, trees were least protected from the wind.

Environmental determinants of growth. -- Positive correlations between June

rainfall and annual increment suggest that June precipitation was an important regulator of

27

radial growth in all stages but the foredunes. Wind stress, which may be most extreme in
foredunes is known to increase radial growth (Telewski 1995) and in this case, radial
growth, as a response to wind stress, may obscure the precipitation-based variation across

annual ring widths.

The moderate positive correlation between radial growth in wetpanne trees and _
precipitation suggests that radial growth was limited by June precipitation in even the
wettest of the four seral Stages. Observation during the field seasons of 1993 through 1996
showed that wetpannes soils were typically saturated, if not inundated during the month of
June, yet the tree-ring analysis suggests that soils in wetpanne sites dried out in the driest
years. Though annual radial increment was smaller in wetpannes than other Stages, a
negative correlation with precipitation would have been expected if waterlogged and
anaerobic soils were limiting growth. There is no indication of this in the present analysis,

however.

While P. nigra growth was correlated with precipitation in three seral stages, basal
area increment (BAD data indicated that biotic interactions and subsequent light limitation
may also be regulating tree growth. Four patterns of growth were evident in BAI data for
tree cores from three sites in each seral stage (Figure 2-12), and these patterns represent the
dynamics of growth commonly experienced by trees in naturally occurring forest stands
(Oliver and Larson 1990). The first growth pattern (A) is that of free or open growth,
which is characterized by an upward trend in BAI. over time as the tree gets larger and is
able to garner more energy. Crown closure has not yet occurred so the trees are not in
competition for light. Trees in sparsely planted stands would be expected to exhibit this
type of growth pattern. The next growth pattern is that of competition (B), which occurs
after canopy closure and is evidenced by a rapid decrease in BAI over time as light becomes

more and more limiting. Trees in more densely planted stands would exhibit this growth

28

pattern earlier. A third (C) is that of canopy closure and subsequent gap formation. Here
the BAI decreases as a response to competition for light, and then increases rapidly again
following an opening in the canopy. This pattern might be expected for a small proportion
of the trees in the more densely planted stands. The fourth growth pattern (D) is that of
trees in the understory which have always been limited by light and have grown slowly

throughout their lives.

Two to four of these growth patterns were exhibited by trees from each site in all
seral stages, perhaps as a response to their specific position in the stand. AS might have
been predicted, however, trees from the sparsely planted foredunes and inland blowouts
showed primarily pattern A, that of open growth. Understory and canopy closure patterns
were second most common in inland blowout sites (each pattern was represented in six of
28 trees), which were more than two times as dense as foredune stands. Trees in the more
densely planted forest edge and wetpanne stands were mostly of pattern D (understory
trees) with pattern C (gap formation) the second most common in both stages. Only one of
56 trees cored in wetpanne and forest edge sites exhibited the open growth pattern. These
data suggest that intraspecific competition for light is a significant limitation to growth in

forest edges and wetpannes, but less so in inland blowouts at SDSP.

Pinus nigra establishment in even-aged plantations

Pinus nigra height, dbh and, stem volume in plantations at Secrest Arboretum in
Wooster, OH were comparable to that of Kellogg Forest (Table 2-7 (Burns and Honkala
1990); Table 2-3), but trees in both experimental forests were considerably taller than even

the tallest trees in forest—edge sites in the dunes (11.1 1" 0.8m). Stem volume on the dunes

was 0.5-0.75 that of 31-year-old stands at Secrest Arboretum. Management practices,

including herbicide and fertilizer use, may have given these plantation trees an advantage

29

over those in the dunes, although the differences are great enough that other factors, such
as moisture availability and soil type, may have been involved. These differences in size
for trees of similar age suggest that both experimental forests were of higher Site quality

than any seral stage in the dunes.

When compared with even-aged P. nigra plantations in France, however, stem
volume per tree at SDSP was 25 to 75 percent higher than that of trees of the highest yield
class of 30-year-old European stands (Table 2-7). Stand densities in France were two to
nine times higher than on the dunes, however, and this may account for the greater stem

volume of individual trees at SDSP.
Growth analysis: native habitat vs. SDSP

Climatic camparisans.--“Climate matching,” or comparison of climatic and
latitudinal variables of native and exogenous ranges is not always a good predictor of
invasion potential (Mack 1996), but may give some baseline indication of habitat
compatibility. The climate at SDSP is quite comparable to that of P. nigra ’s native range.

Annual precipitation averages 911m at SDSP, well within the range of 600-1100mm for
native European sites (V idakovic 1991). Minimum temperature at SDSP (-15°C) is 15°C

higher than in the northern range of P. nigra’s native distribution. Additionally, SDSP is
located at the midpoint of the of native latitudinal range, though the European continent is
subject to the moderating effects of the bodies of water surrounding it. The climate at
SDSP appears to match well with that of the native distribution of P. nigra, and this

undoubtedly contributes to the successful performance of the species at SDSP.

Primary establishment of native P. nigra .» Native populations of P.nigra var.

pallasiana growing in mixed stands with Abies equi-trajani in Turkey Show similar

30

Table 2-7. Comparison of height, dbh, stem volume, and stand density for even-aged P. nigra plantations
in Secrest Arboretum, Wooster, OH (Burns and Honkala 1990), and southwest France (V annier 1984), and
for uneven-aged native stands in Turkey (Ata 1989). Yield classes“ are given only for France.

 

 

 

 

 

 

 

 

 

Yield Class (111) (cm) (m3) (trees/ha)
Wooster, OH
25-year 18.0 11.8 0.15 nodata
31-year 17.8 14.5 0.18 nodata
40-year 22.4 17.1 0.34 no data
France (30-year)
Yield class
Class 4 3.9 8.3 0.010 4549
Class 3 6.0 10.2 0.024 3760
Class 2 8.1 12.1 0.047 3067
Class 1 10.2 14.0 0.079 2509
France (40-year)
Class 4 5.9 10.2 0.024 3782
Class 3 8.4 12.4 0.051 2995
Class 2 11.0 14.6 0.093 2327
Class 1 13.5 17.5 0.162 1812
Turkey
30-year 6.5 9.0 0.021 no data
40-year 9.5 1 1.5 0.049 no data

 

I"Yield class 1 represents the most productive conditions and yield class 4 the least.

31

patterns of establishment success to P. nigra at SDSP (Table 2-3, 2-7). Ata (1989)
performed stern analyses on 30 individuals to determine average height and dbh at ten-year
intervals and found a 30—year average height of 6.5 m and average dbh of 9 cm, and a 40-
year average height of 9.5 m and average dbh of 11.5 cm. This range encompasses
weakest (foredune) to strongest (wetpanne) P. nigra performance at SDSP (Table 2-3).
Mean stem volume, however, was five to seven times greater on the dunes than in 30-year
old trees in native stands in Turkey. This comparison suggests that primary establishment
of P. nigra in all seral stages in the novel environment is as successful as it is in its native
habitat. Ata (1989) indicated, however, that P. nigra was eventually overtopped by the
faster growing A. equi-trojani and excluded from these mixed forests over time. Pinus
nigra also grows in pure stands in its native range, and though no data were available for
comparison, it is possible that the reduction in interspecific competition improves measures

of establishment success in pure relative to mixed Stands.

Stand densities of planted P. nigra in all seral stages on the dunes were similar to
stand densities found in naturally occurring pure stands of P. nigra in Greece. Vergos
(1985) defines five distinct phases of stand evolution in a study of pure stands of P. nigra
in NW Greece (the forest of Krania-Monahition), including young forest stage (3809
trees/ha), optimum stage (904 trees/ha), aging stage (331 trees/ha), decomposition stage
(267 trees/ha) and regeneration stage (1687 trees/ha). Foredune stand densities at SDSP
were most similar to the decomposition stage densities, while forest edge and wetpanne
stand densities were 1.3 times that of the optimum stage. Inland blowouts were more than
three quarters that of optimum stage stand densities. As the planted trees age at SDSP, a
decrease in stand density may be expected until regeneration becomes a significant factor.
If P. nigra populations behave on the dunes as they do in natural stands in Greece, the

density of regenerating foredune stands may be expected to be six times that of their current

32

values, while forest-edge and wetpanne stand density may only increase by about 45
percent. As previously indicated, the stand densities observed at SDSP seem to have little
effect on height, dbh or volume of P. nigra growing on the dunes, though height shows a

slight positive correlation with stand density.

Conclusions

Comparable height, dbh, and stem volume of trees of similar age in the native
region and in the Lake Michigan sand dune environment provides evidence that P. nigra
has the potential to invade the fragile dune ecosystem. Within the dune system, P. nigra
was highly tolerant of a variety of conditions, but as evidenced by significant intra-seral-
stage variation, responded to environmental variables on a scale smaller than defined by

seral stage.

An assessment of the differential success of P. nigra among seral stages is
criterion-dependent. For example, tree height versus stem volume measurements lead to
different conclusions regarding the primary establishment success of P. nigra among seral
stages at SDSP. Height measurements indicate that P. nigra was more successful in forest
edge and wetpanne seral stages (Table 2-8). Height growth rates were declining in forest
edge trees, and not in wetpannes, however, therefore wetpanne trees appear to have been
most successful in primary establishment. Foredunes, with the shortest trees and declining
height growth rates, may be interpreted as least conducive to pine establishment. Stem
volume did not vary across seral stage, however, which suggests that P. nigra established
well in all these communities. Pine growth in all but foredune seral stages was limited by
June precipitation. Biotic interactions which may have resulted in light limitations also
appeared to regulate growth in wetpanne and forest edge stages, but not in inland blowouts

or foredunes (Table 2-8).

33

Table 2-8. A comparison of measures of P. nigra establishment success among four seral Stages at SDSP.
Postulated growth regulators are also listed.

 

 

Height Low High High Low

DBH N 0 difference No difference No difference N 0 difference
Stem volume No difference No difference No difference No difference
Height growth rate Declining Declining Increasing Increasing
Annual radial High Low Low High
increment

Limits to Growth Wind exposure (?) June precipitation June precipitation June precipitation
Biotic interaction Biotic interaction

 

Pinus nigra has been successful through the primary establishment phase of
invasion in foredune, forest edge, wetpanne, and inland blowout seral stages at SDSP.
With the knowledge that it has passed through this critical phase in the invasion process
and may have the potential to persist and invade the dune system, it becomes important to
determine the consequences of its presence. Even though none of the seral stages
investigated excluded the pine from primary establishment, are they differentially affected
by its presence? In addition, P. nigra has recently begun to reproduce and recruit new
offspring in the dune system, thereby passing into the next phases of the invasion process.
A thorough understanding of the variation in reproduction and recruitment across seral

stages will be essential in predicting long-term consequences of this introduction.

34

 

 

 

 

Range of Pinus nigra. I. P.n. ssp.:alzmanni: a) Rn. maurcranico, b) Rn. hispanica. c) Rn.

pyrmaica, d) P.n. cevenncnsr‘s; 2. P.n. ssp. paireriarra : a) P.n. carsr'cmm. 1)) P41. calabn'ca. c) P.r1.

barren; 3. P.r1. ssp.nigricans: :1) P.rl. auslriaca. b)P.rr. dolmmr’ro. c)!’.rl. boxniacu. d)I’.r1. pimlit'u. c)
l’.r1./rulquricu, 1‘) l’.rr.barlalicrr,‘ 4. l’.rl. ssp. pallasiana: 11) [’71. ('(II'UIIHIHIUI, b) PJI‘. taurma. c) I’.rr.
purl/Ira, d)P.n. rararica (from Svoboda. 1953).

Figure 2-1. Native distribution of Pinus nigra in Europe and Asia (adapted from Vidakovic 1991; based on
a map from Svoboda, 1953). Shaded areas indicate occurrence of P. nigra.

 

35

 

0-.--..
.----------C--....--------.

NDllANA

l-‘rc. r.—.\1Ar or rm: 5110111: or: 1.111111 MICHIGAN. SCALE, r:2,85o.ooo.

1'1ng-2. WonofshrdydtefiwgehmkDunceSQtePukhnoaflanshueofukeMichigen,
Allegan, County, MI (from Cowles 1899).

36

 

Figure 2-3. Aerial photograph of the study Site at Saugatuck Dunes Natural Area. and on adjoining private
property, in Allegan County. Michigan (1988 M1 DNR). Pinus nigra plantings are evident as dark lines in
blowout areas and as scattered patches in foredunes. Sampling sites were located throughout four seral
stages in areas with and without Austrian pine. Boundary of study area is demarcated with bold lines. One
cm = 120m.

37

 

Austrian Pine DBH in Four Seral Stages

 

 

DBH (cm)

 

 

l Average

 

 

Foredune Forest Wetpanne
Edge Blowout
Seral Stage

 

 

 

Figure 2-4. Mean diameter at breast height :1: SE of planted P. nigra in five sites in each of four seral stages
at SDSP, Allegan Co., MI (n=5)."

*Data taken from ten trees in each of five sites in forest edge, wetpanne, and inland blowout seral stages.
Only three trees were measured in each of the five foredune sites because of their relatively small area.

38

 

Austrian Pine Height In Four Seral Stages

 

 

 

 

 

 

 

 

1 4
1 2 Site
A 1 o . ‘ I 1
E \o. ' V I 2
: 8 § ; § I 3
'5» 6 {\Q ; I § ' R 4
SI SI S as
4 §' §' § u Average
2 S. §l S
.. s R .. s
o . .
Foredune Forest Wetpanne Inland
Edge Blowout
Stage

 

 

Figure 2-5. Mean height :1: SE of planted P. nigra in five sites in each of four seral stages at SDSP.
Allegan Co., MI (n=5).‘

*Data taken from ten trees in each of five sites in forest edge, wetpanne. and inland blowout seral stages.
Only tree trees were measured in each of the five foredune sites because of their relatively small area.

39

 

Average Austrian Pine Parabaloid Stem
Volume in Four Seral Stages

0
.h

 

O
(A)
01

 

   

O
or

 

 

 

Stem Volume (m3)

 

Foredune Forest Wetpanne Inland
Edge Blowout
Seral Stage

 

 

 

Figure 2-6. Mean stem volume t SE of planted P. nigra in five sites in each of four seral stages at SDSP,
Allegan Co., MI (n=5).*

*Data taken from ten trees in each of five sites in forest edge, wetpanne, and inland blowout seral stages.
Only three trees were measured in each of the five foredune sites because of their relatively small area.

40

 

 

 

 

15 7
I
141 V
v v
13" V
‘V
12- V? 7' E I IVY
v' '
11- V I I I I I
It, In I:
10- '7 VI Inn-I I II' I
A E I l I I
E, 9- I I I: I'll I -ll II
155 u a u I:
Q g- V I I' II III
‘1’ v
7 VI I: m '7: II III
I y I: II I I
5 In! I
v I a I I
5- an: I! I
' II
44! , IV
I I I I y
3.1: '
II
2‘3—7‘ ' I ' 1 ' I ' l ' u r 1
5 10 15 20 25 30 35

DBH(cm)

Figure 2-7. Scattergrarn of relationship between dbh and height in P. nigra planted in four seral stages,
where x=foredune, y=forest edge, z=wetpanne, and o=inland blowout. Measurements taken June 1994 and
1995 at SDSP, Allegan Co., MI.

 

41

 

Average Austrian Pine Stand Density
in Four Seral Stages

 

 

 

1?
E
O
0
2
E:
To
C
8
Foredune Forest Wetpanne Inland
Edge Blowout

Seral Stage

 

 

 

Figure 2-8. Mean stand density (:1: SE) of P. nigra in four seral stages at SDSP, Allegan Co., MI (n=5).“

*Data taken from total inventories of planted P. nigra in five sites in each seral stage. Significantly
different densities are denoted with different letters (p<.0001, 'Ilukey-Kramer HSD).

42

due ooh we assume 5 8 e8: 33 area? we con—.52 6:5 3.88 830.63

= 28205 down Engage Co 803 BEES Bob mo cocoa 2:. ions» .58 025 mew—.358 05 we :80 E 2 e5 0:528 E 80... 2 80¢ econ—=23 9:32..

...:z .8 58.2 .anm a case .88 as a use apes mm e 232. one and: a-" name.

 

 

6589251111
anew «mocomllul

 

 

 

_ 692m atom—

 

 

sozoa 225-111

2533.... 1 let I

(w) 11151011

.. ..Nw

 

rv»

one...» .83 .50... c.
on...» 322‘ 22o... on... 5.322 emaco><

 

 

43

II%
Sosa-m .525 III
2:38.51?
omem aooaoulel
9:620“. lel

 

 

 

 

mm

«P

.28 ..o 0.953 a :0 .593 a. .. 3.302. 3..» 5630.9 9a.... ..o 2...? .8. 2.. 8.. 55% a.

.m.m oz . one 3.: .o 82:53 5 3 e8: 83 «Ev—.3 .o 398:2 $5882.. one» 3.. a. mouse. .88 .58 5 ENE .& 5 mm a :58» 2.32.. 502 ....-N 2.5....

eu<

in
I
O

 

 

 

 

m
m
'A\ .a a
‘Illlrilllsi m
‘ e...
3
.mm

momuum 33$ .50“. c. 539.0 2.90... cues.

.090 00.0 ..o 80058 :0 ad 09.: 33 «we... 00.: .0 009802 .00..» 805 0. 052.96 0.00.0 .0 02.05: 08.8..
0... ..o 8:002. 005.008.. 0. «8.0. 02... ..o ..000 88.. :83 803 8.8 025 ...:O down..." .88 30303 .05.... 0:0 60:090.: .0300 60.6.. 0. 00..» 02... .0 ..000
Be... 00.80 c. EB. 000.0. San. $368... BT03. 3.. 00.0.3.8 2.0880... .0005 0003. .833 .Eom Be. 0. EM... .0. .5. mm H 532» 1.08 000.2 .. .-N 83....

.9005 539.0 to .0302:

on 0. mm mm 2 3 om 0. o» m» o. ..p or o. m m o. p
u u u o

1.. .OO»

4.

-5255 90.51? H com
0000902. If .
0000 .005... Incl I -

05.520“. '6' O O V

s‘ . com

 

 

 

 

 

l
(tu ) luaureroul lenuuv afiererw

1 com

 

. 00h

00905 .0._0m .50". c. 00.0 0030.3
0. 0.020.... ..00>-m :. E0805... .0::c< 0m0._0><

45

 

 

 

 

 

a) Open growth b) Competition
BAI BAI Crown
closure
//
Time Time
c) Gap formation (1) Understory growth
1
BAI BAI

 

 

 

 

Time Time

Figure 2-12. Diagrammatic representation of patterns of BAI (basal area increment) of P. nigra across time
at SDSP, Allegan Co., MI.

CHAPTER 3
BIOTIC AND ABIOTIC ALTERATIONS OF FOUR SAND DUNE
HABITATS BY THE INTRODUCED PINUS NIGRA

INTRODUCTION

Plants alter the environments in which they live. They occupy space, utilize soil
moisture and nutrients, filter incident radiation, and add carbon to the soil, to name just a
few of their effects. These changes in the abiotic environment, in turn, influence the plant
communities of which these individuals are a part, and over many years, stable, self-

sustaining communities are assembled and maintained.

When non-native plant species are introduced into novel habitats, they, too, alter
their environments, sometimes in ways that the native Species cannot tolerate. Loope
(1992) and Williamson (1996) argue that plant introductions most often have little or no
effect upon the biota of their recipient community. When they are successful, however,
invaders can impact all levels of ecological organization of the recipient system. They alter
population dynamics of native species by inhibiting germination or reducing growth and
recruitment into larger size classes (Walker and Vitousek 1991, Equihua and Usher 1993,
Hughes and Vitousek 1993). Effects of invaders on ecosystem function include changes in
geomorphological processes, alteration of biogeochemical and hydrological cycles, and
alteration of fire regimes (Macdonald et a1. 1989 and references therein). Invasive species
also alter community composition and structure. For example, in eastern deciduous
forests, the introduced Lonicera tatarica reduces herb cover, tree seedling density, and
species richness, most likely by shading out native Species (Woods 1993). Similarly,
Rhododendron ponticum is thought to inhibit recruitment of native forest species in Britain
by dramatically reducing light intensity and producing a thick litter layer (MacDonald et a1.

46

47

1989). Though many references are made to effects of invaders upon plant communities,

clear documentation of this phenomenon is lacking (Woods 1993).

This paper documents the effects of an introduced pine (P. nigra) on both abiotic
and biotic components of a sand dune system. Four seral stages, or dune habitats
(foredunes, forest edges, wetpannes, and inland blowouts) were investigated, and

population, community and ecosystem level effects of the pine introduction are discussed.

Invasive species may have important consequences in dune systems for a number
of reasons. Continuous endogenous disturbance in the form of sand movement is an
inherent characteristic of sand dunes which maintains the vigor of the native vegetation
(Ranwell 1972). Any permanent change in the natural disturbance regime of the dune
system would constitute an altered geomorphological process, an ecosystem level effect
which could have irreversible consequences for the maintenance of the dune system. For
example, the Spread of European beachgrass (Ammophila arenaria) on the coastal dunes of
California has altered dune topography, replaced native dune species and reduced Species
diversity (Mooney et a1. 1986, Buell et a1. 1995). The introduction of a large and
potentially invasive conifer into a dune system where it represents a novel life form may
alter sand movement patterns and disrupt the disturbance regime of the system. This, in
turn, might have serious consequences for unique dune plants such as the federally
threatened Pitcher’s thistle (Cirsium pitcheri) which requires 70% open sand for
germination and survival (McEachem 1992).

Additionally, invaders may influence dune succession. The endpoint of ecological
succession in dune systems is thought to be influenced by initial conditions (Olson 1958,
T. Poulson personal communication). The presence of a large, non-native evergreen in

early successional stages may have the potential to accelerate the rate of succession or alter

48

its trajectory by increasing nutrient inputs, developing soils, and redistributing water in the

soil column (Ovington 1950, Anderson 1987).

N on-native pines, planted in dense stands, or plantations, are known to alter abiotic
components of terrestrial ecosystems. Clough (1991) found reduced soil pH, Mg, K, and
Ca, and increased Al in Pinus elliotti and Pinus taeda plantations relative to the indigenous
medium-moist-coastal-platform—type forest in the South Cape, South Africa. In addition,
litter under pine canopy was four to five times that of the indigenous forest. In a
comparison of Pinus radiata plantations and native eucalypt forest in SE. Australia, Feller
(1983) documented decreased levels of nitrogen as well as base cations in plantation soils.
The reduced fertility of plantations soils is likely due to the leaching of base cations elicited
by decreased soil pH (Clough 1991). Water distribution in the soil column is also altered
by pine plantations. In a British dune system forested with Pinus nigra ssp. laricia, water
holding capacity was increased under pine cover, but the top layers of the soil were drier
and soils reached the wilting point four months earlier in the growing season than in

unforested dunes (Ovington 1950, Wright 1955).

Biotic effects of pine plantations include the inhibition of understory cover
(Cowling et a1. 1976, Hill and Wallace 1989, Richardson et al. 1989) and a reduction in
species diversity (Cowling et al. 1976, Richardson et a1. 1989, Lugo 1992). In other
systems, however, pine plantations may have the reverse effect. Understory cover and
species diversity were increased in serpentine systems forested with pines (Chiarucci
1995), and species diversity was increased in dune systems forested by pines (Ovington
1950, Hill and Wallace 1989).

The projected effects of an introduced pine on a dune system are numerous and

potentially irreversible. This study seeks to answer the following questions regarding the

49

effects of the pine on three separate components of the dune system.

1) The disturbance regime: How have the pines altered sand movement patterns
relative to areas without pines?

2) Abiotic characteristics: How have light intensity, soil moisture, pH, and soil
development within and around stands been altered by the pine introduction?

3) Community structure: How have the pines altered species diversity,
composition and community structure under their canopy and in the surrounding

habitat?

Three different outcomes of the introduction of P. nigra on the sand dunes of Lake
Michigan are possible:
1) The presence of the introduced pine will facilitate the growth of native species because
of favorable changes in microenvironment under pine canopy. Vegetation cover will be
increased, particularly in blowouts and adjacent to deciduous forest, where cover is sparse
under natural circumstances.
2) The presence of the introduced pine will inhibit the growth of native species because of
unfavorable changes in microenvironment under pine canopy. Vegetation cover will be
suppressed due to reduced light levels and competition with the pines for limiting
resources.
3) The presence of the introduced pine will inhibit the growth of species native to each
dune habitat, but will provide new microenvironments in which species not present in the

adjacent dune habitat can establish.

Field observations suggested that either the second or third predicted outcome was
most likely at SDSP. To determine the effects of P. nigra on the dunes, vegetation

composition and structure and abiotic characteristics were compared under pines and in

50

adjacent areas with similar vegetation in four seral stages.

METHODS

Study site
This investigation took place at Saugatuck Dunes State Park, as described in

Chapter 2.

Biotic efifects of P. nigra

To determine the effects of Austrian pine on dune community composition and
structure, measurements of various plant and community characteristics were made in 1m
X 2m quadrats at l-m intervals along 20m X 2m belt transects established in foredune,
forest edge, wetpanne, and inland blowout sites at SDSP. Most measurements were taken
June-July 1994, with the exception of wetpanne sites, which were sampled in June and
July 1995. Five additional transects were sampled in forest edge and foredune sites in July
1996 to increase the sample size. All transect locations were randomly selected from a

larger set of possible sites with the aid of an aerial photograph.

Eight to ten belt transects (20m x 2m) were established at the edge of Austrian pine
stands in four seral stages. The transects ran perpendicular to the border between native
and pine vegetation (Figure 3-1). Up to 10m of the transect was located under pine canopy
and 10m in the adjacent vegetation. Some variation in transect length was necessary in
foredune and inland blowout Sites, however, as pine stands did not extend the full 10m of
the transect length. Transects always extended 10m into the native vegetation, but were 2-
10m in length under the pine canopy. Data were collected in 2m x 1m quadrats at each l-m

interval along the transect.

51

Austrian pine Bordering
canopy (P) vegetation (N)

\

10m 5m 0m 5m 10m

   

 

Figure 3-1. Diagrammatic representation of 20-m belt transect at the interface between pine and native dune
vegetation. All measurements were taken at l-m intervals. Multiple transects were located in each seral
stage.

To compare P. nigra dune communities with dune communities unaffected by the
introduced pine, transects were also established in the four seral stages in sites lacking P.
nigra (Table 3-1). In foredunes, five transects were located at the interface between stands
of Papulus deltaides and native grasses which occur in the foredune. Transects were also
established in 10 forest edge sites lacking P. nigra. Ten meters of the forest edge transect
were located in inland blowout vegetation abutting the native deciduous forest (analogous
to the portion of the transect under P. nigra canopy in pine sites). The remaining 10111 of
the transect were located under the native deciduous canopy. The border of the forest edge
was located by identifying large native trees which were probably present at the time of the
pine plantings. In wetpannes, 10 transects between native Jack pines (Pinus banksiana)
and bordering wetpanne vegetation were also established. In blowouts, 10 1m x 2m
quadrats were randomly located in 10 sites, because stands of native trees were lacking.

(A map showing locations of all transects was filed with the Michigan Department of

52

Natural Resources in a 1994 Progress Report. The report is available at the Saugatuck
Dunes State Park office in Holland, MI and with the Natural Heritage Program office in
Lansing, MI. Additional location information is available from the author.)

Table 3-1. Vegetation types present in each portion of the transects in sites lacking Austrian pine. P

indicates the portion of the transect analogous to that characterized by the presence of Austrian pine can0py
in Austrian pine sites. N is the adjacent vegetation, similar in sites with and without Austrian pine.

   

Seral Stage P: first 10m of transect N: Adjacent vegetation (last

 

10111 of transect)
Foredune Papulus deltaides canopy Foredune grasses
Forest Edge Blowout vegetation Native deciduous canopy
Wetpanne Pinus banksiana canopy Wetpanne vegetation

Inland Blowout 10 lX2-m random plots measured

 

In each 1m X 2m interval of established belt transects, cover of all vascular plant
species was determined for grasses using ocular estimates and by taking crown area
measurements of each forb species or woody seedling <1 cm dbh. Stem counts of all non-
grass species, and the height and dbh of each woody plant >1cm dbh were also recorded.
Plant species identification was verified using Voss (1972, 1985, 1996).

Disturbance regime: sand movement

In three of the four seral stages, observations of sand accumulation patterns around
stands of Austrian pine were made to determine the effects of the trees on sand movement.
In July and August of 1994, bamboo stakes (approximately 1m tall) were driven into the
soil with 40cm remaining above ground. Net sand movement was measured against these
stakes in September 1995, approximately one year later. Three series of stakes, 1m apart
from each other, were established at the edge of the pine canopy on windward and leeward

edges of pine stands in foredune, forest edge, and inland blowout sites. Wetpannes were

53

not measured because of the lack of open sand surrounding Stands in this seral stage. Each
series consisted of three stakes: one each at canopy edge, 1m, and 5m into the unplanted
region (see Figure 3-2). To compare the effects of the introduced pine and a native tree on
sand movement, a duplicate set of stakes was established adjacent to native cottonwood

stands (Papulus deltaides ).

 

Figure 3-2. Measurement of sand movement patterns around P. nigra and Papulus deltoides stands. Sand
movement in both windward and leeward directions was determined. Each x represents a stake, used to
measure sand movement.

54

Abiotic efiects of P. nigra

In each of the four seral stages, a variety of physical factors were measured that
were regarded as important to indigenous plants and which were likely to be altered by the

presence of Austrian pine, including the solum depth, soil moisture, and light intensity.

Sail prafile.--Soil profiles were examined during the summers of 1995 and 1996
for indication of the depth to which organic matter had been incorporated and soil
development had taken place. Three to six soil profiles were examined at each of two
locations (five meters into canopy and five meters into bordering vegetation: see Figure 3-
1) along each of three to five established transects in sites with and without Austrian pine in
all four seral stages. In Austrian pine, cottonwood foredune, and Jack pine wetpanne sites,
soil profiles were examined at a point on the transect five meters under canopy and at
another point five meters into the bordering native vegetation. In forest edge sites lacking
pines, soil profile samples were taken at five meters into the blowout abutting the

deciduous forest vegetation and five meters into deciduous forest cover.

To examine the soil profile, soil was exposed to a depth of at least 20 cm with a
trowel or a soil core, and horizon designations and depths were determined in each exposed
soil column. Soil samples of the top five cm of the soil column were collected and tested
for physiological pH with an Oakton pH tester kit. Soil samples were stored in a freezer
before testing,which was completed within 13 days of sample collection. Minimum,
median and maximum pH values were calculated and compared within and among seral

stages.

Sail moisture.» Soil moisture content was determined monthly from May - August

1995 for four seral stages. Soil samples were collected to a depth of five cm with a soil

55

core from points five meters into canopy and five meters into bordering vegetation on each
of three of the ten established transects in sites with and without P. nigra in all seral stages .
All soil was collected between 6:30 A.M. and 9 A.M., before significant drying and plant
transpiration had taken place. Soil moisture content was determined gravimetrically. Water

content of soil was compared within and among pine and control sites.

Light intensity.--To determine the effect of P. nigra stands on light levels, light
intensity readings were taken with a Decagon Ceptometer at each one-meter interval of
established 20—meter transects in all seral stages. Measurements were taken on clear days
in July and August 1994 and 1995 between 11:00 A.M.and 2:30 P.M., when the rays of

the sun were overhead.

ANALYSIS

Biotic efl'ects

Understory cover.» Total percent cover of the understory vegetation for each l-m
interval along each transect was calculated from forb and woody seedling crown area
values and percent cover values for grasses. Understory vegetation included all vascular
plants rooted in the plot with stems of less than 1cm dbh. Cover values for each one-meter
interval were averaged across all transects, and compared among sites with and without P.

nigra in each seral stage, and among seral stages.

Species richness and diversity-Species richness per 2m2 plots (1mX2m transect

intervals) and both Simpson (D) and Shannon-Weaver (H’) indices of diversity were

calculated for understory vegetation using percent cover values in l-m intervals on each

56

transect as a measure of abundance. Both diversity indices include a measure of evenness

of species distribution in the community. The calculations for each follow.

D = l/Zpiz, and

H’ = -2piln pi
where pi is the proportion of individuals of Species i in the total sample of

individuals.

Importance values.» The relative importance of each Species in the understory of
the community was assessed for each seral stage and at each two-meter interval of the
transect by summing relative frequency and relative cover values for each Species in each
two-meter interval along transects of the same seral stage (Brower et a1. 1993). Relative
density was not included in the calculation of importance, because stem density of grasses
could not be determined. Species importance values were compared along transects and

changes in community composition were identified in this manner.

Community similarity.»To further determine the changes in Species composition
along the 20-m transect gradient from P. nigra stands into the native vegetation, a

quantitative index of similarity (ISMO) was calculated using mean percent-cover values for

each species in each 2-m interval along transects. The index is a modification of
Sorenson’s similarity index which, in addition to similarity in species composition, also
includes measures of species abundance (cover). The index was first applied by Motyka
(Mueller-Dombois and Ellenberg 1974) and is calculated as follows:

57

1sMo = 2Mw/(MA + MB) * 100

where Mw = sum of the smaller quantitative values of the species common to the
two plots,
MA = percent cover of all species in A, and

MB: percent cover of all Species in B.

Similarity indices compared each two-meter interval on the transect with the first
and last interval of the transect, as the first interval represents the core of pine community
and the last interval represents the core of the neighboring dune community. Similarity

indices were calculated for communities lacking pine in each seral stage as well.

Lifefarm analyses.» Understory species were classified by life form (graminoid,
vine, forb, Shrub, evergreen tree or deciduous tree) and compared across two-meter
intervals of transects. Shrubs were defined as woody species which never reach the
canopy, and included Camus stalontfera, Hudsonia tamentasa, Lanicera sp., Prunus

pumila, Prunus virginiana , Ptelea mfaliata, and Salix sp..

Woody species : stem density and size structure.» Stem density of woody species
zlcm dbh was determined for 5m intervals in each transect and averaged across all ten
transects within seral stage. No sapling or adult-sized woody Species other than P. nigra
occurred in foredune and inland blowout stands with pine. Similarly Papulus deltaides
was the only woody species beyond the seedling stage in foredune cottonwood stands,
and for this reason, graphs of stem densities in these seral stages are not shown. Stem
densities were compared for Sites with and without P. nigra in forest edge and wetpanne

seral stages.

In forest edge sites, two species, Sassafras albidum and Quercus sp. exhibited a

58

range of Size classes across transects in sites with and without Austrian pine, as did P.

banksiana in wetpanne sites. The abundance of each species was determined for six DBH
classes: S = seedlings <lcm DBH, Class 1 = 1-2.4cm, Class 2 = 2.5-4.9cm, Class 3 = 5-
9.9cm, Class 4 = 10-19.9cm, and Class 5 = >20cm, and total counts of individuals in each

size class were compared within species and seral stage.

Disturbance regime: sand movement

Sand accumulation and erosion in control plots and P. nigra stands were compared
within and among seral stages to determine average cm of overall movement, average cm of
gain and loss, and average percent of stakes gaining sand, losing sand, or stable. The
above measurements for windward and leeward stakes were also compared within and
among foredune P. nigra and Papulus deltaides stands. Chi square tests were used to
determine the differences in stakes gaining, losing, and stable, within and among seral

stages.

Abiotic efl’ects

Soil moisture.» Variation in monthly soil moisture data for May - August 1995
within seral stage was analyzed with a repeated measure MANOVA, where the soil
moisture values from each of four months were treated as the y response variables. Site
type (Austrian pine (A) or lacking Austrian pine (C) ) and sample location (under
pine/cottonwood canopy (P), or in bordering native vegetation (N)) and their interaction
were modelled as fixed effects, and site nested within site type was treated as a random
effect. Replication was not sufficient to test 1995 forest edge seral stage data in this

manner.

59

Light intensity.» Light intensity was averaged across each one-meter interval of all
transects of the same type and compared among sites with and without P. nigra within
each seral stage. In addition, Spearman rank conelations between light intensity and
species richness, as well as light intensity and total understory cover were calculated to

determine the association between light and biotic variables.

Solum thicknes .» Soil development was compared within each seral stage using a
two-way ANOVA with site type (Austrian pine (A) or lacking Austrian pine (C) ) as one
factor and sample location (under pine/cottonwood canopy (or in analogous locations
lacking pine in forest edge Sites) (P), or in bordering native vegetation (N)) as the other
factor. The two main effects and their interactions were modelled as fixed effects in the

AN OVA.

RESULTS

Biotic efiects of Austrian pine

Understory caver.»Growth of understory vegetation was inhibited under P. nigra
in all four seral stages (Figures 3-3 and 3-4). Understory cover in pine stands averaged
8% i 1.5 over all seral stages compared to 45% :1; 4.4 in the portion of transects bordering

the pine stands.

In foredunes, the effect of Austrian pine clearly differed from that of cottonwoods
(Figure 3-3a). Understory cover was 6.5 to 40 times higher in the first seven meters of
transect under cottonwood stands than under Austrian pine canopy (refer to Figure 3-1 for

locations on the transect). Understory cover increased rapidly in the final two meters under

60

pine canopy, but was not equal to that of cottonwood Sites until the tenth meter beyond the
pine stand. In contrast, understory cover was not at all inhibited by cottonwood stands; 65

- 82% of the ground was vegetated across the length of transects in cottonwood sites.

Effects of pine canopy on understory cover were least pronounced in the forest
edge seral stage where cover was two to ten times higher in sites without pine than under
pine canopy (Figure 3-3b). Cover ranged from 6 - 24% under pine canopy, versus 28 -
59% in open dunes bordering the native forest canopy. Understory cover declined under
deciduous canopy in sites without pine but increased under deciduous canopy abutting pine
stands. Variance in understory cover under forest edge pine canopy was high due to an
anomalous transect with a dense cover of bracken fern found nowhere else in Austrian pine

forest edge sites.

Pinus nigra stand effect also differed from that of the native pine, P. banksiana, in
the wetpanne habitat (Figure 3-4a). Understory cover was 3-13 times higher under the
native pine stands than under P. nigra canopy. In addition, average cover in the bordering
wetpannes was higher in the Jack pine sites until the seventh meter into the wetpanne,
where Austrian pine wetpanne cover was 20% greater. Cover was very low under pine
canopy in this seral stage, despite the fact that cover exceeded 100% in the portions of the

transect bordering pine stands (due to multiple layers of understory cover) .

Understory cover was lowest under pine canopy in inland blowout sites, where it
ranged from 0.2 - 4% (Figure 3-4b). Cover in the areas bordering pines was only 9-26%,
but pine effect may have extended beyond the 10m sampled, Since randomly selected inland

blowout plots lacking Austrian pine averaged 24 to 45% in vegetation cover.

Species richness and diversity.» Understory Species richness was increased relative

61

to cottonwood sites in Austrian pine foredune sites, but Simpson and Shannon diversity
showed less of a difference between Austrian pine and cottonwood sites in foredunes
(Figure 3-5, Table 3-2). In forest edges, species richness was slightly reduced under pine
canopy relative to open blowout bordering the native forest, but Simpson and Shannon

diversity indices differed little among sites with and without pine (Figure 3-6).

The greatest difference in species richness and diversity within a seral stage was
between Austrian and Jack pine wetpanne sites (Figure 3-7). Richness was two to four
times higher under Jack pine canopy despite the low vegetation cover. Shannon diversity
was 1.5 to 6.7 times higher under Jack than Austrian pine canopy and Simpson diversity
ranged from 0.98 to 2.49 under Austrian pine canopy and 1.94 to 3.89 under Jack pine
canopy. Though species richness in inland blowouts was already low, it was decreased

even further by the presence of P. nigra canopy (Figure 3-8).

Species composition.»

W.» In both cottonwood and Austrian pine foredune sites,
Ammaphila breviligulata was among the five most important species across transects
(Table 3-3). Calamavilfa langifolia , another early successional graminoid, was important
only in Austrian pine sites, however. It became important under pine canopy four meters
from the edge of the Stand, and maintained its position among the top five importance
values to the end of the transect, ten meters beyond the pine border. Quercus sp. occurred
only under pine canopy in foredunes, while Artemesia campestris and 0enathera biennis

were important only beyond the canopy.

In the first 10111 of forest edge sites, the composition of species with the top five
importance values was almost completely different among sites with and without Austrian

pine (Table 3-4). The graminoids Andrapogon scaparius, Ammaphila breviligulata,

62

Table 3-2a. Average diversity indices across transects in foredune and forest edge seral stages. S= Species
richness per 2m2 area, D: Simpson’s index, and H’= Shannon-Weaver index. In foredunes, wetpannes and
blowouts, intervals commence under pine or cottonwood canopy and progress out towards the bordering
vegetation so that interval one is located at the edge of pine/cottonwood stand. In forest edge sites, intervals
commence under pine canopy or in blowout vegetation in sites lacking pines and progress out towards the
bordering deciduous vegetation. Interval one is the final meter before the forest vegetation.

 

Seral Stage Interval 10 9 8 7 6 5 4 3 2 l

 

Foredune Diversity

 

 

 

 

 

 

 

Index

P.nigra S 3.00 5.00 2.33 2.33 2.33 2.40 2.43 2.71 2.38 2.63
D 2.46 3.64 1.49 1.08 1.46 1.28 1.38 1.09 1.11 1.27
H' 0.97 1.42 0.35 0.15 0.44 0.31 0.28 0.27 0.21 0.34

Cottonwood S 1.75 1.25 1.00 1.00 1.00 1.25 1.60 1.80 1.20 1.00
D 1.00 1.00 1.00 1.00 1.00 1.00 1.12 1.23 1.00 1.00
H' 0.00 0.00 0.00 0.00 0.00 0.00 0.13 0.17 0.00 0.00

Forest edge

P.nigra S 2.70 3.10 2.89 2.89 2.67 3.89 2.90 3.50 3.50 3.30
D 1.65 2.06 1.83 1.72 1.67 1.78 1.88 1.97 1.67 1.77
H’ 0.58 0.82 0.72 0.54 0.64 0.69 0.59 0.67 0.55 0.58

Lacking S 3.60 3.50 4.10 4.20 4.10 3.40 2.90 2.90 3.10 3.40

P.nigra D 1.74 1.64 1.94 1.97 2.05 1.72 1.54 1.97 2.02 1.65
H' 0.58 0.49 0.75 0.75 0.83 0.56 0.44 0.72 0.75 0.64

Wetpanne

P. nigra S 1.10 1.70 1.90 1.80 2.20 2.90 3.60 3.70 4.40 5.70
D 0.98 1.15 1.42 1.22 1.49 1.67 2.07 1.73 1.60 2.49
H‘ 0.17 0.15 0.38 0.29 0.48 0.63 0.85 0.56 0.68 0.97

Jack pine S 5.70 5.50 5.30 4.50 5.60 5.70 7.80 7.70 8.60 10.1
D 2.15 2.43 1.94 2.46 2.50 2.67 3.47 3.10 3.50 3.89
H' 1.05 1.00 0.84 1.05 1.06 1.24 1.38 1.34 1.39 1.46

Inland

Blowout

P. nigra S 0.67 0.78 1.50 1.00 1.60 1.60 1.20 1.50 2.50 2.50
D 0.46 0.67 0.97 0.78 1.06 1.09 0.97 0.90 1.64 1.44

H' 0.07 0.20 0.25 0.12 0.36 0.34 0.29 0.45 0.69 0.51

 

63

Table 3-2b. Average diversity indices across transects in wetpanne and inland blowout seral stages. S:
species richness per 2m2 area, D: Simpson’s index, and H’= Shannon-Weaver index. Interval one
commences just beyond pine canopy and the transect progress out ten meters further into the bordering
vegetation.

   

Seral Stage Diversity 1 2 3 4 5 6 7 8 9 10

 

 

 

 

 

 

 

 

Index

Foredune

P.nigra S 2.50 1.63 1.75 1.50 1.86 1.86 1.63 1.75 2.00 1.75
D 1.16 1.05 1.02 1.00 1.15 1.04 1.05 1.15 1.27 1.18
H’ 0.25 0.09 0.05 0.01 0.12 0.09 0.09 0.15 0.26 0.17

Cottonwood S 1.00 1.00 1.20 1.20 1.20 1.20 1.20 1.40 1.20 1.60
D 1.00 1.00 1.02 1.00 1.20 1.00 1.00 1.08 1.00 1.03
H' 0.00 0.00 0.04 0.00 0.14 0.00 0.00 0.09 0.00 0.07

Forest edge

P.nigra S 3.50 2.80 2.70 3.90 3.40 4.10 4.70 5.20 5.20 5.80
D 1.52 1.67 1.61 1.61 2.28 1.90 1.80 1.84 2.47 2.68
H' 0.43 0.51 0.49 0.53 0.64 0.61 0.61 0.68 0.90 1.01

Lacking S 3.40 3.70 3.67 4.44 5.60 5.10 4.90 5.70 4.70 5.20

P.nigra D 1.92 1.91 1.46 1.70 2.57 2.28 2.45 1.86 2.51 2.53
H’ 0.74 0.77 0.49 0.63 1.00 0.94 1.02 0.79 1.03 1.07

Wetpanne

P. nigra S 6.91 7.60 9.10 9.80 9.00 9.60 11.3 11.2 11.1 10.7
D 2.75 2.93 3.41 2.71 3.09 2.91 3.29 2.90 4.70 4.14
H' 1.18 1.23 1.36 1.23 1.32 1.27 1.42 1.36 1.74 1.64

Jack pine S 9.60 9.60 9.90 9.00 8.50 8.80 8.30 9.80 7.60 9.80

3.19 3.46 3.40 2.65 3.07 2.71 3.20 3.68 2.78 3.20

H' 1.41 1.43 1.41 1.01 1.20 1.16 1.20 1.36 1.15 1.34

Inland

Blowout

P. nigra S 2.88 2.56 3.11 3.22 3.22 2.67 2.89 2.33 2.33 2.33
D 1.65 1.95 1.44 1.72 1.72 1.43 1.91 1.26 1.37 1.1 l
H' 0.51 0.64 0.44 0.59 0.46 0.46 0.64 0.29 0.44 0.34

Lacking S 2.50 3.20 3.10 3.40 3.10 3.80 3.30 3.70 4.50 3.60

P. nigra D 1.14 1.90 1.90 1.70 1.37 1.55 1.97 1.69 1.94 1.81

H’ 0.50 0.62 0.79 0.61 0.53 0.68 0.65 0.60 0.69 0.60

 

00.4300; 00:00.0“. 30:00 0022.00 00.0000
0* 0 0 2. 0 n v 0 N F
E55 $5.5
9.9222 E922 wig E’s—2 Hug-22 mag—22 wags—2 mag—22 E2 mtg—22

I; $0002 @682 «$082 $682 E0002 $682
555$ 553$...
0000300200 00:00.0“.
00.0.00; 00:00.0... 30:00 06.: 00:...
0. 0 0 n o m 0 o N F
n=E._>
W30 W80
520:2 220E2 $20.52 s20._.m2 $20.52
204420 204420 204420 204420 204420 204.20 204420
32 $002
@6002 @6002 mg I)? E082 «$0002 «$082
E452 m>42m2
0.79-n. 092d
mam-222 $522 $22 E2 H.922 mag-22 E22 mtg—)2 mtg—22 wig-22
mmdw amdm amdm a a 3
20695 20695 2062:
.35 g
3.952
05.: 00:... 00:00.0“.

23% 02>
20.39.32: 2.52.5
000.50 0200.002
0:500 0002.".

000% ......>

0.50.0 0.050000
03000500 0.00622
0.6.300. 052,050.00
.00 03.00.02

000.040 0200.002
0.0;. 030.2

05.: 000E
2039.320 0.2855
.00 0002.00

000000 00.50020:
.00 000.020

040.050 02000.2

.002 0:0 32 Eod 0.00 00:80»? 260: unto—:09 05 a. .8050 05 L40 once 05 50.: :5. 9:000 000
30:00 00.0 ..0 000300300 000:: 9500.0 4038:. E-~ 0 8000800.. 03:5: :05 400000.: 8.8 9.0.0 08.... 0.5028 c. 00.00% 50:09:. 0.»: 00,—. .m-m 030k

65

Table 3-4. Top five important species in forest edge sites along 20-m transect. Each number represents a
2-m interval, starting under pine canopy or in blowout vegetation and ending 10m into native deciduous
canopy.

66

§§§§§

Efififi

33333333

33

93.59... 0.308..
.8 880

.3 x30

2...an SE8
888.. 85.50.00
8023 88....

...... 528......
«8.9.5.. at...
2259: SS...
8800. 092.888.
.00 0000.00

204.20 204.20 204.20 204420 204.20 0.2.08. 3.30500

05.0902. 0.5.0 30E
53. 2. >850...
or a 2. n
.099.
E 29.6 E .2950
32: 0,2: 0.2:
32: 0.2: 32...
g g g g g

333

E2;
9:502:
gm

3%

532.32.. 0.2855

80.82. 0.20.620
05.0350

30200 0:002.

0.08.000 x25...

67

Panicum virgatum and Calamavilfa langifolia, as well as the dune shrub Hudsonia
tomentosa, were of high importance in sites lacking Austrian pines, while none of these
species appeared in the list of species of high importance in pine sites until the border
between pine and deciduous canopy, when Panicum virgatum attained this status. In
Austrian pine sites, woody species such as Sassafras albidum, Quercus sp., and Acer sp.

were important, both under pine canopy and in the bordering deciduous canopy.

In wetpannes, many of the important species in jack pine sites, such as Rumex
acetosella, Spiraea tamentasa, and two unidentified graminoids, were unimportant in
Austrian pine sites (Table 3-5). In contrast, the shrubs Camus stolonifera and Salix sp.
were important in the wetpannes bordering Austrian pine canopy, but unimportant in Jack
pine sites. Additionally, Solidago sp. was important throughout the transects in Austrian
pine sites, but was present only under pine canopy in Jack pine sites. The flora of Jack and
Austrian pine sites was quite different, and that of Austrian pine sites more closely

resembled blowout vegetation.

Austrian pine canopy in inland blowouts provided a colonization site for the woody
shrubs, Prunus virginiana and Prunus pumila , as well as the small evergreen forb,
Chimaphila maculata which were unimportant in the bordering blowout (Table 3-6).
Andropogon scoparius was equally important under canopy and in the bordering blowout,
but the common blowout species Hudsonia tamentasa, Ammaphila breviligulata, and
Lithospermum canescens were only among the top five importance values in the bordering

blowout.

Liam-m." The pr0portion of species of a particular life form varied across
transects and in sites with and without pine (Figures 3-9 to 3-14). Generally, graminoids

were reduced under pine canopy in all seral stages. In foredune cottonwood sites, only

68

Table 3-5. Top five important species in wetpanne sites along 20-m u'ansect. Each number represents a 2-
m interval, starting under pine canopy and ending 10m into native wetpanne vegetation.

69

«00.0“. 0:30.00
0

a
Edam—REE
00.908

m522u £522.. 6522a £522m

0

0%

E050

Err—5
2895

>028 05.0 0:0...

E
28.2:
3202092020 32330303030330

2008

gaggnfldggggg

Sign.

o w a w
2895 2895 209.75
a 0.0.8 a

92020 32020
APE)
29—3..
«.35 n35 n35
204420

000042

g2

042020
E

204420 204420 204420 204420 204420 204420 204420
1522; «.525.
E2 E2 g2 E2 E2 @622 E42 0:2 a... 0
00022 00022 00022 00022 08942 000042 0: 0

«.5220.

$30.??ng

003 «00.0.... 03.0 0:0.

02.2.10 022.10
099?. 026m
Eng

ESE—25.521295

mad

m522m

n85

£522.. m522m 1522a

00000000000

000000

7O

0..

3.00.03...

204420

05220

:0..0.000> 50305

a

220....4
E2

5.00.03...

204420

05220

N

00220
E .502
204420

09.20
$42.02

0.0.: 00:.0 0:002 0.00303 :. 00.0000 E00008. .008 020 .

00:00 0.0.: 00:.0

v

00022
025.10
2823

0

204420
510

80042
02.310

N

00.00
15220
000022
022.10

920.920.9660

000305 0:08 :. 0.0.: 00:.0

0.0000800 0.008002
0:300:00 80800005...
00.09.32: 0.200882

82.2.0000? 80::000E00
000.580. 0.5003.
.00 800.:00

as... 2002

0:835. 05380.00
0.800 00:20

.00 0000.60

.8309; 800.:00
.00000000 50000.0:2
0.0.0008 0.200820
000000 000.200.:3
00.: 00:.0

0:0.:.9.> 00:20

.000. 80.0 0.00 0000.000.» .0030... 2:. :5. 050:0
0:0 30:00 0:.0 0.0:: 0:000..." 4038:. E-N 0 5:000:08 .0050: 0000 000055 8.3 08.0 08.0 .0030... 0:0.:. :. 00.0000 050.508. 3.. 00... .0-m 0.00...

71

three life forms were present, and all portions of the transects were dominated by
graminoids (Figure 3-9a). In Austrian pine foredune sites, evergreens, shrubs and vines
were also present, though not abundant (Figure 3-9b). Graminoids made up an
increasingly larger portion of the understory cover under pine canopy with increasing
proximity to the interface with foredune vegetation, and then dominated the understory

beyond the canopy.

In the forest edge seral stage, life form distribution differed most among sites with
and without Austrian pine. In sites lacking Austrian pine, graminoids made up the largest
portion of the understory cover until eight meters into the deciduous canopy, where they
were replaced by forbs (Figure 3-10). Additionally, deciduous tree species and shrubs
increased under deciduous canopy. In Austrian pine sites, little change in life form
distribution took place across the transect (Figure 3-11). Deciduous tree species were the
dominant life form throughout, and were exceeded only by graminoids at the junction of
pine and deciduous canopies. Forbs made up approximately ten percent of the understory

throughout, but shrubs increased slightly under deciduous canopy.

In Jack pine wetpanne sites, forbs were much more prevalent than graminoids
under pine canopy, and they decreased closer to the wetpanne (Figure 3-12). In contrast,
graminoids made up the greatest proportion of the understory vegetation in Austrian pine
wetpanne sites. Additionally, in the wetpannes bordering Austrian pine canopy, shrubs

made up a much greater proportion of the vegetation than in Jack pine wetpanne sites
(Figure 3-13).

In inland blowouts, graminoids increased across the transect, while forbs decreased
(Figure 3-14). Understory cover was much more evenly distributed among life forms

under Austrian pine canopy than in the open blowout.

72

W." In foredune cottonwood sites, understory vegetation under
cottonwood canopy differed little from that in the bordering foredune (83-99.8% similarity,
Figure 3-15a). This contrasted sharply with foredune pine sites, where the similarity of
vegetation under pine canopy to that present 8-10 m from the canopy edge was only 5 -
54% (Figure 3-15b). Once beyond the pine canopy, however, similarity to the final 2-m
interval was high (89 - 94%).

In forest edge sites without pine, species similarity with the first interval was more
than 50% until the fourth meter into deciduous canopy (Figure 3-16a). The transition to a
different flora did not occur until the sixth meter into deciduous canopy, where similarity to
the final interval exceeded 30%. This pattern was quite different from that of forest edge
pine sites (Figure 3-16b). The middle four meters of these transects appeared to have a
different flora than either of the extremes, as exemplified by the low similarity of these

intervals to either the first or final interval of the transect.

Measures of similarity in wetpanne sites differed for Jack and Austrian pine sites as
well (Figure 3-17). Species composition under Jack pine canopy was much more similar
to that of the final two meters of the transect (8 - 46%) than under Austrian pine canopy in
Austrian pine sites (2 - 15%). In addition, similarity within Jack pine wetpannes beyond

pine canopy was higher than that within Austrian pine wetpannes (67 - 74% vs. 41 - 61%).

Species composition in inland blowout sites differed dramatically from that under
pine canopy (4-11% similarity to the first interval, Figure 3-18). Additionally, the middle
six meters of the transect was quite different than either of the extremes, suggesting that
this community was populated by different species than either the core of the canopy or the

core of the blowout.

73

Woody species: stem density and size structure--

Wedges." Mean stem density of woody species 21cm dbh was 2.5 to
3.5 times higher under pine canopy than in the comparable blowouts lacking pine, adjacent
to native forest (Table 3-7). In addition, a wider range of species was represented in pine
sites (Figure 3-19, 3-20). In the 5m abutting native deciduous forest vegetation in forest
edge sites where P. nigra was present, five different woody species were represented in
size classes 21cm dbh. Of the three species represented in sites lacking pines, one,
Papulus deltaides, is found only very rarely in mature deciduous forest vegetation. In the
first 5m under deciduous canopy, six species were present in sites with pine and three in
sites without pine. By the final 5m under native deciduous canopy, eight species were

present in size classes beyond the seedling stage in sites with and without pine.

Table 3-7. Mean stem density (# stems/m2) :1: SE (n=10) of all woody species 21cm DBH across four 5-m
intervals in forest edge and wetpanne seral stages. Data from 1994-1996.

 

Transect Interval“
l 2 3 4

Forest Edge 0.14 :I: 0.04 0.19 :1: 0.05 0.16 :1: 0.04 0.41 :l: 0.10
Austrian pine

Forest Edge 0.04 :t 0.02 0.08 :t 0.04 0.13 i 0.05 0.37 :l: 0.10

 

 

Lacking pine

Wetpanne 0.20 :t 0.05 0.18 i 0.07 1.12 :1: 0.05 0.34 :l: 0.13
Austrian pine

Wetpanne 0.65 :l: 0.09 0.50 :1: 0.10 0.13 i 0.05 0.21 :1: 0.09
Jack pine

 

" 1: first 5m under pine or in blowout, 2: 5m abutting deciduous forest or wetpanne, =5m deciduous
forest or wetpanne abutting pine or blowout, 4: last 5m in deciduous forest or wetpanne.

Size structure data for two species, Quercus sp. and Sassafras albidum, indicate
that the seedling component of these populations was 2.5 to 16 times greater in sites where
Austrian pines were present (Table 3-8). The representation of individuals 21cm dbh was

only slightly higher under Austrian pine canopy than in blowouts abutting deciduous

74

forest. In the first five meters of the adjacent deciduous forest, however, the populations
of Quercus in size classes other than seedlings in Austrian pine sites are double those of
sites without Austrian pine. The non-seedling portion of S.albidum populations is only
half that in sites without Austrian pine in the first 5m under deciduous canopy, but these
proportions are reversed in the final 5m of the transect, where 14 individuals are present in

Austrian pine sites as compared to the seven in sites lacking Austrian pine.

Table 3-8. Size structure of two tree species in 5-m intervals across transects in forest edge sites with and
without Austrian pine. Values represent number of individuals in each DBH class. Data from 1994 and
1996.

   

U)

Aust. pine sites ites lacking Aust. pine

 

 

 

 

 

 

l
DBH class“ E DBH class

Transect s 1 2 3 4 :s 1 2 3 4 5
Interval“ :
Quercus sp. :
1 55 1 o o o E 13 1 o o o o
2 55 1 o o 1 :7 2 o o o o
3 59 2 o 1 3 :7 1 2 o o o
4 137 1 2 2 3 E40 1 1 2 5 1
Sassafras :
albidum :
1 10 o o o o :0 o o o o o
2 48 o 1 o 0 i3 0 o o o o
3 121 2 1 o 0 E22 7 o o o o
4 136 3 5 5 1 j59 4 3 o o o

 

*DBH classes: 8: seedling (<1cm DBH), l= 1-2.4cm, 2= 2.5-4.9m, 3= 5-9.9cm, 4= 10-19.9cm,
5= >20cm

“Transect intervals: 1: first 5m under pine or in blowout, 2: 5m abutting deciduous forest. 3=5m
deciduous forest abutting pine or blowout, 4: last 5m in deciduous forest.

mum Jack pine wetpanne sites, the density of woody species
zlcm DBH decreased along the gradient into wetpannes and reached its lowest point in the

first five meters of the wetpanne (0.13 lmz, Table 3-7). The pattern in Austrian pine sites

75

was quite different, however; stem density was very low under pine canopy and increased

nearly tenfold in the first five meters of the wetpanne, due to the high stem density of
Salix sp. (0.9/m2, Figure 3-22). The number of species represented in the shrub/tree layer

was highest (four) in the center of the wetpanne in both Jack and Austrian pine sites

(Figure 3-21, 3-22).

Pinus banksiana seedlings did not occur in the first 5m interval under Jack or
Austrian pine canopy, but increased from that point on into the wetpanne proper in Austrian
pine sites (Table 3-9). Seedling population sizes reached their peak before the wetpanne in
Jack pine sites and decreased in the wetpanne proper. Individuals 21cm dbh were most
prevalent in the Jack pine stand and were reduced to 1/8 that number in the wetpanne

proper. Sapling populations reached their peak in the wetpanne in Austrian pine sites.

Table 39. Size structure of Pinus banksiana across transects in Austrian and Jack pine wetpannes. Values
represent number of individuals in each DBH class. Data from 1995.

     

Jack pine

Aust. pine sites sites

 

 

 

 

:
DBH class“ E DBH class

Transect s* 1 2 3 4 :s 1 2 3 4
Interval“ :
Pbanksiana :
1 o o 1 1 3 i0 15 5 24 12
2 23 2 1 1 o :41 7 13 10 12
3 45 2 5 o :31 2 2 1
4 so 2 4 o 1 E32 1 1 o

 

*DBH classes: S=seedling (<1cm DBH), l= 1-2.4cm dbh, 2: 2.5-4.9cm, 3= 5-9.9cm, 4= 10-19.9cm,

5= >20cm

"Transect intervals: 1: first 5m under pine, 2: 5m abutting wetpanne, 3=5m wetpanne abutting pine, 4:
last 5m in wetpanne.

76

Disturbance regime: sand movement

Average erosion ranged from no erosion in forest edge sites lacking pine to 3.9cm
on the leeward edges of inland blowout P. nigra stands (Table 3-10). Sand accumulation
ranged from 1.8cm in leeward P. deltaides stands in foredunes and in forest edges lacking
pine to 3.6cm at leeward edges of inland blowouts. There was a general trend of accretion
from 1994 to 1995, only leeward inland blowout P. nigra stands had more stakes losing

sand than gaining.

Number of stakes gaining, losing and stable varied across seral stages around P.

nigra stands (X24,231=13°411 p=0.009). Sand movement patterns did not vary within
stage among P. nigra stands and sites without pine (foredune X22,104=0.44, p=0.80;
forest edge X22.86=1.22, p=0.54; inland blowout X22,58=1.66, p=0.44), nor did they

vary among P. nigra and Papulus deltaides stands in foredunes (X22.130=2.82,
p=0.245).

Furthermore, leeward to windward sand movement patterns did not vary in
foredune and forest edge pine stands, but proportionally more stakes were gaining sand on

windward sides of both foredune Papulus deltaides and P. nigra stands in inland blowouts

(x22 32=5.31, p=0.005; x22 48=7.56, p=0.02).

Abiotic effects of P. nigra

Light intensity.-- Light intensity was greatly reduced under Austrian pine canopy
cover in all four seral stages (Figures 3-23, 3-24). In foredune stands, light intensities
were 1.5 to six times higher under cottonwood canopy than under pine, but were the same

in the bordering foredune vegetation (Figures 3-23a). Light intensity under cottonwood

77

cover was 25 to 91% that of mean values in neighboring vegetation, however.

In forest edge transects with pine, light intensity was consistently low under both

pine and deciduous canOpy and ranged from only 14 to 123 microeinsteins*m-2"‘s'l

Table 3-10. Sand movement around pine stands in three seral stages during the period of 1994-1995.

 

 

 

 

 

 

 

Site Type Stake Avg cm Avg cm Avg cm % % % 11
Location overall loss gain Losing Stable Gaining

Foredune
No Pine 1.3 -1.0 1.9 10.0 30.0 60.0 10
P. nigra leewmd 1.9 -15 2.9 16.7 27.1 56.3 48
P. nigra windward 1.4 -1.6 2.1 18.0 34.0 48.0 50 -
Papulus leeward 1.2 -l.6 1.8 22.2 33.3 44.4 18
deltaides
Populus windward 2.4 -2.3 2.4 16.7 0.0 83.3 18
deltaides
Forest Edge
No Pine 1.2 0.0 1.8 9.3 14.0 76.7 3
P. nigra leeward 2.0 - l .5 2.4 20.5 20.5 59.1 43
P. nigra windward 2.3 -1.4 3.3 0.0 33.3 66.7 44
Inland Blowout
No Pine 2.0 -2.0 2.1 18.5 25.9 55.6 10
P. nigra leeward 2.8 -3.9 3.6 52.0 8.0 40.0 . 27
P. nigra windward 2.5 -2.9 2.4 20.0 10.0 70.0 25

 

(Figure 3-23b). In forest edge sites lacking pine, however, light intensities gradually
decreased with proximity to the deciduous canopy (1670 to 88 microeinsteins*m-2*s'l),

but were still seven times higher than in P. nigra forest edge transects at five meters into the
deciduous forest vegetation. Light intensity was up to 58 times higher in sites without

Austrian pine than in sites with pines in the forest edge seral stage.

78

Light intensity was reduced under both Jack pine and Austrian pine canopy in
wetpanne sites, but was still 1.2 to 4.8 times higher under Jack pines (Figure 3-24a). In
addition, light intensities in Austrian pine sites did not equal those in Jack pine sites until

eight meters from the edge of pine canopy.

In inland blowout sites (Figure 3-24b), light intensity was very low under pine
canopy, but increased 20-fold by the third meter from the canopy edge.

W.--Soil development was four and 11 times greater under pine
canopy (P) in foredunes and blowouts, respectively, than it was in the bordering native
vegetation (N) (Table 3-11). Solum thickness under cottonwood (7.3 3; 2.1 cm) also
exceeded that of the area bordering the stand (4.3 1; 1.0 cm). In wetpanne and forest edge
sites, however, solum thickness did not differ significantly between Austrian pine canopy

and bordering vegetation.

In comparing sites with and without Austrian pine within seral stages, only forest
edge sites differed in solum thickness (Table 3-11). Soil was far less developed in sites
lacking pine than in sites where pines had been planted, and a significant interaction term

indicates that P and N solum thickness varied among sites with and without Austrian pine.

Sam." The highest pH of the top 5 cm of soil was found in the foredune
sites under pine and cottonwood stands with pH up to 7.8 (Table 3-12). The lowest pH
values were found in wetpanne sites, under both Austrian and Jack pine canopy (4.3 and
3.9, respectively). In addition, pH values of 4.2 were seen under deciduous canopy
bordering Austrian pine and under Austrian pine canopy in inland blowout sites. Minimum
pH is lower under pine canopy than under bordering vegetation in foredune, wetpanne, and

inland blowout seral stages. In forest edge sites, however, minimum pH is lower under

79

Table 3-11. Solum thickness in four seral stages under canopy and in bordering vegetation. Significance of
main effects and interactions in two-way ANOVA for each within-seral stage comparison is shown. Data
from 19954996.

 

 

 

Bordering Pine, Two-Way-
Vegetation (N) Cottonwood, or ANOVA
Lacking Pine Significance
(P)

Seral Stage and Mean Solum 11 Mean Solum r1 (A/C)* (P ** A/C* PIN
Stand Type Thickness (cm) Thickness (cm)
(NC) 1; s.e. i s.e.
(A) Foredune 1.8 i 0.5 5 7.6 i 1.9 5 NS 0.01 NS
Austrian Pine
(C) Foredune 4.3 i 1.0 4 7.3 i 2.1 4
Cottonwood
(A) Forest Edge 8.0 i 1.3 5 8.7 i 0.5 5 0.001 NS 0.055
Austrian Pine
(C)ForestEdge 5811.5 5 1.61 1.1 5
Lacking
Austrian Pine
(A) Wetpanne 8.1 i 0.2 3 9.9 i 2.7 3 NS NS NS
Austrian Pine
(C) Wetpanne 12.1 1 2.1 3 8.6 i 3.2 3
Jack Pine
(A) Inland 0.6 i 0.2 4 6.8 1 0.7 4 0.00002
Blowout
Austrian Pine

 

* A=Austrian pine sites, C= sites without Austrian pine
** P=Under pine or cottonwood canopy, or open blowout bordering native deciduous forest, N=in native
vegetation bordering Austrian pines, cottonwoods, or native deciduous forest.

80

Table 3-12. Minimum, median, and maximum pH in four seral stages. Data from 1995.

 

 

Bordering 5 Pine, Cottonwood,

Vegetation E or Lacking Pine

(N)** E (1))"
Seral Stage and Min Median Max n Min Median Max 11
Stand Type (A/C)‘ pH pH pH 5 pH pH pH
ForeduneAustrian 6.0 6.9 7.4 9 5.4 6.8 7.8 9
Pine 5
Foredune 6.4 6.8 7.1 5 E 6.9 7.6 7.8 6
Cottonwood 5
ForestEdge 4.2 6.1 6.9 9 4.5 5.2 6.2 9
Austrian Pine 5
ForestEdge 5.0 5.4 7.3 9 4.6 5.8 7.5 9
Lacking Pine 5
WetpanneAustrian 4.9 6.2 6.6 9 4.3 5.1 7.1 9
Pine 5
Wetpannelack 4.9 5.7 6.6 9 3.9 4.9 6.4 9
Pine 5
Inland Blowout 4.9 5.4 6.0 9 4.2 5.0 6.4 9
Austrian Pine g

 

" A=Austrian pine sites, C= sites without Austrian pine
*"' P=Under pine or cottonwood canopy, or open blowout bordering native deciduous forest, N=in native
vegetation bordering Austrian pines, cottonwoods, or native deciduous forest.

81

the bordering deciduous vegetation (4.2) than under the pine canopy (4.5).

Soil moisture.-Repeated-measure MANOVA showed a significant site type effect
in foredunes and wetpannes (Tables 3-13 to 3-15). Pinus nigra sites were drier than P.
banksiana sites in wetpannes, while P. nigra sites had higher soil moisture than Papulus
deltaides sites in foredunes (Figures 3-25 to 3-26). Soil moisture was higher under canopy
than in the bordering vegetation in foredune and inland blowout sites, but the reverse was
true in wetpanne sites. Time effect was significant only in foredunes, where soil moisture

decreased as the growing season progressed.

Biotic correlations with light intensity. -- Understory cover was highly positively
correlated with light intensity in wetpannes,‘inland blowouts and in foredune Austrian pine
sites (rho: 0.53 to 0.68; Table 3-16), but was not correlated with light intensity in
foredune cottonwood sites. Within seral stages, correlations between understory cover and

light intensity were stronger for sites with Austrian pine than those without.

The association between two measures of species diversity (species richness and
Simpson’s index) and light intensity were not as strong as those between light intensity
and understory cover. Species diversity decreased with increasing light intensity in both
forest edge site types and in foredune Austrian pine sites, but was positively correlated with

light intensity in all other circumstances.

82

Table 3-13. Repeated measure MANOVA for foredune soil moisture across the 1995 growing season.

 

 

 

 

 

Lambda Num Den
Between Subjects
A/C* 0.313 13.1468 1 6 0.0110
PIN *" 0.131 39.5245 1 6 0.0008
AlC*P/N 0.524 5.4448 1 6 0.0584
Site(A/C) 0.318 2.1492 6 6 0.1870
All Between 0.066 9.5065 9 6 0.0063
Intercept 0.045 126.5514 1 6 <.0001
Within Subjects
Time 0.071 17.3738 3 4 0.0093
Time‘A/C 0.238 4.268 1 3 4 0.0974
Time‘P/N 0.489 1.3936 3 4 0.3667
Time‘A/C‘P/N 0.371 2.2607 3 4 0.2235
Time‘Site(A/C) 0. 1 16 0.7464 18 1 1.799 0.7206
All Within 0.018 1.3460 27 12.324 0.2980
Interactions

 

" A=Austrian pine sites, C= cottonwood sites
“ kUnder pine or cottonwood canopy, =in native vegetation bordering Austrian pines or cottonwoods.

83

Table 3-14. Repeated measure MANOVA for wetpanne soil moisture across the 1995 growing season.

 

Exact F DF DF

 

 

 

 

Source Wilks Prob>F
Lambda Num Den

Between Subjects

All Between 0.104 4.9130 7 4 0.0715

Intercept 0.078 47.201 1 l 4 0.0024

A/C“ 0.275 10.5204 1 4 0.0316

PIN " 0.224 13.8291 1 4 0.0205

AIC‘PIN 0.430 5.2993 1 4 0.0828

Site(A/C) 0.458 1.1856 4 4 0.4365

Within Subjects

All Within 0.013 1.0681 21 6.293 0.5068

Interactions

Time 0.183 2.9770 3 2 0.2615

Time‘A/C 0.243 2.0743 3 2 0.3417

Time‘P/N 0.244 2.0665 3 2 0.3426

Time*A/C*P/N 0.226 2.2877 3 2 0.3186

Time‘Site(A/C) 0. 120 0.5704 12 5.583 0.8059

 

"' A=Austrian pine sites, C= sites without Austrian pine

" P=Under pine canopy, N=in native vegetation bordering pines.

84

Table 3-15. Repeated measure MANOVA for soil moisture in inland blowout sites across the 1995
growing season.

 
      

 

Source Wilks Exact F DF DF Den Prob>F

 

 

 

 

Lambda Num
Between Subjects
All Between 0.2470928 12.1883 1 4 0.0251
Intercept 0.0476379 79.9668 1 4 0.0009
PIN " 0.2470928 12.1883 1 4 0.0251
Within Subjects
All Within 0.1647099 3.3809 3 2 0.2366
Interactions
Time 0.3448 1.2668 3 2 0.4697
Time‘P/N 0.1647099 3.3809 3 2 0.2366

 

* P=Under pine canopy, N=in native vegetation bordering Austrian pines.

Table 3-16. Correlations between light intensity and understory cover, species richness/2m2, and
Simpson’s diversity index in four seral stages. Values are Spearman’s rho. Data from 1994-1996.

 

Seral Stage and Undastory Richness Simpson’s n

 

StandType rho p rho p rho p
Foredune 0.53 *"* -0.30 " -0.23 * 94
Austrian Pine

Foredune -0.24 NS 0.33 * 0.33 * 40
Cottonwood

ForestEdge 0.34 ** -0.19 NS -0.33 " 60
Austrian Pine

ForestEdge 0.25 ** -0.31 ““ -0.25 “ 118
Lacking Pine

Wetpanne 0.68 *"* 0.53 "** 0.442 "“ 200
Austrian Pine

Wetpanne 0.57 *"'" 0.39 *""‘ 0.24 *** 200
Jack Pine

Inland Blowout 0.65 "** 0.54 "“ 0.52 ““ 89
Austrian Pine

 

* 0.01< p50.05, " 0.001 < p500], *“ 0.0001 < p50.001, *"" p50.0001

D I S C U S S I O N
Geomorphological efi’ects: Disturbance regime

The P. nigra trees were planted in 1956—1972 at SDSP to stabilize the shifting
sand. According to the results presented here, however, sand movement patterns did not
vary in areas with Austrian pines compared to areas without pines within each seral stage.
The sites sampled showed slight overall accumulation, but the presence of the pines did not
appear to significantly alter the sand movement patterns of the dune system. Buckley
(1987) found a reduction in sand transport rate with increasing plant cover and with

increasing k, as described by the following equation:

q = B [V(1-kC)-Vt]3

where q is sand transport rate, B is Bagnold’s constant, V is wind velocity, k is a constant

dependent upon plant geometry, C is percent plant cover, and V1 is threshold wind

velocity. Though the constant for the geometry of the pines (k) would be expected to be
larger than that of typical dune plants in foredune, inland blowout and forest edge sites,
vegetation cover tends to be reduced around the pine stands, and this may account for the
lack of difference in sand movement patterns in areas with and without pine. It is important
to consider the changes in sand movement patterns over the life span of the trees as well,
because k and C will change as the pines grow taller and as vegetation surrounding them is
gradually reduced. In addition, sand movement patterns may vary across years due to
differences in climatic conditions and wind speeds, and the time period measured in this
study may not necessarily be representative of those in the past. This information provides

an argument against using Austrian pines as stabilizers of the dune system.

85

86

Population and community level efi'ects

F oredunes.--The inhibition of understory cover under Austrian pine canopy in
foredunes indicates that the changes in microenvironment brought about by the pines were
unfavorable to the vegetation that normally occupies this dune habitat (Figures 3-3 and 3-
4). The effect of pine upon community structure extended beyond the pine canopy since
understory cover was inhibited relative to that of cottonwoods up to ten meters from the
pine stand. At first glance, this inhibition of understory cover directly under Austrian pine
canopy might be attributed to the 32-84% decrease in light intensity (Figure 3-23). The
high correlation between light intensity and understory cover in pine sites and the absence
of the same correlation in cottonwood sites in the foredunes, however, might be interpreted
to suggest that a factor or suite of factors other than light (but correlated with light in
Austrian pine sites) is responsible for reduced understory cover. It is also possible that a
minimum level of light intensity is required by the understory vegetation. Because the light
intensity under cottonwoods is 1.5-6.2 times higher than under pines, the grasses below
cottonwoods may be receiving adequate light for growth even at the reduced levels relative
to the light levels received by the vegetation beyond the canopy. Light intensity under pine
canopy, however, may be too low to support the maintenance and growth of the
understory. In addition, species that occur beneath the pine canopy are always in deep
shade, whereas the deciduous habit of the cottonwoods would allow for full light

intensities in the early growing season when grasses are first emerging.

Though establishment of typical foredune species was inhibited by the presence of
Austrian pines, there was an increase in species richness in pine sites (Figure 3-5, Table 3-
2). This suggests that the pines provided a microenvironment conducive to colonization by

non-foredune species. Species that were normally competitively excluded by the dominant

87

Ammaphila breviligulata, or limited by low soil moisture in typical foredune sites were then
able to colonize the pine understory. The altered species composition under pines included
a higher proportion of deciduous trees and shrubs absent from the cottonwood sites. This
might suggest that the pines will accelerate the rate of ecological succession to forest. None
of these young woody individuals had developed into saplings, however, perhaps because
they had only just recently been dispersed to the site. More likely, they could not compete
with the dominant pines beyond the seedling stage. If the latter is the case, the plant
communities under the pines in foredunes will continue to be impoverished until the pines

are removed.

Species composition beyond the pine canopy was also altered; Calamavilfa
langifolia, a foredune species less tolerant of burial (Olson 1958), was of high importance
under the pines and up to 10m beyond. This species did not occur in or around
cottonwood stands. This suggests that sand movement was decreased by the presence of
the pines and had encouraged the colonization of a species representative of a later stage of
succession. The presence of buried A horizons in soil under the cottonwood canopy but
not under pines also suggests a stabilizing effect of the pines across time, despite data to the

contrary in 1994-1995 sand movement studies (Table 3-10).

Forest Edge.-- While the growth of understory cover did not increase under
Austrian pine canopy in forest edge sites, it was only slightly suppressed relative to growth
under native deciduous forest canopy (Figure 3-3b). Species richness and diversity were
only slightly reduced under pine canopy relative to the deciduous forest as well (Table 3-2).
The similarity of abiotic conditions under pines to those of the bordering deciduous forest
may account for the similarity in understory cover and species diversity across transects;
soils in both understories were developed to 8cm, soil pH was similar, though slightly

lower in deciduous forest, and light intensities did not differ (Tables 3—11 and 3—12, Figure

88

3-23).

The effect of Austrian pine on plant community composition was least dramatic in
forest edge sites. The relatively high similarity in species composition and life forms
across most of the transect, was most likely a reflection of the similar abiotic conditions.
The only place where the biotic composition varied was where the Austrian pine canopy
met the forest canopy (Figure 3-16). Here, an increase in graminoids and a reduction in
deciduous tree species was most likely due to an increase in light intensity where the two

canopies met.

The Austrian pines appear to have created a microenvironment similar to that under
deciduous canopy further into the blowouts, and tree and shrub species from the native
forest have begun to colonize and establish under pine cover. Tree and shrub densities
(Figure 3-7) indicate that native deciduous forest species were dispersing to and
establishing under pine canopy. The size structure of these individuals (Table 3-8)
suggests that Quercus sp. and Sassafras albidum were beginning to recruit into the sapling
size class as well. Sassafras albidum is a clonal tree, which relies heavily upon vegetative
reproduction to invade new areas (Bosela 1995). The root connection to the parent plant
gives the new shoot the carbohydrate reserves it needs to become established under the
shade of the pine canopy. In addition, this species has large leaves which may allow it to
collect incident radiation at very low light levels (Bazzaz et a1. 1972, Barnes and Wagner
1981). Quercus is a large-seeded species which benefits from extra food reserves present
in the seed in early germination and establishment. Additionally, the frequent mammalian
dispersal of acorns may allow for establishment further from the parent, as reflected by the
large number of seedlings present 10m distant from the native deciduous trees. Both of

these characteristics may aid its entry into the understory of the introduced pine.

89

In sharp contrast, microenvironmental conditions and species composition of native
blowout vegetation in forest edge sites lacking pine were quite different from the
neighboring deciduous forest. Higher light intensity and decreased soil development may
be responsible for the increased understory cover and greater prevalence of graminoids in
these sites (Figures 3-3b, 3-10). Quercus and S. albidum seedlings were present, but in
much lower numbers than in pine sites (Table 3-8). Population age structure of these
species indicated that seedling recruitment was very low, and that recruitment into the
sapling stage did not occur except under deciduous canopy. This suggests that either seeds
were not dispersed to these sites or that predation and/or environmental conditions limited

their germination and establishment.

The contrast between sites with and without pine provides evidence that Austrian
pine plantings at the forest edge act to change the pattern of succession in areas that would

otherwise be vegetated by blowout species.

Wetpannes.-- The wetpanne habitat provided an opportunity to compare the effects
of the introduced pine with those of a native pine on abiotic and biotic components of the
wetpanne. Both pines inhibited the growth of understory vegetation relative to the
bordering wetpanne, but species richness and cover under Austrian pine canopy were much
lower than under Jack pine (Figure 3—7,Table 3-2). Higher light intensities under Jack pine
may be due to the narrower crowns of this species and to the age structure of the Jack
pines; they circle the wetpanne in different heights with the shortest trees at the edge of the
wetpanne and the tallest trees deeper into the stand. Both understory cover and richness are
strongly positively correlated with light intensity (Table 3-16) indicating that higher light
intensities may be responsible for the higher cover and diversity of the native stands . The
Jack pine wetpannes occurred further from the beach than the Austrian pine wetpannes,

however, which might suggest that they have had a longer time in which to develop a more

9O

diverse plant community.

Austrian pine sites were significantly drier than Jack pine sites (Table 3-25), and
this may have contributed to the differences in the understories of the two site types. The
shrubs Camus stolonifera and Salix sp. made up a much larger component of wetpanne
vegetation in Austrian pine sites (Figure 3-22), and more upland species such as Solidago
sp. were more prevalent under the canopy and in the adjacent wetpanne than in Jack pine
sites (Table 3-5). Though the differences in soil moisture and vegetation suggest that the
Austrian pines are drying down the wetpannes, the initial conditions of these wetpannes is
not known. It is possible that the sites where Austrian pines were planted were always
drier than the Jack pine sites sampled. In studies of the effects of forestation on British
dune systems, it was found that over a period of twenty to forty years, the water table was
lowered in two dune systems planted with P. nigra var. maritima (Ovington 1950, Hill and
Wallace 1989). It was unclear, however, whether this was a result of increased

transpiration by the pines or of other factors.

The continued drying of the wetpannes may mean the successional replacement of
unique dune communities as shrubs continue to encroach and the character of the sites
changes. The wetpanne communities surrounded by Austrian pine are in a later stage of
succession than those surrounded by Jack pine, but it is unclear if this is simply a response
to initial conditions or a response to increased transpiration and a lowering of the water

table by the introduced species.

Inland blowouts.-- As in other seral stages, understory cover and species richness
were depressed under P. nigra canopy in inland blowouts, but the changes in microclimate
encouraged colonization by non-blowout species, as in foredunes (Figures 3-4 to 3-8,

Tables 36). Two- to 32-fold differences in light intensity were likely responsible for the

91

decrease in cover (see Table 3-16 for correlation), while increased soil moisture under pine
canopy might have encouraged the germination and colonization by the increased woody
component of the vegetation (Figure 3-14). Woody species seed sources are quite a
distance away, but the majority of the woody species present (Prunus pumila, Prunus
virginiana) are bird-dispersed. The increased number of perching sites in pine stands
relative to open dune habitat may account for increased dispersal to these sites. As in
foredunes, the woody species that colonize the site under pine canopy have not recruited
into the sapling class. Time lags in dispersal and/or competition with Austrian pines for
moisture, nutrients, and light during establishment may limit recruitment into the sapling
class. If the latter is the case, as is also possible in foredune sites, cover and species

diversity will remain low until the pines are removed.

Assessment of Austrian pine efl'ects

The actual outcome of Austrian pine introduction on each seral stage is summarized
in Table 3-17. The growth of native species is facilitated only in forest edge sites, where

blowout species are lost from the community and forest species successfully replace them.

Inhibition of native vegetation. «The presence of Austrian pines inhibited the
growth of native vegetation in all four dune habitats (Table 3-17). Pine introductions have
often been associated with a reduction in understory cover (Cowling et a1. 1976, Hill and
Wallace 1989, Richardson et a1. 1989). Lugo (1992) found a 50% reduction in understory
biomass in tropical pine plantations (Pinus caribaea) relative to secondary forests of similar
age in Puerto Rico. Other introduced species have had similar effects. In eastern
deciduous forests, Woods (1993) attributed a reduction in understory cover associated
with an invasive shrub, Lanicera tatarica, to seasonal competition for light. At this site,

evergreen herbaceous and Vining or sprawling species were more abundant beneath the

92

invader, suggesting that they were more tolerant of reduced light.

Enrichment or impoverishment of soils by conifers may also influence vegetation
cover. In a serpentine system with poor soils, Chiarucci (1995) found increased
understory cover under pine canopy accompanied an increase in organic matter content and
an improved microenvironment for plant establishment. Generally, however, forestation
with conifers alters soils by decreasing soil pH and levels of base cations (Ca, K, Mg),
increasing exchangeable aluminum, and altering organic matter content, which may have
negative consequences for the native vegetation in these sites (Ovington 1950, Wright
1958, Hill and Wallace 1989, Clough 1991, Chiarucci 1995).

Table 3-17. Actual outcome of P. nigra introduction and projected future trajectory of native dune
populations in four seral stages at SDSP.

 

 

 

Potential Foredune Wetpanne Forest Edge Inland

Outcome Blowout

1. Facilitate native No No Native woody forest No

species. species successful

2. Inhibit native Understory Drying: Loss of Blowout species Loss of cover

species. growth suppressed cover and species (graminoids) reduced, and species

diversity cover slightly lower diversity

3. New colonists Unsuccessful: Successful: Successful: Unsuccessful:

in new habitat! colonize but do Vigorous shrub Deciduous tree colonize but do
not grow growth in wetpanne growth under pine not grow

canopy

Projected trajectory Depauperate w/o Loss of unique Facilitation of forest Depauperate w/o

gap formation wetpanne flora species gap formation

 

* Includes native woody and other species not present in the dune communities adjacent to P. nigra stands.

Reduced species diversity.-- At SDSP, Austrian pines reduced species diversity in
all dune habitats but the foredunes, where it was slightly increased. Species diversity has
been shown to both decrease (Cowling et a1. 1976, Richardson et a1. 1989) and increase
(Ovington 1950, Hill and Wallace 1989, Chiarucci 1995) under pine canopy, depending

93

upon the initial conditions. In Great Britain, the new species entering a dune system
forested with pines included those common in waysides and woodlands (Hill and Wallace
1989). Reduced light intensity may be associated with the decrease in diversity; species
richness was negatively correlated with overstory cover in sclerophyllous communities in
Australia (Sprecht and Sprecht 1989). Increased diversity under pine canopy may reflect a
release from competition in communities dominated by one or several vigorous species
disadvantaged by pine cover (Ovington 1950), or as a result of enrichment of poor soils

(Chiarucci 1995).

Austrian pine and succession. --At SDSP, the presence of Austrian pines was
associated with colonization by species not present in the adjacent dune communities (Table
3-17). In foredunes and blowouts, the increased woody component of the understory
vegetation will not result in the eventual overtopping of the Austrian pines by native woody
species because these individuals never transitioned beyond the seedling stage. Their
presence in the seedling pool was not interpreted as an acceleration of ecological
succession. The presence of Austrian pine at the edge of native deciduous forest, however,
appeared to facilitate an invasion by native deciduous species which may eventually replace
P. nigra as dominants in the community. In this dune habitat, the introduced pine
accelerated the rate of succession. Austrian pines may have been responsible for the drying
of wetpannes and the acceleration of succession in this seral stage, as represented by the

increased shrub component in vegetation bordering the pines.

Conclusions

It is clear that the introduction of Austrian pine into the dune system has had serious

consequences for community structure. The magnitude and direction of Austrian pine

effects varies by seral stage, however, and this is a response to the interaction of the pines

94

with the initial conditions of each dune habitat. Further experimental work will be
necessary to elucidate the factors controlling these differences in the outcome of invasion,
though relative development of soil and proximity to seed source may play an important
role. The results of this study suggest that precautions must be taken in assuming
uniformity of the effects of an invasive species on community structure. The consequence

of invasion are likely to vary with initial conditions.

95

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a) Understory Cover In Foredune Sites
Under Tree Canopy Bordering Foredune Vegetation
.. i -r
%iJEI %{ ,Eiffi i ' 1
5 7° 1
>
o 60 4.. l
0 I
.. 50 ..
g 40 a :
3 30 r: l —O—Austrian Pine
“' 20 -_ '
l - -I-- -Cottonwood
1 0 it _ |
o-h-T.:::'l.al+filalli
O to to v or 1- co in h a:
Interval (m)
b) Understory Cover In Forest Edge Sites
8 0 Pine Canopy Deciduous Canopy
or '
I
70 Blowout Vegetation : I Austrian
a. 60 ' pine
g e ' --l---Lacking pine
8 50 l \ I
I
E 40 ' e I ‘ i
2 30 i \ J' o o
a i u E
20 | 0 ' QE’
10 , I ’
o . r . 4 l 4 + l : llfl re 4 4 4
0 co to st or v- co in h a:
Interval (In)

 

 

 

 

Figure 3-3. Understory cover 1 SE in a) foredune sites with Austrian pine or Cottonwood and b) forest
edge sites with and without Austrian pine. Data from 1994 and 1996.

96

 

 

 

 

 

 

a) Understory Cover in Wetpanne Sites
180 Under Pine Canopy I Bordering Wetpanne Vegetatior
1 60 :
.. 140 I
e l
5 120 : i i .
100 ' - . I
I I - o
'5 I . i E. I . i. ..
o 8 0 .i r
2
o 6 0 E I
a.
4 O 1' ' +Austrian pine
20 ' --l---Jack inc
0 - Xi ft—{“ 1 i l 1 l p
0 cc to v N v- an ID ix a:

Interval (m)

 

 

 

b) Understory Cover In Inland Blowout Sites

50 r Under Pine Canopy Bordering Blowout Vegetation
45*
40.4.-
35‘
30‘
25‘
20‘
15*
10‘
5+
0*“ wvrrwiliiiiii%i

l
r

‘— 0') I!) IN a) to
F

T I I 1

Percent Cover

T

 

10
8
6
4
2

Interval (m)

 

 

 

Figure 3-4. Understory cover :1: SE in a) wetpanne sites with Jack or Austrian pine (1995 data)
and in b) inland blowout sites with Austrian pine (1994 data).

doa— 28 32 ES... 85 .mofin 5584‘ be 383338 55 «3% 2.2.28 5 30855 5.3 828 Mmfi «finance—fit 860% can: .n-m «Emma—

 

 

:5 3:8...

I? I?
0 6 8 I. 9 9 V 8 Z I i 3 8 v 9 9 I. 8 6 0
v n a. 4 n n . n t. n n .4

 

 

 

 

 

 

n n n n n n _ n c
_
_ 4 m6
_
m ._ +lmr . .. hiin’lm W\I|.al.llw i _. m
I ._ _ _ I.-
. x w . _ / . \ L. m; m
_ a 1 #
_ , i N 8
_ 4 d
uoozcozoo-ln.l _ .. m.~ «
65a 5322+ M
_
_ m z
_ w
u .e ma 2
r
e
_ e
_ 11 v
cozfiomo> 0:320“. actoEom " 3280 we; 895 l mé
_
_ -. m

 

«26 2533... :— mmeccoE «0.03m

 

 

63— e5 32 Eat San 6:3 533< 5055 e5 53 v.83 02.0 628 5 80855 8.8 9.28 Mmfi «Bigot 360% :32 .04” 0.5mm

 

:5 .95....

 

 

 

 

 

 

 

O 6 8 I. 9 9 .7 8 7v Ir Ir 74 8 .7 C. 9 I. 8 6 O
annnnirn+n_++nennln+o
_
_
_ T _.
_
_ W
_ N u
_ -. u
_ I"
n t. I. l a n
on... uc_xoe._lll.l -_ a _ a
6:3 chma<+ \ lm\ _ ...w I \ .. M.
\ n ..n v .6
. w
a _ Z
\ I \\ n y m B
J
I
z x _ u
.. _
_ r o
" cozflouo> 30266
3050 maoagooo c_ 8 >350 SE 695

 

megm emem “mono". E mmecgoE «£0on

 

 

.32 Bob 3.5 6.2.3 5523 .5 x02. 55 use 25903 5 909.55 8.8 $88 mmfl ~E§Eot 860% :32 gum 853m

 

:5 .25....

 

 

 

 

 

 

m 6 8 I. 9 9 .7 8 7n l. Ir a 8 .7 9 9 I. 8 6 lop
p p r r _ h r .r Ll L p + a? + L. + + i » b o
q q Jl 4 q a q q n a _i«
_
_
_ N
_
_ m
u . n
_ \ / .9
_ 1 o a
9.3 xomwlllul _ \x r u
we... cetfia<+ _ m.
\ z _ \ . m m
x , x _ m
. \ .r x I: u
_ e
_ I
_
_ Y N w
_
macaque; octoecom _ >3ch 62E .695 e v F

 

$55 053925 E «moccoE 3.03m

 

 

100

.48. 52.. see .8... 5.522. 5.3 Bee... 3.... .836... .65.... e. 335.. 52 see. man «538%.. 8.8% as: ...-m use".

 

 

.5 .25....

 

 

l l
O 6 8 I. 9 C.- .7 8 79 It Ir 3 8 .7 C.- 9 I. 8 6 O
. . . . + . ti 4 r . e+ + 1+ . 4 i . 4. n _ o
_
n m...
_ W
_ ... F e
_ w
_
. ...
_ .m P
_ 8
_ w
_ w
_ T w e
. N
_ m.~ m.
_ Z
3
_ a m M
_ B
_
u m.»
_
_ - 4
co:2oao> 50265 acteEOm .8980 SE .625

62...... 50265 2.2:. E «mezzo—m eoBenw

 

 

101

 

a) Percent of Total Understory Cover of
Three Life Forms in Foredune Cottonwood

 

 

 

 

 

Sites
Under Cottonwood I Bordering Foredune Vegetation
Cano
100 W
_ 3 '
a > I
"' o 80 ‘
1‘3 o :
'5 > so ~ I "a" Deciduous
... g I --A---Forb
5 2 4° 7 ' +Graminoid
2 g '
e 20 . I
a g '0."'-- I
0 WW
.- N co 4 In co r~ no a: 0

Location on Transect (2-m Intervals)

 

b) Percent of Total Understory Cover of Six
Life Forms In Foredune Austrian Pine

 

 

 

 

Sites
Under Pine Canopy - Bordering Foredune Vegetation
100 '
a. 90 I
— 0 i
53 3 80 I ---O-- Deciduous
'- 0 70 I —I—Evergreen
Q6 3 60 '. l —'b'"F0rb
.. 2 5° ,o ‘~ ' —-X—-Graminoid
5 " 4° ‘ ' —-o--sn b
0 I6 30 '0 s‘ I I'll
3 u 0 r, I ---I---Vine
9- 5 2 o ,’\. .. l
1 O ' n..— " A ‘.‘§ a i
0 “== ‘~" ' 'T ' ' -
1- N 0') V’ In CD I\ 0 O) 0

Location on Transect (2-rn Intervals)

 

 

 

Figure 3-9. Life form composition of understory cover in a) cottonwood and b) Austrian pine sites in
foredunes. Data from 1994 and 1996.

.32 v5 33 Sec 8.5 .SE 5534‘ 9.38— 8% engage E 3.60 beacon: we 528950 .58 £3 .on 95mm—

 

Eoiouc. E.«v 802.2... :o .5233
3 m a s o n v m N F

 

 

102

 

was Ill...
nacw - lo. I
22.590 IT
ecu. - - .4- I
«3:280 - -0. - -

 

 

 

 

.ICAOO botsaopun "to; to tueojod

 

3930 326.30

cozfiouo> 52505 I cop

02E 5.322 act-om.— mozw omum 52o".
:. «Econ. oh... o>_.... ..o .950 5233...: .29.. Co Eooeod

 

 

103

.82 v5 32 Eod 85 .33 5523. .23 «3% omunlmPé .960 Dew—025 «o action—Eco Show 05 ._ Tm 23E

 

 

 

2.5 iii:
nacw .. lo. I
205880 +
90.... u l4- I
coocacgm +
«3:280 - -0. - ..

 

 

 

E «scon— 3... 5m Co .830 3293:: .39.. .o Eooeod

32:8... E-«v 83:2... co cozaoo...

 

   

 

o. a o s o m w m m P
in! 1 III-III r w k J C
\\ O|\|C\ " \ 1.4V m
<\ \ . d . on W
\ _ I m...
x u i
0” \ _ ..x O” m
ll \ 32.4---. .I.
6 s. _ .. n
.s _ to. s 0.? u
_ ... .. m
..---o. t .. .. a
Q. coo. \ss‘ ... N -1 on O
I! \\\ _ it Q. a“
0.. _ II .Q. o
n .0000 as if om M
_ ... u
_
_ >350 ocE Boa: I o x.

>350 9.63280

«25 omom 52o“. 2...". 22:23

 

 

.82 ea. 55 see. 85%; a... “.8. ... .28 basses 3 Sage Ea 0.... .NZ 5a....

 

52:8... 8.5 33:2... :0 :ozaooa

 

104

 

 

 

 

 

o .. o m h m m c m N w
ill .l'tll a - Jill! ....Iulilr ll.l.....i.l ......iiI: O
.s.. (€u.rroil “tuitikrlrno
\QQ b" 'q\"\-II 0 I \\¢ I OF
"I. \\ I .0
”0 Infilqihllld _ [I] \\ m
0’.’I _ I: O N 0
n. _ 1R m
SSIITI .7. I, on o
DDbP—mIIOII _ 40’! I,
sesamélxl _ .4! 3. m
acuihi _ I- .9
c029o>w+ “ ./ o m M
326.30 --..-- . I ..o.
z .u
_ O
_ I. I cm 3
_ z m
_ . a
_ «It 1: ON .1
_ .5 IA.
co.§ouo> 3:362. 9:628 _ 980 9...... 895 I ow
_

«25 055925 9...... sow...
:. «Eon. 3... 5m .0 .250 329.225 .So... no Emeon.

 

 

.32 Sat Sun 68% 053903 2.3 533< .5 ~38 beam—ova: mo 328.“an .58 £3 .m_-m 8&5

 

105

32:8... 5.3 83:2... :o 5:30..
or m m x. m m v m w F

 

 

 

 

JOAOO Monupun ":01, go wound

 

 

 

cop

05> ......lll... " j o 0
seam . lo. I _ , o x.
EosEEG+ _
eon.--.<-l _ , cm
:3. 93m IIII “
303230 .. -9. - .. cozgmuo> 058925 unto—tom _ >850 3E 59.5 o m
_
_

mozm 05n— cu...m:< 2.53.25
5 2:3“. 0:.— x.m no .550 3223:: .20... .0 =50on

 

 

106

.vafl Boa 8.5 £86 .5233 v.35 2.3 5534‘ E .353 toga—E: no 52.9.8800 .58 £3 .E-m 0.53...—

 

EECSE 5.3 «002.8... :0 5:30..

 

 

 

 

 

 

 

 

co_§muo> 50305 actoubom

or m m N w m ¢ 0 N _.
I | -q‘ I ‘4 0|. ll. II‘III‘III! + IQIIU O
4-V..\.€.-nunl. lul I --.... ..- 0.. .--... \\\

oo o. 0’0 _ Kr I'll c 'lllfl 11 OF
ll \\\ ,4.\fl\l&../ . m
// \\\.lo‘ _ a, ‘.‘I\I*ONM

q _ a \\ /
_ / .\..\4\ . an m.
nasmioi " (\ m.
Bosseolilu _ ow m
Eon—unaul. _ n
2099m>w+ _ 1r om u
maoagooo :9.-- . M
_ 1, cm W
_ u
_
_ 11 ON 0
o
_ A
_ I, cm .1.
_
_ >3ch 2.5 80:: -. cm

«35 50255 Una—:— :.
2:3". 0:.— w>_..._ :. .530 32953:: ,5 .520“. cmEu><

 

 

107

 

a) Similarity In Species Cover in Cottonwood

Foredune Transects
Bordering Foredune Similarity to
Under Cottonwood Canopy Vegetation Interval 10
,-

       

 

 

 

 

 
    
 
 
 

 

 

Q | . .
- fl Similarity to
g x 80 ' Interval 1
a '8 70 1i :
E 5 60 T I
o
in 2‘ 50 + I
“i: -t l
3 s :3 .
= E “r ,
'3 a 20 JL I
a 10 T I
o i i i i ' i + i + I
v- N 0) V In (D h a) O) 0
Location on Transect (2m Intervals)
b) Similarity In Species Cover in Austrian Pine
Foredune Transects
Under Pine Canopy Bordering Foredune Vegetation
1 00 i . '0-
-—-' '-~ I
a 90 ‘ I .I"""" ' Similarity to
1: 80 + | . Interval 10
8 5 70 «~ 3"
r: '0 I I
2 5 60 " o |
o I
(n 3- 50 ~~ f I
s 5 4o ~~ .
E E 30 ~» '
‘3 a 20 1* I Similarity to
S 1 0 AL .,I ' Interval 1
o «.,.-«7' I i i l i .
v- N (V) V to (D l‘ O O) 0

Location on Transect (2m Intervals)

 

 

Figure 3-15. Similarity in community composition and cover across a) Cottonwood and b) Austrian pine
foredune transects. Solid lines indicate similarity of the community to that of interval one, in the core of
the Cottonwood/Austrian pine community, and dotted lines indicate similarity of the community to that of
interval 10, in the core of foredune vegetation. Data from 1994 and i996.

 

108

 

a) Similarity in Species Cover in Forest Edge
Sites Lacking Austrian Pine
Blowout Vegetation Deciduous Canopy I

,' Similarity to
" Interval 10

  
   
   
  

   
 

I
i
!
l
l
l
l
I
|
I
i / Similarity to

interval 1
1 O T”
o 1 l 11 % I 1 l 1 1

%
v- N m V ID ‘0 h 0 O) O
,-

Similarity index

 

Sorenson's Modified

Location on Transect (2m Intervals)

 

 

b) Similarity in Species Cover in Austrian Pine
Forest Edge Sites

  

 

 

 

|
1 O 0 Pine Canopy | M0008 Canopy ,-
u 90 “ I '
g X 80 -. : ,’ Similarity l0
1, «8 7O __ '/ Interval 10
o g |
2 — 60 4 I
p a 50 '1’ I
c - -_ l
3 5 40 |
s E 3° ;. ..-
3 a 20 - Similarity to
m 1 0 4” Interval 1
o I IL I I ' I I I I A
v- N to 1' lo to IN no a: :3

Location on Transect (2m intervals)

 

 

 

Figure 3-16. Similarity in community composition and cover across a) forest edge sites lacking pine and b)
Austrian pine forest edge transects. Solid lines indicate similarity of the community to that of interval one,
in the core of the Blowout/Austrian pine community, and dotted lines indicate similarity of the community
to that of interval 10, in the core of deciduous forest vegetation. Data from 1994 and 1996.

109

 

Similarity of Species Cover in Jack Pine

 
 
   

 

 

8)
Wetpanne Sites
Simlarity to
' Interval 1O
1 O 0 Under Pine Canopy I Bordering Wetpanne I
. I ,
.3 . :3: . ,
fi
"2 8 70 ~~ : p—-"’"'_'-'
E 5 60 ~~ +o
.“ 3‘ 50 * I
§ 5 40 w I
t: ”E" 30 T I Similarity to
o
3 a 20 ‘I” I Interval 1
in 1 O
O T i i L i I i I , i 4
- N co v to to r~ no 0) 0

Location on Transect (2m Intervals)

 

 

 

b) Similarity in Species Cover in Austrian Pine

Wetpanne Transects
Similarity to
interval 10

     
  
  
 
 
  

 

 

1 i9): . Under Pine Canopy I Wetpanne Vegetation ’0.
'2 so I» ' I"
o x l '
a 0 70 “r I '
t: '0 l
2 E 60 I ' "4"”
o l ’—
0) z; 50 i I o"
g :5 40 I I
E " 30 *~
I. .2 J
at: a: 20 - Similarity to

10 q Interval 1
0 LM" "' n i ‘ 4‘
.- N to v to to N no 0’ ‘3

Location on Transect (2m intervals)

 

 

 

 

 

Figure 3-17. Similarity in community composition and cover across a) Austrian pine and b) Jack pine
wetpanne transects. Solid lines indicate similarity of the community to that of interval one, in the core of
the Austrian/Jack pine community, and dotted lines indicate similarity of the community to that of interval
10. in the core of wetpanne vegetation. Data from 1995.

110

 

Similarity In Species Cover Across
Austrian Pine Inland Blowout Transects

   
   
  
  
 

 

 

|
1 O 0 .

9 O . Under Pine : Bordering Blowout t
E x e o .. camp” I Vegetation - Similarity to
:g '8 7 O T : l Interval 10
a s 6 O 1" I I. 'l - ~ - "
.2 a 50]? :IO -‘.'0’
o h 4 O I.
g g 3 0 fl : Similarity to
g a 20 .. Interval 1
‘0 1 0 - — r'

O T + l T T filfi l I fl]
'- N to v to to N no 0) :3

Location on Transect (2m Intervals)

 

 

 

Figure 3-18. Similarity in community composition and cover across Austrian pine inland blowout
transects. Solid lines indicate similarity of the community to that of interval one, in the core of the
Austrian pine community, and dotted lines indicate similarity of the community to that of interval 10, in
the core of blowout vegetation. Data from 1994.

111

68— v00 vam— BE..— 300 6:33 0.35% u <9“—

.8855 3.32m u 255 saga swam u 00$. .5 3% u mmo< gorse: as: u 25.. 233% arena. u 30 guns 330 u 500
.9520.» BEER H mm”?— gubesau 3:56 u .500 .8»? 35% u 23 500053 0309.088 0:3 5534‘. 8200 3% .00—NV 000.: ..o .3200 .2-..” 0.53m

 

 

5.0+
SEEIOII

gill
awo<--.¢.l
gill
._<<m|o|

sEix...
wwtillii
58+

 

zaiéi

 

 

 

A0322... 5-3 «000:2... :o 0233..

 

 

 

 

 

39.00 0:26.000

002m 020m .003". 050 0030.2 5 3.0:0o 00....

 

n N 9
\il , sfl‘ \ No.0
--VNII: ‘ a
\ u
\\ i #0 o 8
\ _ n...
\ _ ‘
\\ /_ : mod
K / n
_ . . s
u I: i mo 0 We
" /.. ..o N
_ I; w
_ ...-I- l7w
_ Iii/nil 4 N...o
_ ll
_
_
_

 

 

112

.03. .000 32 52.. 30G 6:280... 0.355%

u 0%: garage 358.... n 25$ é 0% "08¢. 0.3.58... 33.0 u 500 seats...» 3.: u 23.. .8383. 33%..» 002 $3.. 8005
u in daubesgq 3500 u .58 :3. .30..an u .630 .00... 00.0.0.3 95.00. 002.0 00000.. 0. 0.0005... 98.0 3.... 80. N v 002. ..o b.0000 .84” 0.33....

 

 

mmtloll
H5". --x.--

siniai
58:0...
000<-..o-l

 

 

.2023... 8.3 0002.2... co 5:003

 

 

 

 

 

 

m N w
l‘ 00 lfi \\\ o
It \\-
lll \\ _ ‘l‘I-III‘
\\l‘lell “\‘\ o
..l... .u... \\\ I «o o
\ .\

\0 \\e!‘T 0000- s lllll \lll\\ll “N
\+\\ . .X ....... ., 3.0 m

_

\
\ \ _ m
\ \ _ u
.\ xx _ ., 00.0 m
\

N. \\ i _ A
\. \ _ fl
0. \x _ 0 00.0 w
\ _ Z
\ _ l\

\\ _

\ _ 4. ..o
\ _
._\ _ co..0.0uo> 2.6266
30:00 0:26.000 _ N F .o

 

0:.0 00:00.3
95.00.. 00:0 020m .003“. :. 00.0000 00.... .o 5.23

 

 

113

.32 as. 3.00 .0 3.50 n 080 ..e. 03. u 0am 50.33. 33.00"
000.. 600.3000 000i .I. BE 600.0038... 0050.0 u .000 00000903 00... 0.00.. 0. 0.000000 000.0 200 80. N V 00.0000 >083 .0 5.0009 .30 0.50.".

 

E0380. 5-0. 000000.... 0o 0o=00o._

 

 

 

000m - II. I
0000- Ix. I
0000 IT

090 +
.0o0 lo: I

 

 

 

(am/smote a) Ausuop use"

 

 

00000.03 00000.8

0.0
00000302.

00:. 0.00.. 0. :mo EopM 00.00am >0oo>> .o 3.0000

 

 

 

114

.08. 520 3.5 0.. 0.00 u 02.0 .20.: 350"
.50 600.3000 0.05% n 00.0 0000.00.00... .3509 u .000 00000903 00.0 00.0002 0. 90800.0 m00.0 300 08. N 0 00.0000 >000? ..0 5.0009 .Nu-m 000w."—

 

?.0>..0.0. 6-3 000000.... 00 00:000..
0 m N ..

G _ .lll" _ “‘ O
“HIIIIIIIIIIIQIII III IIIIIII 'Iul'l .

 

 

 

 

“3“!"
ooo

(am/swan g) Ausuop use"

 

03+

...-fill? I
080+
.80 I? I

 

 

 

‘Q‘Q‘I
ooo

... 5.6

If
:0
o

 

00000.02. 00:00.8 0.00000 00.0

0)
0

000000003
00.0 00.393 0. ...mo 53M 00.000w >000; .0 3.000o

 

 

 

115

 

a) Light Intensity in Foredune Austrian Pine
and Cottonwood Stands

2500 0 Under Tree Canopy Bordering Foredune Vegetation

M
O
O
O

|
I
I : I
l ,
I
I

1500 r. i

 

 

+ Austrian pine
- -l- - Cottonwood

I ‘ '
1000" / V
. \ /l ‘

 

 

 

Light Intensity
(mlcroelnstelnslmZ/s)

|

I

I

500 {r‘ ,
- I
o.II..%%4H

1
0 <0 <0 1- 01 v- co IO r~ o
F

T-

1
I

Interval (m)

 

b) Light lntenslty in Forest Edge Sltes

Deciduous Canopy

 

 

 

 

1 8 0 0 Blowout Vegetation
1 O 0 O I
4 0 0 .

1600
1400 - E
I +Austrian pine
\ / xi --l--Lacking pine
800 4
200 - I E : l i g
0

2000 Pine Canopy or
1 200 1
I
I
600 J I
I
I
Interval (rn)

Light Intensity
(mlcroelnstelnslmZ/s)

 

1O
8
6
4
2
1
3
5
7
9

 

 

 

Figure 3-23. Light intensity 1: SE across a) foredune cottonwood and Austrian pine transects and in b)
forest edge sites with and without Austrian pine. Data from 1994.

116

 

a) Light Intensity In Austrian and Jack Pine
Wetpanne Sites

1 2 0 0 Under Pine Canopy Bordering Wetpanne Vegetation

—L
o
o
o

 

l
I
I
I
800- '\£ Ii~ I
l i ’ '2' +Austrian pine

 

 

 

 

Light Intensity
(mlcroelnetelnelmZ/s

 

600 q I --I--Jack pine
4001 :
200- gig-ii I
0.4.11I4TIIIIHI%II+IIII
o m ‘0 V N 1" ('3 IO N a)

Interval (m)

 

 

b) Light Intensity In Inland Blowout Sites

2500 "

Under Pine Canopy Bordering Blowout Vegetation

 

N
O
O
O

01
O
O

 

Light Intensity
(mlcroelnsteinslmzls)

Interval (m)

 

 

 

Figure 3-24. Light intensity 1 SE across in a) Jack and Austrian pine wetpanne sites (1995) and in b)
inland blowout sites with Austrian pine (1994).

117

 

8) Soil Moisture In Foredune Austrian Pine
and Cottonwood Sites

 

—-¢— Vegetation bordering pine
+ Foredune Pine

- - ID - - Vegetation bordering cottonwood
- -I- - Cottonwood

   
  

_-_ *0

 

s—

 

 

SoII Moisture
(9 water! 9 wet soil)

 

Month

 

 

b) Soll Moisture In Wetpanne Sites

 

—-A-—Wetpanne vegetation bordering Austrian pine
+Wetpannne Austrian pins

0
O
01
o

 

 

 

 

 

 

:__~ -45 ‘ ---D-- Wetpanne vegetation bordering Jack pins
8 0.4 ‘r “a --I--Jack Pine
g; 0.35 ._ \\\ "_.v"'"-- ‘-~.
5 3 0.3 + 2.“ ”..-" ~-""~o
330.25» 3‘0"".
= g 0.2 «I
.3 '£0.15
3 0.1
0.05

 

 

 

 

 

 

Figure 3-25. Mean monthly soil moisture in a) foredunes and b) wetpannes across the 1995 growing
season. Note difference in scales.

118

 

 

 

 

 

 

 

 

 

Soll Moisture In Inland Blowout Sites
—-A— Blowout vegetation bordering pins

3 + Blowout Austrian Pine

3 «ma-- Blowout lacking Austrian pins
2 ’6
3 3
e
'8 or
E \

I-
g 3
w E

a 0.4

V —- - 0 ' 'D ' ' - ------ o .

~ ‘ o u. - o """" III '5
JULY AUGUST
M o nth

 

 

 

Figure 3-26. Monthly soil moisture in inland blowout sites across the 1995 growing season.

CHAPTER 4
REPRODUCTION AND RECRUITMENT OF
THE INTRODUCED PINUS NIGRA ACROSS DUNE SUCCESSION:
DEMOGRAPHIC AND COMMUNITY CONSIDERATIONS

INTRODUCTION

The concept of habitat choice has recently gained favor among plant ecologists,
despite the fact that it has traditionally been applied only to organisms which are mobile
throughout their lives (Bazzaz 1991, Schupp 1995). It is during the motile dispersal phase
of the life history of a plant that “habitat choice” occurs, or more technically, is imposed
upon the plant by its habitat (Bazzaz 1991, Schupp 1995). Seeds are dispersed into an
array of habitats surrounding the parent plant and the selective forces of the environment
upon them shape the evolution of an optimal dispersal-genuination-establishment life
history strategy (Schupp 1995). Invasive plant species present a unique opportunity to
study habitat choice, because they are new to the system and have yet to be evolutionarily
influenced by it.

Among the factors that influence germination and subsequent 0stablishment of
seedlings are light (Augspurger 1984, Pons 1992, Gray and Spies 1996), litter (Facelli
1991, Silman 1996), understory vegetation (Maguire and Forman 1983), moisture (Streng
et a1. 1989, Laman 1995), and rnicrotopography (Harper 1977, Hamrick and Lee 1987).
Since most mortality in long-lived perennials occurs in early life history stages, the
interactions between the early life history stages of a plant and its environment may
determine its eventual adult distribution (Harper 1977, Bullock et a1. 1996, Silman 1996).

Recruitment limitation can occur prior to seedling germination, however. Seed

119

120

production can be constrained by flower set, pollination and subsequent fertilization, floral
abortion, and finally embryo abortion (Owens 1995). Limitation at any of these steps will

reduce seed production and subsequent recruitment.

An understanding of recruitment limitation in the context of non-native plant
species introductions is of particular importance if we are to determine the potential for their
invasion into a novel habitat. In order for a plant to be a successful invader, it must be
able to reproduce and replace itself. While it is important to study the reproductive
characteristics of an invader to determine its reproductive output, invasion success in a
variety of habitats will ultimately be determined by both the reproductive output of the
invader and the environmental factors of the receptor community which regulate

survivorship of young seedlings.

Invasion ecology

Generalized characteristics common to successful invaders have been proposed and
used as early indicators of potential spread (Baker 1974, Crawley 1986, Mouiton and
Pimm 1986, Noble 1989). Rejmanek and Richardson (1996) found three characteristics of
pines to be predictive of their invasion potential: small seeds, short juvenile period and
short mean interval between large cone crops. The latter two characteristics imply rapid
early reproduction, while small seeds are often associated with high seed output, long
dispersal distance, or early germination (Rejmanek 1995). While this approach has been
quite successful in retroactively predicting the invasive potential of pine species, it ignores

the interactions with the environments into which these pin0s were introduced.

Other workers have identified characteristics correlated with community invasibility;

these include frequent exposure to disturbance, low species richness, low plant cover, and

121

favorable interactions of the invader with the biota of the receptor community (Crawley
1986, Fox and Fox 1986, Orians 1986). Communities invasible by pines, in particular,
provide limited competition with regenerating pine seedlings, are lacking in species of a
similar growth form, and are exposed to intermediate to severe disturbance which results in
the reduction of plant cover (Richardson and Bond 1991). These community studies
disregard the characteristics of the potential invaders, however. Though few studies have
combined investigation of both invader and invaded community (but see D’Antonio’s 1993
study of Carpobrotus edulis invasion in three coastal communities), it is generally accepted
that a thorough understanding of the invader must be combined with an understanding of

the characteristics of the receptor community in making predictions about invasion.

Population projection models

The projection matrix model approach combines demographic information about the
invader with information about the community into which it has been introduced. The
response of the life history stages of a species to its environment is embedded in matrix
transition probabilities which are determined from empirical data. The construction of
Lefkovitch or stage-based transition matrices facilitates comparisons of future invasion
success of non-native species in a variety of habitats by projecting population sizes across
time and by determining population growth rates. Matrix elasticities, which pinpoint the
proportional contributions of each life history stage to the population growth rate (Caswell
1989, McPeek and Kalisz 1993, Silvertown et a1. 1993), may offer further insights into the
importance of each life stage to the persistence of a species in a particular habitat.
Lefltovitch matrices have been used to predict the success of rare plants in different
microhabitats (Kephart and Paladino 1997), but only rarely have they been used to project

invasion potential for non-native woody plant species (1. Parker pers. com.).

122

Research objectives

This study seeks to determine not only the reproductive output of a potentially
invasive pine, Pinus nigra, on the sand dunes of Lake Michigan, but also to evaluate its
demographic potential in various receptor communities (foredunes, forest edges,
wetpannes, and inland blowouts). Abiotic and biotic factors such as light, soil moisture,
litter, understory cover, and species richness vary throughout the dune system and among
different seral stages and may constrain the establishment of P. nigra in the dunes.
Because the environmental conditions which regulate seed production vary among seral
stages as well, individual trees may differ in their ability to produce seed, depending upon
the seral stage into which they were introduced. Variable seed production, coupled with
the differing array of microsites available for seedling establishment in these seral stages,

may translate into differential invasion success among successional stages in the dunes.

The present time period is a critical one in which to study invasion of P. nigra into
this system, because most of the 26,000 trees planted in the 1950’s and 1960’s have now
reached maturity, and the oldest recruits (10 years of age) are not yet reproductive.
Analysis of the current size structure of the population of recruits can give us an indication

of future invasion patterns in different seral stages in the dunes.

The objectives of this study were document P. nigra reproduction, germination,
seedling establishment, and subsequent transitions to larger size classes to project its
invasive potential. In addition, seedling survivorship and density were examined in
relation to a variety of biotic and abiotic factors to identify the environmental filters which

regulate its survivorship and growth.

123

The following questions formed the basis for this study:
1) Does the reproductive output of P. nigra , as measured by seed production and
germinability, vary among dune habitats and across years?
2) How do P. nigra population growth rates var)I in different seral stages, and which
stages of its life cycle contribute proportionally more to these population growth rates?
3) What are the key biotic and abiotic determinants of seedling survivorship and how do

these vary among seral stages?

To answer these questions, reproductive output of P. nigra was monitored across
two years in four different seral stages on the dunes. In addition, seedling survivorship
and density were monitored in canopy and gap plots in three seral stages across three
growing seasons. Biotic and abiotic variables were measured in each of these seral stages

and correlated with successful seedling establishment.

Study species

Native to Europe and northem Africa, P. nigra grows to 30m in height and reaches
300 years of age (V ergos 1985, Vidakovic 1991). It typically reaches reproductive
maturity by 15-20 years of age, but has been shown to produce cones as early as four years
of age (V idakovic 1974). Maximum cone production takes place from 60 to 90 years of
age. Male cones produce copious pollen which is wind.dispersed in May and June (Burns
and Honkala 1990). Female cones remain receptive to pollination for three days, but not all
the new female cones on a tree are receptive at the same time. The window of time for
successful pollination is quite brief, nonetheless, as is typical of pines. Fertilization takes

place 13 months following pollination, and cones ripen the following fall.

Trees produce cones every year under normal conditions, but do so most

124

abundantly every three to four years. Austrian pine cones each produce on average 30-40
sound (filled) seeds, approximately 50% of which are gerrninable. Poor cone or seed
crops may be a result of bad weather during pollination or losses to insects and rodents

during seed development (Vidakovic 1974).

METHODS

Study site

This investigation took place at Saugatuck Dunes State Park, as described in

Chapter 2.
Reproductive output

Seed production in 1995.» Austrian pine cones were counted and collected from
trees in Fall 1995 upon ripening. Planted adult trees were randomly selected in five
permanent sampling sites until five reproductive trees were located. All cones present on
each of the five reproductive trees were counted in each of the four seral stages described in
Chapter 1. Up to five second-year cones per reproductive tree were harvested for seed

counting and laboratory germination trials. For trees with fewer than five cones, all
available cones were harvested. Cones were oven dried at approximately 42°C for up to

48 hours. Intact seeds were removed from the open cones, counted, and stored at room
temperature until germination trials began. Seed production per reproductive tree was
determined by multiplying the average number of seeds per harvested cone by the number
of cones per tree. These values were averaged for each site and multiplied by the
proportion of reproductive trees in that site to give a mean number of seeds produced per

tree in that site. Stand densities varied across seral stages (see Chapter 2), so it was

125

important to compare seed production per unit area among seral stages as well. Stand
density values (trees/m2) were multiplied by seed production per tree to yield seed

production/m2. Values were compared across stages and years.

Laboratory germination trials.-- Seeds for germination trials were de-winged and
soaked for 20 minutes in a 10% bleach solution to surface sterilize them, and then
submersed for 20 minutes in distilled water to allow for seed imbibition. Half of the seeds
from each cone were placed on moistened Whatman’s 1 filter paper in petri plates using
sterile technique and squirted with distilled water daily, or whenever condensation was no
longer present in the petri plate. Germination trials were run for 14—1 8 days under constant
light and temperature. Percent germination was calculated for each petri plate (representing

a cone) and compared.

Seed production in 1994.“ In Fall 1994, cone and seed production as well as seed
germinability were determined in a pilot study, similar to the 1995 study, but with a slightly
altered sampling protocol. In 1994, all ten of the permanently marked trees in three of the
established sampling sites in each of four seral stages were examined for the presence of
cones (see Chapter 2 for experimental set-up). Because the foredune stands sampled
contained as few as three trees, sample size was limited to three trees in foredunes. Cones
were counted for each tree and up to five second-year cones were collected from each
reproductive tree, as previously described. Seed counting and germination trials were
performed as in 1995, except that all, rather than half of the seeds from each cone were

placed in individual petri plates for germination trials.

126

Demography of naturally regenerating Pinus nigra at SDSP

Pinus nigra is currently regenerating in the foredune, wetpanne, and inland
blowout seral stages at SDSP, but very little regeneration was evident under pine canopy
and in adjacent native deciduous forest vegetation in the forest edge sites. (Sampling was
not conducted on the other side of forest edge pine stands, at the interface with inland
blowout vegetation, however.) Therefore, intensive studies of seedling distribution and
regeneration were concentrated in foredune, wetpanne, and inland blowout sites. In each
of these three seral stages, the extent of P. nigra regeneration was determined in open areas
and under P. nigra canopy cover. Twenty 1m X 1m quadrats were randomly located in
plots with <50% P. nigra overstory (classified as “gap plots”), bordered by P. nigra, and
in adjacent plots under P. nigra canopy, where overstory cover was >85% (classified as
“canopy plots”). Three replicates each of gap and canopy plots were sampled within each
of the three seral stages. In inland blowout sites, however, no true gaps surrounded by P.
nigra were available for sampling. Alternatively, the open areas bordering pine stands
were sampled and hereafter referred to as "gap plots." Seedlings were too sparse in inland
blowout " gap plots" for the selection of random quadrats, therefore entire populations of
seedlings were inventoried at these three locations. Estimates of density for randomly
located quatrats were based on the area sampled, and for inland blowout gaps, on the total
area inventoried, and all comparisons were made on a per unit area basis. Number of
Austrian pine seedlings and saplings d2 years old (offspring of the planted trees) was
recorded for each quadrat, and age (as estimated by number of whorls with needles),
crown area, and height was recorded for each seedling in each quadrat. Each individual
was inconspicuously marked, and survivorship and growth were monitored at annual
intervals in the growing seasons of 1995 and 1996. New seedlings were measured and
marked each year.

127

Microsite variables

Biotic and abiotic variables were measured for each quadrat to determine which
microsites were most conducive to establishment and growth of naturally regenerated P.
nigra individuals. Biotic variables included species richness (number of species per m2
plot) and ocular estimates of percent understory cover, and both were recorded for each
quadrat in July 1994. The proportion of quadrat area covered by needle and/or deciduous

litter was also recorded at this time, as was litter depth.

The abiotic variables, including light intensity, canopy openness, soil moisture, and
soil pH were measured for each quadrat or a sub-sample of the quadrats in 1996. On a
clear day in July, between 11:30 A.M. and 2:30 P.M., measurements of light intensity
were taken in each quadrat with a Dekagon Ceptometer. Spherical densiometer readings
were also taken in four cardinal directions for each quadrat, as a measure of canopy

openness.

Soil moisture was measured gravimetrically at depths of 0-5cm and 10-15cm in
10—15 quadrats in canopy and gap plots within each of the three foredune and three
wetpanne sites in July. Up to five quadrats of each of three P. nigra seedling density
categories were selected in each plot for soil moisture measurements: L designated no
seedlings, M 1-2 seedlings, and H, >2 seedlings. Soil samples were collected with a soil
core or a trowel between 6:30 A.M. and 9 A.M., before significant plant transpiration and

drying had taken place and soil moisture content was determined gravimetrically.

Soil pH was determined in the laboratory with an Oakton pH meter, for each of the
soil samples taken for the soil moisture readings, above. Three readings were taken for

each soil sample and minimum values for each sample were compared.

128

Mean values for each of the above microsite variables were calculated and compared

for canopy and gap plots in each seral stage.

ANALYSIS
Reproductive output

To compare seed production per tree across seral stages and among years, mean
seed production values for 1994 and 1995 were log transformed to normality and analyzed
with a one-way AN OVA where the seed production in each of the eight combinations of
year and seral stage were compared. Year was not heated as a separate effect since the
limited sample sizes of the 1994 pilot study precluded such a comparison. Multiple
comparisons were performed with Tukey’s all-pairs comparisons. Seed production per

unit area was compared in the same manner.

Seed germination values were arcsine transformed (arcsine*square root of
germinability) to normality, and were compared separately for 1994 and 1995. A
multi-level nesting design was used for 1995 arcsine transformed data to determine the
source of the greatest variation in germinability. Trees were nested within site, sites were
nested within stages, and stage was treated as the main effect. All the nested effects were
modelled as random effects, while stage was modelled as a fixed effect, because it was not
selected at random by the researcher. Data for each year were also pooled by stage and
compared with a one-way AN OVA. Multiple comparisons ('I‘ukey’s all pairs comparisons)
were examined for 1995 data.

Demography

Seedling density.-- To determine which seral stages and light levels were most

129
conducive to the establishment and growth of Austrian pine seedlings and saplings, the
density of seedlings (#lmz) in three seral stages and in gap and canopy plots was analyzed

across three years (1994-1996) with a Repeated Measure MAN OVA. Each year was
treated as a response variable (y), and stage, light, and stage*light interactions were
modelled as fixed treatments. To further determine the effect of light and seral stage upon
seedling establishment, density of seedlings older than two years of age were separated
from the total density of seedlings and tested across three years with a Repeated Measure
MANOVA in the same manner.

Repeated measure MAN OVA compares three characteristics of the profiles of
seedling density across the three sampling dates: “parallelism”, “levels”, and “flatness”
(von Ende 1993). Parallelism is determined by examining the time*stage and time*light
interactions, and significant differences of these interaction terms would indicate that
density does not change in a parallel fashion across time among stages, or alternatively,
among canopy and gap plots across time. If the time*stage interaction is significant, the
significance of stage effects and stage*light interactions must be interpreted with caution
(von Ende 1993). In similar fashion, if the time*light interaction is significant, all light
main effects and interactions must be interpreted with caution. “Levels” are the differences
in magnitude of the main effects of both stage and light treatments. The significance of
these main effects would suggest that seedling densities vary among stages or among light
treatments across years. Finally, the time effect tests for “flatness” (von Ende 1993), or

the change in density across time.

Seedling height and survivorship.-- Mean height of juveniles (>1year old) was
compared across seral stages in 1996. Height data from all three sites within each seral
stage were log transformed to normality and compared using a nested two-way ANOVA,

where stage, light, and stage*light interactions were modelled as fixed effects, and

130

Site(stage) was modelled as a random effect. The same nested ANOVA model was used to
compare arcsine transformed seedling survivorship data for two transition periods

( 1 994— 1 995 and 1995— l 996).

Transition matrices.-- Lefltovitch, or stage-based transition matrices were used to
determine projected population growth rates in three seral stages. Matrices were
constructed using age and height class data for three years under two light levels in three
seral stages for a total of 12 matrices. P. nigra life history was categorized into five life
stages: germinant (g), juvenile 1-3 G 1-j3) and adult (a) (Figure 4-1). The germinant life
stage (g) was represented only by newly germinated individuals in their first growing
season, while the juvenile life stages were determined by height ( j 1<10cm,
10cmsj2<25cm, j3>25cm). None of the juveniles were reproducing. Adults (a) were
defined as mature, reproductive trees and were lumped as one size category, because
effects of planting density on seed production overshadowed any size based differences in
fecundity. In this study, adults were represented only by planted trees, because none of the
recruits had yet reached reproductive maturity. All individuals in g and j categories were
naturally regenerated. Seeds of P. nigra are known to be viable in controlled conditions for
up to 10 years (Burns and Honkala 1990), but data on seed bank survival of P. nigra in
natural habitats were unavailable, so seed bank transition probabilities were not included in

the model.

Several transitions were possible for each of the life history stages of P. nigra
(Figure 4—2). Individuals in the g category could either die or transition to a j category.
Although transitions from g to jl were most common, transitions from g to j2 were also
observed. Individuals in the j life stage could die, remain in the j life stage, or transition to
the a life stage. Stasis or progression from small to large j’s was most common, but burial

by litter or tissue loss by herbivory could result in retrogression from larger to smaller j

131

categories. Adults (a) could either die or remain as adults. They also produced seed each

year .

Actual transition probabilities were determined from field plots in each seral stage.

The fecundity estimate (Fa) was determined by multiplying the mean per-tree seed

production by the mean density of first-year seedlings in each seral stage and at each light
level in each year of the study. This value was divided by the estimated mean seed
production per square meter in each seral stage to yield a proportion of individuals that

became germinants. The (#seedlings/m2)*(#seeds produced/m2)l values were used in lieu

of laboratory seed germination probabilities, because laboratory values ignore the filtering
that takes place in the field with seed predation, moisture limitations, damping off, etc.
Estimated seed production values, incorporated into the previous equation, were calculated
from seed production data collected for each seral stage in 1994 and 1995. Because seed
production values for trees in each seral stage were similar across years, they were

averaged together and used in matrices for both 1994 to 1995 and 1995 to 1996 transitions.

Pu and Pji transition probabilities were determined by pooling data from all three

replicates of each of the six possible combinations of light level, and seral stage in each
year. The frequency of transition from one category to the next was divided by the total
number of individuals in the first category at time one to yield its transition probability to
the second category at time two. Because none of the recruits of the P. nigra population
had yet reached reproductive maturity, an empirical transition probability from j3 to adult
(Pm) was impossible to calculate. Instead, a probability of .04 (1/25) was used since j3

trees are typically at least seven years of age and will likely reach reproductive maturity
within 25 years, suggesting that at least 105 of them are transitioning to the adult life stage

annually. One percent mortality was included in the model for both the j3 life stage and the

132

adult life stage, though these probabilities were also impossible to determine empirically. It
is important to note that the transition from j3 to adult encompasses a large size range, and

that stasis within the j3 category for up to 20 years would not be unexpected.

The population growth rates (lambda) were determined for each matrix, and
elasticities, which describe the relative contribution of each life history stage to the
population growth rate (Caswell 1989, Kephart and Paladino 1997) were also examined
using RAMAS-Stage (Ferson 1991).

To project population sizes across time, initial population sizes for each life stage
were needed. These values were simply the number of individuals in each category at the
start of the simulation. The initial number of adults present in each seral stage was set at
five, because this approximated stand density estimates in the sampling area. Initial seed
input was determined by multiplying the seed production per tree by the initial number of
adults in each seral stage, because the actual seed input was not known. These calculations
assume that adult populations and seed inputs were the same in both canopy and gap plots,
but this overestimates the initial adult populations in gaps. Since estimates of seed
dispersal were not available, they were not included in the model, though they would

clearly improve the accuracy of the resultant projections.

Microsite variables

To determine the differences in microsite variables in canopy and gap sites and
among seral stages, each of the factors recorded was examined using a nested ANOVA,
where stage, light, and stage*light were modelled as fixed effects and site(stage) was

modelled as a random effect.

133

Nonparametric Spearman rank correlations between P. nigra seedling densities,
survivorship, and the environmental variables of understory cover, species richness, needle
and deciduous litter, litter depth, light intensity, canopy openness, soil moisture, and soil

pH were also calculated.

All analyses were performed using JMP software (SAS 1995).

RESULTS

Reproductive output: seed production and laboratory germinability

Average seed production per tree in the foredunes was almost 10 times higher than
that of wetpanne and inland blowout trees, and 20 to 40 times greater than that of forest
edge trees, depending upon the year (Figure 4-3). Though log seed production varied

significantly among stages, it did not differ across years.

On a per unit area basis, seed production in foredunes was only three times higher
than that of the other seral stages in 1995 (Figure 4-4). Seed production did not vary
among the remaining three seral stages in 1995. As a result of smaller sample sizes, 1994

seed production values did not vary statistically among seral stages.

Laboratory germinability was lowest for foredune seed in both 1994 and 1995
(Table 4-1), and while forest-edge seed had the highest germinability (89%) in 1994,
wetpanne seed was most gerrninable in 1995 (18%). Differences in germinability across
seral stages were not statistically significant in 1994 due, in part, to small sample sizes, but
in 1995, foredune and inland-blowout seed was significantly less gerrninable than
wetpanne and forest-edge seed. Germinability varied dramatically across years, with an

134

average of 67% in 1994 and only 10% in 1995.

When 1995 data were pooled by seral stage and compared, gerrninabilities varied
across seral stage of seed origin. A multi-level nested AN OVA indicated, however, that the

greatest proportion of variation in germinability was among trees within sites (Table 4—2).

Table 4—1. Laboratory germinability for Austrian pine seed collected from four seral stages in 1994 and
1995. Germination trials performed with seeds from individual cones: n is the number of cones for which
trials were run in each seral stage. Germinability per cone was significantly different across years, but also
among seral stages in 1995.

 

 

Blowout

1994 1995 Significant

Seral Stage Mean (96) 8.5. n : Mean (96) SE. n difference
Foredune 59.8 5.2 30; 8.1 1.3 98 b‘
ForestEdge 88.9 6.6 3§ 4.5 3.0 30 ab
Wetpanne 67.6 4.4 38§ 18.2 3.3 58 a
Inland 69.6 5.2 331 6.0 1.4 78 b

 

I"Means marked with different letters differ significant in germinability among stages in 1995 (p=.05, Tukey
all pairs comparison HSD).

Table 4-2. ANOVA for 1995 arcsine transformed germination data. Germination values were determined
for individual cones collected from trees in foredune, wetpanne, forest edge, and inland blowout seral stages.

 

 

5mm SS MS 1215 Emilia It
Stage 0.95503 0.31834 3 1.9468 0.1563
Site(Stage) 2.31647 0.16546 14 1.1792 0.3248
Tree(Stage,Site) 6.03599 0.14037 43 2.6085 <.0001
Error 10.816385 0.053813 201

Demography

Seedling density.-- Seedling densities were 1.5 - 4 times greater in foredune canopy

plots than in those of other seral stages (Figure 4-5), depending on the year. Foredune

135

gap seedling densities were also higher than those in other gaps (1.3 - 45 times more

dense), except in 1996, when wetpanne gaps had the highest densities.

Repeated measure MAN OVA indicated that total density did not change in a parallel
fashion across time among stages, though it did have the same slope among light treatments
across time (Figure 4-6, Table 4-3). The non-parallel nature of the profiles of seedling
density in seral stages across time indicates that the significance of stage effects and
stage*light interactions must be interpreted with caution (von Ende 1993). Stage effect was
significant, however, suggesting that seedling densities varied among stages across years.
Density did not differ among canopy and gap plots across years, but seedling density

changed across time, as suggested by the significance of the time effect.

Table 4-3. Repeated measure MANOVA for density of all P. nigra seedlings in three sites within each of
three seral stages and at each of two light levels across the growing seasons of 1994-1996.

 

 

 

 

 

Effect Wilks Exact F DF DF p
lambda Numerator Denominator

Between subjects
Stage .493 6.16 2 12 .014
Light .999 .0001 1 12 .992
Stage*Light .921 0.52 2 12 .608
Within Subjects
Time .309 12.3 2 11 .002
T'rme‘Stage .272 5.04 4 22 .005
Time‘Light .766 1.68 2 l l .232
Time*Stage*Light .743 0.88 4 22 .492

 

To examine the relationship between light and establishment of seedlings beyond
the second growing season, a repeated measure MAN OVA on older seedling densities (>2
years of age) was performed (Table 44). Profiles of seedling density in both seral stage

136

and light treatments were not parallel across time, that is, densities did not change in the
same direction and the same magnitude over three growing seasons (Figure 4-7). Neither
stage nor light treatments, nor the interaction between them, differed significantly in
seedling density. Contrary to the analysis of total density, densities of >2-year-old
seedlings did not vary among stages, nor light treatments. But the mean densities were
much less similar across canopy and gap plots than in the total density analysis. This
would indicate that light levels play a more important role in the survivorship of seedlings
beyond the second growing season than they play in the initial germination and recruitment
of one to two year seedlings. Finally, the time effect on seedling density was highly
significant, indicating that seedling densities varied across the three growing seasons. The
direction of the time effect differed from that of total seedling density; the density of
seedlings >2 years old increased with time, while total seedling density declined.

Table 4-4. Repeated measure MANOVA for density of P. nigra seedlings greater than two years of age
within foredune, wetpanne and inland blowout seral stages and at each of two light levels (campy and gap)
across the growing seasons of 1994-1996.

 

 

 

 

 

Effect Wilks Exact F DF DF p
Lambda Numerator Denominator

Between subjects
Stage .648 3.26 2 12 .074
Light .782 3.35 l 12 .092
Stage*Light .849 l .07 2 12 .374
Within Subjects
Time .370 9.37 2 1 l .004
Time‘Stage .342 3.90 4 22 .015
Time‘Light .919 0.48 2 1 1 .629
Time‘Stage‘Light .757 0.76 4 22 .525

 

Seedling size.-- Juvenile seedlings (more than one season old) growing in

wetpannes and inland blowouts in 1996 were two to four times taller in gap than canopy

137

plots (Figure 4-8, Table 4-5). Log height of juvenile seedlings in foredunes did not differ

across gap and canopy plots, however, nor did height vary with seral stage.

Table 4-5. ANOVA Table for log height of seedlings in foredune, wetpanne and inland blowout seral stages
and in canopy and gap plots in established quadrats in 1996. 181 seedlings were measured.

   

 

Source SS MS Num DF F Ratio p
Stage 0.47401 0.237 2 2.3318 0.1442
Site(Stage) 0.74141 0.12357 6 1.8501 0.0922
Light 3.051 17 3.051 17 1 45.6820 <.0001
Light‘Stage 1.66273 0.83137 2 12.4472 <.0001
Error 1 1.28775 0.066791 169

 

Survivorship.-- Seedling survivorship was two to three times higher in gap plots
than under Austrian pine canopy (Figure 4—9, Table 4—6). In canopy plots, survivorship
ranged from 19% to 46%, while survivorship in gaps was higher at 33% to 100%.
Among seral stages, survivorship was highest in inland blowout gap plots (93-100%) and

lowest in foredune canopy plots (19 - 36%).

Survivorship varied across years as well; a higher proportion of seedlings survived
from 1995-1996 than from 1994-1995 in all but wetpanne gap plots. A nested ANOVA of
arcsine transformed seedling survivorship for 1994—1995 indicated that seral stage and light
effects and their interactions were significantly different (Table 4-7). Survivorship did not
vary significantly among stages in 1995-1996 (Table 4-8). The pattern of interactions
differed across the two time periods as well. In 1994 - 1995, survivorship in canopy and
gap plots in foredunes did not differ, but by 1995-1996 survivorship in foredune canopy
plots was significantly lower than in gap plots. Wetpanne canopy and gap survivorship
were not statistically different in 1995-1996, but they were in 1994-1995. This indicates

that the conditions in 1994-1995 were more conducive to survivorship in wetpanne gaps,

138

and less so in foredune gaps than were the 1995-1996 conditions.

Table 4-6. Seedling survivorship in two growing seasons in three seral stages and in canopy and gap
plots. N represents the number of quadrats which contained seedlings for which survivorship was
determined across the time period. Contrasts between arcsine transformed seedling survivorship within
each seral stage were performed within each time period to determine significant differences. Means marked
with an asterisk were significantly different among light treatments within seral stage (p<0.05).

 

1994- 1995 1995- 1996
Seral Stage Light Mean SE. n Mean SE. n
Foredune Canopy 0.185 0.034 39 g0.358* 0.093 23
Gap 0.328 0.070 28 go.638* 0.087 18
Wetpanne Canopy 0255* 0.075 31 §0.462 0.132 13
Gap 0889* 0.048 2710.554 0.075 27
Inland Canopy 0285* 00.073 33 g0.417* 0.149 12
Blowout Cap 0933* 0.040 301.000* 0 30

 

 

 

1

 

Table 4-7. ANOVA Table for 1994-1995 arcsine transformed survivorship of seedlings in foredune,
wetpanne and inland blowout seral stages growing in canopy and plots (light) in established quadrats.
Survivorship was calculated for 188 quadrats.

 

 

Source SS MS DF F Ratio p
Stage 9.007 4.503 2 10.940 0.009
Site(Stage) 2.495 0.416 6 1.673 0.130
Light 23.518 23.518 1 94.611 <.0001
Light’Stage 6.256 3.128 2 12.585 <.0001
Error 43.749 0.249 176

 

Structure and dynamic .-- Population size structure of the new recruits was very
similar in canopy plots of all seral stages; structure of foredune gap populations was also
similar to canopy plots (Figure 4-10, population sizes (n) in Table 4-9). In 1994, the
highest proportion of individuals in the population were germinarlts, but in the two
successive years the population shifted to a majority of individuals in the smallest juvenile

size class (<10cm in height). No seedlings greater than 100m in height were found in any

139

of the canopy plots, though a small percentage of individuals (0.6% - 6.8%) exceeded
10cm in the foredune gaps.

The population was much more evenly distributed among life history stages in
wetpanne and inland blowout sites. Nearly equal proportions of individuals were found in
j2 and j3 classes in 1995 and 1996 in these sites. In addition, the proportion of individuals
>25cm in height was highest in these gap plots (1.5% - 22.7% of the population). The
proportion of germinants was 1.6 to 46 times higher in 1994 than in any successive year,
'mdicating that 1994 was a relatively high recruitment year for P. nigra.

Table 4-8. ANOVA Table for 1995-1996 arcsine transformed survivorship of seedlings in foredune,

wetpanne and inland blowout seral stages and in canopy and gap plots in established quadrats. Survivorship
was calculated for 123 quadrats.

 

 

Source 88 MS DF F Ratio p
Stage 2.723 1.363 2 3.882 0.070
Site(Stage) 2.120 0.353 6 1.063 0.389
Light 5.707 5.707 1 17.168 <.0001
Light‘Stage 2.424 1.212 2 3.646 <0.030
Error 36.897 0.332 111

 

Table 4-9. Population size of regenerating P. nigra in three years in three seral stages in gap and canopy
plots. No adults were included.

 

1994 1995 1996 Area sampled (m2)
Seral Stage Gap Canopy Gap Canopy Gap Canopy Gap Canopy
Foredune 147 155 S7 37 51 18 56 55
Wetpanne 61 55 51 10 32 ll 61 60
Inland Blowout 36 65 35 13 36 6 620’ 60

 

"' Total inventories were taken for entire blowout gap plots due to low seedling densities.

Projection matrices.-- Transition matrices for 94-95 and 95-96 P. nigra populations

in canopy plots indicated that no transitions beyond the jl life stage had yet occurred (Table

140

4-10 and 4-11). Matrices for gap plots showed that survivorship of seedlings increased
with size. For example, survivorship of gerrninants in wetpanne gaps was 73%, but was

100% for jl's. In addition, inland-blowout-gap individuals had the highest survivorship.

Projected population growth rates (lambda) varied with seral stage and with year
where the probability of remaining an adult was 99% (Table 4-12) and yielded values that
would indicate rapidly increasing populations (1.201) to slightly declining populations
(0.997) . These values were calculated only for gap plots because seedlings never grew
taller than 10cm in canopy plots. From 1994 to 1995, all populations were increasing
(lambda >1), and wetpanne sites had the highest population growth rates. In 1995-1996,
however, population growth rates indicated that the populations were in decline. Growth
rates could not be calculated in 1995-1996 for inland blowout populations because no
gerrninants were present in 1996, thereby blocking the future population growth.

Population growth rates were approximately 0.02 lower for projections in which

P a = 0.95. In the transition from 1994 to 1995, wetpannes still had the highest population

growth rate, but foredune and inland blowout populations were increasing as well (1.028
and 1.022). From 1995-1996, growth rates in foredune and wetpanne were nearly equal
and indicated that the population was declining.

A comparison of projected population sizes among seral stages in 1994-1995
indicated that if conditions remained as they were during this time period, the population of
wetpanne sites would exceed their initial population sizes by 10,850 times in 50 years
(Table 4-13). Foredune populations were projected to increase only 27 times in 50 years,
while inland blowouts were projected to increase five-fold. If conditions from the 1995-

1996 growing season were to continue, populations were projected to increase by 8% in 50

141

m
oooo
oooo

oouo
0

m
oooo
oooo

N.o.n
0

n
oooo
oooo

mmmfi...
0

N N. 2 3 0000
03.0 80.. 08.. 33.0 >053...
03.0 - - - a
03.0 300 - - ....
02.0 30.0 300 - 0.

0 30.0 20.0 33.0 .0.

m. a. ... w

000 00030:. 000.5

0 a 3 on 00.5
03.0 930 000.. 300 came...
9.0.0 - - - a
03.0 030.0 - - m.

- 4:0 300 30.0 N.

- $0.0 0.8.0 0.0 .0.

m. a. ... m

000 00000.03

0 0 a 0: 00.5
03.0 03.0 02.0 03.0 :58...
03.0 - - - a
03.0 03.0 - - m.

- 03.0 33.0 - 0.

- 003.0 80.0 3.0 .0.

m. N. ... 0

000 0000000".

n
oooo
oooo

god.

a

n
oooo
oooo

Mbmfi
0

am
Good
Good

mww.mo
0

m.

o
o

N.

8050820 80050

0
02.6

:00
..

000000 00030:. 000.5

o

m..

N.

w.
oom.o

00
...

0.00000 00000003

Ma

. 000000 0000000".

o
o

a.

3
bnwo

0.3.0
0..

0.0.0 00m 000 000000 0. 8&3... .0000 00.0.. 0. 80.00.0000 000.00 .303 w0000000w00 000 000.0000 000.0000 30.000.

mm
3.6

$26
w

hm
Rod

goo
w

.3
0.56

0.56

m

0005 00.000000 0. 00.00000. 0000 0.0 .000 0.0.. 0.008.000.0000 0. 000.0 ..o 5.0000 0000.... 000.0 05 000.» mu: .0 .00 :0 000 000.0> 0.000 .020...

.020.

tam “Oh.
a

.210 030,—.

142

n m. 0.
ooo.. ooo.. 3.0.o
ovoo - -
o0...o N006 -
oo~.o ammo .Nvo

.. - and

0. 0. ..

00m. .0030... 000.0.

m w. o0

o hcwo oovo
ovoo .. -
om0.o ..0..o -

- no 20

- - no

0. 0. ..

0000000903

. . on
800 o00.o N006
ovoo - -
on0.o ovoo -

.. omoo mvoo

- oo0.o 0N0.o

0. 0. ..

000 0000000...

.oo

villi—fl

.00....

23% .OH.
a

m
o00.o
o00.o

0.m...
0

00.0.5380 80.0000
0000. 000000.00 0. 00.00.00 0000 0.0 00.. 0.0.. 0.00000080000 0. 0000 .0 b.0000 000.0 000.0 0... 000.0 mu0 .0 .00 ..0 20 000.0> ..000 .00.0...,

o o w.
o o 03o
- - mmmo
0. 0. ..
000000 .0030... 000.0.
o o 0.
o 0 00nd
- - 00nd
0. 0. ..
000000 00000.03
o o m.»
o o 0 . 0o
- - 0.0.o
0. 0. ..
000000 0000000".

.oo

.00....

>000 .0...

0.0.0 000 000 000000 0. 0000.0 .0000 00.... 0. 000000.000 9.0.0 .30... 000000003. ..0. 000.000. 000.000.. 000730. .. ..v 0.00...

143

years in foredune sites and decrease to 1/4 of their initial size in wetpanne sites. If the
projection was run for a longer period in foredunes, however, population size would
decrease, since lambda <1.

Table 4-12. P. nigra population growth rates in gap plots in three seral stages as calculated from projection

matrices from 1994-1995 and 1995-1996 transitions. Inland blowout population growth rates were
inestimable for the second time period because none of the seeds transitioned to juveniles during this time.

 

 

 

 

1994-1995 1995-1996
Seral Stage where Pu=.99 where Pu=.95 where Pn=.99 where Pu=.95
Foredune 1.047 1.028 0.997 0.970
Wetpanne 1.201 1.190 0.998 0.971
Inland Blowout 1.038 1.022 - -

 

Table 4-13. Projected population size in three seral stages calculated from transition matrices for 1994-1995
and 1995-1996, where Pn=0.99.

 

 

 

Growth Rate

Years Elapsed Foredune Wetpanne Inland Foredune Wetpanne
Since Start Blowout :

Initial 151 66 413 100 67

5 675 196 46 123 19

15 876 1152 62 116 18

25 1289 7270 88 114 18

so 4029 716000 221 108 18

population 1.046 1.201 1.038 0.997 0.998

 

Elasticities indicated that the greatest proportional contribution to the growth rate in
all seral stages is derived from the transitions in the j3 and adult life stages (Table 4—14).
In the time period from 1995-1996 when survivorship in the earlier life history stages was
lower, transitions in the j3 and adult life stages contributed to more than 90% of the

population growth rate. In the 1994-1995 period in wetpanne and inland blowout sites,

144

nmwd

hood

Vmwd

wood

95 30303 255
Bed -
mm _ .o -

.. Sod Sad

- «86

m. ...
900 0:590?

256 -
N2 .e -

- 386 hood

- Sod N86

m. a

900 05—020..“—
ummfiwmma

flew—0m 0:5 85 meta. owns 0.:— 23.3.50»

movd

Quad
0

ovmd

God
a

band

awed
a

mmod - -
whod ewcd -
wood Nm—d cvcd

.. Sad wncd

a .8... a
95 30305 “Eu—5

god .. .
2.0—.0 god ..
- :36 £90
1 £55 2:6

m a. a
95 053903

mmod - -
vwmd awed -
- Need mmod

- weed emod
m. n. a
900 05608"—
mi

mmcd

05 8 802mg: 0: 803 205 0050 £508 900 30305 053 32.32 05 08 0338a .0: 33 0.050320 00 5303030
.36 n =33 5 9:50.22 mo 3:53er 05 80:3 ._ 7v 23 c3. 030,—. Sec 803mg .038 82038.0 :5.

dadnam 20:3 .803 mam E momma .88 085 5

88%? Ex: 3% $026808 5. 86:35 52.253 820% Ea 324.8. 5.. 4826035 .43. 63a...

145

flow—om 0:5 3:: meta: 2:05:20»
2 :08 Eat 28230.: e: 203 205 0050. .5202 .30 25305 .0525 @3733 05 .8 0330.2— 8: 33 82058.0 .8 coma—sofio
.35 u :58 5 055222 .6 £32.62 65 22. as .2800 ._ 3. Ba 21.. 65a... 82.2 8.8%.? .822 32236 .2...

$3 Sod - - - a
- 83 :3 - - a
- Rod 3:... Sod -
- - n8... ~85 Sod :

m3: - m
a m. a. a w
%o .8320 .0525 205265 225
Ste 23 - - - a 33 ~85 - - - a
- Sad 23 - - m. - 33 «mod - - a
- - :od 23 - a - - $3 82. Sod w.
- - - v85 23 a - - :55 :3 ”3.0 :
Sod - - - - 0 ~86 - - - - m
a a «..2 a w a a m z w
900 0:590? 95 0:590?
42$ 23 - - - a 23 Rod - - - a
- Etc 23 - - a - £3 53 - - m
- - 886 :3 - a. - - «85 Rod - w.
- - 52. 8o... Sod : - - mood Sad .85 :
Sod - - - - a mood - - - - m
a n. n. a m a w. a. 2.. a
000 0562a 900 05:20....—
qmmmwwfl mg

.maduam 20:3 32a 9% E mowfim :28 025 E
mafia—amen ENE 33m w::20=0w2 28 000302 .8282. 32.32 :5 mam—.432 8.: $020302“: .m 7.... 030,—.

146

transitions in the j3 and adult life stages contributed less than half to the population growth
rate. When the probability of remaining an adult was decreased to 95%, the proportional
contribution of the j3 and adult life stages were 2-10% less than when PM =99% (Table 4-

15).

Microsite variability

Community attributes." Quadrats located within canopy plots in all three seral
stages had a high percentage of needle litter cover (87-88%), very little deciduous litter
cover (0.1-7%), and a low percentage of bare soil (0.7- 2.7%) (Table 4—16). Gap quadrats
varied across seral stages in litter attributes. Foredune-gap quadrats had 11 times more
needle litter cover than inland blowouts, while wetpanne-deciduous-litter cover was three
times that of foredunes and seven times higher than inland-blowout deciduous-litter cover.
Bare soil was most prevalent in inland-blowout—gap quadrats (85%), and very low in both
foredune and wetpanne gap quadrats (5-6%).

Table 4—16. Average % needle and deciduous litter cover, and % bare soil in canopy and gap plots in three
seral stages. N refers to the number of m2 quadrats sampled. Significant ANOVA results for Stage.
Site(Stage), Light and Light‘Stage interactions are listed as w=stage effect, x=site(stage) effect, y=light
effect, and z=light‘stage effect.

*
Ground Cover Needle Litter EDeciduous LitterE Bare Soil

 

 

 

(96) (8») (%1
Campy Gap {Campy Gap ECanOPy Gap
Foredune 85.2 64.8 0.7 22.3§ 2.67 4.82
Wetpanne 83.1 17.9 7.1 71.3§ 2.69 6.45
Inland Blowout 87.1 5.9 0.1 7.6 0.69 84.67
n 174 145 3 174 145 174 134

 

T

Significance wyz E wxyz 5 wxyz

 

147

Soil moisture and pH measurements were taken only in foredune and wetpanne
sites (Table 4-17). Soil moisture ranged from 0.72 to 1.43% in canopy plots in wetpannes
and foredunes, similar to values in foredune gap plots (0.72 - 0.74%). Wetpanne soil
moisture in gap plots was 10-19 times higher than that of canopy and foredune gap plots
(11 - 13%). In the top 5cm of soil, pH was .3 to .4 units lower in canopy than gap plots
and 0.2 to 0.8 units lower than at a depth of 10-15cm. Additionally, foredune soil was 0.4
to 0.6 pH units more basic than wetpanne soil.

Table 4.17. Average soil moisture and soil pH at two depths (0-5cm and 10-15cm) in canopy and gap plots
in foredune and wetpanne seral stages. N refers to the number of m2 quadrats sampled Significant

ANOVA results for Stage, Site(Stage), Light and Light‘Stage interactions are listed as w=stage effect,
x=site(stage) effect, y=light effect, and z=light‘stage effect.

fl

 

 

 

 

Soil Factors Soil 5 Soil 5 Soil pH at 5 Soil pH at

Moisture : Moisture at : 0-5cm E 10-15cm

at 0-5cm 5 10-15cm 5

1%) 5 (96) 5 5

Canopy Gap g Canopy Gap : Canopy Gap 3 Canopy Gap
Foredune 1.16 0.74 0.72 0.72 5.65 5.92 6.44 6.51
Wetpanne 1.43 13.48 0.91 10.52 5.11 5.56 5.89 5.91
n 75 75 75 73 77 74 77 74
Significance xyz gxyz Ewy ENS

 

Light intensity and canopy openness were higher in gaps than canopy plots, but
differed across stages in gap plots (Table 4-18). Foredune gap quadrats received the
lowest light intensities and had the least open space in the canopy, while wetpannes were
characterized by intermediate light intensities and canopy openness. Inland blowout gaps
had three times the light intensity of foredune gaps and 2.5 times the campy openness.

Biotic factors varied across canopy and gap plots and also among stages (Table 4-
19). Species richness was lowest in canopy plots, butwetpanne canopy quadrats

contained three times more species than inland blowouts. Wetpanne gap species richness

148
was four times higher than both foredune and inland blowout gap quadrats which averaged
2.5 species per m2. Understory cover was low in canopy plots (0.4 - 5.2%), but foredune

canopy plots had more vegetative cover than did wetpannes. Inland blowout gap cover

was minimal (8%), but foredune and inland blowout cover was high (52 - 80%).

Table 4-18. Average light intensity and canopy openness in canopy and gap plots in three seral stages. N
refers to the number of m2 quadrats sampled. Significant ANOVA results for Stage, Site(Stage), Light and
Light*Stage interactions are listed as w=stage effect, x=site(stage) effect, y=light effect, and z=light‘stage
effect.

 

 

 

 

 

Light Factors Light Intensity 5 Canopy

(microeinsteins‘m'zs) - 5 Openness (%)

Campy Gap Campy Gap
Foredune 116.62 464.58 8.08 33.04
Wetpanne 80.88 773.90 4.83 56.10
Inland Blowout 84.83 1456.43 E 6.69 85.12
11 172 146 174 147
Significance wxyz iwxyz

 

Community correlates with survivorship.» Strong correlations suggested that litter
and light were both important to the survivorship of P. nigra seedlings during 1994-1995,
but that the amount of bare ground in the plot was most important in regulating
survivorship during the period of 1995-1996 (Table 4-20). From 1994-1995, survival was
strongly negatively correlated (-0.624) with the amount of needle litter in the plot and
positively correlated with light-related factors (light intensity (0.523) and canopy openness
(0.611)). These same factors were also significantly correlated with survivorship the
following year (95-96), but the association was not as strong. Deciduous litter,
litter depth and soil factors (pH and soil moisture) were not significantly correlated with
1995-96 survivorship, and were only weakly so in 1994-95.

149

Table 4-19. Average species richness and understory cover in canopy and gap plots in three seral stages. N
refers to the number of m2 quadrats sampled. Significant ANOVA results for Stage, Site(Stage), Light and
Light‘Stage interactions are listed as w=stage effect, x=site(stage) effect, y=light effect. and z=light‘stage
effect.

Vegetation

 

       

Species Richness

 

 

 

 

E Understory Cover

(#species/mz) (%)

Canopy Gap Canopy Gap
Foredune 0.95 2.64 5.20 52.35
Wetpanne 1.63 8.80 3.89 80.41
Inland Blowout 0.44 2.23 g 0.35 8.17
n 174 147 163 145
Significance w y z : w x y z

 

Table 4-20. Spearman rank conelations between Austrian pine seedling survivorship and abiotic and biotic
variables. Soil moisture and pH values were taken for a subsample of quadrats in only foredune and inland
blowout sites, therefore sample sizes are smaller for these correlations. It indicates the number of quadrats
used for the correlation analysis.

 

Survivorship 1994- 1995 Survivorship 1995- 1996

 

 

 

 

Variable Spearman Rho p n Spearman Rho p n
Needle Litter -0.6242 “* 169 -0.4096 ”* l 1 1
Deciduous Litter 0.4704 “" 169 ‘ 0.1073 NS 111
Litter Depth ~0.2430 " 96 -0.2330 NS 56
Bare Ground 0.4156 "* 169 0.5960 "* 111
Species Richness 0.4295 "* 169 -0.0503 NS 111
Understory Cover 0.4104 *” 169 -0.0329 NS 1 11
Canopy Openness 0.6111 “W 169 0.4734 ““ 111
Light Intensity 0.5246 ”"' 186 0.4219 “* 122
Soil Moisture 0.3607 “ 88 0.2036 NS 51
(O-Scm)

Soil Moisture 0.4275 "‘“ 88 0.2280 NS 51
(10-15cm)

Soil pH (0-5cm) 0.0744 NS 88 0.0152 NS 51
Soil pH (10-15cm) -0.0783 N S 88 0.0607 NS 51

 

” 0.0013 p <0.01
*" p < 0.001

150

Correlations between the number of seedlings present in each plot and
environmental factors were also examined, and were found to be much weaker than those
with survivorship. In 1994, seedling counts were weakly positively correlated with needle
litter (rho = 0.20) and negatively correlated with deciduous litter (rho = -0.23). None of
the other correlations with 1994 seedling counts were statistically significant. Weak
correlations (rho = -0.29 to 0.369) were evident between both 1995 and 1996 counts and
all environmental factors excepting soil moisture and soil pH. The strongest correlations in
1995 were with canopy openness (rho = 0.338) and light intensity (rho = 0.249), and in
1996 with light intensity (rho = 0.369), needle litter (rho = -0.296) and bare ground (rho =
0.297).

DISCUSSION

Assuming tree seedlings were planted after two or three years of growth,
introduced P. nigra individuals at SDSP were between 25 and 42 years of age during the
time of this study, and the older trees were approaching the peak of reproductive maturity
(60-90 years of age). Though many barriers separated seed production from actual
germination and establishment of young P. nigra individuals, many seeds were

germinating, and some seedlings were becoming established.
Reproductive output

The earliest filter in pine regeneration is at the level of seed production, which can
be constrained by cone set, pollination, and embryo abortion. Cone set has been shown to
be limited by environmental factors including light intensity, temperature, water stress and

nutrient availability (Owens 1995), all of which vary among successional stages in the

151

dunes. Pines with fewer actively growing branch tips will produce fewer cones. Foredune
P. nigra stand densities were significantly less than those of wetpanne and forest edge
stands (Chapter 2), and the shading that occurs in the denser stands may limit the number
of growing branch tips and inhibit cone initiation (Table 4-18). This may begin to account
for the 10 to 40-fold greater per-tree seed production in foredune stands relative to

wetpannes and forest edges (Figure 43).

Because P. nigra seed production did not differ across years, it is assumed that
neither 1994 or 1995 was a mast year. Heavy cone crops are typically produced every
three to four years in P. nigra. (V idakovic 1991).

Pollination limitation can occur following cone set, but does not become evident
until the germination phase of the life cycle. Unpollinated ovules may develop normally
until the date of fertilization, but the seeds that develop will not be filled and will not
germinate (Owens 1995). Cold, wet weather during the short time window of pollen
receptivity in the spring (three days) may inhibit pollen cloud formation which is essential
for significant seed set to occur in pines (V idakovic 1974). Climatic differences may be the
cause of the differences in seed germinability across years in the dunes (Table 4-1),
although an examination of monthly precipitation and temperature records for May and
June of 1993 and 1994, indicates that weather patterns were quite similar in these years
during the time at which pollination would have taken place for the 1994— and 1995-
collected cones. The majority of ungerrninated seeds from both 1994 and 1995 were
hollow (Leege personal observation), however, which suggests that pollination was
unsuccessful for ungerrninated seeds in these years. Perhaps other as yet unidentified

factors are acting to influence these differences in seed gerrninablity across years.

If pollination does occur, the next constraint to seed production in conifers is

152

embryo abortion, which may occur as a result of self-inviability (Owens 1995). Multiple
fertilizations may occur in pines, but most self-fertilizations result in post-zygotic
abortions. Seeds with aborted embryos will be filled but will not germinate. In foredunes
where the P. nigra individuals are distributed in small groups and located in the path of the
westerly winds off the lake, pollination limitation is likely severe. The only pollen source
for these individuals is the small stands which they inhabit. If pollination does occur, the
probability of selfing is much higher, which may result in the abortion of embryos
following fertilization. The combination of pollination limitation and embryo abortion may
be responsible for the significantly lower germinabilities of foredune seed (Table 4—1).
Wetpanne and forest-edge seeds were most germinable and because stand densities in these
seral stages are much greater, pollination limitation and the likelihood of selfing are
probably much less.

The greatest source of variation in germinability was among trees within sites
(Table 4-2). This may reflect genetic differences among trees, or differential pollination
success as a result of location in the stand. Trees on the windward side of the stand are
less likely to experience successful pollination than those on the leeward, since pollen will
be dispersed in the direction of the prevailing wind.

Demography
Population structure and demography give a picture of past events which have
shaped the population, but can also give an indication of its demographic future (Bullock et
al. 1996). This information can be used in predicting invasion potential of P. nigra in the

dune system.

Seedling density and survivorship.-- Pinus nigra seedling densities varied among

153

successional stages and were highest in foredunes (Figure 4-5). This variation across stage
is likely in response to the higher seed input from the foredune trees, but also reflects
microsite differences in conditions conducive to germination and survivorship. Total
seedling densities did not vary between canopy and gap plots, but densities of established
seedlings greater than two years of age tended to be higher in gap plots (Figure 4-5),
though this result was not statistically significant (Table 4-4). Seedling height and
survivorship, which are both higher in gaps (Figure 4-8, Table 4-6), corroborate this
finding and suggest that the “gap” classification encapsulates characteristics which are
conducive to successful P. nigra regeneration. The comparison of total seedling densities
with densities of older seedlings suggest that gap and canopy plots are equally conducive to
dispersal and germination, but that gaps, due to their differing environmental conditions,

are more likely to allow establishment of the pine seedlings over time.

The higher seedling densities in foredunes are not due to higher survial;
survivorship was lowest in this seral stage for both 1994-1995 and 1995-1996 (Table 4-6).
Foredunes have a greater average canopy cover than either of the other two seral stages,
and lower average light intensity, as well as more needle litter (Tables 4-16 - 4—19), and

these characteristics may negatively impact young seedlings.

Pinus nigra seedling survivorship was significantly higher in gaps than in canopy
plots (Table 4-6 to 4-8). Gaps generally provide high light intensities, but low soil
moisture, while plots under pine canopy provide low light intensities, but high moisture
due to cooler temperatures and a deep layer of nwdle litter. Soil moisture is likely
particularly important to germination and survivorship through the first season, but light
availability may be more important in regulating establishment and growth in subsequent
years. Additionally, seedling roots may be unable to penetrate the deep litter layer (subject
to drying) to reach the soil under pine canopy and are consequently unable to become

154

established after the first year.

Wetpanne gaps present a departure from the traditional forest gap, in that soil
moisture levels are more than ten times that of other gap plots (Table 4-17). Therefore
moisture limitation is not a consequence of occurrence in wetpanne gaps. Alternatively,
wetpanne gaps are inundated for a period of time each growing season and survivorship

may instead be inhibited by waterlogged and anaerobic soils in the early stages of growth.

Size structure." Population size-structure data suggest that all canopy plots were
alike, regardless of the seral stage in which they were located, but that populations in
foredune gaps were most similar to those in canopy plots (Figure 4-10). The even size
structure of wetpanne and inland blowout gap populations suggests higher survivorship
and stronger recruitment into larger size categories than that of foredune populations.
These data indicate that the formation of gaps is essential to the regeneration of this species,
but that in foredune sites, regeneration within gaps may be limited. In addition, the non-
forested areas around the pine stands are very likely vulnerable to invasion, because they

possess characteristics similar to those of gap plots.

Also of importance was the variation in the proportion of the population in the
germinant life history category across years (Figure 4-10). The proportion of gerrninants
was high in all seral stages in 1994, but was reduced to a small proportion of the
population in successive years. This may be due to differences in seed input or
germinability across years, but is most likely a response to differing climatic conditions
during the critical time period during which germination took place. Water availability
through the germinant life stage is well known to be a limiting factor in successful
establishment. The conditions in 1994 were very wet at the beginning of the surmner,

while the two successive years were drier and not as conducive to germination and early

155

survival of young pine seedlings. In 1995, seeds were highly germinable but the climatic
conditions were not supportive of early survival. In 1996, the early summer conditions
were very wet, but germinablity of seed was much lower, therefore limiting successful
recruitment into the germinant life history stage. Regeneration of P. nigra appears to be

constrained at different stages of its life history in different years.

Matrix projections indicate that population growth rates were highly variable across
years but that those of wetpanne P. nigra populations were always highest (Table 4-12).
Foredune and inland blowout growth rates were similar, but because initial population sizes
were lower in inland blowouts, projected population sizes in 50 years were also much less
(Table 4-13). Kephart and Paladino (1997) found significant spatiotemporal variation in
growth rates of a rare perennial in two microhabitats and suggested that long term
monitoring must be an essential part of a successful management strategy. Long-term
projections of population size of P. nigra in the dunes, however, must be modified by the

effects of successional change as well as by density dependent effects.

Because P. nigra population growth rates were so variable during the two
transition periods for which they were detemlined, it is important to examine the
relationship between climatic variables and population fluctuations. Based upon weather
data from the years examined, it is hypothesized that June precipitation > 12.4cm (high) is
too high for successful wind pollination, while June precipitation < 7.8cm (low) is too low
for successful germination and seedling establishment. Successful recruitment and
survivorship, and consequent population growth of P. nigra (lambda >1) would be
expected to be associated with low June precipitation during pollination followed in two
years by high June precipitation during germination and establishment of seedlings.
Alternatively, unsuccessful recruitment and population decline would be expected to be

associated with high June precipitation during pollination followed in two years by low

156

June precipitation during germination and establishment of seedlings. Climate data for the
period of 1955-1997 for the Holland weather station, 13 km from SDSP, indicate that 12%
of these series of years met the previously established criteria for poor conditions, and
16%, the criteria for good conditions for P. nigra population growth. If past weather
patterns are predictive of the future, a net increase in P. nigra population size will be

expected over time.

Elasticities indicate that the rate of transition from the largest juvenile G3) to the
adult life stage contributed the most to growth rates in these populations (Table 4-14 -4—
15). This is in concordance with many other woody species (Silvertown et. al 1993)
where stasis and progression in larger size categories contributes proportionally most to the
growth rate. A hypothetical four percent reduction in survivorship for adults shifted the
highest elasticities from the adult life stage to the j3 life stage and decreased the growth rate
in all seral stages (Table 4—15). This can be interpreted to suggest that the adult stage is the
most sensitive life history stage for P. nigra in the dunes and thereby the most important
stage to control. The gradual removal of adults from the population is projected to slow
population growth and may stem the invasion of this species in the dunes. Adult removal
will increase light intensity, however, and may create conditions more conducive to pine

seedling germination and growth.

Microsite variables

As is common among conifers, Pinus nigra seedling survivorship was closely
correlated with light (Table 4-16). Gray and Spies (1996) found a strong positive
relationship between light and conifer seedling survivorship in the Pacific Northwest, but
also found that high soil surface temperatures and high transpiration rates in the most

exposed sites resulted in carnbial death and seedling desiccation. This would suggest that

157

survivorship would peak at moderate light levels, but this pattern was not evident at SDSP.

The importance of light to seedling survivorship varied with the year, however. In
1995-1996, bare soil was the environmental factor most highly correlated with
survivorship of P. nigra seedlings in the dunes (Table 4-16). Light appeared to play a
more important role in establishment during years of high recruitment of gerrninants, but in
years of low germinant recruitment, the presence of bare soil may facilitate root penetration
to available water, which is likely more limiting in these years. Allen and Lee (1989) found
bare soil to be a favorable microsite for P. nigra invasion in tussock grassland in New
Zealand, but Tilman (1997) found no significant effect of bare soil with respect to

community invasiveness in a native oak savannah in Minnesota.

The relative unimportance of soil pH in P. nigra seedling survivorship indicated that
the acidic needle litter input of the pines is not unfavorably altering conditions for pine
regeneration. Reduction in soil pH may have negative impacts upon the native vegetation,
however, and may act to leach available nutrients (Clough 1991). Soil pH was lower

under canopy than in gap plots, and higher in foredunes than wetpannes (Table 417).

Despite their lower seed production, higher species richness and understory cover,
wetpannes were most susceptible of all seral stages to P. nigra invasion in the three years
studied. It has long been held that areas of low understory cover and low species richness
would be most susceptible to invasion (Crawley 1986, Fox and Fox 1986, Orians 1986),
but this study has established otherwise in the dune system. Tilman (1997) found that
invasibility was limited by increased species richness in oak savannah. The highest

number of species in a 1m2 plot in this study was 14 (in dune wetpannes), and 296 of 321

plots studied in all seral stages had fewer than 10 species within them. The oak savannah

158

study investigated a majority of plots with 12-25 species (Tilman 1997), and perhaps the
narrow point of overlap in the range of species richness in the two studies represents a
threshold value at which the effect of species richness upon community invasibility
changes. In addition, high species richness is just one of a suite of characteristics unique to
wetpannes in the dune system. The combination of factors that comprise this dune habitat
are conducive to P. nigra success. Not coincidentally, perhaps, is the fact that P.
banksiana, the only pine native to the three seral stages investigated in this portion of the
study, is located almost exclusively in wetpanne sites in this dune system. Preliminary
studies suggest that the success of this native pine in seral stages other than the wetpanne is

limited by low soil moisture and seed predation (Lichter pers. com.).

Conclusion

This study suggests that the recruitment and spread of P. nigra at SDSP is limited
by different factors in different years. Environmental factors play a different role in the
filtering process depending upon the size of the pool of gerrninants and the climatic
conditions. Light availability, composition and cover of litter, and biotic factors such as
understory cover and species richness are more highly positively correlated with seedling
survivorship in years of high recruitment, while the availability of bare ground is the
highest positive correlate with survivorship in years of low recruitment. The variability of
the effects of environmental factors upon P. nigra seedling survivorship across years
emphasizes the importance of long term monitoring in attempting to predict the
invasiveness of introduced plant species. Matrix elasticities indicate that changes in the
transition probabilities in the adult and largest juvenile stages are the most likely to alter
population growth rates. Consequently, control of these life history stages is of greater
consequence than management of smaller size classes. Despite the fact that management

efforts ought to focus on reducing the survivorship of the larger size classes, this study

159

indicates that the eventual distribution of P. nigra will largely be determined by its

interactions with the environment during the early stages of its life.

This study also documents the successful reproduction and recruitment of P. nigra
in three seral stages on the dunes of Lake Michigan. The introduced pine is clearly gaining
a foothold in its novel environment, though differentially so in foredunes, wetpannes and
inland blowouts. High P. nigra population growth rates in wetpannes in 1994-1995
suggest that this seral stage is most prone to invasion by the pine. Foredunes and inland
blowouts appear to be less invasible, though their populations were increasing in 1994—-
1995 as well. The indication of successful P. nigra proliferation in the dune environment
presented in this chapter signals the transition from non-native species introduction to

biological invasion.

op
CO-‘G 9:303
63/

Figure 41. Life cycle for Pinus nigra where terms indicating the stages of the life cycle are
defined as follows:

s = seed

g = germinant (first season seedling)
jl = 0-10cm seedling

j2 = 10-25 cm seedling

j3 = >25 cm seedling

a = reproductive tree

3 ’ ' ‘ ‘ I"Tag
Stage 311+] jl ngl ijj] Pj2jl - -

.12 ngz lejz Pj2j2 Pj3j2

J3 - - P3253 Pj3j3

a ' ' ' Pj3a I)aa

Figure 4-2. Matrix model for P. nigra. Life cycle terms are defined in Figure 4~1. P
refers to the probability of transition from the life cycle stage at time t to the life cycle stage
at time t +1. F is a fecundity coefficient that describes the adult contribution of offspring to
time t + 1 during time t. The 8 stage was not included because seed bank transition
probabilities were not known.

 

161

 

Average Seed Production per Tree in
Four Seral Stages

1800
1600
1400
1200
1000
800
600
400
2001*

 

Average it Seeds per Tree

Foredune Forest Wetpanne Inland
Edge Blowout

Seral Stage

 

 

 

Figure 4-3. Mean 1994 and 1995 seed production per tree :tSE in Austrian pine stands in four seral stages.
Means were calculated from cone counts from up to five reproductive trees. and seed counts in up to five
cones per reproductive tree in five sites per seral stage. 1995 (n=5). Data taken from ten trees in each of
three sites per seral stage in 1994 (n=3). Bars demarcated with different letters denote statistically
significant differences in seed production among stages and across years (p=.05. Tukey all pairs comparison
HSD).

162

 

Austrian Pine Seed Productlon In
Four Seral Stages

 

.1994
81995

Seeds/m2

 

 

 

 

Foredune Forest Wetpanne Inland
Edge Blowout
Seral Stage

 

 

 

Figure 44. Mean seed production per unit area :tSE for 1994 and 1995 in Austrian pine stands in four seral
stages. Means were calculated from cone counts from up to five reproductive trees, and seed counts in up to
five cones per reproductive tree in five sites per seral stage, 1995 (n=5). Data taken from ten trees in each
of three sites per seral stage in 1994 (n=3). Significance testing was done only for 1995 data. Bars
demarcated with different letters denote statistically significant differences in 1995 seed production among
stages (p=.05, Tukey all pairs comparison HSD).

163

68.4%. e. omen .88 nose 5
0:0 :33 Em: :80 0. 00:0 02:. E 09.5.08 ..0 0020000 ..0 003.03. 00.. 0.09 .50... 00» 000 .8050 5 0000.0 .200 02:. 5 mm H b.0000 mam—000m .ntv 0.5m."—

 

uua .0> 30:00

 

 

 

 

 

 

 

tuo>
O O O
a e e
U U U
o o o
d d d
M m. m.
6 6 6
6 6 6
9 S 9
xx... “a“ a. o
. . m... K
e
r —. m
m
T m;
m
. N u
m.
r m.N ‘
526.0 Bee.- . m m
00:00.02; W
0.5020“.- r m o
- 0.

 

0:650:60 20... 65 30.5
0095 .20m 000:... c. 3.000.. 95000....

 

 

 

164

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘Stage I
5
4—
d
"c’
as 3‘
m
2 u
a)
...I
8. 2-
S
<0 4
1- k \
O I I L
1994 1995 1996
YResponses
Stage 1994 1995 1996
D 2.79246032 1.57608605 0.69484127
M 1.00634921 1.06200919 0.43214286
0 0.57340153 0.1992155 0.11456638
u.__ _———___J
r—— ——————————
ICIG
5
4—
2 T
as 3"
0
2 .
9 2
Q '1
o d
1-
O l I
1994 1995 1996
YResponses
CIG 1994 1995 1999
C 1.62592593 0.87865497 0.30740741
G 1.28888145 1.01288552 0.52029293

 

 

 

Figure 4-6. a) Seedling density across three years in three seral stages: foredune (D), wetpanne (M) and
inland blowout (0). Least square mean seedling densities are plotted against year.

b) Seedling density across three years in two light conditions: canopy (C) and gap (G). Least square mean
seedling densities are plotted against year.

165

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r--— a
Stage
1.2_
1.0-
g .-
g 0.8:
5 0.6-
_l —I
g, 0.4"
2 -
0.0—# l
1994 1995 1996
YResponses
Stage 1994 1995 1998
D 0.25396825 0.35417711 0.57420635
M 0.27539682 0.43652882 0.39047619
0 0.02423274 0.0491998 0.10188522
_____.4
ICIG I
e__4
1.2
1 O‘-4
. '1
5 0.8‘
g I
‘3 03‘ AJ/J
0 0.4"l
‘ ‘ l” J
o 0.2-‘ / -
0.0 l l
1994 1995 1996
YResponses
CIG 1994 1995 1998
C 0.05925926 0.18518519 0.24814815
G 0.30980595 0.37475197 0.46289703

 

 

L

Figure 4-7. a) Density of >two-year-old seedlings across three years in three seral stages: foredune (D),
wetpanne (M) and inland blowout (0). Least square mean mdling densities are plotted against year.

b) Density of >two-year-old seedlings across three years in two light conditions: canopy (C) and gap (G).
Least square mean seedling densities are plotted against year.

166

Mean 1996 Height of Juvenile
Austrian Pine Seedlings ( _>_ One
Year Old) In Three Seral Stages in
Canopy and Gap Plots

 

 

 

30 T
E 25 ~
3
_, 20 ..
‘5.
'5 15 1
a:
c 10 ‘-
8
a 5 1
o J + . _
Foredune Wetpanne Inland
Blowout

Seral Stage

Figure 4-8. Mean height :1: SE of juvenile Austrian pine seedlings (21y) in foredune, wetpanne and inland

blowout seral stages in canopy and gap plots (n=3 sites). Seedlings measured in three sites in each seral
stage in 1996.

602-32 000 32;va 6000000. 9:305 0.5 00>0 30.0 no» 0:0 30.30 E noun: .800. 0005 5 mm H nEEoztam m==voom .0... 0.5mm

 

008w _Eow
50306 002:. 0000903 000020".

 

 

167

 

 

82-82 933
82.0.02 30-
82002 .3880
32-08— .3250-

 

 

 

. Nd

Bulnuuns uoluodord

. wd
. m6

 

 

20:. 000 .000 30:00 .... mom-3m .20..." 00::
0. 0007mm? 0:0 33-003 .0. 0_._20>_>._:m

 

 

168

052.0202 23 3.0.02... 2.. .0 252 avg £89038. .883. . 0.0.2.

.3 005000.00 0.03 0000.0 0:002... 0.0 000 .000000 0538» .00... 50:. 5 0.000505 00.05.0000 3300 .3 E00 00.000050. 003 A3 000.0 E05000» 00. 0.053
000 3.: O m-_ 0:002... .3. 005.0000 “030:0. 00 0.... 0000.0 @805 0.5..— .Goo..00o_v 0000000 0530..» 02... ..0 0000 5 0000.0 D000... 0.... .50.. 5 0.000.505
:0 .0 0000.500... .000200. 005 0.0—0 .8000“. 000 00» 5 0000.0 .0000 00.5. 5 000050000 05.. 05.0030 05.200000. 5 0.0.0.50 05m 910 .070 00.0.0.

 

000.0 00.0.: 0....

6 8 7. l 5

cl
3!
l

 

 

 

 

 

 

 

o o o 0
uonundod
to uomodord

 

30000 000 P
00000.03

 

 

 

000.0 00.0.... 0....

000.0 0.0.0.... 0....
n... 6 Ne. _la.

g

[0
If

e!
a
o

«D-

 

“i
o

:0 v
d o
uonelndod
to uoluodou

.r ad

30000 00.0 -- .
00:00.0“.

  

 

uouelndod
so uoluodord

 

 

 

 

300.00

2.02.05 000.0.

 

 

 

 

CHAPTER 5
GENERAL CONCLUSIONS

This investigation provides an early diagnosis of a pine invasion on the sand dunes
of Lake Michigan. The findings presented indicate that P. nigra clearly has the potential to
threaten the ecological integrity of the dune system at SDSP. Pinus nigra has been
successful in primary establishment, reproduction, and recruitment phases of the invasion
process, and has therefore joined the ranks of the 10% of introduced species that escape
cultivation and become naturalized (Williamson 1996).

Pinus nigra survived and grew equally well in foredunes, forest edges, wetpannes
(ephemeral dune ponds), and inland blowouts and exhibited a broad ecological tolerance
for diverse environmental conditions. The primary establishment phase of invasion did not
exclude the pine from any of the dune habitats into which it has been planted, but
established individuals appeared to be most growth limited by spring drought, even in

wetpannes.

It is interesting to note that P. nigra exhibited ecological responses at different
spatial scales in the different phases of early invasion. While primary establishment success
did not correspond with the environmental differences among seral stages in the dunes,
reproduction success clearly differed among seral stages. Pinus nigra seed production per
tree and per unit area was highest in the foredunes and lowest in the forest edges, and
appeared to be negatively correlated with stand density. The open-grown trees of the
foredunes produced heavy crops of cones, whereas many of the densely planted forest
edge trees did not reproduce at all. Seed germination varied from 10% to 67% across the
two year period (1994-1995) , perhaps as a response to weather conditions during

pollination.

169

170

Pinus nigra seedling recruitment has also been successful in three of four seral
stages at SDSP. Seedlings were extremely rare under pine canopy and in adjacent native
forest in forest-edge sites, so this seral stage was not investigated further. Seedling
densities were highest in foredune sites, most likely as a response to high seed input, but
survivorship was lowest in this seral stage. Within stages, survivorship was positively
associated with open or sparse canopy conditions, here referred to as gap plots, though
environmental correlates with survivorship varied depending upon the size of the pool of

new seedlings. In years with high numbers of germinants, light, understory cover, and
species richness/m2 were all positively correlated with survivorship, but in years of smaller

germinant cohorts, the presence of bare soil was most highly associated with survivorship.

Pinus nigra had already significantly impacted native dune communities at SDSP,
even though the second generation of individuals had not yet matured and the invasion was
still in its earliest stages. Its effects were extensive; it reduced light intensity, increased soil
development, and generally inhibited understory cover and species richness under its
canopy. Composition of the pine understory community was often less than 20% similar

to that of the adjacent dune community.

The population projection matrices suggest that Pinus nigra will persist on the
dunes in foredune, wetpanne and inland blowout seral stages. Population growth rates
varied across the years of 1944-95 and 1995-96, from rapidly increasing (1.201) to
decreasing (0.970), but were always highest in wetpannes. At the very least, Pinus nigra
populations are expected to remain stable in foredunes and inland blowouts; at most, they
are expected to increase rapidly in wetpannes, expanding to 17 times their original numbers

within 15 years.

The wetpanne habitat appeared to be most threatened by the invading pine.

171

Evidence of more successful primary establishment in this seral stage, as measured by
height growth, suggested that wetpanne conditions were more conducive to pine growth
than those in other seral stages. Reproduction was intermediate, but seedling survivorship
was high, and matrix projections indicated that population growth rates were highest in
wetpannes as well. The conditions conducive to the survivorship and growth of the parent

trees appeared to aid in recruitment as well.

Wetpanne communities were already heavily impacted by the presence of the
invader. Low light intensity, understory cover and species richness were evident under P.
nigra canopy, but the environmental effects of P. nigra extended even into the wetpannes
themselves. Pinus nigra threatened to alter the character of wetpannes by driving a
conversion to a more terrestrial community, as suggested by the increased woody
component in the flora, and by drying the soil, though present soil moisture content may
reflect intrinsic differences between P. nigra and P. banksiana wetpannes. If these trends
continue, the unique flora of the wetpanne habitat may be lost beneath the shade of the
growing P. nigra recruits and as a response to the continued drying of the soil by increased

transpiration of the trees.

Pinus nigra did not appear to be recruiting in the forest edge seral stage. Though
primary establishment, or growth of the planted trees was successful, the reproduction and
recruitment phases of the invasion process were not. The impact of P. nigra on understory
cover and species richness was less pronounced in this seral stage, in that they were least
inhibited of any seral stage, relative to sites without P. nigra. The presence of the pine also
appeared to facilitate succession by extending the abiotic conditions of the native forest into
the adjacent blowout. Woody tree species had recruited into larger size classes under pine
canopy, and in the absence of significant pine regeneration, were projected to eventually

dominate and replace the pines.

172

The prospects for P. nigra invasion and impact in inland blowout and foredune
seral stages were similar. Primary establishment was successful, and near-equilibrium
population growth rates suggested that the pines present would replace themselves.
Impacts on the native dune communities were extreme, in that understory communities
were nearly absent relative to sites without P. nigra. The woody tree and shrub species
present under pine canopy rarely grew larger than seedlings and were not expected to

replace the pines without significant gap formation.

Analysis of projection matrix elasticities suggests that the large juvenile and adult
life-stages contributed most to population growth rates. Any alteration in survivorship of
particular life-stages will in turn influence population projections and may be used to
manage invasive pine populations. For example, if the survivorship of adult and large
juvenile life- stages was curtailed by girdling or tree removal, population growth rates
would be expected to decrease. Tree removal would result in gap formation, however, and
data from this study show that gap characteristics are highly correlated with seedling
survivorship. Increased survivorship would feed back to increase population grth rates,
potentially negating the consequences of the reduction in adult survivorship. Therefore
control of P. nigra populations at SDSP must include reducing survivorship of both adults

and juveniles.

Considering that P. nigra appears not to be playing an important role in dune
stabilization, and that its removal from the dunes would likely not be harmful, the
following management recommendations are advised. For additional information, please
see the full management plan for P. nigra at SDSP, to be filed with the Parks and
Recreation Division of the Michigan DNR in Fall 1997.

Gradual removal of the adult pines by girdling is recommended for foredune and

173

blowout stands. Regenerating saplings around these stands should be clipped at ground
level every five years to control the spread of the pines. As the girdled pines die and light
penetrates the canopy, the dominant species from the adjacent vegetation should be planted
under the former pine canopy to facilitate the successful restoration of the dunes to their

original state.

The same recommendations apply for wetpannes, but because this seral stage is
most threatened by the presence of P. nigra, removal of the adult pines is much more
urgent. It is not known if P. nigra stands are altering the hydrology of the wetpannes,
however. Therefore trees should be girdled rapidly in several wetpannes, but left
untouched in several others. Water levels in girdled and ungirdled stands should be
monitored to determine if P. nigra is responsible for the drying of the wetpannes. It may
not be possible to reverse the trajectory of succession in these wetpannes, but it will be
important to discern the role that P. nigra plays in their apparent drying. Regenerating
saplings should be clipped at ground level every 3 years, as population growth rates in
wetpannes are high. Other woody species in the wetpanne (abundant due to the presence
of P. nigra) should also be thinned, to reduce transpiration levels and to begin to reverse

the current successional trajectory .

Forest edges are least threatened by P. nigra. Because little pine regeneration is
evident in the deciduous forest adjacent to P. nigra stands, and because deciduous woody
species are recruiting under the pine canopy, these P. nigra stands should be left as they
are. The blowout areas on the other side of the pine stands should be monitored every five

years for evidence of pine regeneration and any new recruits clipped at ground level.

The success of these management practices should be monitored every three years.

Permanent transects should be established in girdled locations to follow the response of the

174

native vegetation to tree death. In addition, regenerating P. nigra population sizes should
be monitored to determine the success of girdling and clipping in curtailing population

growth.

Follow-up studies might include further analysis of the species that colonize pine
sites but are absent from surrounding dune vegetation. This study suggested that animal
dispersed species and those with clonal growth patterns were more successful in
establishing in the presence of P. nigra. It would be interesting to further examine seed
dispersal, germination and recruitment limitations of native woody species under pine
canopy in all seral stages, and in so doing to begin to explain the mechanism of P. nigra

impact and effect on the process of dune succession.

Knowledge of potential dispersal distances is essential for projecting the spread of
an invader. It would be interesting to overlay projected seed shadows of P. nigra on actual
seedling distributions, in order to examine patterns of seedling survivorship and growth
and to predict the extent and direction of spread. Seed predation and herbivory may also be
significant factors in the success of P. nigra on the dunes, and an examination of these

interactions would further explain population dynamics of the invader.

It is ironic that the presumed stabilization advantage of the P. nigra introduction has
not come to pass, yet at the same time, the non-native pines have become a major ecological
force which threatens the integrity of the dune system. The pines have already altered
native community composition, and with their continued proliferation, will have lasting and
potentially irreversible effects. This study has clearly documented the transition from
introduction to invasion, and has begun to identify the consequences of “biological

pollution” in the sand dunes of the Great Lakes.

LIST OF REFERENCES

LIST OF REFERENCES

Allen, RB. and W.G. Lee. 1989. Seedling establishment microsites of exotic conifers in
Chinochloa rigida tussock grassland, Otago, New Zealand. New Zealand Journal of
Botany 27:491-498.

Anderson, M. 1987. The effects of forest plantations on some lowland soils: 1. A second
sampling of nutrient stocks. Foresz 60(1): 69-85.

Ata, C. 1989. Silvicultural characteristics of Abies equi-trojani Aschers-Sinten and growth
relationships between Pinus nigra Arnold var. pallasiana End] and Abies equi-trajani in
mixed natural forests of Turkey. Forestry. 62(3):285-296.

Augspurger, CK. 1984. Light requirements of neotropical tree seedlings: a comparative
study of growth and survival. Journal of Ecology 72:777-795.

Baker, H.G. 1974. The evolution of weeds. Annual Review of Ecology and Systematics
521-24.

Barnes, B.V. and W.I-I. Wagner, Jr. 1981. Michigan trees. University of Michigan Press.
Ann Arbor. 384 pp.

Bazzaz, EA. 1986. Life history of colonizing plants: some demographic, genetic, and
physiological features. pp. 96-110 In: H.A. Mooney and J .A. Drake, eds. Ecology of
biological invasions of North America and Hawaii. Springer-Verlag, New York. 321 pp.

Bazzaz, FA. 1991. Habitat selection in plants. American Naturalist 137: 8116-8130.

Bazzaz, R.A., V. Paape, and W.R. Boggess. 1972. Photosynthetic and respiratory rates of
Sassafras albidum. Forest Science 18:218-222.

Bosela, MJ. 1995. The mode of origin of root buds and root sprouts in the clonal tree,
Sassafras albidum. Master’s Thesis. Michigan State University. 59 pp.

Brock, J .H. 1994. Tamarix spp. (salt cedar), an invasive exotic woody plant in arid and
semi-arid riparian habitats of Western USA. pp. 27-44 _In: de Waal, L.C., L.E. Child, P.
M. Wade, and J. H. Brock (eds.) Ecology and management of invasive riverside plants.
John Wiley and Sons, Chichester. 217 pp.

Brower, J .E., J .H. Zar and C.N. von Ende. 1993. Field and laboratory methods for
general ecology, 3rd ed. Wm. C. Brown Publishers. Dubuque, IA. 237 pp.

Buckley, R. 1987. The effect of sparse vegetation on the transport of dune sand by wind.
Nature 325:426-428.

175

176

Buell, A.C., A.J. Pickart, and J .D. Stuart. 1995. Introduction history and invasion
patterns of Ammaphila arenaria on the north coast of California. Conservation Biology
(6): 1587- 1593

Bullock, J .M., J. Silvertown and B. Clear Hill. 1996. Plant demographic responses to
environmental variation: distinguishing between effects on age structure and effects on age-
specific vital rates. Journal of Ecology 84:733-743.

Burns, RM. and B.H. Honkala. 1990. Silvics of North America: 1. Conifers.
Agricultural Handbook 654. USDA. Washington, DC. vol. 1, 675 pp.

Caswell, H. 1989. Matrix population models. Sinauer Associates, Inc. Sunderland, MA.

Chiarucci, A. 1996. Species diversity in plant communities on ultramafic soils in relation to
pine afforestation. Journal of Vegetation Science 7:57-62.

Clough, ME. 1991. Comparison of aluminum, basic cations and acidity in the soil of
indigenous forests and pine stands in the southern Cape. South African Journal of Plant
and Soil 8(4) 177-183.

Cowles, HQ 1899. The ecological relations of the vegetation on the sand dunes of Lake
Michigan. Part 1. Geographical relations of the dune floras. Botanical Gazette 27:95-117.

Cowling, R.M., E.J. M011, and BM. Campbell. 1976. The ecological status of the
understorey communities of pine forests on Table Mountain. South African Forestry
Journal 99: 13-23.

Crawley, J. J. 1986. The population biology of invaders. Philosophical Transactions of the
Royal Society of London B 314:711-731.

Cronk Q.C.B. and J .L. Fuller. 1995. Plant invaders. Chapman and Hall. London.
241pp.

D’Antonio, CM. 1993. Mechanisms controlling invasion of coastal plant communities by
the alien succulent Carpobrotus edulis. Ecology 74(1):83-95.

D’Antonio, CM. and PM. Vitousek. 1992. Biological invasions by exotic grasses, the
grass/fire cycle and global change. Ann Review of Ecology and Systematics 23:63-87.

Environmental Protection Agency and Government of Canada. 1995. The Great Lakes: an
environmental atlas and resource book. An atlas of the Great Lakes. 3rd ed. 46pp. EPA
Publication 905-B-95-001.

Ewel, J .J . 1986. Invasibility: lessons from South Florida. pp. 214-230 In: H.A. Mooney
and J.A. Drake (eds.) Ecology of biological invasions of North America and Hawaii.
Springer-Verlag, New York. 321 pp.

Equihua, M. and MB. Usher. 1993. Impact of carpets of the invasive moss Campylopus
introflexus on Calluna vulgaris regeneration. Journal of Ecology 81: 359-365.

177

Facelli, J.M. 1994. Multiple indirect effects of plant litter affect the establishment of woody
seedlings in old fields. Ecology 75(6): 1727-1735.

Person, S. 1991. RAMAS/Stage: generalized stage-based modeling for population
dynamics. Applied Biomathematics, Setauket, NY.

Fox, MD. and B]. Fox. 1986. The susceptibility of natural communities to invasion. pp.
57-66 In: R.H. Groves and J .J . Burdon, eds. Ecology of biological invasions.
Cambridge University Press. Cambridge.

Fritts, H.C. 1976. Tree rings and climate, vols. I and II. Academic Press, Inc. London.
565 pp.

Gray A. N. and T.A. Spies. 1996. Gap size, within-gap position and canopy structure
effects on conifer seedling establishment. Journal of Ecology 84:635-645.

Harnrick, J .L. and J .M. Lee. 1987 . Effect of soil surface topography and litter cover on
the germination, survival and growth of musk thistle (Carduus nutans). American Journal
of Botany 74(3):451-457.

Harper, J .L. 1977. Population biology of plants. Academic Press. London. 892 pp.

Hepting, G.H. 1971. Diseases of Forest and Shade Trees of the United States. U.S.
Department of Agriculture Forest Service - Agriculture Handbook No. 386. Superintendent
of Documents.

Hill, MD. and BL. Wallace. 1989. Vegetation and environment in afforested sand dunes
at Newborough, Anglesey. Forestry 62(3)249-267.

Hobbs, R.J. 1989. The nature and effects of disturbance relative to invasions. pp. 389-
405. In J. A. Drake et al. (eds.) Biological invasions: a global perspective. John Wiley and
Sons Ltd. Chichester. 525 pp.

Hobbs, R.J. and LP. Huenneke. 1992. Disturbance, diversity and invasion: implications
for conservation. Conservation Biology 6(3)324-337.

Hughes, F. and PM Vitousek. 1993. Barriers to shrub reestablishment following fire in
the seasonal submontane zone of Hawai’i. Oecologia 93:557-563.

Husch, B., C.I. Miller, and T.W. Beers. 1993. Forest mensuration. 3rd ed. Krieger
Publishing Company. Malabar, Florida. 402 pp.

Kaufman, L. 1992. Catastrophic change in species rich freshwater ecosystems. Bioscience
42(1 1):846-858.

Kephart, SR. and C. Paladino. 1997. Demographic change and microhabitat variability in
a grassland endemic, Silene douglasii var. oraria (Caryophyllaceae). American Journal of
Botany 84(2):]79-189.

Laman, T. 1995. Ficus stupenda germination and seedling establishment in a Bomean rain
forest canopy. Ecology 76(8): 2617-2626.

178

Lanner, RM. 1985. On the insensitivity of height growth to spacing. Forest Ecology
Management. 13:143-148.

Lehotsky, K. 1972. Sand dune fixation in Michigan 30 years later. Journal of Forestry.
March 1972: 155-160.

Lodge, D.M. 1993. Biological invasions: lessons for ecology. TREE 8(4):133-137.

Loope, LL. 1992. Overview of introduced plant species in national parks and biosphere
reserves of the United States. pp. 3-28 In: C.P. Stone, et al. (eds.) Alien plant invasions
in native ecosystems of Hawaii: management and research. University of Hawaii Press.
Honolulu, Hawaii. 887 pp.

Lugo, A. 1992. Comparison of the tropical tree plantations with secondary forests of
similar age. Ecological Monographs 62(1): 1-41.

MacDonald, I.A.W., L.L Loope, M.B. Usher, and O. Hamann. 1989. Wildlife
conservation ad the invasion of nature reserves by introduced species: a global perspective.
pp. 215-255 In: J. A. Drake et al. (eds.) Biological invasions: a global perspective. John
Wiley and Sons Ltd. Chichester. 525 pp.

Mack, RN. 1990. Catalog of woes. Natural History. March. pp.44-53.

Mack, RN. 1992. Characteristics of invading plant species. pp.42-46. In: C.P. Stone,
et al. (eds.) Alien plant invasions in native ecosystems of Hawaii: management and
research. University of Hawaii Press. Honolulu, Hawaii. 887 pp.

Mack, RN. 1996. Predicting the identity and fate of plant invaders: emergent and
emerging approaches. Biological conservation 78: 107-124.

Maguire, DA. and R. T.T. Forman. 1983. Herb cover effects on tree seedling patterns in
a mature hemlock-hardwood forest. Ecology 64(6): 1367-1380.

Matziris, DJ. 1989. Variation in growth and branching characters in black pine (Pinus
nigra Arnold) of Peloponnesos. Silvae Genetica 38(3-4)77-81.

Matziris, D. 1993. Variation in cone production in a clonal seed orchard of black pine.
Silvae Genetica 42(2-3): 136-141.

McEachem, AK. 1992. Disturbance dynamics of Pitcher’s thistle (Cirsium pitcheri)
populations in Great Lakes sand dune landscapes. Ph.D. Dissertation. University of
Wisconsin, Madison. 186 pp.

McPeek, M.A. and S. Kalisz. Population sampling and bootstrapping in complex designs:
Demographic analysis. pp. 232-252. In: S. M. Scheiner and J. Gurevitch eds. Design and
analysis of ecological experiments. Chapman and Hall. New York. 445 pp.

Mills, E.L., J .H. Leach, J .T.Carlton, and CL. Secor. 1993. Exotic species in the Great
Lakes: a history of biotic crises and anthropogenic introductions. Journal of Great Lakes
Research 19(1):1-54.

 

179

Mooney, H.A., S.P. Hamburg, and J.A. Drake. 1986. The invasions of plants and
animals into California. pp. 250-272. _In: H.A. Mooney and J .A. Drake (eds.) Ecology
of biological invasions of North America and Hawaii. Springer-Verlag, New York. New
York. USA. 321 pp.

Moulton, MP. and S.L. Pimm. 1986. Species introductions to Hawaii. pp. 231-249 _In:
H.A. Mooney and J.A. Drake, eds. Ecology of biological invasions of North America and
Hawaii. Springer-Verlag, New York. New York. USA. 321 pp.

Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and methods of vegetation ecology.
John Wiley and Sons, New York. 547 pp.

Musil, CF. and GR Midgley. 1990. The relative impact of invasive Australian acacias,
fire and season on the soil chemical status of a sand plain lowland fynbos community.
South African Journal of Botany 56(4): 419-427.

Noble, 1. R. 1989. Attributes of invaders and the invading process: terrestrial and vascular
plants. In: J .A. Drake et al. (eds.) Biological invasions: a global perspective. 525 pp.

Oliver, CE. and BC. Larson. 1990. Forest stand dynamics. McGraw-Hill, Inc. New
York. 467 pp.

Olson, J. S. 1958. Rates of succession and soil changes on southern Lake Michigan sand
dunes. Botanical Gazette 119(3):125-170.

Orians G.H. 1986. Site characteristics favoring invasions. pp. 133-148 in H.A. Mooney
and J .A. Drake, (eds.) Ecology of biological invasions of North America and Hawaii.
Springer-Verlag, New York. New York. USA. 321 pp.

Ovington, J .D. 1950.The afforestation of the Culbin sands. Journal of Ecology 38:303-
319.

Owens, J .N. 1995. Constraints to seed production: temperate and tropical forest trees. Tree
Physiology 15:477-484.

Paton, DC. 1993. Honeybees in the Australian environment: does Apis mellifera disrupt
or benefit the native biota? BioScience 43(2):95-103.

Peterson, J .M. and E. Dersch. 1981. A guide to sand dune and coastal ecosystem
functional relationships. Michigan State University Cooperative Extension
Service/Michigan Sea Grant Program. Extension Bulletin E-1529. MICHU-SG-81-501 .

Pons, TL. 1992. Seed responses to light. pp 259-284. In Fenner, M. ed. Seeds: the
ecology of regeneration in plant communities. CAB International. UK. 373 pp.

Pysek, P. 1994. Ecological aspects of invasion by Heracleum mantegazzianum in the
Czech Republic. pp. 45-54. In L.C. deWaal, L.E. Child, P.M. Wade. and J .H. Brock.
eds. Ecology and management of invasive riverside plants. John Wiley and Sons.
Chichester. 217 pp.

180

Pysek, P., K. Prach and P. Smilauer. 1995. Relating invasion success to plant traits: an
analysis of the Czech alien flora. pp.39-60. In P. Pysek, K. Prach, M. Rejmanek, and M.
Wade. eds. Plant invasions - general aspects and special problems. SPB Academic
Publishing. Amsterdam. 263 pp.

Pysek, P. and Pysek, A. 1995. Invasion by Heraculeum mantegazzianum in different
habitats in the Czech Republic. Journal of Vegetation Science 6: 711-718.

Ramakrishnan, RS. and PM. Vitousek. 1989. Ecosystem-level processes and the
consequences of biological invasions. pp. 281-300 In: J .A. Drake et al. (eds.). Biological
invasions: a global perspective. 525 pp.

Ranwell, SD. 1972. Ecology of salt marshes and sand dunes. Chapman and Hall.
London. 258 pp.

Rejmanek, M. 1995. What makes a species invasive? pp. 3-14 In P. Pysek, K. Prach, M.
Rejmanek, and M. Wade. (eds.) Plant invasions - general aspects and special problems.
SPB Academic Publishing. Amsterdam. 263 pp.

Rejmanek, M. and D. M. Richardson 1996. What attributes make some plant species more
invasive? Ecology 77(6): 1655-1661.

Richardson, D.M. and W. J. Bond. 1991. Determinants of plant distribution: evidence
from pine invasions. American Naturalist. 137(5):639-668.

Richardson, D.M. , I.A.W. Macdonald and 6.0. Forsyth. 1989. Reductions in plant
Species richness under stands of alien trees and shrubs in the fynbos biome. South Africa
Forestry Journal 149: 1-8.

Riffle, J. W. 1989. Field performance of Ponderosa, Scots, and Austrian pines with
Pisotliuthus tinctorius ectomycorrhizae in prairie soils. Forest Science. 35(4)935-945.

SAS Institute Inc. 1995. JMP, Version 3.1. Cary, NC.

Schupp, E.W. 1995. Seed-seedling conflicts, habitat choice, and patterns of plant
recruitment. American Journal of Botany 82(3):399-409.

 

Silman, M. R. 1996. Regeneration from seed in a neotropical rain forest. Ph.D.
Dissertation. Duke University. 151 pp.

Silvertown, J. M. Franco, I. Pisanty and A. Mendoza. 1993. Comparative plant
demography-relative importance of life-cycle components to the finite rate of increase in
woody and herbaceous perennials. Journal of Ecology 81 :465-476.

Singer, F.S. W.T. Swank, and E.E.C. Clebsch. 1984. Effects of wild pig rooting in a
deciduous forest. Journal of Wildlife Management 48(2):464-473.

Sprecht, R.L. and A. Sprecht . 1989. Species richness of sclerophyll (heathy) plant
communities in Australia- the influence of overstorey cover. Australian Journal of Ecology
37: 337-350.

 

181

Streng, D. R., J .S. Glitzenstein and RA. Harcombe. 1989. Woody seedling dynamics in
an east Texas floodplain forest. Ecological Monographs. 59(2): 177-204.

Stokes, M.A. and TL. Smiley. 1968. An introduction to tree-ring dating. U.of Chicago
Press. 73 pp.

Sukopp, H. and U. Starfinger. 1995. Reynoutria sachalinensis in Europe and in the Far
East: a comparison of the species ecology in its native and adventive distribution range.

pp. 151-159 In: P. Pysek, K. Prach, M. Rejmanek, and M. Wade. (eds.) Plant invasions
- general aspects and special problems. SPB Academic Publishing. Amsterdam. 263 pp.

Telewski, F.W. and A.M. Lynch. 1991. Measuring growth and development of stems. In
J .P. Lassoie and T.M. Hinckley (eds.) Techniques and approaches in forest tree
ecophysiology.

Telewski, F.W. 1995. Wind-induced physiological and developmental responses in trees.
pp. 237-263 In: M. P. Coutts and J. Grace (eds.) Wind and trees. Cambridge University
Press. 485 pp.

Tilman, D. 1997. Community invasibility, recruitment limitation, and grassland
biodiversity. Ecology 78(1):81-92.

Tyser, R.W. and CA. Worley. 1992. Alien flora in grasslands adjacent to road and trail
corridors in Glacier National Park, Montana. Conservation Biology 6(2)253:262.

Vanniere, B. ed. 1984. Tables de production pour les forets francaises. Centre national de
recherches forestieres. 2nd ed. Nancy. 159 pp.

Vergos, Von St. 1985. Strukturen und entwicklungsdynamic naturlicher schwarzkiefern-
walder Nordwest-Griechenlands. Forstarchiv 56:78-82.

Vermeij, G. J. 1996. An agenda for invasion biology. Biological Conservation 78:3-9.

Vidakovic, M. 1974. Genetics of European black pine. Annales forestales anali za
sumarstvo 6(3):57-86.

Vidakovic, M. 1991. Conifers - morphology and variation. Graficki Zavod Hrvatske.
Zagreb. 756 pp.

Vitousek, PM. 1990. Biological invasions and ecosystem processes: towards an
integration of population biology and ecosystem studies. Oikos 57: 7-13.

Vitousek, P.M., L.L Loope, and CR Stone. 1987. Introduced species in Hawaii:
biological effects and opportunities for ecological research. TREE 2(7)

Vitousek, PM. and LR. Walker. 1989. Biological invasion by Myricafaya in Hawaii:
plant demography, nitrogen fixation, ecosystem effects. Ecol. Monogr. 59:247-265.

von Ende, C. N. 1993. Repeated-measures analysis: growth and other time-dependent
measures. In: S. M. Scheiner and J. Gurevitch (eds.) Design and analysis of ecological
experiments. Chapman and Hall. New York. 445 pp.

182

Voss, E. G. 1972, 1985, 1996. Flora of Michigan, 3 vols. Cranbrook Institute of
Science. Bloomfield Hills, MI.

Walker, LR. and RM. Vitousek. 1991. An invader alters germination and growth of a
native dominant tree in Hawaii. Ecology 72(4): 1449-1455.

Wheeler, N.C., H.B. Kriebel, C.H. Lee, R.A. Read and J .W. Wright. 1976. Fifteen-year
performance of European black pine in provenance tests in North Central United States.
Silvae Genetica 25 (l):1-6.

Widrlechner, M.P. et. al. 1992. Performance of landscape plants from Yugoslavia in the
North Central United States. Journal of Environmental Horticulture lO(4)l92-198.

Williamson, M. 1996. Biological invasions. Chapman and Hall. London. 244 pp.

Williamson, M. and A. Fitter. 1996. The varying success of invaders. Ecology 77(6):
1661- 1666.

Witkowski, ET. 1991. Effects of invasive alien acacias on nutrient cycling in the coastal
lowlands of the Cape Fynbos. Journal of Applied Ecology 28(1): 1-15.

Woods, K.D. 1993. Effects of invasion by Lanicera tatarica L. on herbs and tree seedlings
in four New England forests. American Midland Naturalist. 130262-74.

Wright, J .W. and W. 1. Bull. 1962. Geographic variation in European Black pine - two
year results. Forest Science 8(1):32-42.

Wright, T.W. 1955. Profile development in the sand dunes of Culbin forest, Morayshire.
I. Physical properties. Journal of Soil Science 6(2):270-283.

 

‘C

”1111710111111“