PM»?

.. .. .. rug“

F.
r:

3‘

122;.
a E

’9‘") .

I

 

ma _

u
l

g.

:1.

THESSS

   

’3

 
    
   

LIBRAR‘!
Michigan State
University

This is to certify that the

dissertation entitled

A comparison of productivity and related traits
for European larch (Larix decidua Miller) and red pine (Pinus resinosa Ait.)
across a site quality gradient in the Great Lakes region

presented by
John Philip Ger] ach

has been accepted towards fulfillment
ofthe requirements for

_H§s_t;r__ degree in Science

 

Date

MSUI': an Affirmative Action/Equal Opportunity Instilulion 042771

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cz/CIHC/DaieDuepSS-p. 15

 

 

A COMPARISON OF PRODUCTIVITY AND RELATED TRAITS FOR
EUROPEAN LARCH (Larix decidua Miller) AND RED PINE (Pinus resinosa Ait.)
ACROSS A SITE QUALITY GRADIENT IN THE GREAT LAKES REGION

By

John Philip Gerlach

A THESIS

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

MASTER OF SCIENCE
Department of Forestry

2001

ABSTRACT
A COMPARISON OF PRODUCTIVITY AND RELATED TRAITS FOR

EUROPEAN LARCH (Larix decidua Miller) AND RED PINE (Pinus resinosa Ait.)
ACROSS A SITE QUALITY GRADIENT IN THE GREAT LAKES REGION

By
John Philip Gerlach

Productivity and related traits were compared for European larch (Larix decidua
Miller) and red pine (Pinus resinosa Ait.) in 27 forest stands, per species, that collectively
represented a site quality gradient. There were two objectives: 1) determine site factor vs.
productivity relationships for each species, and 2) examine the interrelationships of soil,
leaf/canopy, and productivity characteristics over a range of resource availabilities for
species that differ in leaf life span and related traits.

Productivity (site index and overstory annual net primary production) were most
strongly related to the single variables of soil moisture (sign of the coefficient +) and
nitrogen availability (+). Using stepwise multiple regression models, the combination of
growing season temperature (July T +, growing degree days -) and site water balance (+)
accounted for 71% of the variation in site index for European larch while growing degree
days (-) and foliar %N (+) explained 32% of the variation in red pine site index.

Larch had greater leaf N, specific leaf area, leaf area index, and productivity than red
pine and leaf N was the single species trait most strongly related to productivity for both
species. Over a gradient of increasing soil moisture holding capacity, productivity and
leaf N increased for both European larch and red pine and the rate of increase was
modestly greater for European larch. Despite these interactions, there was no trade—off in

productivity since European larch maintained higher productivity over the entire gradient.

ACKNOWLEDGEMENTS

I would like to first thank my major adviser, Dr. Michael B. Walters, for his
guidance and friendship since our time in Minnesota. His vision was invaluable in the
development and completion of the project. The members of my advisory committee,
Drs. James Hart, Richard Kobe, and Donald Zak, are gratefully acknowledged for their
advice and expertise throughout my tenure at Michigan State.

The core project is the brainchild of Gary Wyckoff; without his commitment to
forestry research and the financial assistance of Mead Corporation, the project simply
would not have been possible.

Many land owners and cooperators deserve thanks for assisting us in finding stands
of red pine and, the elusive, European larch. In particular, Bill Cole and Ken Ellidt
opened the doors to Canadian sites; landowners or managers included Greg Kowalewski;
Bob and Edna Zellar; Charlie Becker; Dan Hartman and Tim Tollefson, Heidi Brunkow
and Dave Olson; Greg Edge; Lew Castle; Mark Jacobs, Dan Gilmore, and Egon
Humenberger; Steve Williams and Steve Scheers; Mr. Lathrop, Michael Rose, Paul
Robertson, and James Barnes; Martin Neumann; Zoltan Kovacs; John Osmik, Graham
Davis, and Len Munt; Bob Penwell; and Steve D’eon, Peter Copis, and Craig Robertson.
The friendly folks at UPTIC (Ray Miller, Brad Bender, Kile Zuidema, Bill Cook, and
Ingrid Klotz), the TRC (Paul Bloese and Randy Klevickas), and the Cloquet Forestry
Center (John Blanchard and Bob Stine) provided key resources and time throughout the

project.

iii

I would like to collectively thank fellow graduate students and other personnel in the
department who helped in many different ways throughout the course of the project.
Tony Fox deserves particular thanks for his indispensable assistance in the field and
encouragement from beginning to end.

. Lastly, I would like to express heartfelt thanks to my family for their support and
encouragement. In particular, my wife Stafia deserves a special thanks as without her

patience and support, I would be lost.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. vi
LIST OF FIGURES ......................................................................................................... viii
CHAPTER 1
Predicting Productivity of European larch and red pine in the Great Lakes region ........... 1
INTRODUCTION ........................................................................................................... 4
METHODS .................................................................................................................... 10
Field Sites ............................................................................................................................... 10
Site Productivity ...................................................................................................................... 11
Predictors of productivity ....................................................................................................... 15
Statistical analyses .................................................................................................................. 3 1
RESULTS ...................................................................................................................... 33
DISCUSSION ................................................................................................................ 47
REFERENCES .............................................................................................................. 53
W
A comparison of site factors, productivity, species traits, and their interrelationships for
European larch and red pine in the Great Lakes region .................................................... 62
INTRODUCTION ......................................................................................................... 64
METHODS .................................................................................................................... 73
RESULTS ...................................................................................................................... 75
DISCUSSION ................................................................................................................ 95
REFERENCES ............................................................................................................ 1 02

LIST OF TABLES

Table 1.1. Stand location (nearest town), mean age, trees per hectare, diameter at breast
height (dbh) , total height, and aboveground production (ANPPo - 1999) for the 54

study stands. ............................................................................................................... 34
Table 1.2. Definition of terms and units. ......................................................................... 36
Table 1.3. Measures of productivity for European larch and red pine species. ............... 37

Table 1.4. Physical soil characteristics in 27 stands each of European larch and red pine.

.................................................................................................................................... 38
Table 1.5. Mean chemical soil characteristics of species across stands. ......................... 39
Table 1.6. Vegetation characteristics of European larch and red pine in all stands. ....... 39
Table 1.7. Topographical characteristics of study stands by species ............................... 40
Table 1.8. Mean climatic characteristics for 1998 — 1999 across the study area ............. 40

Table 1.9. Pearson product-moment correlation coefficients for measures of productivity
and predictor variables reported by species. .............................................................. 43

Table 1.10. Strongest multiple regression equations for predicting productivity of
European larch and red pine. ...................................................................................... 45

Table 1.11. Multiple regression equations for 'simple' measures of predicting
productivity (8125) for European larch and red pine ................................................... 46

Table 1.12. Comparison of biomass production data for European larch and red pine... 48

Table 2.1. Comparison of characteristics for European larch and red pine on a

moderately high fertility site in Wisconsin.T .............................................................. 65
Table 2.2. Pearson product-moment correlation coefficients for measures of physical soil
attributes and productivity by species. ....................................................................... 75
Table 2.3. Definition of terms and units. ......................................................................... 76

Table 2.4. Analysis of variance for regressing 8125 by physical soil attributes, available
soil water holding capacity (AWC), and their interactions. Mean Square, P value,
and adjusted R2 are reported ....................................................................................... 77

Table 2.5. Pearson product-moment correlation coefficients for measures of physical soil
attributes and rates of soil nitrogen mineralization and nitrification. ........................ 79

vi

Table 2.6. Pearson product-moment correlation coefficients for leaf/canopy traits and
measures of productivity (total increment 1999, ANPPMOQI; site index at 25 yrs,
S125). ........................................................................................................................... 80

Table 2.7. Least squares model for the effects of foliar attributes, species, and their
interactions on site index at base age 25 (3125.) Mean Square, P value, and R2 are
reported ....................................................................................................................... 82

Table 2.8. Analysis of variance for testing species, leaf nitrogen (Nlcafma), and their
interactions in predicting change in foliar carbon isotope discrimination (A). Mean
Square, P value, and R2 are reported. ......................................................................... 83

Table 2.9. Least squares models of site index at base age 25 (S125) as a function of
carbon isotope discrimination (A), and leaf nitrogen (N leagms) for European larch,
and SLAHW, and Nleafmass for red pine. ...................................................................... 84

Table 2.10. Analysis of covariance for foliar attributes, species, and their interactions in
predicting ANPPo-tota1. Mean Square, P value, and R2 are reported. ......................... 85

Table 2.11. Least squares models predicting total increment (ANPPo-mm1) as a function of
i) canopy mass and leaf nitrogen mleafmss), and ii) carbon isotope discrimination (A)
and specific leaf area of litter (SLAh-ner) and their respective interactions. Mean
Square, P value, and R2 are reported. ......................................................................... 89

Table 2.12. Model statistics for fitting productivity (site index at base age 25) by
physical soil characteristics with least squares regression. ........................................ 90

Table 2.13. Analysis of variance for leaf/canopy attributes as predicted by available
water holding capacity (AWC), species (SPP), and their interactions. Mean Square,
P value, and R2 are reported. ...................................................................................... 92

Table 2.14. Least squares regression results for vegetation characteristics versus AWC
for each species. Model R2, P value, and coefficients of slope (m) and intercept (b)
are reported by species. .............................................................................................. 93

vii

LIST OF FIGURES

Figure 1.1. Study locations (11 =15) containing the 54 study stands. ............................... 10

Figure 2.1. Hypothesized relationship between stand productivity and site quality (as
defined by productivity being a function of site factors — e. g. soil texture, moisture,
etc.) ............................................................................................................................. 69

Figure 2. 2. Hypothesized responses of foliar and canopy characteristics and productivity
to site quality (~moisture/nut1ient gradients) for European larch (“larch”) and red
pine (“pine”). .............................................................................................................. 71

Figure 2. 3. Hypothesized relationship between productivity (ANPPo) and foliage mass
(either mass or area basis) and foliage N per unit ground area. ................................. 72

Figure 2.4. Hypothesized result of multiple regression ANPPo and [(foliage mass and
foliage N per unit ground area)] and 13 C/ 12C]. ........................................................... 72

Figure 2.5. Area—based leaf nitrogen (Nleafma) in relation to site index (S125) ................ 81

Figure 2.6. Mass-based leaf nitrogen (N1eaf-mass) in relation to site index ($125) for
European larch and red pine. ...................................................................................... 81

Figure 2.7. Senesced tissue specific leaf area (SLAmCr) in relation to site index (S125) for
European larch and red pine. ...................................................................................... 81

Figure 2.8. Foliar carbon isotope discrimination (A) in relation to site index (S125) for
European larch and red pine. ...................................................................................... 81

Figure 2.9. Actual S125 in relation to Predicted SI25 for European larch. Predicted values
from two factor regression of S125 by carbon isotope discrimination (A) and mass-

based leaf nitrogen (Nleafmss). .................................................................................... 81
Figure 2.10. Area-based leaf nitrogen (N]eaf.m~ea) in relation to foliar carbon isotope
discrimination (A) ....................................................................................................... 83
Figure 2.11. Leaf area index (LAI) in relation to 1999 total biomass increment (ANPPO-
total). ............................................................................................................................. 86
Figure 2.12. Canopy specific leaf area (SLAcanopy) in relation to 1999 total biomass
increment (ANPPO-tota1). .............................................................................................. 86

Figure 2.13. Canopy biomass in relation to 1999 total biomass increment (ANPPo_tota1). 86

Figure 2.14. Litter specific leaf area (SLAimer) in relation to 1999 total biomass
increment (ANPPo-tota1). .............................................................................................. 86

viii

Figure 2.15. Mass-based leaf nitrogen (N1eaf-mass) in relation to 1999 total biomass
increment (ANPPo-tota1). .............................................................................................. 86

Figure 2. 16. Carbon isotope discrimination (A) in relation to 1999 total biomass
increment (ANPPo-tom1). .............................................................................................. 86

Figure 2.17. Actual ANPPo-total in relation to predicted ANPPo-tom1 values for European
larch. Predicted values from two factor regression of ANPPMom by canopy biomass
and mass-based leaf nitrogen (Nleafmass). .................................................................... 86

Figure 2.18. Actual ANPPMom in relation to predicted ANPPo.tom1 values for red pine.
Predicted values from two factor regression of ANPPMM by carbon isotope
discrimination (A) and litter specific leaf area (SLAlim). .......................................... 86

Figure 2.19. Available water holding capacity (AWC) in relation to site index (base age
25, S125). ..................................................................................................................... 91

Figure 2.20. Productivity/site moisture (AWC) in relation to area-based leaf nitrogen
(N1eaf.ama) for European larch and red pine. ................................................................ 93

Figure 2.21 . Productivity/site moisture (AWC) in relation to mass-based leaf nitrogen
(N leafmass) for European larch and red pine ................................................................. 94

Figure 2.22. Productivity/site moisture (AWC) in relation to annual nitrogen loss for
European larch and red pine. ...................................................................................... 94

Figure 2.23. Productivity/ site moisture (AWC) in relation to canopy biomass for
European larch and red pine. ...................................................................................... 94

Quilted.

Predicting Productivity of European larch
and red pine in the Great Lakes region

Abstract Site factor-productivity relationships were examined for EurOpean larch
(Larix decidua Miller) and red pine (Pinus resinosa Ait.) using twenty-seven, even-aged
monoculture plantations for each species distributed across the Great Lakes region.
Plantations for each species were geographically paired and averaged 36 and 38 years old
for European larch and red pine, respectively. Productivity was estimated as site index
obtained from stem analyses and as overstory annual net primary production (ANPPO)
obtained with field measurements and published allometric biomass equations. The
range and averages for physical soil characteristics and climate were similar for Eur0pean
larch and red pine stands, but European larch stands had less forest floor depth, higher
soil temperatures, and greater rates of nitrogen mineralization and nitrification. Average
site index (base age 15) and ANPP0 values were, on average, 42 and 31% greater,
respectively, for European larch than red pine. The single variables most strongly
associated with productivity for both species were indices of soil moisture (sign of the
coefficient +) and nitrogen availability (+). The relationships of combinations of soil,
leaf/canopy, and climate characteristics to productivity were determined with multiple
linear regression (MLR). For European larch, growing season temperature (+ July T, -
Growing degree days) and site water balance (+) were the best combined predictors of
productivity, accounting for 71% of the variation in site index. For red pine, growing
degree days (-) and foliar %N (+) explained 32% of the variation in site index. Since
climate variables were strong predictors of productivity, site effects on productivity were
isolated by extracting climate effects from the data (i.e. using the residuals of the MLR
relationship of productivity vs. climate as the predicted variable). For climate residuals,

water balance (+) and foliar 13C/ 12C ratio (-) explained 48% of the variation in European

larch site index and foliar nitrogen (%N, +) explained 28% of the variation in red pine
site index. For ANPPO, a combination of soil (A horizon depth (+), pH (-), organic matter
(~) and growing degree days(—) variables explained 51% of variation in ANPPo for red
pine. For European larch ANPPO, 27% of the variation in ANPPo was explained by
nitrification rates (+) and specific leaf area (+). In general, these relations demonstrate
that both temperature and site/vegetation variables related to water and nitrogen
availability can explain a large percent of the variation in red pine and European larch
productivity across the Great Lakes region, and that each species and measure of

productivity respond uniquely to these variables.

INTRODUCTION

European larch (Larix decidua Miller) and red pine (Pinus resinosa Ait.) have long
been associated with reforestation and paper/pulp production due to their rapid growth on
some sites (McComb 1955, Wilde et al. 1965, Valade 1998). However, productivity
varies markedly across sites and little is known about the site factors responsible for this
variation, especially for European larch in the Great Lakes region (Carmean 1975).
Qualitative evidence suggests that European larch productivity is highly sensitive to site
quality whereas red pine is less sensitive. Thus, red pine may have greater productivity
on poor sites than European larch and vice versa on rich sites. Quantifying comparative
site-productivity relationships for European larch and red pine may be important for
silvicultural planning, if maximizing productivity on any given site is the focus. The
goals of this study were to 1) measure the site-productivity relationships for two
commercially important species for a collection of stands in the Great Lakes region, and
2) quantify below- and above-ground site variables that drive variation in productivity.

A popular method for determining the relationship between site factors and growth is
the use of multiple regression equations that use some combination of soil, vegetation,
topography, and climate variables as predictors of species productivity (usually estimated
as site index: the height of a dominant/co-dominant tree at a given base age) for data
collected across a broad range of sites. Several researchers studying various species have
had mixed success in developing strong predictive relations between site index and a host
of both below- and above-ground site variables. Some studies have found only

moderately strong relationships between site factors and site index despite exhaustive

sampling efforts (e.g. Gaiser 1950, Broadfoot 1969, Payandeh 1986, Harding et a1. 1985,
Monserud et a1. 1990). Conversely, Wang (1995) explained 83% of the variation in white
spruce (Picea glauca (Moench) Voss) site index by using combinations of soil
chemical/physical, understory vegetation, and tree foliage chemistry variables. Chen et
al. (1998) explained nearly 80% of the variation in aspen (Populus tremuloides Michx.)
site index with chemical and physical soil properties alone, and using a combination of
soil chemical and climatic variables, nearly 94% of variation was explained. There may
be several reasons for variability in site-productivity relationships: site index is an
imprecise estimate of productivity, and the predictor variables chosen may not be the
factors that truly drive variation in productivity.

One way to improve the determination of site-productivity relationships is to
simultaneously assess site index and alternative measures of site productivity. One such
alternative is aboveground net primary production (ANPP; Aber and Melillo 1991). In
the Great Lakes region, and across broad site productivity and stand composition
gradients, ANPP has been found to be highly correlated with nitrogen (N) mineralization
and soil texture (Pastor et a1. 1984, Reich et al. 1997) as well as forest floor nitrogen
(%N) and carbon (C)fN ratios (Fassnacht and Gower 1997). For closed canopy forests,
ANPP of the overstory tree species (ANPPO) is an estimate of productivity per unit
ground area. For forestry, it is unlikely that ANPPo will replace site index as a measure
of stand productivity since the latter is easy to determine and almost universal in usage.
However, our basic understanding of the factors most important in driving variation in
productivity across forested landscapes will be strengthened by the simultaneous

consideration of site index and ANPPo as indices of forest productivity.

Another reason for the often relatively weak relationships found between collections
of site variables and site index is that some of the important factors driving variation in
productivity may have been overlooked (Monserud et a1. 1990). In the Great Lakes
region, and throughout much of North America, forest productivity is likely to be
physiologically limited by: an over-abundance or lack of water; mineral nutrients such as
N; the base cations magnesium (Mg), calcium (Ca), and potassium (K) in some regions;
and/or temperature. It is rare that site-productivity studies consider a set of predictor
variables that encompass these potential limitations. Several studies have found at least
moderately strong relationships of soil chemistry, foliage chemistry, physical soil
properties, climatic and topographic characteristics with forest productivity, but few have
simultaneously considered these variables (but see Wang 1995, Fassnacht and Gower
1997, Chen et al. 1998). Examples of specific variables found closely related to forest

productivity include:

0 soil/forest floor nutrients
nitrogen (Fassnacht and Gower 1997, Reich et al. 1997, Chen et a1. 1998)
potassium (Fassnacht and Gower 1997, Chen et al. 1998)
calcium (Chen et a1. 1998)
phosphorous (Alban 1972)

° foliar nutrients
nitrogen (Wang 1995, Hebert and Jack 1998, Chen 1998)
iron/potassium/magnesium (Truong 1975a, Truong 1975b)

- depth of mineral soil to an impermeable layer (Gaiser 1950, Coile 1952, Aird and
Stone 1955, Gilmore 1992, Wang 1995)

° silt+clay content (Stoeckeler 1948, F assnacht and Gower 1997)

- topography (e.g. slope, elevation, aspect) (Day 1947, Brown and Duncan 1990,

Monserud et al. 1990, Chen et al. 1998)
° temperature or precipitation (Cook 1941, Grier and Running 1977, Gholz 1982,
Monserud et al. 1990, Churkina and Running 1998)

Efforts have been made to identify factors limiting red pine growth with soil-site
studies. In a soil-site study involving 200 sites in Connecticut, Hickock et al. (1931)
found that no single soil variable was well correlated with site index at age 15. They also
noted that nearly all sites used would have been classified as “favorable” for tree growth
and that it was only on the sandiest sites, which had lower water holding capacity, that
height growth productivity declined. In Massachusetts, Mader and Owen (1961)
concluded that red pine height and volume growth was significantly related to soil
organic matter, nutrient supply, and texture. In the Great Lakes region, van Eck and
Whiteside (1963) found that red pine growth in lower Michigan was uniform on sites
across 27 soil series in lower Michigan, but high clay content was associated with a
sharper growth decline than expected after 25 years. Also in lower Michigan, Hannah
and Zahner (1970) found that red pine stemwood production as well as site index of jack
pine (Pinus banksiana Lamb.) and bigtooth aspen (Populus grandidentata Michx.) were
significantly higher on soils with fine textural banding due to an increased supply of
moisture and nutrients. Wilde et al. (1965) found soil texture and soil nutrient
concentrations had the strongest influence in predicting site index for pine plantations in
Wisconsin. Alban (1972) showed that red pine site index in Minnesota plantations was
significantly related to soil P content.

For European larch, there are no site-productivity studies for the Great Lakes region.

However, in the northeast United States, Aird and Stone (1955) and Gilmore (1992)

found that soil depth was the most reliable site variable in predicting site index,
accounting for 82% and 57% of the variation in larch growth, respectively.

Notably, most of these site characteristics are indices, and not direct measures of
nutrient/water availability and temperature. For example, the combination of soil depth
and texture are important indices of 3 trees soil water environment. Texture largely
determines water-holding capacity while shallow soils can restrict root access to water
and/or result in water logging, either of which negatively impact productivity (Pritchett
and Fisher 1987). Thus, site-productivity studies often use several variables that are
indirectly related to stand productivity. If site variables are chosen that collectively
comprise, and closely index, the actual variables driving differences in productivity (e. g.
nutrients, water, temperature), then these combinations of site factors may more closely
predict productivity across forest landscapes (e.g. Wang 1995, Chen et al. 1998).

The objectives of this study were to determine the relationships of site variables to
European larch and red pine productivity and their interactions using a collection of fifty-
four stands (twenty-seven of each species) distributed across the Great Lakes region. The
predictor variables selected were considered to be close indices of the
resource/environmental variables responsible for limiting growth. Variables also were
selected with regard to the ease of collection in hopes that simple, applied silvicultural
tools could be developed (e.g. species selection criteria in plantation establishment given
certain site characteristics to increase productivity.) Predictor variables included soil
physical (depth to impermeable layers, texture, color, available water holding capacity,
bulk density), soil chemical (rates of N mineralization/nitrification, pH), foliar chemical

(carbon isotope ratios, %N), regional climate (water balance, growing degree-days), and

site topography (aspect, slope) characteristics. The specific hypothesis that was tested
was: Some combination of the site factors and tree characteristics (e.g. soil texture, soil
depth, water balance, growing degree-days, foliar nitrogen concentrations) in multiple
regression will explain a large proportion of the variation in productivity for both

European larch and red pine.

METHODS

Field Sites
The study consisted of paired stands, twenty-seven of each species, distributed in the

Great Lakes region from north-central Minnesota to eastern Ontario (See Figure 1.1.).
Stands were geographically paired ranging from directly adjacent to one another to <30
kilometers (km) apart. Selection criteria for stands included closed canopy, monoculture
plantations from ~15-50 years old that were even-age and uniform in height and
diameter. Stands with records or evidence of past damage (e.g. Eretlzizon dorsatum
dorsatum (Linnaeus), porcupine; Pristiphora erz'chsonii (Hartig), larch sawfly) were
avoided. The collection of stands was intended to span a site fertility gradient with local

pairs subject to similar climate and on sites of similar glacial history.

 

 

 

 

100 0 100 200 300 400 500 Miles
‘ _~.

 

 

 

 

Figure 1.1. Study locations (11 =15) containing the 54 study stands.

10

Field data were collected in 1999 and 2000 from one 10 m radius plot in each stand.
Samples were collected from within that plot systematically using predetermined random

distances from plot center at random compass bearings.

Site Productivity

In each stand, productivity was estimated with two indices: site index and overstory

aboveground net primary productivity (ANPPO).

Site index. Site index (the height of co-dominant trees at a given index age) was
determined by stem analyses on three felled undamaged, co-dominant trees. Stem
disks (5 cm thick 'cookies') were collected at the base, breast height (1.37 m), and
every 1.5 m from breast height to the top of the tree. Disks were labeled, grouped by
tree, returned to the laboratory, and then sanded until all rings were readily visible.
Ages for all cookies were determined by counting rings along two or more radii until
a common age was determined for a given cookie. Disks with slow radial growth
(i.e. tight rings) were cross-dated to identify any potential 'locally absent' rings
(Stokes and Smiley 1996) using WinDendro 6.5c Software (Regent Instruments,
Quebec, QC, Canada).

Before constructing a height by age curve for each stand, error in partitioning
the bole was addressed because the true total height associated with the ring count of
a given disk is most-likely higher than the sampled height (Dyer and Bailey 1987).
First, for all sections except the section above than the highest sampled cookie, a
method developed by Carmean (1972) was applied. This method assumes that l) on

average, a cookie will be sampled in the middle of a given year’s height growth, and

11

2) height growth per year between cookies is constant. This method has been shown
to be the most accurate (Dyer and Bailey 1987) and is described by Newberry (1991)

as follows (Eq. 1.1.):

[%i,0_n{wj (1.1.)

’ i 2 (Ir-n+1)

Hi,- = estimated height for growth ring j based on point i
h,- = height at the ith section point
r,- = number of growth rings at the ith point
j = grth ring number (with pith as starting point)
For the section above the highest sampled cookie, the following correction afier

Newberry (1991) was applied to avoid underestimating the corrected height of the

ring associated with the highest cookie (Eq. 1.2.):

 

 

(n-mi-9-5)i+(j_1)x[ (hm—hi) ] (1.2.)

With the corrected data, a height versus age graph was plotted for each stand.
Using IMP statistical software (SAS Institute, Cary, NC, USA), a sigmoid curve and
a second order polynomial were fit using the Nonlinear and Fit Y by X platforms,
respectively. Both methods utilize least squares regression to fit a curve to the data.
The curve that produced less error (sum of squares error) was selected as being the
best descriptor of height growth for a given stand. Richards’ (1959) nonlinear

function, a flexible option for summarizing growth data, was used in fitting the data

12

non-linearly. The equation, as described by Monserud (1984a), was as follows (Eq.

1.3.):

H =b,x(1—e"’2"‘)b3 (1.3.)
H = total height minus 1.37 meters (height at breast height)
A = age at breast height
e = base of the natural logarithm
b,- = parameters estimated for each stand
To avoid error associated with plantation establishment or anomalous early
growth unrelated to site characteristics, the index age was based on breast height age
rather than basal cookie age (see Carmean 1972, Monserud 1984a, Monserud et al.
1990, Gilmore 1992). The resulting index curves represent height growth over time
and at one or more base/index ages (Carmean 1975, Monserud 1984b). Final
selection of a base age for a stand was set at <5 years from calculated base age (e.g.

21 year old stand extrapolated to base age 25 but site index for a 20 year old stand

reported at base age of 15).

ANNPO. Overstory annual net primary production (ANPPO) also was calculated as a
measure of productivity. It was estimated as stem/branch biomass increment plus
foliage production of the overstory species (either European larch or red pine). To
estimate stem/branch biomass increment, published species-specific allometric
equations of stem and branch mass as a function of diameter at breast height were
used for European larch (Gower et al. 1993) and red pine (Perala and Alban 1994 as
cited by Ter—Mikaelian and Korzukhin 1997). Equations for European larch were

developed in southwestern Wisconsin in three, 28-year old replicate stands using 15

13

trees (Gower et al. 1993). As reported by Ter-Mikaelian and Korzukhim (1997),
Perala and Alban (1994) harvested 69 trees (dbh range 2-46 cm ) from the Upper
Great Lakes region (e.g. Minnesota, Michigan) in the development of the latter set of

equations. The equations were as follows (Eq. 1.4. and 1.5.):

Gower et al. 1993
logloY = a + blong (1.4.)
Y = component dry mass (kg)
X = tree dbh (cm)
a, b = regression coefficients for different components/species
Perala and Alban 1994 as cited by Ter—Mikaelian and Korzukhim 1997
M = a1)” (1.5.)
M = component dry mass (kg) 1
D = tree dbh (cm)
a, b = regression coefficients for different components
Dbh for 1999, 1998, and 1997 was measured on the cookie collected at breast
height for stem analysis by subtracting ring widths from the field measurement of
dbh in 2000. Ring widths were measured using WinDendro 6.50 Software (Regent
Instruments, Quebec, QC, Canada). Stern and branch biomass increments were
calculated as the difference of biomass accumulation between 1999-1998 (i.e. 1999
increment) and 1998-1997 (i.e. 1998 increment).
Leaf litter was collected in three 61 x 61 x 15 cm traps in November 1999.
After being returned to the laboratory within three days, litter was dried at 70 °C for

48 hours. Litter was then separated for each trap to remove non-targeted species’

litter (i.e. not European larch or red pine litter.) Samples were then re-dried at 70°C

14

to a constant mass and weighed. Litter produced per trap [grams (g) per 0.3721
square meter(m2)] was extrapolated to megagrams per hectare (Mg/ha) and averaged
across traps for a stand value.

Foliage biomass production based on 1999 litter mass was added to 1998 and
1999 stem plus branch increment to estimate ANPPO. This value was expanded to an
areal basis (i.e. Mg ha") with tree density measurements. (For the latter see
Vegetation Characteristics below.) Mean annual increment (Mg ha“ yr", MAI) was

calculated as total biomass accumulation in 1999 divided stand age.

Predictors of productivity

Five general categories of variables assumed to be important correlates of
productivity were collected. These were physical + chemical soil prOperties, and

vegetation, topographic, and regional climate characteristics.

Physical soil characteristics. In August 1999, a composite (n=18) soil sample that
included friable litter (~OJOa; Brady and Weil 1996) was collected from 0-15 cm
with an Oakfield-type sampler. A second composite soil sample (n=4) was collected
from 15 to 150 cm using a bucket auger. Where 150 cm in depth was not attainable
due to impervious substrate, the depth attained was recorded. All soil was field
sieved to 4 mm, mixed thoroughly to promote homogenization, and a subsample was
collected. Samples were air dried in the field, air dried at the laboratory (within 10
days of original sampling), and then sieved to 2mm. Litter layer (0,, Ca, 06) depth
(cm) also was measured in each stand and averaged (n=9).

Soil samples from the two strata described above were used to determine the

percent sand/silt/clay using a modified hydrometer method (Grigal 1973). A

15

proportionate amount of soil was taken from each sample stratum to total ~ 75 g and
~ 50 g air-dry soil for sandy and clayey soils, respectively. Also, a 10 g sample was
weighed and dried at 105 °C for 24 hours to determine the oven-dry/air-dry ratio
used to convert air dried samples used in textural analyses to an oven dry mass.

One hundred milliliters (m1) of a 5% sodium hexametaphosphate [(NaPO3)6]
solution, a dispersing agent, and ~ 200 ml of deionized water was added to each
sample and allowed to stand for 30+ minutes. Samples were then mixed for exactly
5 minutes using an electric mixer. The mixed solution was transferred to a settling
jar and the total volume was increased to 1000 ml. The solution was stirred
thoroughly and two measurements were taken using an ASTM No. 152H hydrometer
with a Bouyoucos scale in g per liter (1). The first measurement, taken at 40 seconds,
measures the sand content. The second measurement, taken at 4 hours, measures
clay content. After each measurement, a hydrometer reading in a ‘blank’ settling jar
containing 100 ml of dispersing agent and 900 ml deionized water was taken to
adjust for temperature and the viscosity of the dispersing agent. The “corrected”
reading is the initial reading minus the subsequent blank reading. Temperature was
also monitored with a StowAway®Tidbit® temperature data logger (Onset Computer
Corporation, Bourne, MA, USA) every 4 seconds in the ‘blanlc’ jar to ensure
changes in solution temperature were detected. Textural class (e.g. sandy loam vs.
loamy sand) was defined following Soil Survey Staff (1975) protocol. Calculations

of each textural component was as follows (Eq. 1.6., 1.7., and 1.8.):

% sand = [ 1 — (corrected 40 secbnd reading/sample oven dry weight) ] x 100 (1.6.)

16

% clay = (corrected 4 hour reading/sample oven dry weight) x 100 (1.7.)

% silt = 100— (% sand + % clay) (1.8.)

In May-June 2000, a 0.5 x 0.5 x 0.75m deep soil pit was dug in each stand in a
central, representative location. Excavated soil from the pit was sieved to 4
millimeters (mm). Volume of coarse fragments >4 mm was determined via
displacement of water in a calibrated bucket. Composite measures (across pit faces)
for thickness (cm) and color (Munsell® Soil Color Charts 1992, Kollmorgen
Instruments Corp., Newburgh, NY, USA) of the A horizon as well as depths (cm) to
impermeable layers and/or C horizon were taken. Bulk density (Db, dry soil g/ cubic
centimeter (cm3)) was calculated using the core method (Blake and Hartge 1986).
Five mineral soil core samples of a known volume (15 cm depth @ 4.7625 cm diam
= 267.2 cm3 per core) were collected using a split core sampler with a slide hammer
attachment and composited. At the laboratory, soils were dried (105 °C) to a

constant mass and weighed. Bulk density for the stand was calculated as (Eq. 1.9.):

Db = Mass of oven-dry soil (g)/volume (cm3) (1.9.)

Soil organic matter was determined in mineral soil collected during bulk density
sampling. Sample oven dry (105°C) mass was determined. Samples were then
ashed in a 400°C muffle furnace for eight hours to determine loss of carbon (Ben-
Dor and Banin 1989). After being removed from the muffle furnace and prior to
weighing, samples were cooled over CaClz for 12 hours. Percent weight loss was

calculated as: (mass/05 — mass400)lmass105. Assuming that organic matter is 1.72

17

times the amount of organic carbon in the soil (Brady and Weil 1996) and that
carbon was 48% of the volatilized material (V itousek 1982), organic matter was

calculated as follows (Eq. 1.10.):

Soil organic matter = % ash x 0.48 x 1. 72 (1.10.)

Available soil water holding capacity (AWC), the difference in soil water
content at field capacity (FC, -0.01 MPa) and permanent wilting point (PWP, -1.5
MPa), was estimated from % sand and % clay using the equations of Harding and

Grigal (1984) (Eq. 1.11., 1.12., and 1.13.):

PC = 27.85 + 0.14 (00 clay) — 0.27 (% sand) (1.11.)
PWP= 10.69+0.I6(%clay)—0.11(%sand) (1.12.)
AWC=FC—PWP (1.13.)

AWC on a volumetric basis was calculated by multiplying AWC (g water (g dry
soil)'1) and soil bulk density (g/cm3) and then subtracting the volume of soil occupied
by coarse fragments.

Soil temperature was monitored in 30 stands with StowAway®Tidbit®
temperature data loggers (Onset Computer Corporation, Bourne, MA, USA) every 8
minutes from April to July 2000. One logger per stand was placed at a depth of 7.5

cm in a representative location within the 10 m fixed-radius plot.

Chemical soil characteristics. Soil pH was measured on a sample with the 0-15 and

15-150 cm strata composited (Thomas 1996). Ten g of soil and 10 ml of deionized

l8

were mixed, allowed to stand for 30+ minutes, and measured with a calibrated
Consort P601 pH meter (Consort Instruments, Belgium) by submersing the electrode
into the supernatant while simultaneously stirring the solution with a glass rod.

Rates of N mineralization were calculated in three separate incubations: 1) one
28 day ex situ incubation for all stands with soils collected in August 1999 (the top
stratum, 0-15 cm, as explained above in Physical soil characteristics); 2) one 35 day
in situ incubation for 30 stands in May-June 2000; and 3) one 28 day ex situ
incubation for 30 stands with soils collected in May 2000.

In August 1999, soils were air dried upon collection in the field and transported
back to the laboratory within eight days. Two full laboratory incubation replicates
were performed. Two ‘initial’ and two ‘final’ 10 g samples per stand of air dry soil
were re-moistened four days prior to the start of the incubation due to the spike of
microbial activity that occurs after re-wetting. This period was intended to allow a
stabilization of the soil as air drying results in a portion of the microbial biomass to
become lysed, thus that biomass N is a portion of the potentially mineralizable pool
(Bartlett and James 1980, Campbell et al. 1993). The remaining two samples per
stand were kept moist and incubated in a dark, humid environment at 25°C +/- 1°C
for 28 days.

In May 2000, an in situ incubation was begun in 30 stands using the pipe
method (Raison et a1. 1987) to minimize soil disturbance (Binkley and Hart 1989).
Eight polyvinyl chloride (PVC) pipes measuring 5 cm in diameter x 18 cm deep
were driven to a depth of 15 cm. The tops were covered with duct tape to prevent

rain or debris from entering the pipe. A small slice in the tape was cut with a knife

19

for ventilation. Adjacent to each pipe, another pipe was driven to 15 cm and soil
was extracted by simultaneously pulling and twisting the pipe, ensuring the soil core
was intact for the entire 15 cm. These eight cores of soil were sieved to 4 mm and
thoroughly mixed to promote homogenization. One composite sample was
collected, double-bagged in polyethylene bags to prevent drying, and kept at ~1-2°C
during transport back to the laboratory. The in situ ‘final’ samples were excavated at
35 days, transported within 3 days, and extracted in the laboratory.

In addition to the 35 day in situ incubation, a 28 day ex situ incubation was
begun in May 2000 with ~10 g sample per stand for the same 30 stands. Incubated
samples were stored in a similar fashion to the 1999 ex situ samples: constantly
moist in a dark, humid environment at 23°C +/- 1°C for 28 days.

All extractions of ammonium (NH?) and nitrate (NO3') were canied out
following Mulvaney (1996). One hundred ml 2 M KC] and 10 g soil were combined
in a 250 ml Erlenmeyer flask, shaken on an orbital shaker for 1-2 hours, and filtered
through Whatrnan #42 filter paper after 30 minutes of settling. Samples were kept
cold (1-2°C) until analyzed for NH4+ and N03' with a continuous- or segmented-
flow AutoAnalyzer (01 Analytical, College Station, Texas, USA). ‘Initial’ samples
(T0) were extracted to determine standing pools of NH; and N031 ‘Final’ samples
(Tp) were extracted to determine the amount of NH4+ and N03' produced during the
incubation. Oven dry mass (g) was determined after drying (105°C) a subsample
(~10g) of soil and calculating the air-dry (1999)- or fresh(2000)-/oven-dry ratio.
Calculation of the NIL;+ and N03” concentrations on a gravimetric basis after

Robertson et al. (1999) was as follows (Eq. 1.14.)

20

ugion/g soil=(Cx V)/W (1.14.)
C = concentration of ion in extract (mg/l)
V = volume of extract (KCl plus moisture in soil sample) (m1)

W = oven-dry mass of soil (g)

NOTE: ug g'1 = ppm = mg kg"1

Net N mineralization was calculated as the difference between NI'I4+ + NO3’
concentrations of the incubated (‘final’) and original (‘initial’) soil samples, or
(N114+ + NO3’)F — (NH; + N03")0. Net nitrification was calculated as the difference
between nitrate concentrations of the incubated and original soil samples, or (NO3')F
— (NOg')o. To express rates for both net mineralization and nitrification on a daily

basis, each was divided by the total incubation duration in days.

Vegetation Characteristics. Stand, canopy. and leaf characteristics for each stand were

measured as follows.

Basal area and trees per hectare. Stand basal area (BA, m2 ha“), the cross sectional
area of trees at dbh, was determined as the average of four 20 basal area factor
(BAF) variable-radius plots using a CRUZ-ALL angle gauge (Forestry Suppliers,
Inc. Jackson, MS, USA). Average BA was converted to m2 ha’l from the following

calculation after Avery and Burkhart (1994) (Eq 1.15.):

BA per acre = (total # of trees tallied/4 plots) x 20 BAF (1.15.)

21

Trees per ha was also determined for each stand. The average BA per tree was
determined for the three selected plus trees. Total trees ha'l was then calculated on

the basis of the BA ha”1 as calculated above.

Leaf life span. Leaf life span was determined for red pine by counting the cohorts of
needles present on a subsample of branches in July and August 2000. For European
larch, a leaf life span of approximately six months was assumed (Gower et al. 1993,

Reich et al. 1999) but not needed in any canopy calculations.

Foliar nitrogen. Foliar samples were collected during the growing season after
nutrient concentrations had stabilized (Myre and Camiré 1996) and sampled
consistently with regard to crown position between trees. For all collections, needles
were separated from branches immediately and only needles were retained. Samples
for 1999 were collected using a shotgun to retrieve sun-exposed, canopy branches
from two of the three selected co-dominant trees. Larch samples consisted of short
shoot needles. Pine samples consisted of second year foliage.

Samples for 2000 were collected from the co—dominant trees felled for stem
analyses. Canopies of two trees were vertically stratified into thirds by canopy
length. Needles were proportionately collected between and within these strata and
composited (i.e. each stratum selected proportionately to volume of canopy and
needles from each branch sampled proportionately to abundance of foliage on that
branch). Larch samples consisted of short shoot needles. Each cohort for pine (n=3
or 4) was sampled separately. Collected samples were stored at 1-2 °C during

transport. After being returned to the laboratory, samples were dried at 70 °C.

22

 

Subsamples were taken for specific leaf area (SLA) and subsequent measures as
described below (See Can0py leaf area and mass and Nitrogen retranslocation.)
Other subsamples were ground using a tissue pulverizer (Kinetic Laboratory
Equipment Co., Visalia, CA, USA) and analyzed for total percent N (%N) with an
elemental analyzer (Carlo Erba NA-1500 Series II or NC-2500; CE Elantech,

Lakewood, NJ, USA). This was expressed on a mass (Nleaf.mass, mg g'l) and area

(N karma, mmol m'2 x 102) basis.

Photosynthesis. Rates of light saturated photosynthesis (Amu) were measured in the
field using a portable photosynthesis system in an open-system configuration (L1—
6400, LI-COR, Lincoln, NE, USA). Measurements were taken in the field during
July and August 2000, the point in the growing season when needles are considered
to have firlly developed and nutrient concentrations have stabilized (Myre and
Camiré 1996). Measurement protocol included the following settings: incoming
stream flow rate equal to 400 umol s", ambient carbon dioxide (C02) concentrations
set at 370 umol C02 mol'1 air, and saturating photon flux densities of 1500 umol
quanta m'2 3'1 via an enclosed LED light source in the cuvette. Following Ellsworth
and Liu (1994), measurements were taken on foliage from a subset of branches
removed from sun exposed areas of canopies. Stems from three trees were
rehydrated after removal by cutting them underwater prior to taking measurements.
Measurements were taken on short shoots for European larch and second year

foliage for red pine.

23

Canopy leaf area and mass. Leaf area was measured using WinSeedle 5.0 (Regent
Instruments, Quebec, QC, Canada) for litterfall (1999) and live samples (2000). The
samples were then dried at 70° C and weighed to determine dry mass (g). (See
Foliar nitrogen and ANNPO above for live and senesced foliage sampling protocols,
respectively.) SLA was then calculated as fresh area (cmz) per unit dry mass (g).
Leaf area index (LAI) was calculated by two methods: 1) converting litter mass
to canopy mass by correcting for mass lost during senescence and then converting to
LAI via SLA for the live canopy, and 2) extrapolating the area of the litter caught
without any correction factor assuming that the area of live and senesced tissue is the
same. LAI was also measured in the field via LAI 2000 Plant Canopy Analyzers
(PCA; Li-Cor, Inc., Lincoln, NE, USA) in August 1999. For both calculation
methods, LAI was calculated for each litter trap in a stand and averaged. The

following equations summarize the calculations (Eq. 1.16. and 1.17.):

MLcan—av
SLA x [ll-oo- x ___]

 

 

can—avg
1)LAI = "" x cohorts (Eq 1.16.)
Trap...“
SLAcan-avg = average canopy SLA (cm2 g"). .
Ml.a,,,.a,,g = live canopy foliage mass per length (mg mm'l)
litter = litter collected in fall 1999 (g)
ML”, = senesced needle foliage mass per length (mg m")
T rapma = area of litter trap (cm )
cohorts = number of cohorts counted in July/August 2000
*NOTE: SLAca,,.avg and MLWPWg calculated across all red pine cohorts.
2)LAI = SLA”! X 1m” x cohorts (Eq. 1.17.)
Traparea
SLAm = senesced needle SLA (cm2 g")

24

litter = litter collected in fall 1999 (g)
Trapma = area of litter trap (cmz)
cohorts = number of cohorts counted in July/August 2000
Standing canopy biomass also was calculated for each stand by correcting for
the mass loss of live to senesced foliage for litter captured in fall 1999 and then

extrapolating to Mg ha". This was done for each trap and then averaged for the

stand. The equation is as follows (Eq. 1.18.):

ML _
Biomass = litter x —M x cohorts (Eq. 1.18.)
In
Biomass = standing (‘fresh’) canopy mass (Mg ha”)
litter = litter collected in fall 1999 (g)
Ml.a,,,.avg = live canopy foliage mass per length (mg mm'lf
ML”, = senesced needle foliage mass per length (mg m”)
cohorts = number of cohorts counted in July/August 2000

'NOTE: MLammg calculated across all red pine cohorts.

Nitrogen retranslocation. Live canopy foliage (1999 and 2000) and senesced tissue
(1999) were used to determine percent N retranslocated. (See Foliar nitrogen and
ANNPO above for live and senesced foliage sampling protocols, respectively.) First,
on a mass basis, mass per unit length for both live and senesced foliage was
determined for both species using WinSeedle 5.0 (Regent Instruments, Quebec, QC,
Canada) to measure needle length. These samples were then dried at 70°C to a
constant mass and weighed. Average mass per unit length was then calculated. The

percent N retranslocated (% retransmass) was determined as follows (Eq. 1.19.):

25

(ML xN )—(MLl.xN.)

 

%retrans = can can I t I" x 100 (1.19.)
mass ML x N
can can
MLm = live canopy foliage mass per length (g mm‘lf
Ncan = live canopy needle nitrogen content (mg g”')‘
MLm = senesced needle foliage mass per length (g m")
Nm = senesced needle nitrogen content (mg g' )

‘Note: ML“,n and N“,” for oldest red pine cohort.

On an area basis, this was determined using nitrogen concentrations (g kg") and
the specific leaf mass (SLM - g cmz, the inverse of SLA) for both live and senesced

foliage. The % retransarea was calculated as follows (Eq. 1.20.):

 

(N xSLM HN. xSLM .)
%retrans 2 can can [It 1" x100 (1.20.)
area Ncan x SLMcan
Nam = live canopy needle nitrogen content (g kg")1
SLM“, = live canopy needle SLM (g cm2 g")I
N”, = senesced needle nitrogen content (g kg”)

SLM”, = senesced needle SLM (g cm2 g’l)
INote: N00,, and SLM“,n for oldest red pine cohorts.

The calculations above which involve a live needle/senesced needle mass per
unit length correction factor (e.g. LAI equation #1, standing canopy biomass, and %
retransmass) assume that needle length does not change when needles senesce. It also
assumed throughout that red pine senesces only the oldest cohort each season.

Total foliar N lost in 1999 was calculated by converting litter N (mg g") per
litter trap area (0.3721 m2) to N loss (kg ha") and averaged across litter traps for a

stand measure.

26

Water use efficiency. Carbon isotope discrimination (A) was determined on all live
foliar samples (n=54) collected in July and August 2000 and a subsample (n=10) of
those collected in August 1999. Samples collected in 2000 were composited across
cohorts providing one sample per stand. Samples for 1999 were only short shoots
for European larch and second year cohort needles for red pine. All samples were
collected fresh, immediately separated from the branch, and stored at 1-2°C until
being returned to the laboratory. At the laboratory, samples were dried at 70°C and
ground using a tissue pulverizer (Kinetic Laboratory Equipment Co., Visalia, CA,
USA). A 2-3 mg subsample was then prepared and analyzed for carbon isotOpe
content using a isotope ratio mass spectrometer (IRMS, Finnigan Delta Plus,
Bremen, Germany) interfaced with an elemental analyzer (NC2500, CE Elantech,
Lakewood, NJ, USA). The molar abundance ratio of carbon isotopes (”C/12C) of
CO2 was measured against a reference calibrated against the standard, PDB-
belemnite (Belemnitella americana). The carbon isotope composition (5) relative to
the PDB standard in parts per thousand (%o) was calculated using the sample and

standard ratios, Rsampie and Rsmdard, respectively: (Eq. 1.21.):

 

o'3c —( R‘“””’"" -l]x1000 (121)
sample — R ° '

3 tan dard

The molar abundance ratio was then expressed as the carbon discrimination by
the leaf (A) in %o. As noted by Ehleringer (1991), data presented in such a way
“directly expresses the consequences of biological processes”. The following

equation was used in calculating A (Eq. 1.22.):

27

13 13
A = a Cair - 5 Csample (1 22 )
1+6'3C . '

 

sample

The carbon concentration of air (513Cair) was assumed to be —8%o following
Farquhar et al. (1989) as cited by Marshall and Zang (1994), Ehleringer (1991),

Schlesinger (1997), Sparks and Ehleringer (1997), and Lambers et al. (1998).

Topographic characteristics. Stand topographical characteristics were measured in
May-June 2000. Data collected included percent slope using a clinometer. If slope
was >0°, description of slope (Grigal et al. 1999) and aspect of the plot equal to the

azimuth as measured via a compass were also recorded.

Regional climatic characteristics. Daily precipitation and temperature data for 1998
and 1999 were obtained from weather stations (n=18) 1~30 km in distance from
respective stands (Environment Canada 2000, National Oceanic and Atmospheric
Administration 2000). With these data, total annual precipitation and daily
temperature records were obtained. From temperature data, growing degrees days

(sum °C-day; GDD) were calculated as (Eq. 1.23.):

Tempmax + Tempmin
2

 

GDD =[ ]— Tempbm (1.23.)

where T empmax and T empmm are daily maximum and minimum temperatures (°C),

respectively. The base temperature ( T empbase), the temperature below which the

28

species of interest are expected to be dormant, was set at 5 °C (Perala 1985). If the
mean temperature for a day was less than T empbm, it was set equal to T empbm and
the equation was evaluated as zero GDD for that day (“Method 2”, McMaster and
Wilhelm 1997).

Water balance (WB) was estimated for each stand as the difference between the
annual moisture potential (precipitation, PRECIP; soil water storage capacity equal
to field capacity minus percent coarse fragments, F C) and outputs of moisture

(potential evapotranspiration, PET) (Stephenson 1990) as (Eq. 1.24.):

WB = PRECIP + FC -— PET (1.24.)

Each component is commonly expressed in cm of water addition'or loss (e.g. Grier
and Running 1977, Jenny 1980, Gholz 1982, Yin 1993, Hebert and Jack 1998).
PRECIP and PET are summed for the year while PC is a constant capacity for each
stand. Precipitation data described above were used to calculate annual sums of
PRECIP. PC was calculated using Eq. 8 as described in Physical soil characteristics
above after Harding and Grigal (1984). PET was calculated after Bonan (1989) and

is described therein as follows (Eq. 1.25.):

 

—1
ZXE 380 E
p I: (305 ]+(ez—el)] X[ a+( )+ (e- er)+550]x s ( )

E, = PET (cm)‘
E = elevation of measuring weather station (m)

T a = mean monthly air temperature (°C)
e 2 = saturation vapor pressure (mbar) at mean maximum daily temperature
(°C) of the warmest month of the year

29

e 1 = saturation vapor pressure (mbar) at mean minimum daily temperature
(°C) of the warmest month of the year

e: = 33.8639[(0.00738Ta + 0.8072)8 — 0.000019 1.8T, + 48| +0.0013161”
R, = mean monthly solar radiation (cal cm”2 day' converted to MJ m2 day‘l)

Note:

Data were converted to cm from original cal cm’2 day'1 by the Jenson-Haise

method (Jenson et al. 1990) as cited by Hogg (1994). Dividing the RS by A, the
latent heat of vaporization (24.54 MJ m'2 cm“), provides PET in cm.
Vapor pressures were calculated after Bosen (1960) per Bonan (1989).

it

Solar radiation data were available for six locations in the United States from
1961 — 1990 Normal Data (National Renewable Energy Laboratory 2000) and one

location in Ontario for 1998 — 1999 (Environment Canada 2000). All data were

global radiation data collected on a flat-plate collector with a tilt angle of 0°.

30

Statistical analyses
JMP (4.0) statistical software (SAS Institute, Cary, NC, USA) was used to examine

relationships between indices of stand productivity, site predictor variables, and their
interrelationships.

Pearson product-moment correlation analyses were performed to relate soil,
vegetation, and climate characteristics to site index (base ages 15, 25, and 50) and annual
biomass increment (ANPPMoml). The Pearson product-moment correlation coefficient (r),
a measure of the strength of the linear relationship between each pair of variables (Sokal
and Rohlf 1969), between predictor variables also was examined for evidence of
collinearity. The linear relationship between potentially confounding variables (e.g.
average stand age and/or climate) and productivity also was assessed in bivariate plots
and strength, of pairwise correlation coefficients. The residuals of those bivariate
relationships were used in some analyses when regressing productivity by predictor
variables to ‘extract’ that source of error.

Multiple regression models (MLR) were constructed using soil variables (% clay,
available soil water holding capacity, bulk density, % coarse fragment, depth of
A/C/forest floor horizons, % organic matter, pH, rates of mineralization/nitrification),
climatic variables (growing degree days 1999, water balance, average July temperature
1998-1999), vegetation variables (carbon isotope discrimination, foliar nitrogen, specific
leaf area) or all (n = 17) of the above as predictor variables. To conserve degrees of
freedom in the regression procedures, only predictor variables which had values for all
stands were selected (n = 27 of each species). Separate models were constructed for
European larch and red pine in four analyses with the above predictors and the predicted

variable (Y) defined as follows:

31

1) Y = productivity data as Site Index (15 & 25) ‘

2) Y = residuals of Site Index (15 & 25) by climate

3) Y = residuals of ANPPMoml by average stand age ‘

4) Y = residuals of (ANPP0-,0,a1 x AGE residuals) by climate

'NOTE: Climate variables included were precipitation and temperature to
‘extract’ effect of climate. Water balance was retained as a predictor due to it
being calculated with site variables.

The ‘mixed’ selection method (SAS Institute, Inc. 2000) was used in variable
selection as it has been shown to be a the most successful selection method compared
with ‘forward’ and ‘backward’ selection (Miller 1990) and an “objective screening
procedure” (McClave and Dietrich 1991). The entry/exit alpha (or) was set at 0.10 as
recommended by Draper and Smith (1981). After the variable selection process, a least
squares regression was performed to determine the amount of variation (adjusted for
degrees of freedom, R32) in the predicted variable that was explained by the predictor
variable(s). Residual plots were examined for variance heteroscedasticity. Final model
mean square error (MSE), R32, and p-value were recorded.

Another set of MLR models was run using a subset of ‘simple’ (i.e. easy to
measure/collect) soil variables to predict site index. Variables selected included each
component of texture (sand, silt, and clay) grouped with bulk density, pH, and horizon
(e. g. A, C, forest floor) depth. Selection of the best subset of variables in these three sets

of four variables followed the ‘stepwise’ methods described above. European larch and

red pine site index were modeled separately.

32

RESULTS

General Stand Characteristics

For the study stands, the mean age (years) ranged from 12.5 years to 67 years for
European larch and 11.5 to 65 years for red pine (Table 1.1.). For both species, most of
the younger stands were in the northern portion of the study area including Michigan’s
Upper Peninsula, northern Wisconsin, and Minnesota. Most of the older stands were in
southern Michigan and southern Ontario. Average age was similar for both species; 36.5
and 37.8 years old for European larch and red pine respectively. Average stand diameter
and height for European larch (24.6 cm and 19.5 m) were greater than for red pine (23.1
cm and 17.2 m) whereas stand density was much greater for red pine (963 trees ha'l) than
for European larch (619 trees ha'l). Despite red pine having greater stand density and
overstory production in 1999 (ANPPo-tota1, Mg ha”) was on average 85% greater in
European larch than red pine stands (Table 1.3.). In general, ANPPmotai (Table1.1.) and
site index (not shown) for both species was greatest for young stands at sites in the

northwest part of the study area.

33

Table 1.1. Stand location (nearest town), mean age, trees per hectare, diameter at breast
height (dbh) , total height, and aboveground production (ANPPo - 1999) for the 54 study
stands.

 

 

stand spp location mean trees per mean dbhI mean heightI ANPPo
' age hectare (cm) (m) (Mg ha")
1 larch Augusta, MI 59 421.1 30.6 (0.6) 24.8 (0.5) 4.6
2 pine Augusta, MI 46.5 800.6 29.6 (1.6) 21.3 (0.4) 1.6
3 larch Augusta, MI 60.5 523.7 32.5 (3.0) 24.9 (1.1) 7.4
4 pine Augusta, MI 48.5 764.9 29.0 (1.4) 20.8 (0.2) 2.6
5 pine Bangor, WI 27 1008.6 22.8 (1.6) 15.7 (0.1) 3.7
6 larch Bangor, WI 21 666.6 24.8 (0.6) 19.8 (1.1) 9.9
7§ larch Bangor, WI 34 397.1 28.1 (6.7) 24.3 (1.0) 7.9
8 pine Bangor, WI 36 1441.2 20.1 (0.4) 19.4 (0.5) 2.5
9§ larch Bangor, WI 34 227.6 36.7 (1.6) 23.3 (0.7) 5.8
10 pine Bangor, WI 35 1179.1 20.7 (1.9) 18.4 (0.8) 2.8
11 larch Argonne, WI 17 607.5 21.9 (2.0) 14.4 (1.1) 9.7
12 pine Argonne, WI 20.5 713.1 21.2 (1.4) 11.8 (0.3) 6.1
13 larch Libby, MN 15 312.1 17.6 (5.2) 13.3 (1.8) 5.2
14 pine Libby, MN 13 2063.2 12.1 (1.9) 8.5 (0.2) 7.0
15 larch Cloquet, MN 17 1135.1 15.1 (2.3) 12.6 (1.6) 8.8
16 pine Cloquet, MN 17.5 1615.1 17.3 (0.7) 10.3 (0.7) 7.7
17 larch Winter, WT 17.5 386.4 21.2 (2.5) 13.7 (1.5) 4.3
18 pine Winter, WI 17 1415.6 15.4 (1.4) 10.0 (0.2) 6.7
19 larch Brampton, MI 13 701.0 17.0 (1.6) 11.5 (0.6) 11.0
20 larch Brampton, MI 12.6 789.1 18.2 (2.2) 10.9 (0.6) 9.2
21 larch Brampton, MI 13 800.8 17.6 (0.75) 11.9 (0.4) 14.5
22 pine Trenary, MI 27 950.5 22.7 (3.6) 17.1 (0.9) 4.0
23 pine Cornell, MI 21.5 1362.0 19.4 (0.7) 12.2 (0.3) 5.6
24 pine Cornell, MI 11.5 1195.7 15.6 (1.8) 7.6 (0.4) 8.2
25 larch Newberry, MI 39 352.3 29.4 (3.1) 18.6 (1.4) 3.1
26 pine Newberry, MI 38 562.2 22.2 (1.1) 18.0 (0.3) 4.4
27 larch Turkey Point, ON 49 343.7 30.3 (4.7) 24.2 (0.9) 2.6
28 larch Turkey Point, ON 56 431.2 24.1 (2.4) 19.9 (1.1) 2.3
29 pine Walsingham, ON 36 831.4 23.6 (1.4) 17.8 (0.4) 2.7
30 pine Turkey Point, ON 65 431.5 35.4 (1.6) 25.7 (1.0) 1.2
31 larch Fonthill, ON 58 557.6 33.5 (2.6) 28.6 (1.2) 4.1
32 pine Fonthill, ON 29 502.4 20.2 (0.8) 15.1 (1.1) 1.4
33 pine Elizabethville, 0N 35.5 1501.5 20.4 (1.8) 15.9 (0.7) 2.2
34 pine Elizabethville, ON 28.5 1863.4 17.5 ( 1.0) 14.3 (0.5) 2.7
35 pine Elizabethville, 0N 33.5 1487.8 17.4 (2.1) 12.0 (0.04) 1.8
36 larch Elizabethville, ON 41 572.5 29.8 (3.3) 24.3 (0.03) 8.6
37 larch F enella, ON 67 389.1 26.6 (3.4) 23.5 (0.2) 3.0
38 larch Fenella, ON 46.5 482.0 30.9 (2.5) 22.2 (1.1) 5.2

34

Table 1.1. (cont’d)

 

 

stand spp. location mean trees per mean dbhI mean height"t ANPPo

a3 hectare (cm) (m) (Mg ha“)
39 larch Petawawa, 0N 35.5 719.6 19.6 (1.5) 20.0 (1.1) 3.6
40 larch Petawawa, 0N 44.5 577.6 24.1 (1.9) 23.5 (1.3) 3.7
41 pine Petawawa, ON 63 312.1 30.3 (5.2) 25.0 (1.5) 1.3
42 pine Petawawa, ON 64 731.8 22.3 (2.2) 20.3 (0.7) 2.1
43 larch Petawawa, ON 24 2049.4 15.6 (0.3) 16.8 (0.5) 8.3
44 pine Petawawa, 0N 61.5 818.4 26.4 (1.2) 23.3 (0.7) 2.8
45 larch Siloam, ON 45 396.5 25.0 (1.4) 17.8 (0.5) 3.1
46 pine Siloam, ON 47 676.7 23.9 (4.5) 19.9 (0.5) 2.2
47 larch Coppins Cs., ON 59.5 434.0 27.1 (3.5) 20.5 (0.6) 4.2
48 pine Coppins Cs., ON 64.5 298.6 32.5 (5.5) 24.2 (2.0) 2.7
49 larch Palgrave, ON 33 926.7 20.9 (3.0) 18.4 (1.6) 8.8
50 pine Palgrave, ON 32 986.7 22.4 (0.5) 16.5 (1.1) 3.7
53 larch Monticello, ON 37 950.4 22.4 (3.2) 20.5 (1.5) 8.6
54 pine Bellwood, ON 51 214.8 34 (0.7) 22.0 (0.8) 1.3
55 pine Bellwood, ON 50 281.8 29.6 (2.2) 21.3 (0.4) 1.5
56 larch Stirton, ON 35.5 570.5 25.2 (3.2) 23.7 (1.3) 3.8

 

Note: Age, dbh, and total height are means of the three selected co-dominant trees
selected for stem analysis. ANPP0 is calculated with published allometric equations
assuming that the co-dominant stem analysis trees are representative of the overall

stand.

I mean (i 1 standard deviation)
§ Two co-dominate trees sampled.

35

Table 1.2. Definition of terms and units.

 

 

Acronym Parameter Units
8115,2150 site index (base ages 15, 25, or 50) m

ANPPMom total overstory annual net primary production Mg ha‘l 1999'l
ANPPMm stem component of ANPPo-total Mg ha" 1999’l
% sand sand fraction soil texture %

% silt silt fraction of soil texture %

% clay clay fraction of soil texture %

AWC available water holding capacity cm3 m'3

D, bulk density of soil g cm'3

pH pH of soil --

% CF soil coarse fragments (> 4 mm) %

Adept}, depth to A horizon cm

Cdepu1 depth to C horizon cm

FEM,“l depth of forest floor cm

% OM organic matter fraction of soil %

NMIN nitrogen mineralization mg kg1 day'l
NIT nitrification mg kg'1 day”1
A foliar carbon isotope discrimination %o

leaf %N nitrogen fraction of leaf %

Nleapmass mass-based leaf nitrogen mg g'1

Nleapma area-based leaf nitrogen mmol m"2 x 102
annual N loss annual nitrogen loss through litterfall kg ha’1

canopy biomass mass of fresh canopy Mg ha'1
SLA.mopy live canopy specific leaf area cm2 g'l
SLAlitter senesced tissue specific leaf area cm2 g'1

LAI leaf area index m2 m“2

JULY average July temperature °C

GDD growing degree days count (°C)
WB site water balance cm

 

36

Productivity

ANPPMom] (stems + foliage) and ANPPOMm (bole and branches) ranged seven fold
within species. Over 75% of annual aboveground production was in bole and branches
for both species. Site index at a base age of 15 (S115) ranged two fold within species and
was approximately 50% greater for European larch than red pine. Sample size was
smaller for site index at base age 25 (S125) and 50 (8150), but values were still greater for

European larch than for red pine. (Table 1.3.)

Table 1.3. Measures of productivity for European larch and red pine species.

 

 

 

European larch red pine
n mean range 11 mean range
mean annuallm°fl°m°m 27 4.66 1.82 - 8.65 27 3.56 1.36 - 6.53
(Mg ha yr ) '
ANPPo-total
(Mg ha_, 19994) 27 6.35 2.32 -1451 27 3.43 1.18 - 8.21 .
ANPPO‘StCm
(Mg ha,, 1999-1) 27 4.84 1.67 — 11.96 27 2.74 0.95 — 6.60
3115 27 13.22 9.36 - 17.61 27 9.34 6.57 - 13.24
(base age 15, m)
3125 18 17.70 13.85 - 22.21 23 14.21 10.24 — 17.69
(base age 25, m)
3150 7 22.88 18.70 — 28.04 10 21.92 18.20 - 23.49
(base age 50, m)

 

Boldface indicates significantly different means (Tukey-Kramer HSD, or = 0.05).

Physical Soil Characteristics

For both species, the majority of the stands were found on sandy loam soils (n = 11
and 10 of 27 stands for European larch and red pine each, respectively.) There was a
weak trend for European larch stands to be on lighter (i.e. lower bulk density), more
developed (i.e. greater horizon depth) soils, with fewer coarse fragments, than red pine

stands. Soil supporting European larch stands also had, on average, greater available

37

water holding capacity (AWC), higher % organic matter, greater spring temperatures, and
lower forest floor depth than red pine stands. However, only forest floor depth and mean
soil temperature, characteristics primarily due to species effects, were significantly

different between species (Table 1.4.).

Table 1.4. Physical soil characteristics in 27 stands each of European larch and red pine.

 

 

 

European larch red pine

mean range mean range
sand (%) 63.90 16.28 — 94.67 67.90 12.81 — 95.99
silt (%) 27.05 3.33 - 63.69 24.20 2 — 55.76
clay (%) 9.05 0.67 — 41.86 7.90 1.34 — 39.79
AWC (em3 m'3) i 7.74 2.44 — 16.82 6.95 2.04 — 17.24
bulk density (g cm'3) 1.18 0.83 — 1.41 1.22 0.9 — 1.51
coarse fragments (%) 3.47 0 — 25.62 4.09 0 — 30.58
depth toA(cm) 15.15 3—33 13.81 4—36
depth to C (cm) 47.52 19 — 80 44.93 14 — 75
forest floor depth (cm) 2.81 1.17 — 5.56 4.53 2.61 — 6.5
organic matter (%) 3.37 1.54 — 6.20 3.32 1.16 — 7.23
mean soil temp (°C) § 5.32 -0.22 — 11.15 4.37 -03 -— 9.98

 

Boldface indicates significantly different means (Tukey-Kramer HSD, or = 0.05).

IAvailable water holding capacity.

§ Mean daily temperature at 7.5 cm depth from 4/13/00 to 4/30/00 for 14 stands of

European larch and 15 stands of red pine.

Chemical soil characteristics

In laboratory incubations, nitrogen mineralization and nitrification rates were greater
for soils from European larch stands than red pine stands (Table 1.5). Rates were greater
in 1999 versus 2000. This discrepancy may be due to the different times the soils were
collected (August 1999 versus May 2000) or by differences in handling protocol (air

dried and remoistened in 1999 versus field-fresh samples in 2000). There was a weak

trend of slightly higher soil pH in European larch than red pine stands. (Table1.5.)

38

Table 1.5. Mean chemical soil characteristics of species across stands.

 

 

 

European larch red pine
mean range mean range
1999 N mineralization (mg kg" day?)1 1.50 0.41 - 3.11 1.15 0.35 -— 5.30
1999 nitrification (mg kg'] day“) 1.63 0 — 3.15 0.94 0 — 3.52
2000 N mineralization (mg kg" day") 1 0.66 -0. 13 — 1.16 0.45 0.05 - 0.97
2000 nitrification (mg kg" day") 0.65 0 — 1.27 0.26 0 — 0.62
pH 5.83 4.64 — 7.39 5.74 4.27 — 7.55

 

Boldface indicates significantly different means (Tukey-Kramer HSD, or = 0.05).

l Removing red pine outliers in 1999 (1 of n = 54) and 2000 (2 of n = 30), rates of N
mineralization in European larch soils were significantly greater than red pine (1999;
mean = 0.989 mg kg" day"; 2000, mean = 0.3816 mg kg" day").

Vegetation characteristics

Compared to red pine, European larch foliage had greater carbon isotope
discrimination (A), mass-based fresh canopy and senesced leaf litter N, canopy and litter
SLA, leaf area index, and %N retranslocated but lower area-based N and mean canopy

biomass versus red pine (Table 1.6.).

Table 1.6. Vegetation characteristics of European larch and red pine in all stands.

 

 

 

European larch red pine
mean range mean rang:

foliar A (%o) § 21.03 19.33 — 22.22 20.02 18.76 — 21.18
foliar nitrogen (%) 1.94 1.26 — 2.66 0.96 0.77 — 1.42
foliar nitrogen (mrnol m-2 x 102) 1.13 0.79 — 1.66 1.91 1.38 —- 2.92
litter nitrogen (%) 0.57 0.30 - 1.09 0.37 0.29 — 0.53
foliar N turnover (kg ha" yr") 12.23 5.65 — 23.45 7.05 1.89 — 14.20
canopy SLA (em2 g") 3 123.86 102.29 — 161.71 36.09 32.02 — 43.24
litter SLA (cm2 g") 3 117.29 90.26 — 145.96 36.25 31.39 — 44.30
leaf area index 4.08 1.99 — 6.24 2.82 0.63 — 4.56
canopy biomass (Mg ha") 3.31 1.72 — 5.46 7.95 1.73 — 14.25
% nitrogen retranslocated 79.55 58.50 - 93.21 72.05 58.16 — 83.08

 

Boldface indicates significantly different means (Tukey-Kramer HSD, or = 0.05).
§Carbon isotope discrimination.
1 Specific leaf area.

39

Topographical characteristics

Nearly half of the stands for each species had a slope of 0%; 60-80% of the
remaining stands were located on a shoulder slope with a mean slope of ~4% and a
southeast aspect. Due to the very skewed distribution (mostly 0’s) and small % slope in
the remaining stands, topographical characteristics were not used in fiirther analyses

(Table 1.7.).

Table 1.7. Topographical characteristics of study stands by species.

 

 

 

European larch red pine
mean range mean range
slope (%) 4.37 0 - 14 4.81 O — 12
aspect (azimuth) 161.33 6 — 346 150.19 18 — 360

 

Climatic Characteristics

The climate across the sites can be generally characterized by cold winters and
warm, humid summers (Table 1.8.). Sites in the northwestern part of the study (e.g.
Minnesota and Wisconsin) were generally coldest (average VJan. temp -—9.83 °C). The
lowest average precipitation and stand water balance were in southwestern Ontario,

763.19 mm and 41.48 cm, respectively.

Table 1.8. Mean climatic characteristics for 1998 — 1999 across the study area.

 

 

mean range
annual precipitation (mm) 833.74 675.3 — 970.15
' January temp (°C) - 7.30 -10.89 - -2.77
July temp (°C) 21.11 18.50 — 22.98
growing degree days 2390.84 1925 -— 2823.06
water balance gm) 48.80 24.70 - 70.41

 

Note: Data are from 18 climate stations <30 km from study areas.

40

 

Predicting productivity

Three approaches were taken to predict productivity: I) investigate the effects of
single variables in predicting ANPPo-m1 and 8125, 2) use combinations of variables in
stepwise multiple linear regression to predict ANPPonom. and S125, and 3) use a subset of

simple to measure variables to predict S125.

Relationships of individual soil, vegetation, and climate variables with productivity.
In general, correlations of site index to soil, climate, and vegetation variables were
stronger for 8125 than for S115 or 8150 (Table 1.9.; See Table 1.2. for definition of terms).
Average July temperature and A were associated with both 8125 and Slso for European
larch. NIT, %clay, and FFdem1 were all associated with 8115 and 8125 for European larch.
No predictor variable was significantly correlated with $115 or S150 for red pine.
Temperature variables (e.g. average July, GDD) were positively correlated with 8125 and
8150 for European larch but negatively correlated with ANPPonomt and the ANPPonotat —
age residuals for red pine. NTT, SLA, and Cdepm and temperature variables were
associated with ANPPo-toml and the ANPP0_,om1— age residuals for European larch and red
pine, respectively.

Pearson correlation analysis was used to examine collinearity among predictor
variables. Some variables exhibited strong co-linearity. AWC was correlated with sand
(r = -0.9711, r = -0.9714) and clay (r = 0.7467, r = 0.8699) for European larch and red
pine, respectively. The correlation between average July temperature and GDD was also

strong (r = 0.9573).

41

 

 

Predictor variables were also correlated with the residuals of a smoothing spline (A =
3000) fit to ANPPonoml versus age since there was a strong negative relationship of
ANPP.-....1 with stand age for both European larch (R2 = 0.473) and red pine (R2 = 0.846)
stands. The correlation between site index and stand age was not significant as is
expected since it is determined retrospectively at a common base age (e.g. 15, 25, 50)
with stem analysis. In general, the same variables that were correlated with ANPPonom.

were correlated with the residuals of ANPPonoml versus age.

42

   

.Ao_.o u do zoom—oboe anaEwm 8:365 wowed—em .288 mo FEES-woo So .m._ 038. 8m
.3388 “Eon 28 :23 :82»:ch do: coomEtoUH .Aooom n 6 053m a «23 E owm one; omega 3 acreaEZam oo £9683;

 

 

 

 

 

43

 

 

mmmd- mmmd ohod onod mm_d- flood mood mood- mmd omod m3 do

mmd oEd- woo .o- mood- oovd- wmod ovd mood mmd- oo_d DQO do

md _ood- mood- omvd- bond- bod vmod m.o wood- vmmd 33 who
vood- momd- omod E _ .o oomd- momd- w o od- om _ .o Sod vmvd aim
oomd 55d 2 md omod- vvod- mood- wood mvmd vwmd bond 2.x.
on o .o- 2: .o mvmd mvmd m H md- Sod- omd- omod- vmmd mmod <
mo o .o w mod- vo _ .o- vvmd- owed- vmbd- mwmd- oood- vwmd obmd mm
mm o .o mhmd de vo o .o- o3 .o- o omd mmvd ovmd mvmd wovd .52 do
woo .o- mvvd mod Rod- ovod m2d vfld Pod- God mood 752 do
3: .o- Sod wood- Rmd- mood wood S md mood wvmd mood 20 .x.
mmod moo .o- vwmd- oomd- ovod om o .o- vovd- howd- mvvd- o _o.o- Lace-om
wood 53.? mmmd- mvmd- m2 .o- vmmd oood o mmd- movd- vomd- 5..er
mood m3 .o- Zamd- ommd- mvmd m2 .o- mmmd- vad- wmmd- oomd- soon/x
mmod m _o.o- Rod- Sod mid mmd vod of .o- mo od- mood- mu .xo
mmo .o ood wood 2 od Sod- mmmd- _oo.o momd vvmd mo md DO
mmmd Sovd momd :ld- Hmmmd- ommd momd ommd oomd oomd U>><
Emd movd oomd oood- mood- mood mood wand \romd oi .o >30 .x.
2md- o_ md mod ommd- “wond- wwmd mmmd mood ovmd Rmd Em Ax.

mmd vwmd- :ld- 3d 2 Md vd- ovmd- mmmd- oomdw oomwd- ocmm ax.

a w o . $8 a m . 58
3m Rom 3m Avaoemmmmzqo am” _I _Momw%mm< 3m 2% 5m Aflnamemhmmzwa %mmm.1_mecw%mm<
05m :82 cmomoSm

 

.86on 3 3:88 moi—at? .88an can bgoosoeo me 8:53.: So $56508 coca—280 EoEoE-oozooa sea-Sum do 033-

 

Multiple Linear Regression. Climate variables JULY and GDD as well as WB (a
combination of climate and soil characteristics), explained the most variation (71.9%) in
European larch 8125 (Table 1.10., Y group 1). For red pine, the combination of GDD and
foliar %N was the best predictive model for S125 explaining over 30% of the variation in
site index (Table 1.10., Y group 1). Removing the effects of climate by using the
residuals of 8125 predicted by JULY, and GDD; A and WB explained 48.2% of the
variation in 8125 residuals for European larch (Table 1.10., Y group 2). Regressed against
the same residuals for red pine, foliar %N alone explained 28.2% of the variation in S125
(Table 1.10., Y group 2). Notably, models predicting 8115 could not be developed in most
cases since no variables met stepwise criteria of P < 0.10 to enter and P > 0.10 to exit.

For European larch, NIT and SLA explained 27% of variation in the ANPP-Mom;
residuals of the ANPPMM versus age relationship (Table 1.10., Y group 3) whereas, for
red pine, the combination of Adepth, pH, % OM, and GDD explained 51.3% of the
variation in ANPPo--om residuals (Table 1.10., Y group 3). An age and climate extracted
data set was obtained by first obtaining residuals from the ANPP-Howl versus age
relationship (smoothing spline fit, A = 3000). The next step was to regress these residuals
using stepwise multiple linear regression on climate variables and finally saving the
residuals of the model as an ‘age then climate extracted’ data set. When “extracting” the
effects of both age and climate from ANPP-gm] for red pine, A and pH were still able to

explain 38.7% of the variation in productivity (Table 1.10., Y group 4).

.83 v so
EmoEcwm 333:2, =< .mzcoo mo coEsmoo So .m; 033. 8m do .I- < .coofiowg n > 638:0 n U dam u m Amofimtg 8860a
mod 25- .oouombxo 038:0 one owe I v 688:5 owm I m 680838 038:0 I m ”3% =« I _ ”CC 283:; 38605 ”m HO Z

 

 

 

 

 

 

 

 

 

mmood and $2 5352 126:3 1 one n 3852.4 < 23
e
-- -- -- $2 1 seammz< -- e22
385v 22 a no Azoégod - 88:85 .. £383 1 283;: + ammo n _seamaz< < 25
.8.
88¢ Em memo Emmoemod + Ezommg + ES. 1 .8452... < :82
$86 of ammo Azéommm + 98m- 1 Em >
ocfi
23¢ 22 mo; Azémzm-éamiEm >
N
£86 3: $4.0 Emmi 1 85mm: + %o.: u Em <
:82
-- -- -- <2 1 Em --
0805 2m 82 Azgoom: 18983 t :3; .1. Em < o o
E
-- -- -- <2 1 Em --
a
585v a: a; 85:: .o + :59de .3583 + Race- ..... Em o
:83
:3 2 .m $3 daemmomcod 1 «3.2 u Em m .<
m mm: mm Eco: 25 an >

 

6:3 to.“ one :82 cmomoSm oo 3352603 wcoomooa 8o mcotmsoo commmewom 032:8 “mowcobm do ._ 038.

45

 

 

To build simpler models intended as silvicultural tools, relatively easy to measure
site factors were used to predict S125. Variables considered included % sand, % clay, %
silt, Db, pH, Adcp-h, Cdcpm, PFC-6pm. Using these variables, % sand and pH were the best
predictors of S125 for European larch. A model with only %clay was the best predictor of

8125 for red pine but explained variance was low (Table 1.11.).

Table 1.11. Multiple regression equations for 'simple' measures of predicting
productivity (S125) for European larch and red pine.

 

 

spp. Model Rf MSE P
larch 8125 = 29.18 - 0.074(%SAND) — 1.19(pH) 0.469 3.56 0.0034
pine 8125 = 13.631 + 0.067(%CLAY) 0.122 2.89 0.0573

 

NOTE: Model components: %SAND — percent sand of soil texture; pH —- soil pH,

%CLAY — percent clay of soil texture. All variables significant (P < 0.10).

Given the potential confounding effects of climate across the geographic gradient of
sites, the same predictor variables were regressed against the residuals of 8125 predicted
by climate for a conservative analysis of soil variables predicting 8125. The amount of
variation explained for predicting EurOpean larch and red pine 8125 decreased (model R,2
= 0.15 and 0.116, P = 0.0602 and 0.0624, respectively). Percent sand and % clay were
the only significant variables selected via stepwise regression for European larch and red
pine, respectively. However, both were significant with or without the effect of climate

in the data.

46

DISCUSSION

Productivity
Compared to red pine, European larch achieved greater net productivity and

greater height grth rates, especially in earlier stages of plantation growth. In only one
study area did the height growth of red pine exceed that of European larch at the age of
15 or 25 (SI 15 for stands #54 @ 16.11 m and #53 @ 14.72 m, respectively.) At 50 years,
in one of three study areas with both species old enough for S150 calculations, red pine
height grth surpassed that of neighboring European larch (21.8 m vs. 18.7m. for stands
#46 and #45, respectively.) However, the mean annual increment (MAI, Mg ha'1 yr'l) for
both species was 2.23 Mg ha’1 yr'l. In other study areas European larch stands generally
had higher rates of grth than red pine at all ages. For example, for two neighboring
~40 year old stands mean annual increment (MAI) for European larch was over 20%
greater than the MAI of red pine stands (stands #36 and #33: 5.8 vs. 4.6, Mg ha'1 yr‘1 at
41 and 35.5 years, respectively).

Measures of productivity for both species for this study are in the range of estimates
found in other studies (Table 1.12.). Results reported from Miller et a1. (1993) are from
European larch plantations on shallow peat soil and had been thinned regularly. Gower
et al. (1993) data are from one study area used in this study (southwest Wisconsin.)

Alban and Laidly (1982) and Schlaegel (1975) data are from Minnesota.

47

Table 1.12. Comparison of biomass production data for European larch and red pine.

 

Productivity Stand Age

 

 

We have) (years) Study
1.82 — 8.65 12 — 67 this study
European larch 6.82 28 Gower et al. 1993
1.77 63 Miller et a1. 1993
1.63 19 Miller et al. 1993
1.36 — 6.53 12 — 65 this study
red pine 5.85 28 Gower et a1. 1993
4.56 43 Alban and Laidly 1982
3.68 40 Schlaegel 1975

 

Predictors of productivity — comparison by species

Although the two species occupied a similar range of soils in terms of physical
mineral soil characteristics, differences in species leaf litter chemistry and leaf habit may
have been largely responsible for species differences in microclimate, organic soil
components, and N mineralization. H0wever, differences in soil are modest compared to
differences in site index and ANPPO. Red pine leaf litter was significantly deeper, a
result of more recalcitrant, lower N material necessary for foliage to persist for 3 to 4
years (Scott and Binkley 1997, Fassnacht and Gower 1999). During leaf off, the
deciduous habit of European larch promotes insolation throughout the winter and early
spring — increasing soil temperature and increasing the amount of moisture infiltration
(Sartz and Tolsted 1976). The combination of higher litter quality (higher N) and higher
spring temperatures promotes higher NMIN and NIT for European larch as warmer,
moister soils are beneficial for the soil microorganisms which control soil N availability
in forest ecosystems (Linn and Doran 1984, Barnes et al. 1998, Zak et al. 1999). Thus,
the significantly different FFdep-h and NMIN/NIT rates are likely species influenced given

the higher Nleapmass, litter N, annual N turnover, and deciduous habit of European larch.

48

Correlation of soil, vegetation, and climate variables with productivity

In general, moisture as well as the physical soil factors controlling moisture were
strong predictors of productivity for both species. The correlation of higher productivity
with greater moisture availability in the largely sandy soils encountered should not come
as a surprise because tree grth and microbial activities providing nutrients are moisture
dependent (Zak et al. 1994). Other studies have reported increased productivity with
moisture in the Great Lakes region (e.g. Pastor et al. 1984, Fassnacht and Gower 1997).
Results of this study also follow Marshall and Zhang (1994) and Kloeppel et al. (1998)
with an evergreen coniferous species (red pine) exhibiting greater water use efficiency
(defined as lower A) compared to a deciduous conifer. This suggests that European larch
uses water less efficiently on a leaf area basis relative to red pine.

There are many reasons why some predictor variables were not correlated with all
of the chosen measures of productivity. First, sample sizes differed between measures
(e.g. site index at different base ages.) An explanation of why $115 was not sufficient in
differentiating site quality may be due to the fact that forest plantations are still relatively
young at 15 years. Variation inherent in site fertility/quality, may not yet be fully
expressed. Resources (e.g. nutrients, moisture) even on relatively poor sites may be
relatively abundant in young, monoculture plantations as compared with older stands
since conditions are right for high mineralization rates and demand is low.

The productivity of both species also was correlated with NMIN and/or NIT.
There was a stronger correlation of productivity with NIT than with NMIN for European

larch, while the opposite was true for red pine. The results for European larch are

49

 

 

opposite of what Zak et‘ al. (1989) observed in nine upland deciduous forests in northern
Lower Michigan. Similar to Zak et a1. (1989), rates of NMIN and NIT from this study
are from ex situ incubations that represent a potential versus a field measure. One
explanation for the stronger correlation between NIT and productivity for European larch
than red pine is that European larch is more efficient in NH} and NOg‘ uptake compared
to some species (e.g. Scots pine, Pinus sylvestris L.) (Malagoli et al. 2000). Thus,
European larch may be able to utilize both forms of inorganic nitrogen more effectively
than red pine.

One of the puzzling outcomes of correlating European larch productivity with site
variables was the consistent negative correlation with Adept-1 or Cdep-h. Aird and Stone
(1955) and Gilmore (1992) found measures of soil depth as their best positive predictors
of European or Japanese/European larch site index, respectively. However, Aird and
Stone (1955) used log transformed soils depth data that excluded soils “characterized by
coarse texture or shallowness to rock”. The predicted variable, $150, was developed
anamorphically from pure or mixed EurOpean and Japanese larch stands spread across
southern New England (mostly New York) with an average age of 20 and 15 years,
respectively. Gilmore (1992) noted that there was no significant correlation between
solum thickness (~depth to C) and 8120, but it was found to be an important factor in site
quality assessment with discriminant analysis.

Lastly, another sharp contrast in the correlation analyses was the positive association
of temperature variables with European larch SI25 productivity but the opposite for red
pine ANPP-mm]. While no clear explanation is available, the answer may be related to

moisture supply for red pine more sensitive in warmer temperatures.

50

 

 

Prediction of productivity in multiple linear regression

Similar to the correlation analyses, connections can be drawn with moisture and
the best predictors of European larch productivity and, to a lesser extent, red pine. The
best model for European larch explained over 70% of variation in 8125 using climate
variables, including an increase in WB. However, this model involves an increase in one
temperature variable (JULY), a decrease in the other temperature variable (GDD), and
greater WB occurring simultaneously. This may be a result of two collinear factors (e.g.
average July temperature and GDD) being used as predictor variables. When ‘extracting’
the effect of climate, an increase in WB and greater water use efficiency (decrease in A)
both involve increased moisture and explain nearly half of the variation in S125 for
' European larch. Regarding the predictor variables for red pine 8125 or ANPPo_total, a
decrease in temperature (GDD) could be related to increases in moisture as
evapotransporation would be lower.

In this study foliar %N was a significant predictor of red pine site index, a result
similar to what Wang (1995) and Chen et al. (1998) found for white spruce and trembling
aspen, respectively. For European larch productivity, the increase of both SLA and NIT
explain 27% of variation in ANPPMmL Potential explanation of this relationship could
be increased productivity related to an increase in the canopy area available for

photosynthesis but also increased inorganic N supply from higher NIT.

51

 

 

Conclusion

On average, European larch was more productive than red pine. The only average
site variables that differed between species were greater FFdepth and lower NMIN/NIT
rates for red pine than European larch. These characteristics may have been strongly
influenced by differences in species foliar chemistry. Compared to red pine, European
larch had higher A, Nleapmass, litter N, annual N turnover, SLAcanopy, SLA-(m, LAI, % N
retranslocated and lower Nleafma and canopy biomass.

In multiple linear regression analyses, up to ~70% and ~50% of the variation in
productivity was explained for European larch and red pine, respectively. In more
conservative analysis, where potential confounding effects were removed (e.g. average
stand age and/or climate), ~40-50% of variation in growth was explained. However,
excluding climate also may remove some of the effects of site variables that covary with
climate, and thus should be interpreted with caution.

While the majority of variation in productivity was unexplained for certain models
using multiple linear regression, moisture and nitrogen avaialability were identified as
important predictors of productivity. For species selection criteria in plantation
establishment, European larch outgrew red pine consistently over a broad range of site

qualities.

52

 

 

 

 

REFERENCES

53

 

 

Aber, JD. and J.M. Melillo. 1991. Terrestrial Ecosystems. Saunders College
Publishing, Philadelphia. 429 pp.

Aird, P. L. and E. L. Stone. 1955. Soil characteristics and the growth of European and
Japanese larch in New York. J. For. 53: 425.429.

Alban, D. H. 1972. The relationship of red pine site index to soil phosphorus extracted
by several methods. Proc. Soil Sci. Amer. 36: 664 - 667.

Alban, DH. and PR. Laidly. 1982. Generalized biomass equations for jack and red pine
in the Lake States. Can. J. For. Res. 12: 913 - 921.

Avery, TE. and HE. Burkhart. 1994. Forest Measurements. 4th ed. McGraw-Hill, Inc.,
New York. 408 pp.

Barnes, B.V., D.R. Zak, S.R. Denton, and SH. Spurr. 1998. Forest Ecology, 4th ed.
John Wiley & Sons, Inc., New York. 774 pp.

Bartlett, R. and B. James. 1980. Studying dried, stored soil samples — some pitfalls. Soil
Sci. Soc. Am. J. 44: 721-724.

Ben-Dor, E. and A. Banin. 1989. Determination of organic matter content in arid-zone
soils using a simple “loss-on-ignition” method. Commun. In Soil Sci. Plant Anal. 20:
1675-1695.

Binkley, D. and SC. Hart. 1989. The components of nitrogen availability assessments in
forest soils. Adv. Soil Sci. 10: 57 — 112.

Blake, GR. and K.H. Hartge. 1986. Bulk Density. In A. Klute (ed.) Methods of Soil
Analysis. Agron. Monog. 9, Part 1. American Society of Agronomy. Madison, WI.

Bonan, GB. 1989. A computer model of the solar radiation, soil moisture, and soil
thermal regimes in boreal forests. Ecol. Model]. 45: 275 — 306.

Bosen, J .F . 1960. A formula for approximation of the saturation vapor pressure over
water. Mon. Weather Rev. 88: 275 — 276.

Brady, NC. and RR. Weil. 1996. The Nature and Properties of Soils. 1 1th ed. Prentice
Hall, Upper Saddle River, NJ. 740 pp.

Broadfoot, W.M. 1969. Problems in relating soil to site index for southern hardwoods.
For. Sci. 15: 354-364.

Brown, J.H. and CA. Duncan. 1990. Site index of red pine in relation to soils and
topography in the Allegheny Plateau of Ohio. North. .1. Appl. For. 7: 129 — 133.

54

 

 

 

Campbell, CA, B.H. Ellert, and Y.W. Jame. 1993. Nitrogen Mineralization Potential in
Soils. In M.R. Carter (ed.) Soil Sampling and Methods of Analysis. Lewis
Publishers, Boca Raton. 823 pp.

Carmean, W.M. 1972. Site index curves for upland oaks in the central states. For. Sci.
33: 3-13.

Carmean, W. H. 1975. Forest site quality evaluation in the United States. Adv.
Agronomy 27: 209-269.

Chen, H.Y.H., K. Klinka, and RD. Kabzems. 1998. Site index, site quality, and foliar
nutrients of trembling aspen: relationships and predictions. Can. J. For. Res. 28:
1743 — 1755.

Churkina, G. and SW. Running. 1998. Contrasting climatic controls on the estimated
productivity of global terrestrial biomes. Ecosystems 1: 206 - 215.

Coile, TS. 1952. Soil and the growth of forest. Adv. Agronomy 4: 329-398.

Cook, DB. 1941. Five seasons’ growth of conifers. Ecology 22: 285-296.

Day, W.R. 1947. On the effect of changes in elevation, aspect, slope, and depth of free-
rooting material on the growth of European larch, Japanese larch, Sitka spruce, and

Scots pine in Mynydd Ddu Forest. Forestry 20: 7-20.

Draper, NR. and H. Smith. Applied regression analysis, 2nd edition. John Wiley &
Sons, New York, 709 pp.

Dyer, ME. and KL. Bailey. 1987. A test of six methods for estimating true heights
from stem analysis data. For. Sci. 33(1): 3-13.

Ehleringer, J .R. 1991. 13C/ 12C Fractionation and Its Utility in Terrestrial Plant Studies.
In DC. Coleman and B. Fry (eds) Carbon Isotope Techniques. Academic Press, Inc.
San Diego. 274 pp.

Ellsworth, D.S. and X. Liu. 1994. Photosynthesis and canopy nutrition of four sugar
maple forests on acid soils in northern Vermont. Can. J. For. Res. 24: 2118-2127.

Environment Canada. 2000. Temperature, precipitation, and solar radiation data.
Ontario Climate Centre. Downsview, Ontario.

Farquhar, G.D., J .R. Ehleringer, and KT. Hubick. 1989. Carbon isotope discrimination
and photosynthesis. Annu. rev. plant physiol. plant mol. biol. 40: 503 - 537.

55

 

 

Fassnacht, KS. and ST. Gower. 1997. Interrelationships among the edaphic and stand
characteristics, leaf area index, and aboveground net primary production of upland
forest ecosystems in north central Wisconsin. Can. J. For. Res. 27: 1058-1067.

Fasshacht, KS. and ST. Gower. 1999. Comparison of the litterfall and forest floor
organic matter and nitrogen dynamics of upland forest ecosystems in north central
Wisconsin. Biogeochemistry 45: 265 - 284.

Gaiser, RN. 1950. Relation between soil characteristics and site index of loblolly pine
in coastal region of Virginia and the Carolinas. J. For. 48:271-275.

Gholz, H.L. 1982. Environmental limits on aboveground net primary production, leaf
area, and biomass in vegetation zones in the Pacific Northwest. Ecology 63: 469-481.

Gilmore, D.W. 1992. Soil-site relationships for European larch plantations in Maine.
University of Maine Thesis.

Gower, S.T., P. B. Reich, Y. Son. 1993. Canopy dynamics and aboveground production
of five tree species with different leaf longevities. Tree Physiol. (12): 327-345.

Grier, CG. and SW. Running. 1977. Leaf area of mature northwestern coniferous
forests: relation to site water balance. Ecology 58: 893 - 899.

Grigal, DE. 1973. Hydrometer method of particle analysis. Minnesota Forestry
Research Note 245. University of Minnesota, St. Paul.

Grigal, D.F., J.C. Bell, R.J. Ahrens, R.D. Boone, E.F. Kelly, H.C. Monger, P. Sollins.
1999. Site and landscape characterization for Ecological Studies. In G.P. Robertson,
D.C. Coleman, C.S. Bledsoe, and P. Sollins (eds), Standard soil methods for long-
term ecological research. Oxford University Press, New York. 462 pp.

Hannah, RR. and R. Zahner. 1970. Nonpedogenetic texture bands in outwash sands of
Michigan: their origin, and influence on tree growth. Proc. Soil Sci. Amer. 34: 134 —
136.

Harding, RB. and DP. Grigal. 1984. Prediction equations for water-holding capacity of
some Minnesota forest soils. Minnesota Forestry Research Note 293. University of
Minnesota, St. Paul.

Harding, R.B., D.F. Grigal, and B.H. White. 1985. Site quality evaluation for white
spruce plantations using discriminant analysis. Soil Sci. Soc. Amer. J. 49: 229-232.

Hebert, M.T. and SB. Jack. 1998. Leaf area index and site water balance of loblolly

pine (Pinus taeda L.) across a precipitation gradient in East Texas. For Ecol.
Manage. 105: 273—282. '

56

Hicock, H.W., M.F. Morgan, H.J. Lutz, H. Bull, and HA. Lunt. 1931. The relation of
forest composition and rate of growth to certain soil characters. Connecticut Agric.
Expt. Sta. Bul. 300.

Hogg, B.H. 1994. Climate and the southern limit of the western Canadian boreal forest.
Can. J. For. Res. 24: 1835- 1845.

Jenny, H. 1980. The Soil Resource: Origin and Behavior. Springer-Verlag, New York.
377 pp.

Jenson, M.E., Burman, RD, and R.G. Allen (Editors). 1990. Evapotranspiration and
irrigation water requirements. American Society of Civil Engineers, New York.
Man. Rep. Eng. Pract. 70.

Lambers, H., F. S. Chapin HI, and TL. Pons. 1998. Plant physiological ecology.
Springer-Verlag, New York. 540 pp.

Linn, D.M., and J .W. Doran. 1984. Effect of water-filled pore space on carbon dioxide
and nitrous oxide production in tilled and non-tilled soils. Soil Sci. Soc. Am. J. 48:
1267-1272.

Kloeppel, B.D., S.T. Gower, I.W. Treichel, and S. Kharuk. 1998. Foliar carbon isotope
discrimination in Larix species and sympatric evergreen conifers: a global

comparison. Oecologia 114: 153-159.

Mader, D.L. and DP. Owen. 1961. Relationships between soil properties and red pine
growth in Massachusetts. Proc. Soil Sci. Amer. 25: 62 - 65.

Malagoli, M., A. Dal Canal, S. Quaggiotti, P. Pegoraro and A. Bottacin. 2000.
Differences in nitrate and ammonium uptake between Scots pine and European larch.
Plant and Soil 221: 1-3.

Marshall, J .D. and J. Zhang. 1994. Carbon isotope discrimination and water-use
efficiency in native plants of the north-central Rockies. Ecology 75: 1887 — 1895.

McClave, J .T. and EH. Dietrich, II. 1991. Statistics, 5th edition. Macmillan Publishing,
San Francisco. 928 pp.

McComb AL. 1955. The European larch: its race, site requirements and characteristics.
For. Sci. 1: 298-318.

McMaster, GS. and Wilhelm, W.W. 1997. Growing degree-days: one equation, two
interpretations. Agric. For. Meteor. 87: 291-300.

Miller, A.J. 1990. Subset Selection in Regression. Chapman and Hall, London. 229 pp.

57

 

 

Miller, J.D., J.M. Cooper, and HG. Miller. 1993. A comparison of above-ground
component weights and element amounts in four forest species at Kirkton Glen. J.
Hydrol. 145: 419 — 438.

Mulvaney, KL. 1996. Nitrogen — Inorganic forms. In D.L. Sparks, AL. Page, P.A.
Helmke, R.H. Loeppert, P.N. Soltanpour, M.A. Tabatabai, C.T. Johnston, M.E.
Sumner. (eds.) Methods of Soil Analysis, Part 3. Chemical Methods. SSSA Book Ser.
5. SSSA and ASA, Madison, WI.

Myre, R. and C. Camiré. 1996. The effect of crown position and date of sampling on
biomass, nutrient concentrations, and contents of needles and shoots in European
larch. Trees 10: 339-350.

Monserud, R.A. 1984a. Height growth and site index curves for inland Douglas-fir
based on stem analysis data and forest habitat type. For. Sci. 30: 943 — 965.

Monserud, R.A. 1984b. Problems with site index: an opinionated review. In J.G.
Bockheim (ed.) Symp. Proc. Forest Land Classification: Experiences, Problems,
Perspectives. NCR-102 North Central For. Soils Com., Soc. Am. For., USDA For.
Serv., and USDA Soil Cons. Serv.

Monserud, R.A., U. Moody, and D. W. Breuer. 1990. A soil-site study for inland
Douglas-fir. Can. J. For. Res. 20: 686-695. '

National Oceanic and Atmospheric Administration. 2000. Temperature and precipitation
data. National Climatic Data Center. Asheville, North Carolina.

National Renewable Energy Laboratory. 2000. Solar radiation Normal Data (1961 -
1990). Golden, Colorado.

Newberry, J .D. 1991. A note on Carrnean’s estimate of height from stem analysis data.
For. Sci. 37: 368-369.

Pastor, J., J.D. Aber, C.A. McClaugherty, and J.M. Melillo. 1984. Aboveground
production and N and P cycling along a nitrogen mineralization gradient on
Blackhawk Island, Wisconsin. Ecology 65: 256-268.

Payandeh, B. 1986. Predictability of site index from soil factors and lesser vegetation in
northern Ontario forest types. Canadian For. Serv. Great Lakes For. Cent. Inf. Rep.
O-X-373.

Perala, DA. 1985. Predicting red pine shoot growth using growing degree days. For.
Sci. 31: 913-925.

Perala, DA. and DH. Alban. 1994. Allometric biomass estimates for aspen-dominated
ecosystems in the Upper Great Lakes. US. For. Serv. Res. Pap. NC-134.

58

 

 

Pritchett, W.L. and RF Fisher. 1987. Properties and Management of Forest Soils. John
Wiley and Sons, New York.

Raison, R.J., M.J. Connell, and PK. Khanna. 1987. Methodology for studying fluxes of
soil mineral-N in situ. Soil Biol. Biochem. 19: 521 — 530.

Reich, P.B., D.F. Grigal, J .D. Aber, and ST. Gower. 1997. Nitrogen mineralization and
productivity in 50 hardwood and conifer stands on diverse soils. Ecology 78: 335-
347.

Reich, P.B., D.S. Ellsworth, ME. Walters, J .M. Vose, C. Gresham, J.C. Volin, and W.D.
Bowman. 1999. Generality of leaf trait relationships: a test across six biomes.
Ecology 80: 1955-1969.

Richards, F.J. 1959. A flexible growth function for empirical use. J. Exp. Botany 10:
290 - 300.

Robertson, G.P., P. Sollins, B.G. Ellis, K. Lajtha. 1999. Exchangeable Ions, pH, and
Cation Exchange Capacity. In G.P. Robertson, D.C. Coleman, C.S. Bledsoe, and P.
Sollins (eds.), Standard soil methods for long-term ecological research. Oxford
University Press, New York. 462 pp.

Sartz, RS. and D.N. Tolsted. 1976. Snowmelt runoff from planted conifers in
southwestern Wisconsin. US. For. Serv. Res. Pap. NC-205.

SAS Institute, Inc. 2000. JMP® Statistics Guide, Version 4. SAS Institute, Inc., Cary,
North Carolina.

Schlaegel, BE. 1975. Yields of .four 40-year—old conifers and aspen in adjacent stands.
Can. J. For. Res. 5: 278- 280.

Schlesinger, W.H. 1997. Biogeochemistry, 2nd edition. Academic Press, San Diego.
588 pp.

Scott, NA. and D. Binkley. 1997. Foliage litter quality and annual net N mineralization:
comparison across North American forest sties. 0ecologia 111: 151 - 159.

Soil Survey Staff. 1975. Soil taxonomy: a basic system of soil classification for making
and interpreting soil surveys. USDA-SCS Agric. Handbook No. 436. US.
Government Printing Office. Washington, DC.

Sokal, RR. and F .J . Rohlf. 1961. Biometry; The Principles and Practice of Statistics in
Biological Research. W.H. Freeman and Company, San Francisco. 776 pp.

59

Sparks, J .P. and J .R. Ehleringer. 1997. Leaf carbon isotope discrimination and nitrgen
content for riparian trees along elevational transects. 0ecologia 109: 362 — 367.

Stephenson, N.L. 1990. Climatic control of vegetation distribution: the role of the water
balance. Am. Nat. 135: 649-670.

Stoeckeler, J .H. 1948. The growth of quaking aspen as affected by soil prOperties and
fire. J. For. 46: 727-737.

Stokes, MS. and TL. Smiley. 1996. An Introduction to T rec-Ring Dating. The
University of Arizona Press. 73 pp.

Ter—Mikaelian, M.T. and MD. Korzukhin. 1997. Biomass equations for sixty-five North
American tree species. For. Ecol. Manage. 97: 1-24.

Thomas, G.W. 1996. Soil pH and soil acidity. In D.L. Sparks et al. (ed) Methods of Soil
Analysis, Part 3. Chemical Methods. SSSA Book Set. 5. SSSA and ASA, Madison,
WI.

Truong, dinh Phu. 1975a. Nutrient requirements of planted European larch in Quebec.
Plant Soil 42: 109-1 17.

Truong, dinh Phu. 1975b. Potassium et magnesium: deux éléments limitant la croissance
en hauteur du pin rouge au Quebec. Can. J. For. Res. 5: 73 — 79.

Valade, J .L. 1998. The potential use of larch in the Canadian pulp and paper industry: A
review. Pulp Paper Can 99: 156-158.

van Eck, WA. and ER Whiteside. 1963. Site evaluation studies in red pine plantations
in Micigan. Proc. Soil Sci. Amer. 27: 709 — 714.

Wang, G. 1995. White spruce site index in relation to soil, understory vegetation, and
foliar nutrients. Can. J. For. Res. 25: 29-38.

Wilde, S.A., J .G. Iyer, C. Tanzer, W.L. Trautmann, and K.G. Watterston. 1965. Growth
of Wisconsin coniferous plantations in relation to soils. Univ. Wis. Res. Bull. 262.

Yin, X. 1993. Variation in foliar nitrogen concentration by forest type and climatic
gradients in North America. Can. J. For. Res. 23: 1587-1602.

Zak, D.R., G.E. Host, and KS. Pregitzer. 1989. Regional variability in nitrogen

mineralization, nitrification, and overstory biomass in northern Lower Michigan.
Can. J. For. Res. 19: 1521-1526.

60

Zak, D.R., D. Tilman, R.R. Parmenter, C.W. Rice, F.M. Fisher, J. Vose, D. Milchunas,
C.W. Martin. 1994. Plant production and soil microorganisms in late-successional
ecosystems: a continental-scale study. Ecology 75: 2333

Zak, D.R., W.E. Hohnes, N.W. MacDonald, and KS. Pregitzer. 1999. Soil temperature,
matric potential, and the kinetics of microbial respiration and nitrogen mineralization.
Soil Sci. Soc. Am. J. 63: 575-584.

61

 

£1212th

A comparison of site factors, productivity,
species traits, and their interrelationships
for European larch and red pine in the
Great Lakes region

62

 

Abstract The objectives of this study were to I) examine the interrelationships of soil
site factors, leaf/canopy characteristics and measures of productivity (site index and
overstory annual net primary production, ANPPO) for European larch (Larix decidua
Miller) and red pine (Pinus resinosa Ait.) over an array of stands that constitutes a broad
landscape scale moisture/nutrient gradient, and 2) compare species productivity for
species with different leaf life spans and associated traits to examine if a trade-off exists
such that red pine (long leaf life span) has greater productivity than European larch (short
leaf life span) on poor sites and vice versa. The relationships between soil, vegetation
characteristics, and productivity were examined using least squares regression. Site
index for European larch and red pine were strongly correlated with indices of moisture
(available water holding capacity - AWC, r = 0.502 and 0.441, respectively) and nitrogen
availability (nitrification, r = 0.415 and 0.372, respectively). Species traits correlated
with site index productivity for European larch and red pine were foliar nitrogen (%N, r
= 0.598 and 0.477, respectively) and specific leaf area of senesced foliage (r = 0.400 and
0.506, respectively). Because the study species affected chemical and organic soil
characteristics, I compared species productivity and related traits over the physical soil
resource availability gradient, AWC. Both species productivity increased with AWC and
the rate of increase was weakly, and not significantly, greater for European larch than for
red pine. Over the entire range of AWC, European larch maintained greater productivity,
indicating that there no trade-offs in productivity due to differences in species traits, as

was predicted.

63

 

INTRODUCTION

On well-drained sites in the Great Lakes region, tree productivity is limited by
nutrient and water supply and by differences in species’ abilities to utilize those
resources. Differences in species’ abilities to use resources may be due to species-
specific variation in physiological and morphological traits. Among these traits are leaf
life span and associated characteristics. In general, species with short leaf life span also
have high light-saturated photosynthetic rates (Amax), leaf nitrogen (Nlcaf), leaf area per
unit mass (specific leaf area, SLA), and canopy nitrogen (N) loss rates, and low canOpy
mass and water use efficiency (WUE), whereas species with long leaf life span have low
Amax, Nleaf, and canopy N loss rates and high canopy foliage mass and WUE (Reich et al.
1992, 1995, 1997a, 1999; Kloeppel et al. 2000). These collections of traits, and observed
abundances of evergreen (long leaf life span) vs. deciduous (short leaf life span) species
across landscape to global-scale fertility gradients suggest that evergreen species have
relative productivity advantages over deciduous species on sites of low fertility/moisture
status and vice versa; deciduous species have productivity advantages on high
moisture/fertility sites. The goals of this study were to consider a species of short
(European larch (Larix decidua Miller), ~6 months) and long (red pine (Pinus resinosa
Ait.), ~40 months) leaf life span and address the following questions: (1) what are the
importance of differences in leaf life span and associated traits (e.g. Nleaf; N turnover; leaf
area index, LAI) to productivity for trees growing across fertility/moisture availabilities,

and (2) do differences in these traits represent a trade-off such that evergreen trees have

64

 

 

greater productivity on poor sites and vice versa. If so, what is the specific nature of
these trade-offs?

European larch and red pine have leaf life spans averaging 6 and 40 months,
respectively. Gower and Richards (1990) speculate that deciduous larch spp. can have
similar productivity in the marginally productive environments that favor the
conservative characteristics of evergreen conifers due to a number of characteristics (e.g.
the deciduous habit, architecture, relatively high rates of photosynthesis). While larch
needles have higher Am”, they also have relatively high rates of leaf and N loss (Gower
and Son 1992) and lower foliage mass compared with evergreen conifers (Gower et al.
1993). Conversely, red pine often has a denser canopy (higher LAI and canopy mass)

and retains its needles longer, but has lower Amax (Reich et al. 1999).

Table 2.1. Comparison of characteristics for European larch and red pine on a

 

 

moderately high fertility site in Wisconsin.T
European larch red pine

leaf life span (growing seasons) -- 3-4 times greater
leaf area index (leaf area/ ground area) -- 1.2 times greater
Annual leaf loss (mass) ~ the same

leaf N concentrations 2 times greater --

annual N losses 1.8 times greater --
maximum rates of photosynthesis (Ammss) 2.5-4 times greater --
specific leaf area (leaf area/leaf dry mass) 3.5 times greater --
overstory biomass production (ANPPO) 1.2 times greater --

 

1‘(Data from Gower et al. 1993 and Reich et al. 1999.)

Common garden studies have shown levels of productivity can be similar for species
with different leaf habits on a given site (e.g. Matyssek 1986, Tyrrell and Boemer 1986,
Gower et al. 1993). Matyssek (1986) found that despite larch spp. maintaining higher
Nleaf and Amax.(mass & area) and lower WUE than spruce, the two species attained similar

stem increment. In another larch and spruce comparison, Tyrell and Boerner (1986)

65

 

reported that both species had similar relative growth rates and grth efficiency (wood
produced per nutrient loss in litter) despite vastly different leaf life spans. Gower et al.
(1993) also reported that productivity, measured as aboveground tree biomass, was not
significantly different for larch, spruce, and pine species despite large species differences
in SLA, leaf N, and LAI. These studies compared fimctionally different species on
similar sites. However, it is unclear how these relationships may change across a broad
range of site fertilities. Because of this limitation, Gower et al. (1993) could only suggest
that differences in leaf traits may result in trade-offs in productivity across a soil
moisture/fertility gradient.

Other studies have aimed to quantify how a single species acclimates to variation in
resource availability by assessing species’ . attributes across a gradient of resources.
Bockheim et al. (1989) measured changes in foliar nutrient concentrations and
productivity of red pine across a precipitation-chemistry gradient in Wisconsin. While
within-site variation was too large to be able to draw conclusions about differences in
precipitation chemistry affects among-sites they did find that needle N was significantly
and positively related to site index. Hebert and Jack (1998) quantified variation in LAI
and site water balance in loblolly pine (Pinus taeda L.) plantations across a historically-
based precipitation gradient in Texas. They reported that anomalous precipitation
patterns just prior to study masked the gradient, but they did find a significant correlation
between LAI and foliar N and P.

Past studies also have investigated multiple species over a range of sites. Pastor et
al. (1984) reported increased productivity (ANPP) across an N mineralization gradient on

a 70 ha island but the design did not 'allow a comparison of the six species along its

66

entirety. Kloeppel et al. (1998) compared physiological traits of Larix spp. (including
European larch) and co-occurring evergreen conifers (not including red pine) at various
locations in the northern hemisphere. They reported that, despite lower WUE, Larix spp.
can have similar rates of productivity compared to those of evergreen conifers. Again, all
species were not present at all locations. In another study, Kloeppel et al. (2000)
compared leaf-level resource use for western larch (Larix occidentalis Nutt.), a deciduous
conifer, and two sympatric evergreen conifers, Douglas fir (Pseudotsuga menziessi,
Beissn., Franco) and lodgepole pine (Pinus contorta Engelrn.) along a resource
availability gradient in Montana. Contrary to Kloeppel et al. (1998), they found that both
types of conifers were equally water limited as measured with carbon isotopes. Kloeppel
et a1. (2000) were able to Show that the deciduous conifer had higher SLA, mass-based
photosynthetic rates and foliar N but evergreen conifers exhibited higher N-use
efficiency. Yin (1993) found that differences in foliar N concentrations were strongly
correlated with climatic factors (e.g. mean July temperature, light availability). Neither
Kloepplel et a1. (2000) nor Yin (1993) measured productivity.

Studies have also assessed the relationships of regional variation in N mineralization
(Reich et al. 1997b), LAI (Fassnacht and Gower 1997), potential net nitrification and
mineralization (Zak et al. 1989), or soil microbial biomass (Zak et al. 1994) with stand
productivity (measured as ANPP or ANPPO) for multiple species. These studies reported
significant and positive relationships between their respective predictor variables and
stand productivity, yet the designs did not permit a separation of species from site.

The Extended COllaboration to Link Ecophysiology And forest Productivity

(ECOLEAP) project is an effort initiated in 1996 to improve the understanding of site-

67

 

level controls on boreal-and sub-boreal forest productivity as extracted by multispectral
satellite data on a landscape scale in Canada (Bernier et al. 1999). While this study is
directed towards estimating forest productivity, it is too coarse in focus to identify
species’ trade-offs.

In review, studies have provided fragments of information, but have lacked a design
that quantitatively assesses interrelationships among soil moisture/fertility, leaf traits, and
productivity for species with contrasting leaf traits. The goals of this study were to
compare leaf and canopy traits of EurOpean larch and red pine and their relationship to
productivity across a broad soil resource gradient in the Great Lakes region. In addition
to furthering basic understanding of the interrelationships of species’ traits, resource
availability, and productivity, this information could be used as a silvicultural tool for
improving productivity through better species selection in plantation establishment.
Pinpointing the bottlenecks of productivity for a species also improves the possibility that
genetically superior stock could be deve10ped for enhanced production. Original

predictions and general hypotheses are:

0 European larch will be more productive than red pine on fertile sites. In addition,
European larch will have higher Nlcaf and Amax than red pine if moisture is not
limiting (Gower et al. 1993). On less fertile sites, European larch will have much
lower Nleaf, Amax, and LAI due to high N turnover (Gower et al. 1993). European
larch will also be more sensitive to droughty sites with a less efficient use of water

(Kloeppel et al. 1998).

68

 

 

o On less fertile sites, red pine productivity will be higher than European larch due to
characteristics leading to greater nutrient conservation: longer leaf life span leading to
greater foliage mass, LAI and WUE, and lower N turnover than European larch
(Gower and Son 1992). Extended leaf life span for red pine will compensate for
lower N uptake rates, thus maintaining productivity by maintaining relatively high
Nteapam, Amuma, LAI, and foliage mass. On high fertility sites, red pine will not have
greater productivity than European larch due to inherent Amax limits. Overall, due to
its plasticity in leaf life span, the productivity of red pine will be less sensitive to

changes in fertility or moisture than European larch.

 

larch

 

Stand productivity

 

 

 

poor rich

Site quality
(productivity as f(x) soil texture, moisture, etc.)

Figure 2.1. Hypothesized relationship between stand productivity and site quality (as
defined by productivity being a function of site factors — e. g. soil texture, moisture, etc.)

69

 

Specific hypotheses that are tested include:

Leaf N will scale closely with light-saturated photosynthesis (Amax)

European larch will have higher ANPPo than red pine on rich sites (see Fig. 2.1.)
because of higher Nteapmass (Fig. 2.2.d) and Amaxmass (1), despite only slightly higher
Armam,a (e) and LAI (h) and a lower foliage mass (g) than red pine.

Red pine will be have higher ANPPo than European larch on poor sites (see Fig. 2.1.)
due to much greater leaf life span (Fig. 2.2.a) and foliage mass (g), moderately higher
LAI (h), Amearea (e), and WUE (j), and lower annual N losses (i), despite much lower

Nleafmss (d) and AmxmSS (1) than Eur0pean larch.

Compared to poor sites, European larch on rich sites will have similar leaf life span
(Fig. 2.2.a), slightly lower SLA (b), lower WUE (i), and greater Nleaf.(ar¢a & mass) (c &
d), Amara,“ & mass) (e & 1), annual foliage turnover, N losses (i), and foliage mass (g)

and area (LAI; h) per unit ground area.

Compared to poor sites, red pine on rich sites will have shorter leaf life span (Fig.
2.2.a), Slightly lower WUE (i), and higher SLA (b), Nleaf.(ar¢a & mass) (C & d), Amax-(area &

mass) (e & f), foliage mass (g) and area (LAI; h) per unit ground area, annual foliage

turnover, and N losses (i).

70

a)

leaf life span

c)

Nleaf-area

Amax-area

8)
foliage
mass/area

i)

annual
N

losses

larch

 

pine

 

larch
/pine

 

pine

%-

 

 

larch

W

 

poor rich

site quality

b)
SLA

d)

Nleaf-mass

>337

max-mass

h)
LAI

1)

PM

 

me

 

y

/pife

 

pine

larch

 

 

pine

\

 

poor rich

site quality

Figure 2. 2. Hypothesized responses of foliar and canopy characteristics and productivity
to site quality (~moisture/nutrient gradients) for European larch (“larch”) and red pine

(C‘pin639).

7l

o ANPPo is a function of the product of foliage (mass (Figure 2.3.a) or area Gigure
2.3.b)) times foliar N (mass or area) per unit ground area and these relationships will
be similar for European larch and red pine.

ANPPo

 

 

   

 

 

fOIiagC maSS X Nleaffnass LAI X Nleafarca

Figure 2. 3. Hypothesized relationship between productivity (ANPPO) and foliage mass
(either mass or area basis) and foliage N per unit ground area.

0 The combination of (foliage mass x Nleafimass) and l3C/lZC ratio explains more
variation in ANPPo than (foliage mass x Nleafmass) alone (Figure 2.4.).

ANPP0

 

 

 

(foliage mass x Nleapmass)
and ‘3C/12C

Figure 2.4. Hypothesized result of multiple regression ANPPo and [(foliage mass and
foliage N per unit ground area)] and 13C/ 2C].

0 Species’ productivity relative to one another will be greater for red pine than
European larch on low fertility sites whereas productivity will be greater for
European larch than red pine on high fertility sites (Figure 2.1)

72

 

lVIETHODS

Measurements were made to determine physical soil characteristics, foliar/canopy
traits, and productivity as described in Chapter 1. These data were used to test and model

interrelationships as well as the success of foliar/canopy traits to predict productivity.

Statistical methods.

Pearson product—moment correlation analyses were performed to relate soil and
vegetation characteristics to site index (base ages 15, 25, and 50) and annual biomass
increment (ANPPO-mm1). The Pearson product-moment correlation coefficient (r), a
measure of the strength of the linear relationship between each pair of variables (Sokal
and Rohlf 1969), between predictor variables also was examined for evidence of
collinearity.

Two-factor models were used to test the effects of pairs‘of predictor variables and
their interactions on productivity. If interactions had P > 0.25, the variances were pooled
(Bancroft 1964). The coefficient of determination (R2) adjusted for degrees of freedom,
Mean Square, and model P value were reported.

Least squares regression models were used to exam the relationship of physical soil
variables with productivity. Variables were chosen with their independence from species
in mind (e.g. % sand instead of forest floor depth since the latter is influenced by
species.) The variable with the largest average coefficient of determination with
productivity among species was then substituted as a gradient across which changes in

physiology/morphology for each species could be compared. This was done by

73

 

regressing leaf/canopy traits (e.g. carbon isotope discrimination, foliar nitrogen, canopy
biomass) by the physical site factor to assess changes in species’ attributes given a

change in inherent site productivity.

74

 

RESULTS

Relations of soil variables with productivity

For both European larch and red pine, site index at base age 25 (8125) was positively
correlated with available water holding capacity (AWC) and the textural characteristics
that determine it (i.e. % sand , silt, and clay) (Table 2.2.; See Table 2.3. for definition of
terms). In addition, rates of nitrification (NIT) and total N mineralization (NMIN)
obtained from laboratory incubations were positively related to S125 for red pine, whereas,
for European larch, only nitrification rates were positively associated with S125. No soil
variables were significantly correlated with red pine ANPP...“ and only NIT and FFdepth

were correlated with European larch ANPPo-total.

Table 2.2. Pearson product-moment correlation coefficients for measures of physical soil
attributes and productivity by species.

 

 

 

European larch red pine
ANPPo-total SI Al\IPPo-total SI
(Mg ha’1 99") 25 (Mg 113.! 994) 25
% sand -0.269 —0.549 0.18 -0.384
% silt 0.246 0.352 -0.229 0.316
% clay 0.207 0.662 -0.066 0.402
AWC 0.316 0.502 -0. 141 0.441
Db 0.244 0.061 0.115 0.06
F chpth -0.442 -0.464 -0.209 -0. 105
% OM 0.248 0.319 -0.277 0.091
99 NMIN 0.041 0.154 -0.077 0.442
99 NIT 0.543 0.415 -0.164 0.372
pH 0.284 -0.285 -0.244 -0.038
WB 0.250 -0.005 0.079 0.307

 

75

Boldface indicates significant correlation (or = 0.10).

To investigate if soil attributes in addition to AWC explain additional variance in

productivity, single factors were combined in regressions with AWC to predict S125.

AWC was chosen as the first variable in these regressions since it explained, on average

Table 2.3. Definition of terms and units.

 

 

Acronym Parameter Units

S115, 25, 50 site index (base ages 15, 25, or 50) m

ANPPMotal total overstory annual net primary production Mg ha'1 1999'l
ANPP“...m stem component of ANPPMotal Mg ha‘1 1999’l
% sand sand fiaction soil texture %

% silt silt fiaction of soil texture %

% clay clay fraction of soil texture %

AWC available water holding capacity cm3 111'3

D. bulk density of soil g cm’3

pH pH of soil --

% CF soil coarse fragments (> 4 mm) %

Adept}. depth to A horizon cm

Cam,“I depth to C horizon cm

FFdepul depth of forest floor cm

% OM organic matter fraction of soil %

NMIN nitrogen mineralization mg kg'1 day'1
NIT nitrification mg kg"1 day"
A foliar carbon isotope discrimination %o

leaf %N nitrogen fraction of leaf %

Nleapmass mass-based leaf nitrogen mg g’1
Nleapam area-based leaf nitrogen mmol In2 x 102
annual N loss annual nitrogen loss through litterfall kg ha'1
canopy biomass mass of fresh canopy Mg ha'1
SLAcanOPy live canopy specific leaf area cm2 g’l
SLAMer senesced tissue specific leaf area cm2 g'1

LAI leaf area index m2 m'2

JULY average July temperature °C

GDD growing degree days count (°C)
WB site water balance cm

 

76

 

 

across species, a high % of variation in S125, and because, compared to % clay and %
sand (other strong predictors of 8125), it was more normally distributed. For red pine,
AWC, % OM, and their interactions explained more variation in 8125 than AWC alone
(R2 = 0.16) and all three terms were significant (P=0.0-1, 0.099, and 0.04 respectively).
However, explained variance was relatively low for the combined model (R2 = 0.26). No
other variable in combination with AWC explained appreciably more variation in red
pine 8125 than AWC alone. In contrast, for European larch, the combination of pH and
AWC explained 49.4% of the variation in S125 compared to 20% for AWC alone and both
AWC (P=0.0012) and pH (P=0.006) were significant.

Although NMIN and NIT rates were important corollaries of S125 for red pine,
inclusion of NMIN or NIT did not markedly increase the amount of variation explained
by AWC alone. This may be due to the generally strong collinearity of NMIN and NIT

with AWC (data not shown).

Table 2.4. Analysis of variance for regressing 8125 by physical soil attributes, available
soil water holding capacity (AWC), and their interactions. Mean Square, P value, and
adjusted R2 are reported.

 

 

 

MS (P>F) RT
% organic matter 7.25 0.099
red pine AWC 19.59 0.010 0.264
%OM*AWC 1 1.76 0.040
pH 34.38 0.006
European larch 0.494
AWC 53.795 (0.0012)

 

NOTE: For interactions P>0.25 variances were pooled (Bancroft 1964). R2 reported are
adjusted for degrees of freedom (Rf).

77

 

Interrelations of soil physical and chemical variables

Nitrogen mineralization and AWC, the soil characteristics most strongly related to
S125 for both species, were themselves strongly positively related (Table 2.5.). This
relationship was stronger for NIT than for total NMIN rates. Of all variables correlated
with NMIN, % OM was the strongest. Percent OM is a characteristic related to both soil
moisture holding capacity (but notably not incorporated into the AWC calculation) and to
the supply of organic material for NMIN. Forest floor depth was negatively associated
with all measures of nitrogen cycling, but was only significant with nitrification for red
pine. In part the weak relations could be an artifact of the laboratory incubation
technique since forest floor material was mostly screened and removed from the soils. In
a mixed model of the effects of % OM, species and their interactions on N
mineralization, species affected NMIN independent of % OM (both effects P < 0.001) but
interactions were not significant (P>0.10). For any given % OM larch sites had higher
NMIN rates, and NMIN increased with % OM for soils from both European larch stands
and red pine stands in parallel. Similarly AWC and Species both affected NIT strongly
(both effects P<0.001), and interactions were not significant (P>0.10). At any AWC,
NIT was greater for soils from larch stands and NIT increased with AWC for soils from

larch and red pine stands in parallel.

78

Table 2.5. Pearson product-moment correlation coefficients for measures of physical soil
attributes and rates of soil nitrogen mineralization and nitrification.

 

European larch red pine
N mineralization nitrification N mineralization nitrification

(mg kg" day") (Ins kJL‘day") (mg kg" day") (mg kg" day")

 

% sand -0.570 -0.519 -0.409 -0.676
% silt 0.617 0.422 0.411 0.635
% clay 0.260 0.493 0.319 0.596
AWC (cm3 m'3) 0.495 0.583 0.3041 0.639
D, -0109 0.029 -0.358 -0.263
pH 0.175 0.499 -0.283 0.318
% OM 0.665 0.407 0.541 0.271
WB 1999 (cm) 0.383 0.134 0.411 0.421
FFd 61cm) -0.230 -0.319 -0.011 -0.346

 

1

Correlation significant (r = 0.393) when one outlier removed. Note: Rates of
mineralization (‘N MIN’) and nitrification (‘NIT’) are from laboratory incubations

(1999). Boldface indicates significant correlation (or. = 0.10).

Relations of vegetation characteristics and productivity-Site index

Several leaf/canopy traits were strongly associated with 8125 and ANPPo-m1 (Table
2.6.). Foliar N was more strongly correlated with 8125 than any other variables. For red
pine, the relationship of foliar N to S125 was strongest when N was expressed on a per
unit leaf area basis (N leaflarea) whereas, for European larch, foliar N expressed on a mass
basis (N leafmass) was most strongly related to S125. Carbon isotope discrimination (A) was
negatively associated with S125 for European larch but was unrelated to $125 for red pine.
Thus productive European larch stands tended to be more water use efficient in
photosynthesis than less productive stands. In general, variation in leaf area per unit leaf
mass (specific leaf area, SLA) was not as closely related to 8125 as was variation in leaf
N. However, SLA determined from needles senesced in autumn (SLAnm) was strongly

related to S125 for red pine. Furthermore, SLAMer was consistently more strongly related

79

 

 

to both S125 and ANPPo-tom1 than was SLA determined from foliage collected in mid-

summer (SLAcanopy).

Out of the four foliar variables related to S125 for red pine and/or European larch,

only the effects of %N were independent of species (Table 2.7., Figs. 2.5., 2.6., 2.7. and

2.8.). For species pooled, % N mass explained 57% of the variation in SI25 (Fig. 2.6. and

data not shown).

Table 2.6. Pearson product-moment correlation coefficients for leaf/canopy traits and
measures of productivity (total increment 1999, ANPPMom]; site index at 25 yrs., 8125).

 

 

 

European larch red pine
ANPPo-total SI ANPPo-total SI
(Mg ha“ 1999“) 25 (Mg ha‘1 1999“) 25

0.224 -0.590 0.345 0. 149

0.284 0.598 -0.056 0.477

-0.005 0.471 -0. 132 0.553

0.440 -0.018 0.174 -0.205

0.444 0.400 0.800 0.506

0.597 -0.046 -0.318 0.155

canopy biomass 0.431 -0.027 -0.364 0.201
annual N loss 0.339 -0.016 -0.371 0.227

 

Boldface indicates significant correlation (a = 0.10).

terms .

80

See Table 2.3. for definition of

 

 

 

 

 

24

21‘

18‘

8125 (m)

15‘

12‘

 

 

 

 

9 . .
0.5 1.0 1.5 2.0 2.5 3.0 3.5
Nmfma (mmol rn'2 X 102)

Figure 2.5. Area-based leaf nitrogen
(N karma) in relation to site index (S125) for
European larch and red pine.

24

 

21‘

—o
00
C
O

8125 (m)
6 5‘.
We
0
O
O

A
A

 

 

 

9 . . . . . .
20 40 60 80 100 120 140 160
SLAlitter (cm2 8")

Figure 2.7. Senesced tissue specific leaf
area (SLAlmer) in relation to site index (S125)
for European larch and red pine.

 

24

21‘

18*

15‘

12‘

 

 

 

 

10 15 20 25 30
Nleafmass (mg g4)
Figure 2.6. Mass-based leaf nitrogen

(Nlcagmass) in relation to site index (S125)
for European larch and red pine.

 

) rrvv frfi vafiifivv
~

211

18‘

15‘ A

12‘

 

 

 

 

9 . . . .
18 19 20 21 22 23
A (%0)

Figure 2.8. F oliar carbon isotope

discrimination (A) in relation to site index
(8125) for European larch and red pine.

 

24

Actual 3125 (m)

 

 

 

0 European larch

A red pine

 

 

 

 

 

14 15 l6 17 18

fi

20 21 22

Predicted 3125 (m) (A and Nleaf mass)

Figure 2.9. Actual S125 in relation to Predicted S125 for European
larch. Predicted values from two factor regression of S125 by carbon

isotope discrimination (A) and mass-based leaf nitrogen (N1caf.mass).

81

 

Table 2.7. Least squares model for the effects of foliar attributes, species, and their
interactions on site index at base age 25 ($125.) Mean Square, P value, and R2 are
reported.

 

 

 

 

 

MS (P>F) R2
Nleapmass 43 .775 (0.001 1)
spp. 4.824 (0.2472) 0.581
Nleafmass*spp- " "
Nympm 45 .661 (0.0013)
spp. 132.767 (<0.0001) 0.541
Nleaf-area*spp- " "
A 12.795 (0.0860)
spp. 126.687 (<0.0001) 0.493
A*spp. 28.145 (0.0109)
SLAMer 24.901 0.0549
spp. 16.47 0.0190 0.467
SLAliuer*spp. 1 1.422 0.1048

 

NOTE: If significant interactions P>0.25 variances were
pooled (Bancroft 1964). R2 adjusted for degrees of freedom.

Although leaf Nleaf.area and Nleaf.mass were strong, consistent predictors of S125 for both
species they were also strongly collinear (Nlmgmalss versus Nleagma — European larch, r =
0.81 P = <0.0001; red pine, r = 0.89, P = <0.0001). These relations were much stronger
than those between SLA and Nleafma or SLA and Nleafimass (P never > 0.03).

There was also a strong relationship between Nleafma and A. In a least squares model
of the effects of species, Nleafma, and their interactions on A, Nleafma was strongly related
to A but species and interaction terms were insignificant (Table 2.8.). If one red pine
outlier was removed, both species and the interactions were still not significant ( P =
0.6115 and 0.5315, respectively) and the model explained over half of the variation in A
(R2 = 0.529). This suggests that Nleaf.area is the primary determinant of variation in A

(Table 2.8.). Thus, there was a decrease in A with an increase in Njeafma independent of

82

 

 

species (Figure 2.10.). ~This can be interpreted as water use efficiency increasing with

Nleafma independent of species.

 

 

A red pine

0 European larch

9- 20.0 ‘

 

 

 

 

 

 

 

0.5 1.0 1.5 2.0 2.5 3.0 3.5

Nleaf area (mmol m'2 x 102)

Figure 2.10. Area-based leaf nitrogen (N magma) in relation to foliar carbon isotope
discrimination (A).

Table 2.8. Analysis of variance for testing species, leaf nitrogen (Nleafma), and their
interactions in predicting change in foliar carbon isotope discrimination (A). Mean
Square, P value, and R2 are reported.

 

MS
(P>F)
Nleaf-area 282
(0.0055)
spp. 0.3508
(0.31 12)
SPP * Nleaf-area 05839
(0.1929)
NOTE: If significant interactions P>0.25 variances were
pooled (Bancroft 1964). R2 adjusted for degrees of freedom.

R2

0.500

For both European larch and red pine, the combination of the two strongest
vegetation corollaries of site index markedly increased the amount of variation explained
in site index over the individual variables. For European larch, A and Nleaf.mass explained

51% of the variation in S125 (Table 2.9. and Figure 2.9.) compared to 31% for A and 32%

83

 

 

for Nleagmass alone (data not shown). For red pine SLAMer and Nleapam accounted for 46%

of the variation in S125 (Table 2.9.), compared to 21% for Nleafma and 20% for SLAMer

alone.

Table 2.9. Least squares models of site index at base age 25 (S125) as a function of

carbon isotope discrimination (A), and leaf nitrogen (N 1eafmass) for European larch, and

 

Eurgrean larch

 

 

 

 

 

 

Predictor MS (P>F) R2
A 24.361 0.0153

0. 5 l 0
Nleafmass 25.413 0.0136

Red pine

Predictor MS (P>F) R2
SLAMer 14.21 0.0088

0.46
Njeafma 16.10 0.0058

 

Note: For both models interactions were pooled since P>0.25 (Bancroft 1964).
Mean Square, P value, and R2 are reported.

Relations of vegetation variables and productivity — ANPP0-,0,a1_

For both species ANPPMotaI, was positively related to several characteristics
measured on the same basis as ANPPOMal, (i.e. per unit area of ground). These included
leaf area index (LAI), canopy biomass, and annual N 1088 (% nitrogen in leaf litter x leaf
litter mass per unit area). As expected, canopy biomass was positively associated with
ANPPomm; for European larch, while, surprisingly, the opposite was true for red pine —

increased canopy biomass was associated with decreased ANPPMOQI (Figure 2.13).

84

SLAMer was the single strongest predictor of ANPPomal for red pine and was a
significant, although weaker corollary of ANPPo-toml for European larch.

Similar to the strong needle-level based vegetation corollaries of $125 (i.e. Nleafma,
Nleapmass, SLAM“), some of the ground-area based vegetation corollaries of ANPPo-total
were themselves correlated (e. g. annual N loss versus canopy biomass - European larch, r
= 0.36 P = 0.06; red pine, r = 0.75 P = < 0.0001).

Although, within species, SLACanolDy is only relatively weakly correlated with ANPPO-
ml, , across species it was a relatively strong predictor of ANPPo-tota1 since high-SLAcanopy
European larch has greater ANPPMom than low-SLAcanopy red pine (Table 2.2., Figure
2.12.). For LAI, canopy biomass, and SLAM.“ there were both strong species effects and
strong interactions with species in their relation to ANPPMM, indicating that these
vegetation characteristics affected productivity differently for European larch and red

pine (Table 2.2., Figures 2.11., 2.13., 2.14.).

Table 2.10. Analysis of covariance for foliar attributes, species, and their interactions in
predicting ANPPMoml. Mean Square, P value, and R2 are reported.

 

 

 

 

 

MS (P>F) R2
Canopy biomass 27.979 (0.0392)
spp. 94.831 (0.0003) 0.321
Can. biomass*spp. 60.136 @0031)
Leaf area index 9.672 (0.1877)
spp. 56.154 (0.0023) 0.409
LAI*spp. 71.950 (0.0006)
SLAcanopy 52.382 (0.0054)
spp. 20.657 (0.0741) 0.346
SLACQopv’kspp -- --
SLAlitter 100.223 (<0.0001)
spp. 76.225 (0.0003) 0.455
SLAline,*spp. 42.076 (0.005 6)

 

NOTE: If significant interactions P>0.25 variances were
pooled (Bancroft 1964). R2 adjusted for degrees of freedom.

85

 

 

(NOTE: The following figure legends are for Figures on next page.)

 

A red pine

0 European larch

——

 

 

 

Figure 2.11. Leaf area index (LAI) in relation to 1999 total biomass increment (ANPPO-

total).

Figure 2.12. Canopy specific leaf area (SLAcanopy) in relation to 1999 total biomass
increment (ANPPo_tota1).

Figure 2.13. Canopy biomass in relation to 1999 total biomass increment (ANPPMoml).

Figure 2.14. Litter specific leaf area (SLAlmer) in relation to 1999 total biomass
increment (ANPPo-tota1).

Figure 2.15. Mass-based leaf nitrogen (N leafmass) in relation to 1999 total biomass
increment (ANPPo-tom1).

Figure 2.16. Carbon isotOpe discrimination (A) in relation to 1999 total biomass
increment (ANPPMmI).

Figure 2.17. Actual ANPPMota. in relation to predicted ANPPMM values for European
larch. Predicted values from two factor regression of ANPPMom. by canopy biomass and

mass-based leaf nitrogen (Nleafimass).

Figure 2.18. Actual ANPPWWI in relation to predicted ANPPMota] values for red pine.
Predicted values from two factor regression of ANPPMotal by carbon isotope
discrimination (A) and litter specific leaf . area (SLAM...)

86

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

: 16“Fig. 2.11. L: 16 Fig. 2.12.
<5 . o a 141 0
g 14 2
: 12‘ T 12 l
:2 104 .2 10.
no 00
2 8- g 8-
a 6- :g e
O u
*7 4- 5 4‘
o a. A
S: 2. D- 2
z
32 o - . < o . , . . .
4 5 20 4o 60 80 100 120 140 160 180
o 2 2 1
Leaf Area Index SLAcanopy (cm g“ )
Fi .2.13. Fi .2.14.
:16 g :1 g
0\ ‘ O '0‘ _. .
g 14 314
:12 ‘ :12 .
.2101 2101
90 on
2 3i A 2 84 a
v A V /
“é 6' A 6 6~
0 O
E 4‘ .:\ mus A A E 4‘ o *0
A 4\\ a.
9g 2‘ 48 A g 2- "
< 0 . r . r . . . < 0 . ‘ f r . r fl
0 2 4 5 3 10 12114 16 20 4o 60 80 100 120 140 160 180
Canopy biomass (Mg ha‘ ) SLAlitter(szg'1)
Fi .2.15. F1 .216
A 16 g A 16 g
'T o 7
§ 144 §14.
«—' 12‘ ~12-
'7 O 'T
2 10‘ . Q 210«
on o g on
g 84 ' . E 8‘ A A A
E 6- . o 1; 5. A ‘2
.9 . o ‘5 A ..
I ‘ . *7; - _,..—-
a? 4 0‘ . $ . a? 4 ALA/'45 A
g 2 9 a. 21 //’A‘“X’ AAA A
< ‘i Z A A A A AA
0 ' 1 Y T ' i T I < 0 I W I I I
‘0 12 14 16 18 2° 32 24 26 23 18.5 19.0 19.5 20.0 20.5 21.0 21.5
NIeafmass (m9 9 ) A (%o)
:16 FI9-2-17- :16 Fig. 2.18.
I . '0‘
§ 141 a 14-
:121 : 12 ‘
2101 iio‘
to
E 3. g 8‘ Ag: A}
E 6i %- SJ /7-\/
e 4‘ a? a,
g 2- % 2‘ /QA%AM
< 0 < o . r . .
' ' ‘ ' ' ' I o 2 4 6 8 1o
2 3 4 5 6 7 8 9 1o .
. . ANPP A&SLA.
Predicted ANPPMota, (canopy biomass & Nleafmass) Predicted O-total ( 1m)

87

 

 

For red pine, an increase in ANPPMOmI was associated with increases in A and SLAM“er
(Figures 2.14. and 2.16., respectively). The amount of variation in ANPPMotal explained
by the combination of A and SLAlmer (variance pooled) was greater than either variable
alone (Table 2.11.), as evidenced by the close relationship between actual values of
ANPPo-tota1 and ANPPMom values predicted by A and SLA“tter (Figure 2.18.). For European
larch the combination of A and SLAMer did not explain more variation in ANPPo-total
than either variable alone (Table 2.11.). For European larch, the combination of canopy
biomass and Nleagmass explained a greater amount of variation in ANPPMoml. than did of
canopy biomass or Nleafmass alone. This is illustrated by the tighter relationship of
ANPPMma] with the ANPPMotal values predicted by canopy biomass and Nleagmass than
with canopy biomass or Nleafmass (Figures 2.13., 2.15. and 2.17.). This relationship is not
surprising since the product of canOpy biomass and Nleaf.mass is the amount of foliar N per
unit ground area, a characteristic which might be expected to scale with productivity.
However, for red pine the combination of canopy biomass and Nieagmass failed to increase

the amount of variation explained for ANPPMota. by canopy biomass or Nleapmass alone

(Table 2.11.).

88

Table 2.11. Least squares models predicting total increment (ANPPMoml) as a fimction of
i) canopy mass and leaf nitrogen (N 163mm), and ii) carbon isotope discrimination (A) and
specific leaf area of litter (SLAnm) and their respective interactions. Mean Square, P
value, and R2 are reported.

 

 

 

 

European larch red pine
MS 2 MS 2
(P>F) R 03>F) R
can mass 58.26 16.259
°py (0.0094) (0.0613)
31.49 1.587
Nleaf-mass (00489) 0.309 (05448) 0146
can. mass* Nleaf.mass -- --
A 19.79 9.89
(0.1217) (0.0122)
57.102 70.023
., 7
SLAM (0.0118) 0275 (< 0.0001) 0'7‘8
A * SLIAKIitter " --

 

NOTE: If significant interactions P>0.25 variances were pooled (Bancroft 1964). R2
adjusted for degrees of freedom.

Comparing species productivity and related traits over physical soil gradients

To compare changes in leaf and canopy traits between species and across soil
resource gradients, a physical soil gradient was identified. Chemical soil variables
including pH, NMIN, NIT, % organic matter and forest floor depth were not used, since
these characteristics are influenced by species. AWC and % clay were, on average, the

best physical soil predictors of S125 (Table 2.2.). AWC was also the physical soil variable

89

most closely associated with ANPPo-tota1 for both species. However, it was not significant
in both cases (data not shown), perhaps because the relation of site variables to ANPP is
partially confounded by age (Chapter 1) and potential stand density effects (this Chapter)

on ANPPo-tota].

Table 2.12. Model statistics for fitting productivity (site index at base age 25) by
physical soil characteristics with least squares regression.

 

 

 

European larch red pine

R2 P R2 P
%sand 0.302 0.018 0.147 0.070
%clay 0.439 0.002 0.162 0.057
bulk density 0.004 0.809 0.004 0.785
AWC 0.252 0.033 0.195 0.035

 

In addition, clay had a skewed distribution, which resulted in a few data points
overly affecting regression results. For these reasons, AWC was selected as the best
single measure of a physical soil productivity gradient over which both species could be
compared (Figure 2.19.). In least squares models of the effects of Species, AWC and
their interactions on leaf/canopy characteristics, Species effects were significant for all
traits (Table 2.13.). AWC significantly affected Nleaf.mass, Meagan, annual N loss, N-leaf
mass, and SLAMer independent of species. For SLAM”, Species x AWC interactions
were significant, indicating that for each species, SLAMer responds differently to the

AWC gradient.

90

 

 

 

 

 

 

 

 

 

 

 

E
m 16
9
VJ
14
‘2 ‘ A red pine
A _
10 A
0 European larch
8 I I r I I f *I7 VI’
0 2 4 6 8 1O 12 14 16 18

Available water holding capacity (cm3 111-3)

Figure 2.19. Available water holding capacity (AWC) in relation to site index (base age
25, S125).

For vegetation characteristics where AWC effects were significant at P<0.2,
regression statistics further illustrate that as AWC increased, Nleapmass and Nleagma
increased in parallel for both species with Nleagmass higher for European larch than red
pine and vice versa with N133):am for red pine (Figure 2.20. and 2.21.). For the two
characteristics that showed significant AWC * species interactions, canopy biomass was
greater for red pine than European larch at any AWC and it increased with AWC for red
pine but was insensitive to AWC for European larch (Table 2.13., Figure 2.23.). For
SLAM“, European larch had higher values than red pine at any AWC and it increased

with AWC for European larch but was insensitive to AWC for red pine (data not shown).

91

Table 2.13. Analysis of variance for leaf/canopy attributes as predicted by available
water holding capacity (AWC), species (SPP), and their interactions. Mean Square, P

 

 

value, and R are reported.
AWC SPP SPP x AWC R2
A 0.002 14.093
(%o) (0.93 77) (< 0.0001) " 0426
mm... 83.32 1230.46 8.712 0 822
(mg g'l) (0.0003) (< 0.0001) (0.2193) '
Ngagma 0.6304 8.256 __ 0 727
(mmol m‘2 x 102) (0.0028) (< 0.0001) °
annual N loss 128.62 316.859 __ 0 420
(kg ha") (0.0035) (< 0.0001) °
litter N 0.025 0.543 __ 0 423
(%) (0.2243) (< 0.0001) '
SLAzcano , 173.344 99738.098 __ 0 945
(cm g ) (0.2256) (< 0.0001) '
SLA“...r 482.59 85533.35 446.39 0 953
(cm2 g") (0.0194) (< 0.0001) (0.0243) °
1.47 21.10
LAI (0.2864) (0.0002) " 0276
canopy biomass 10.69 292.31 1 1.67 0 550
(Mg ha") (0.1357) (< 0.0001) (0.1194) '
leaf life span
. 0.0007 58.58
(# growmg (0.9101) (< 0.0001) ' 0957
seasons)

 

NOTE: If no significant interactions (P>0.25) variance pooled and R2 reported.
otherwise, R2 adjusted for degrees of freedom (R32).

92

 

 

Table 2.14. Least squares regression results for vegetation characteristics versus AWC
for each species. Model R2, P value, and coefficients of slope (m) and intercept (b) are

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

reported by species.
spp. R2 P m b
N ( 4) larch 0.223 0.013 0.391 16.42
mm“ mgg pine 0.376 0.000 0.202 8.213
-2 2 larch 0.122 0.074 0.018 0.991
N‘eaf'mmmom X10) pine 0.215 0.015 0.033 1.681
cam biomass (M ha.) larch 0.000 0.920 -0004 3.35
W g pine 0.104 0.109 0.215 6.43
. 2 -i larch 0.192 0.022 1.385 106.57
SLA‘"(°m g ) pine 0.001 0.866 0.027 36.06
-1 larch 0.092 0.556 0.337 9.617
annualNloss(kg ha ) pine 0.349 0.002 0.394 4.251
Note: Boldface indicate significance (or = 0.10).
g? 3.5
0
: 3.0 ‘ A
E 2'51 A A A “a A red pine
o 2.04 fifiui”’“ '—
E 15 49%;”st AA .4 0 European larch
8 ‘ l A .0. . o o —
3‘3 05 . . . . . ._fi .
Z 0 2 4 6 8 1o 12 14 16 18

Available Water Holding Capacity (cm3 m‘3)

Figure 2.20. Productivity/site moisture (AWC) in relation to area-based leaf nitrogen
(Niagara) for European larch and red pine.

93

 

HHNNU)
O V! O UI,CD

 

Leaf nitrogen (mg 9'1)

0
A %o. e— __. ..
55mm?

 

-—-—-‘A"'
A

 

 

A red pine

0 European larch

 

 

 

 

*1

5 . . . . . -
0 2 4 6 8 10 12 14 l6 18
Available Water Holding Capacity (cm3 m'3)

Figure 2.21. Productivity/site moisture (AWC) in relation to mass—based leaf nitrogen
(Niagms) for European larch and red pine.

 

 

 

 

 

 

 

 

A 25

._ o

' o

2 20 ,.

0" o .
a" 15 :0 A A red pine
2 10- . 0“ //'° 0 European larch
E AA? Airgh/f o """

g 5 {m A AA 0°

(- A

E A A

< O

0 2 4 6 81012141618
Available Water Holding Capacity (cm3 m'3)

Figure 2.22. Productivity/site moisture (AWC) in relation to annual nitrogen loss for
European larch and red pine.

 

 

 

 

 

 

 

 

.1“ 16
13 14« A
A
5124 A A
a 10- 4 AM A ,,’—/’A’ A redpine
g 8. 31:”?sz _
g 6- A .4 .9 A 0 European larch
A
Ci 4‘ O . 0.. o —
g“ 2‘ 0'92) . o' 0
0 024681012141618

Available Water Holding Capacity (cm3 m‘3)

Figure 2.23. Productivity/site moisture (AWC) in relation to canopy biomass for
European larch and red pine.

94

DISCUSSION

Relations of soil variables with productivity

Consistent with other studies in the Great Lakes region, productivity was
positively correlated with soil water (e. g. Pastor et al. 1984, Fassnacht and Gower 1997)
and nutrient supply (e.g. Fassnacht and Gower 1997, Reich et al. 1997b). Zak et al.
(1989) showed that N supply was strongly related to productivity at low N supply, but at
high N supply on morainal sites, productivity became insensitive to variation in N.
Although there were strong correlations of both N cycling and soil water variables with
productivity, N cycling and moisture variables were themselves highly correlated. Thus
it is impossible to determine if productivity is more limited by water, nitrogen, or co-
limited by both. The negative correlation between FFdepm and productivity (ANPPHoml)
for European larch suggests that moisture and nitrogen limitation are inextricably linked
since microbial activity, which is responsible for litter decomposition, is moisture
dependent (Linn and Doran 1984, Barnes et a1. 1998, Zak et al. 1999).

As Sartz and Tolsted (1976) showed, spring infiltration rates of water in European
larch stands can be substantially higher than in evergreen conifer stands. Field data from
(Chapter 1) also show soil temperature to be warmer in critical spring periods where
microbial activity could persist earlier in the season as compared with neighboring

conifer stands.

95

Interrelations of physical and chemical soil characteristics

The interrelations of physical and chemical soil characteristics provided further
evidence of a relationship between soil moisture and both productivity and N cycling.
Measures of potential moisture (e.g. AWC, WB) as well as %OM, which both increases
water holding capacity markedly (Brady and Weil 1996) and provides the short term
source for mineralizable N, were positively associated with nitrogen cycling rates. The
positive correlation between %OM and NMIN also suggests that microbial populations in
these stands are carbon limited (sensu Zak et al. (1994). Similarly, NIT increased with
AWC for both species — also suggesting that moisture was an important determinant of
microbial nitrification. In both cases, NMIN and NIT were greater for European larch
than red pine. Consistently higher NMIN and NIT rates at any given level of AWC or
OM, even in potential laboratory incubations, suggest that soils supporting European
larch stands have greater nitrogen supply capabilities, perhaps due to high quality % OM,
higher leaf litter N, warmer soil temperature, higher spring infiltration, and less lignified,
higher N litter (Chapter 1, Sartz and Tolsted 1976, Scott and Binkley 1997). Fassnacht
and Gower (1999) observed that species on richer sites in northern Wisconsin tended to
have higher N turnover but also higher litter quality (lower ON), and lower nutrient and
organic matter residence time. In this study, European larch supplied, on average, ~twice
as much N annually than red pine but also had higher decomposition (i.e. less forest
floor) and higher rates of nitrogen mineralization.

While European larch stands do have a number of characteristics that could lead to
higher potential NMIN and NIT, the microbial processes ultimately responsible for

mineral transformation could be pH dependent (e.g. activity of bacteria and

96

 

actinomycetes). There were positive, although weak relationships between pH and
NMIN and NIT for European larch (1999: R2 = 0.03 P = 0.382, R2 = 0.249 P = 0.008;
2000: R2 = 0.217 P = 0.079, R2 = 0.307 P = 0.032). Furthermore, the combination of pH

and AWC explained nearly 50% of the variation in $125 for European larch.

Relations of vegetation characteristics and productivity

The vegetation characteristics that best predicted productivity for either species
were measures of foliar N. The increase in productivity from increased foliar N could be
parallel with the increases in productivity correlated with moisture and nitrogen supply
via nitrogen mineralization. Increase N supply, N uptake, and foliar N should then scale
with increased canopy photosynthesis and productivity (Matyssek 1986, Reich et al.
1995; but see Kloeppel et al. 2000). The combination of canopy biomass and N for
European larch explained more variation in productivity than either component alone.
This argues that greater N in a larger light-capturing, photosynthesizing canopy is
positively correlated with productivity. SLAcanopy and SLAMer were strongly related to
productivity across species. This is not surprising since high SLA leaves tend to have
high Nleapmass (at least across species) and Nleaf-mass was a strong determinant of
productivity. More surprising was the fact that SLAMer was more strongly related to
productivity than SLAQanopy both within and across species. This could reflect SLAcanopy
being more variable due to carbohydrates or non-structural material that change in
amount daily or through the season and bear no direct relation to productivity. Senesced
tissue, on the other hand, contains only basic structural components and may be a better

quantitative measure of productivity as predicted by leaf structure. Fassnacht and Gower

97

 

(1997) also found that SLAMer explained a surprisingly high proportion (76%) of
variation in productivity among conifer, mixed conifer-hardwood, and hardwood stands.

For European larch, a decrease in A also was associated with an increase in
productivity — a result opposite of that hypothesized. The decrease in discrimination
against the heavier l3C (lower A) indicates that interoellular C02 is lower, thus high
productivity trees are more efficient in their use of water but less efficient in their use of
N. The fact that Nleaf-mass not only increased independent of A, but contributed to
explaining over 50% of variation in 8125 productivity also suggests that increased
photosynthetic capacity was related to both increased productivity and increased water
use efficiency.

However, it should be noted that, if water were never limiting, a lower
interoellular C02 would result in lower photosynthesis per unit leaf N, and thus lower
photosynthetic nitrogen use efficiency. In contrast to European larch, A for red pine
increased with productivity. It is possible that, for red pine, low water constrained
productivity on some sites as reflected by lower A (i.e. more closed stomata) for low
productivity red pine stands. An increase in productivity for red pine was associated with
other unexpected results: both a smaller canopy and lower LAI. Given the increase in A,
this may suggest that the reduced amount of canopy improved water availability, thus
opening stomata wider and discriminating more against the heavier ‘3 C. Lower LAI with
increased productivity equates to a higher photosynthetic efficiency per unit needle.
However, nitrogen also increased for red pine with increases in productivity. It is
possible that higher N in less (lower canopy biomass) but more exposed (higher SLA)

foliage resulted in higher rates of productivity.

98

Other limitations of the data may give a partial explanation for inconsistencies in
relationships (e.g. canopy biomass for red pine decreased with higher productivity).
ANI’Po productivity was estimated with only three trees per stand. While the stands were
even-aged, monoculture plantations, a sample size of three trees may be too small to
represent the entire stand. Also, area-based measures (e.g. canopy biomass, LAI) were
calculated from litter caught from August to late November. There is a chance some
litter was not captured and all measurements dependent upon litterfall are underestimated.

Furthermore, another caveat for interpretation of relations of area based
productivity (ANPPO) with its determinants is that several of the study stands were
thinned in the past. Although closed canopy conditions were a stand selection criterion, it
is possible that interrelations between ANPPo and vegetation characteristics were affected

by density and past thinning regimes.

Species comparison across a productivity gradient

It was hypothesized that differences in leaf life span and associated traits for two
species would confer an advantage in productivity for each species at different positions
along a soil resource gradient; i.e. difference in species traits would be productivity trade-
offs over a resource gradient. As hypothesized, short leaf life-span species European
larch, had higher Nleaf-mass, SLA, and annual N losses but lower water use efficiency,
Nleafma and canopy biomass than long leaf life-span species red pine. Leaf life span for
red pine was less plastic than expected; 24 of the 27 stands carried three cohorts and only
three carried four cohorts. Also, it was not hypothesized that LAI would be higher for

European larch on more productive sites.

99

 

While species productivities and foliar and canopy traits were affected by
resource availability (i.e. the AWC gradient) these responses were generally parallel for
the species and only weak interactions were observed. Thus there was no support for the
hypothesis for species productivity trade-offs based on differences in leaf traits;
European larch was always more productive than red pine. Even on poor sites, European
larch had higher Nleaf-mass, LAI, and NMIN compared to red pine. As productivity of
European larch productivity increased, not only did Nieapmass increase, but so did water
use efficiency. Given that nitrogen and soil moisture were the strongest corollaries with
productivity, the leaf/canopy traits of European larch thus were able to promote

productivity across the entire site quality gradient of the study.

Conclusion

The strongest corollaries of growth for both species were foliar N, SLAimer, and
soil moisture. Simultaneously considering physical and chemical soil properties, the
supply of moisture is undoubtedly influential in the rate of productivity for both species.
Given the species effects on litter quality, and perhaps on microbial physical
environments, European larch promotes higher nitrogen cycling rates than red pine. This
may be part of the reason that European larch out produced red pine across the entire site
quality gradient. Short leaf life span European larch was associated with high Niear-mass,
annual N turnover, and SLA. The long leaf life span of red pine may be more
advantageous for productivity as stands mature (Reich 1998), but for the relatively young
stands compared, productivity was lower for red pine across a broad site quality gradient

despite it having higher Nleafma, canopy biomass, and water use efficiency.

100

Larch had higher productivity than red pine over the entire site quality gradient.
Thus there were no productivity trade-offs as was hypothesized based on species

differences in leaf and canopy traits.

101

 

 

REFERENCES

102

 

Bancroft, TA. 1964. Analyses and inferences for incompletely specified models
involving the use of preliminary tests of significance. Biometrics 20: 427 — 442.

Barnes, B.V., D.R. Zak, S.R. Denton, and SH. Spurr. 1998. Forest Ecology, 4th ed.
John Wiley & Sons, Inc., New York. 774 pp.

Bernier, P.Y., R.A. Fournier, C.H. Ung, G. Robitaille, G.R. Laroque, M.B. Lavigne, R.
Boutin, F. Raulier, D. Paré, J. Beaubien, and C. Delisle. 1999. Linking

ecophysiology and forest productivity: An overview of the ECOLEAP project. For.
Chron. 75: 417 - 421.

Bockheim, J.G., J.E. Leide, and LE. Frelich. 1989. Red pine growth and chemical
composition of foliage and forest floors across a precipitation-chemistry gradient in
Wisconsin. Can. J. For. Res. 19: 1543 — 1549.

Brady, NC. and RR. Weil. 1996. The Nature and Properties of Soils. 11th ed. Prentice
Hall, Upper Saddle River, NJ. 740 pp.

Fassnacht, KS. and ST. Gower. 1997. Interrelationships among the edaphic and stand
characteristics, leaf area index, and aboveground net primary production of upland
forest ecosystems in north central Wisconsin. Can. J. For. Res. 27: 105 8-1067.

F asshacht, KS. and ST. Gower. 1999. Comparison of the litterfall and forest floor
organic matter and nitrogen dynamics of upland forest ecosystems in north central
Wisconsin. Biogeochemistry 45: 265 — 284.

Gower, ST. and J.H. Richards. 1990. Larches: deciduous conifers in an evergreen
world. Bioscience 40: 818-826.

Gower, ST. and Y. Son. 1992. Differences in soil and leaf litterfall nitrogen dynamics
for five forest plantations. Soil Sci. Soc. Amer. J. 56: 1959-1966.

Gower, S.T., P. B. Reich, Y. Son. 1993. Canopy dynamics and aboveground production
of five tree species with different leaf longevities. Tree Physiol. (12): 327-345.

Hebert, M.T. and SB. Jack. 1998. Leaf area index and site water balance of loblolly
pine (Pinus taeda L.) across a precipitation gradient in East Texas. For Ecol.
Manage. 105: 273-282.

Kloeppel, B.D., S.T. Gower, J.G. Vogel, and PB. Reich. 2000. Leaf-level resource use
for evergreen and deciduous conifers along a resource availability gradient. F unc.
Ecol. 14: 281 - 292.

Kloeppel, B.D., S.T. Gower, I.W. Treichel, and S. Kharuk. 1998. Foliar carbon isotope

discrimination in Larix species -and sympatric evergreen conifers: a global
comparison. 0ecologia 114: 153-159.

103

 

 

 

Linn, D.M., and J .W. Doran. 1984. Effect of water-filled pore space on carbon dioxide
and nitrous oxide production in tilled and non—tilled soils. Soil Sci. Soc. Am. J. 48:
1267-1272.

Matyssek, R. 1986. Carbon, water and nitrogen relations in evergreen and deciduous
conifers. Tree Physiol. 2: 177-187.

Pastor, J., J.D. Aber, C.A. McClaugherty, and J.M. Melillo. 1984. Aboveground
production and N and P cycling along a nitrogen mineralization gradient on
Blackhawk Island, Wisconsin. Ecology 65: 256-268.

Reich, P.B., M.B. Walters, D. S. Ellsworth. 1992. Leaf life-span in relation to leaf,
plant, and stand characteristics among diverse ecosystems. Ecol. Monogr. 62: 365-
392.

Reich, P.B., B.D. Kloeppel, D.S. Ellsworth, and MB. Walters. 1995. Different
photosynthesis-nitrogen relations in deciduous hardwood and evergreen coniferous
tree species. 0ecologia 104: 24-30.

Reich, P.B., M.B. Walters, and D.S. Ellsworth. 1997a. From tropics to tundra: global
convergence in plant functioning. Proceedings of the National Academy of Sciences
USA 94: 13730—13734.

Reich, P.B., D.F. Grigal, J.D. Aber, and ST. Gower. 1997b. Nitrogen mineralization
and productivity in 50 hardwood and conifer stands on diverse soils. Ecology 78:
335-347.

Reich, PB. 1998. Variation among plant species in leaf turnover rates an associated
traits: implications for growth at all life stages. In: H. Lambers, H. Poorter, and
M.M.I. Van Vuuren (eds.), Inherent variation in plant growth. Physiological
mechanisms and ecological consequences. Backhuys Publishers, Leiden.

Reich, P.B., D.S. Ellsworth, ME. Walters, J .M. Vose, C. Gresham, J .C. Volin, and W.D.
Bowman. 1999. Generality of leaf trait relationships: a test across six biomes.
Ecology 80: 1955-1969.

Sartz, RS. and D.N. Tolsted. 1976. Snowmelt runoff from planted conifers in
southwestern Wisconsin. US. For. Serv. Res. Pap. NC-205.

Scott, NA. and D. Binkley. 1997. Foliage litter quality and annual net N mineralization:
comparison across North American forest sties. 0ecologia 111: 151 —— 159.

Sokal, RR. and F .J . Rohlf. 1961. Biometry; The Principles and Practice of Statistics in
Biological Research. W.H. Freeman and Company, San Francisco. 776 pp.

104

 

Tyrrell, LE. and R.E.J. Boemer. 1986. Larix laricina and Picea mariana: relationships
among leaf life-span, foliar nutrient patterns, nutrient conservation, and growth
efficiency. Can. J. Bot. 65: 1570-1577.

Yin, X. 1993. Variation in foliar nitrogen concentration by. forest type and climatic
gradients in North America. Can. J. For. Res. 23: 1587-1602.

Zak, D.R., G.E. Host, and KS. Pregitzer. 1989. Regional variability in nitrogen
mineralization, nitrification, and overstory biomass in northern Lower Michigan.
Can. J. For. Res. 19: 1521-1526.

Zak, D.R., D. Tilman, R.R. Parrnenter, C.W. Rice, F .M. Fisher, J. Vose, D. Milchunas,
C.W. Martin. 1994. Plant production and soil microorganisms in late-successional
ecosystems: a continental-scale study. Ecology 75: 2333-2347.

Zak, D.R., W.E. Holmes, N.W. MacDonald, and KS. Pregitzer. 1999. Soil temperature,

matric potential, and the kinetics of microbial respiration and nitrogen mineralization.
Soil Sci. Soc. Am. J. 63: 575-584.

105

 

 

 

 

 

 

 

r . {9.1: .,.

 

 

 

1 .. .w..~ 37.7:
.3

 

 

2. ....L.
.93". .i

 

:. :1: r
.r Fri. .1
. L.

.

. ; ._..

.5... , , . . ”fining
. :1: . . i z

.. yi
» in}. .

......p.:;;
i.

.r..r .9 ..

.. 1..
3.11.;

. (:35...

K. .

.535.

5. . a.

e .

. a...

rm.
5......5 . .

13¢... v‘vlunrx

.a. . .. .

. .3. !~ .x .e

. . . Err... ,. 53.1.!

.......a. far.»- . . has. .._._.....:

._,

5.3

..,.,_...

>-.r

:
euwum

Cain. J.“
an... . r.2.
. a £53.... _

3.... a
5 u.

.7
a1...

 

(. u