Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/239938436

InfluenceofDisturbanceonTemperateForestProductivity

ARTICLEinECOSYSTEMS·SEPTEMBER2012

ImpactFactor:3.94·DOI:10.1007/s10021-012-9599-y

CITATIONS

17

READS

338

4AUTHORS:

EmilyPeters

MinnesotaDepartmentofNaturalResources

14PUBLICATIONS96CITATIONS

SEEPROFILE

KirkRWythers

UniversityofMinnesotaTwinCities

10PUBLICATIONS212CITATIONS

SEEPROFILE

JohnBradford

UnitedStatesGeologicalSurvey

101PUBLICATIONS1,032CITATIONS

SEEPROFILE

PeterBReich

UniversityofMinnesotaTwinCities

653PUBLICATIONS43,200CITATIONS

SEEPROFILE

Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,

lettingyouaccessandreadthemimmediately.

Availablefrom:JohnBradford

Retrievedon:04February2016

Influence of Disturbanceon Temperate Forest Productivity

Emily B. Peters,1* Kirk R. Wythers,2 John B. Bradford,3 and Peter B. Reich2,4

1Institute on the Environment, University of Minnesota, Saint Paul, Minnesota 55108, USA; 2Forest Resources Department, University

of Minnesota, Saint Paul, Minnesota 55108, USA; 3United States Geological Survey, Southwest Biological Science Center, Flagstaff,

Arizona 86001, USA; 4Hawkesbury Institute for the Environment, University of Western Sydney, Penrith, New South Wales 2751,Australia

ABSTRACT

Climate, tree species traits, and soil fertility are key

controls on forest productivity. However, in most

forest ecosystems, natural and human distur-

bances, such as wind throw, fire, and harvest, can

also exert important and lasting direct and indirect

influence over productivity. We used an ecosystem

model, PnET-CN, to examine how disturbance

type, intensity, and frequency influence net pri-

mary production (NPP) across a range of forest

types from Minnesota and Wisconsin, USA. We

assessed the importance of past disturbances on

NPP, net N mineralization, foliar N, and leaf area

index at 107 forest stands of differing types (aspen,

jack pine, northern hardwood, black spruce) and

disturbance history (fire, harvest) by comparing

model simulations with observations. The model

reasonably predicted differences among forest types

in productivity, foliar N, leaf area index, and net N

mineralization. Model simulations that included

past disturbances minimally improved predictions

compared to simulations without disturbance,

suggesting the legacy of past disturbances played a

minor role in influencing current forest produc-

tivity rates. Modeled NPP was more sensitive to the

intensity of soil removal during a disturbance than

the fraction of stand mortality or wood removal.

Increasing crown fire frequency resulted in lower

NPP, particularly for conifer forest types with

longer leaf life spans and longer recovery times.

These findings suggest that, over long time periods,

moderate frequency disturbances are a relatively

less important control on productivity than climate,

soil, and species traits.

Key words: disturbance; PnET;

NPP; foliar N; N mineralization; Great Lakes.

INTRODUCTION

Net primary productivity (NPP) is an important

component of the terrestrial carbon budget and

is a key factor in determining whether a region,

landscape, or ecosystem is a long-term carbon

source or sink. Temperate and boreal forest eco-

systems, in particular, removed 1.22 Pg C y-1 from

the atmosphere on average from 1990 to 2007,

offsetting 17% of anthropogenic CO2 emissions

from global fossil fuel burning and cement pro-

duction (Pan and others 2011). Slight changes in

forest productivity at global, regional, and local

scales can have substantial effects on net carbon

sequestration rates (Goulden and others 2011) and

consequences for forest products industries (Irland

and others 2001). With a future of changing and

uncertain climate (Solomon and others 2007),

understanding the major drivers of forest NPP is

Received 4 June 2012; accepted 22 August 2012;

published online 23 September 2012

Author Contributions: Emily B. Peters: Conceived of or designed

study, performed research, analyzed data, wrote the paper. Kirk

R. Wythers: Conceived of or designed study, performed research, ana-

lyzed data. John B. Bradford: Conceived of or designed study, analyzed

data. Peter B. Reich: Conceived of or designed study, analyzed data.

*Corresponding author; e-mail: ebpeters@umn.edu

Ecosystems (2013) 16: 95–110DOI: 10.1007/s10021-012-9599-y

� 2012 Springer Science+Business Media New York

95

essential for managing forests to achieve both

ecological and economic objectives.

Climate, tree species traits, and soil fertility are

well established as key drivers of forest productiv-

ity. Climate conditions, notably temperature and

precipitation, control much of the seasonal and

inter-annual variability in NPP because of their

direct influence on the metabolic processes of

photosynthesis and plant respiration (Potter and

others 1999; Twine and Kucharik 2009; Bradford

2011). For example, in a cold climate region, war-

mer temperatures can increase rates of photosyn-

thesis and lengthen the growing season, leading to

higher NPP. Tree species traits at the ecosystem

scale, such as leaf area index (LAI) and foliar N

concentration, also influence NPP by impacting

light interception and photosynthetic efficiency

(Green and others 2003; Ollinger and others 2008;

Reich 2012). Additionally, productivity in cold

temperate forests is positively related to soil N

availability (Reich and others 1997).

However, in most forest ecosystems, natural and

anthropogenic disturbances, such as wind throw,

fire, logging, and cultivation, also exert important

influence over productivity (Foster and others

2003). Following a stand-replacing disturbance,

NPP is typically low for several years due in part to

low LAI, reaches a maximum at canopy closure,

and declines slightly as the stand matures (Odum

1969; Gower and others 1996; Ryan and others

1997; Howard and others 2004; Goulden and

others 2011). The exact magnitude and pattern of

change in NPP following a disturbance, however,

depends on the type, intensity, and prior frequency

of disturbance. Previous studies suggest that more

intense disturbances, such as clear-cut harvests

compared to selectively cut harvests, temporarily

result in lower NPP (Davis and others 2009;

Peckham and Gower 2011). In addition, repeated

stand-replacing disturbances can prevent forests

from reaching maximum rates of NPP (Gough and

others 2007) by triggering nitrogen losses through

volatilization, leaching, or organic matter export

and reducing overall soil fertility (Latty and others

2004). Despite a qualitative recognition that dis-

turbances have long-term impacts on ecosystem

function (Foster and others 2003), there is still no

systematic understanding of the short- and long-

term consequences of disturbances on productivity

in forest ecosystems (Gough and others 2008). As

disturbance regimes around the world shift because

of changes in climate, human management, and

land use (Dale and others 2000), it is important to

understand how forest productivity may respond.

Although the ability to examine disturbance

effects using chronosequence-based field studies is

often constrained by the long lifespan of trees,

small sample sizes, and confounding biophysical

factors among sites (Gough and others 2007;

Goulden and others 2011), ecosystem modeling

approaches offer an opportunity to explore indi-

vidual and combined effects of disturbance history,

type, intensity, and frequency on forest produc-

tivity over time. Ecosystem simulation models

allow exploration of past as well as potential future

effects of disturbance on ecosystem function, and

comparison with empirical data can increase con-

fidence in the trends and absolute values of model

output (Ollinger and others 2002; Balshi and others

2007; Scheller and others 2011

In this study, we used the ecosystem model

PnET-CN to characterize the roles of disturbance

and vegetation type in controlling forest produc-

tivity in the upper Great Lakes region of North

America. Disturbance regimes in this region have

undergone major changes over the nineteenth and

twentieth centuries, moving largely from wind

throw- and fire-dominated ecosystems to those

dominated by harvest (White and Host 2008).

Given the economic and ecological importance of

the region’s forests and the magnitude of historical

and expected future alterations to the disturbance

regime, it is essential to assess the role of disturbance

on forest productivity in the upper Great Lakes

region and other forested areas around the world.

Several prior studies have used PnET-CN to

either broadly account for regional disturbance

history (Pan and others 2009) or to simulate site-

specific disturbance history (Aber and Driscoll

1997; Aber and others 1997; Ollinger and others

2002; Davis and others 2009; Katsuyama and oth-

ers 2009). However, the ability to assess the realism

of these model adjustments is modest, as only two

studies included corroboration data, and did so for a

small number of replicates. One study that explic-

itly corroborated PnET-CN disturbance parameter-

izations compared N losses from six watersheds

with differing disturbance histories (uncut, burned,

typical harvest, and experimental harvests) to

modeled output (Aber and Driscoll 1997). Unfor-

tunately, this comparison did not include data on

productivity. The second study compared NPP and

foliar N output from PnET-CN with measurements

from four forest stands of differing harvest history

(Davis and others 2009). Although these two

studies provide valuable and detailed model cor-

roboration over time, a complementary cross-site

comparison of measured and modeled forest

96 E. B. Peters and others

productivity is needed for multiple stands of dif-

fering composition and disturbance history.

Our main objectives in this study were to: (1)

assess the importance of past disturbances on

aboveground NPP, net N mineralization, foliar N,

and leaf area index at 107 forest stands of differing

types (aspen, jack pine, northern hardwood, black

spruce) and disturbance history (fire, harvest) in

Minnesota and Wisconsin, USA by comparing PnET-

CN model simulations against field measurements;

and (2) assess the relative importance of disturbance

type, intensity, and frequency on forest productivity

rates in northern Minnesota. In this study, we take

advantage of a regional data set of 107 forest stands

with different disturbance histories to better assess

how past disturbances contribute to current forest

productivity by simulating responses with and

without disturbance using PnET-CN.

METHODS

Model Structure

PnET-CN is a forest ecosystem model (www.pnet.

sr.unh.edu/, July 31, 2012) that simulates carbon,

water, and nitrogen cycles at a monthly time step

(Aber and others 1997; Ollinger and others 2002).

The PnET suite of models uses generalized empiri-

cal relationships between foliar N, photosynthetic

capacity, vertical scaling of leaf mass area, and leaf

lifespan (Ellsworth and Reich 1993; Reich and

others 1998) to simulate photosynthesis and tran-

spiration in a multilayered canopy (Aber and others

1996). The canopy model is nested within a model

that estimates respiration, carbon allocation, and

water balances (Aber and others 1995), which is

further nested within a model of biomass accu-

mulation, decomposition, and nitrogen cycling

(Aber and others 1997). PnET-CN allows C and N

to interact and system N to adjust dynamically in

response to litter volume, litter quality, and

decomposition rates.

A strength of the PnET-CN model is its ability to

simultaneously simulate forest responses to many

changing environmental factors, including climate

variability, N deposition, tropospheric ozone,

atmospheric CO2, and land-use history (Aber and

others 1997, 2002; Ollinger and others 2002).

Detailed descriptions of PnET-CN’s structure and

algorithms are available elsewhere (Aber and

others 1997; Ollinger and others 2002). Here, we

describe only the disturbance-related processes in

the model, which were the focus of this study.

PnET-CN contains a set of scenario algorithms in

which physical disturbance events (for example,

wind throw, harvest, fire) can be specified and act

to transfer or reduce pools of C and N, thereby

affecting the long-term photosynthetic rates and

cycling of C and N through the system. It is

assumed that forest type remains the same over

time, as PnET-CN does not simulate changes in

species composition from natural succession or

management (Aber and Driscoll 1997). Disturbance

events are prescribed, rather than randomly gen-

erated, and described using the following four

parameters: (1) the year in which the disturbance

occurs, (2) the fraction of stand mortality caused by

the disturbance, (3) the fraction of dead wood re-

moved by the disturbance, and 4) the fraction of

soil removed by the disturbance. These parameters

are applied at the end of the year in which a dis-

turbance is prescribed. The stand mortality

parameter imposes a mortality fraction to each of

the foliage, wood, root, and reserve biomass and N

pools in addition to the 2.5% mortality applied

annually. Consequently, live biomass and N are

converted to dead biomass and N in the specified

proportion. Dead biomass and N are converted to

foliage, wood, and root litter pools and then

immediately added to the soil C and N pools. The

wood-removal parameter specifies the proportion

of dead wood biomass and N to be removed from

the system completely. The soil-removal parameter

specifies the proportion of soil C and N to be

removed from the system completely. PnET-CN

represents soil as a single ‘‘active soil pool’’ con-

taining plant-accessible N and includes both the

forest floor (O horizon) and the mineral soil layers.

Different types of disturbance events can be

simulated using various combinations of these four

parameters (Aber and Driscoll 1997; Aber and

others 2002; Ollinger and others 2002; Davis and

others 2009). For example, PnET-CN can represent

a wind throw or insect defoliation event using the

stand mortality parameter to transfer biomass to

the forest floor; a harvest using the stand mortality

and wood-removal parameters to simulate the

fraction of dead biomass that is removed or remains

on the forest floor (for example, residues); and a

fire using the stand mortality, wood-removal, and

soil-removal parameters to simulate volatilization

of both biomass and soil (Aber and Driscoll 1997).

Model Inputs

We used measured climate data from the Midwest

Regional Climate Center (http://mcc.sws.uiuc.edu/,

July 31, 2012). For each of the 107 forest stands, we

calculated 10-year monthly means of maximum and

minimum air temperature at 2 m, incoming solar

Influence of Disturbance on Temperate Forest Productivity 97

radiation, and cumulative precipitation from the

nearest weather station. To best match environ-

mental conditions associated with field measure-

ments, climate variables and N deposition rates were

averaged over the 10 years prior to when data were

collected at each stand. Current rates of wet and dry

nitrate (NO3) and ammonium (NH4) deposition

were obtained from the National Atmospheric

Deposition Project (http://nadp.sws.uiuc.edu/, July

31, 2012) monitoring sites at Voyageurs National

Park, MN and Perkinstown, WI, whereas back-

ground N deposition rates (before 1860) were esti-

mated from Galloway and others (2004). We applied

a linear ramp to interpolate between background

and current rates of N deposition. We applied a

nonlinear CO2 ramp that increased atmospheric CO2

concentrations from 280 ppm prior to 1880 to

365 ppm by 2000 (Ollinger and others 2002). We

obtained soil water-holding capacity data from the

National Conservation Resource Service’s Soil Sur-

vey Geographic Database (Matthew Peters, personal

communication).

Vegetation parameters for aspen, northern

hardwood, jack pine, and black spruce forest types

followed the original parameters described in Aber

and Driscoll (1997) and Aber and others (1997) for

northern hardwood, pine, and spruce-fir stands

(Table 1). Parameters changed from their original

values are documented in Table 1.

Corroboration Data

We compared modeling results with field mea-

surements collected from 107 forest stands in

Minnesota and Wisconsin that varied in soil water-

holding capacity and N deposition rates (Table 2).

These stands were located primarily on federal or

university-owned land. Aboveground NPP (ANPP),

leaf area index (LAI), net N mineralization, and

foliar N concentration data were acquired between

1980 and 1998 in several different studies on forest

carbon dynamics (Pastor and others 1984;

Nadelhoffer and others 1985; Fownes 1986; Gower

and others 1993; Reich and others 1997, 2001a, b,

unpublished data). Collectively, these stands have

been previously analyzed (Reich 2012). In total,

there were 30 aspen (Populus tremuloides), 43 jack

pine (Pinus banksiana), and 18 black spruce (Picea

mariana) stands from northern MN (Reich and

others 2001a, unpublished data), and 16 northern

hardwood stands from central MN at Cedar Creek

Ecosystem Science Reserve (Reich and others

2001a), southwestern WI at Coulee (Gower and

others 1993), and southern WI at Black Hawk

Island (Pastor and others 1984) and the University

of Wisconsin, Madison Arboretum (Nadelhoffer

and others 1985; Fownes 1986). Northern hard-

wood stands were dominated by sugar maple (Acer

saccharum), red oak (Quercus rubra), or white oak

(Q. alba). Dominant tree species in each forest type

represented greater than 40% of the total basal area

in each stand.

Field-measured ANPP was calculated as the sum

of wood and foliage production. Wood production

was estimated by applying species-specific allome-

tric diameter growth relationships to all trees in

each stand (Pastor and others 1984; Nadelhoffer

and others 1985; Gower and others 1993; Reich

and others 2001a, b). Diameter growth was deter-

mined from 5-year radial increment cores or a

combination of radial increment cores and repeat

diameter measurements. Foliage production at all

stands was measured using litter trap collections.

Net N mineralization and foliar N concentrations

were determined at each stand. Net N mineraliza-

tion was measured at all sites using in situ incu-

bations. The semi-open tube method was applied at

northern and central MN stands (Reich and others

2001a, b), whereas the polyethylene bag technique

was used at southwestern and southern WI stands

(Nadelhoffer and others 1983; Pastor and others

1984; Lennon and others 1985; Nadelhoffer and

others 1985). Although to our knowledge no

comprehensive comparison has been made of these

two methods, they are among the most similar

techniques used to estimate net N mineralization

and have been combined in prior syntheses (Reich

and others 1997; Mueller and others 2012).

Nonetheless, it is unknown whether and how they

contribute to differences among stands. Foliar N

was calculated as the average of all species-specific

foliar N concentration measurements and weighted

by each species’ fractional contribution to total

basal area.

LAI was measured somewhat differently across

studies. At Black Hawk Island and the UW Arbo-

retum, LAI was estimated by multiplying the lit-

terfall mass of each species by its specific leaf area,

then summing across species (Fownes 1986). At

Coulee, LAI was determined by applying species-

specific allometric relationships of leaf area and

DBH to all trees in the stand (Gower and others

1993). At northern and central MN stands, LAI was

estimated as the average of two methods: (1) lit-

terfall mass was converted to area using measured

specific leaf area and leaf life-span (for evergreen

conifer species) and (2) percent canopy openness

was converted to LAI using a simple light-extinc-

tion coefficient model (Reich and others 2001a, b).

Values derived from each approach listed above

98 E. B. Peters and others

Table 1. Vegetation parameters in PnET-CN Representing Aspen, Northern Hardwood, Jack Pine, and BlackSpruce Forest Types

Parameter Description Aspen Northern hardwood Jack pine Black spruce

AmaxA Intercept (A) and slope (B) for

foliar N–photosynthesis

relationship (nmol CO2

g-2 leaf s-1)

-46 -46 5.3 5.3

AmaxB 71.9 71.9 21.5 21.5

AmaxFrac Amax as fraction of early

morning instantaneous rate

0.75 0.75 0.75 0.75

BaseFolRespFrac Respiration as fraction of max

photosynthesis

0.1 0.1 0.1 0.1

CFracBiomass Carbon fraction of biomass 0.45 0.45 0.45 0.45

DVPD1 Coefficients for power func-

tion which convert VPD to

fractional loss in photosyn-

thesis

0.05 0.05 0.05 0.05

DVPD2 2 2 2 2

f Soil H2O release parameter 0.04 0.04 0.04 0.04

FastFlowFrac Fraction of H2O lost to

drainage

0.1 0.1 0.1 0.1

FLPctN Minimum [N] in foliar litter 0.009 0.009 0.0045a 0.0025a

FolNConRange Max fractional increase in [N] 0.6 0.6 0.7 0.6

FolNRetrans Fraction of foliar N lost as

litter N (litter N/foliar N)

0.5 0.5 0.5 0.5

FolRelGrowMax Maximum relative foliage

growth rate (y-1)

0.95 0.95 0.3 0.3

FolReten Foliage retention time (y) 1 1 2.3 4

GDDFolEnd Growing degree day end of

foliage production

764a 764a 1031a 1400

GDDFolStart Growing degree day onset of

foliage production

332b 332b 332b 300

GDDWoodEnd End of wood production

(growing degree days)

764b 764b 1031b 1400

GDDWoodStart Wood production onset

(growing degree days)

332b 332b 332b 300

GRespFrac Growth respiration, fraction

of allocation

0.25 0.25 0.25 0.25

HalfSat Half saturation light level

(lmol CO2 m-2 s-1)

200 200 200 200

k Canopy light attenuation

constant

0.58 0.58 0.5 0.5

Kho (Ksom) Decomposition constant for

SOM (y-1)

0.075 0.075 0.075 0.075

MaxNStore Max N content in mobile

plant pool (g m-2)

76c 76c 76c 76c

MinWoodFolRatio Wood: foliage C allocation

minimum

0.8a 0.8a 0.5a 1a

NImmobA Coefficients of fraction of

mineralized N remobilized

as function of SOM C:N

151 151 151 151

NImmobB -35 -35 -35 -35

PlantCReserveFrac Plant C fraction reserved after

bud allocation

0.75 0.75 0.75 0.75

PrecIntFrac Fraction of precipitation

intercepted and evaporated

0.11 0.11 0.15 0.15

PsnTMin Minimum temperature for

photosynthesis (�C)

4 4 4 0

Influence of Disturbance on Temperate Forest Productivity 99

Table 2. Site Parameters in PnET-CN Representing 107 Forest Stands in Minnesota and Wisconsin

Parameter Description Value range

Lat Latitude (�) 43.07–48.17

WHC Soil water-holding capacity (cm) 7.8–52.0

NH4dep Total NH4 deposition (g N m2 y-1) 0.189–0.351

NO3dep Total NO3 deposition (g N m2 y-1) 0.228–0.256

Table 1. continued

Parameter Description Aspen Northern hardwood Jack pine Black spruce

PsnTOpt Optimum temperature for

photosynthesis (�C)

24 24 24 20

RespQ10 Q10 of respiration 2 2 2 2

RLPctN Minimum [N] in root litter 0.012 0.012 0.012 0.011

RootAllocA Intercept (A) and slope (B) of

the fine root–foliage rela-

tionship

0 0 0 0

RootAllocB 2.63d 2.63d 2.63d 2.63d

RootMRespFrac Fine root maintenance respi-

ration to biomass produc-

tion ratio

1 1 1 1

RootTurnOverA Quadratic coefficients for fine

root turn over (fraction y-1)

as a function of annual N

mineralization

0.789 0.789 0.789 0.789

RootTurnOverB 0.191 0.191 0.191 0.191

RootTurnOverC 0.0211 0.0211 0.0211 0.0211

SLWdel DSLW as foliar mass increases

above (g m-2 g-1)

0.2 0.2 0 0

SLWmax Specific leaf weight at canopy

top (g m-2)

91e 81c 200 170

SoilMoistFract Exponent in computing the

efficiency of soil respiration

0 0 -1 0

SoilRespA Intercept (A) and slope (B) in

relationship of mean temp

and soil respiration (g C m-2

month-1)

27.46 27.46 27.46 27.46

SoilRespB 0.0684 0.0684 0.0684 0.0684

WLPctN Minimum [N] in wood litter 0.002 0.002 0.002 0.002

WoodLitCLoss Fractional loss of C mass in

wood decomposition

0.8 0.8 0.8 0.8

WoodLitLossRate Fraction transfer from dead

wood to SOM (y-1)

0.1 0.1 0.1 0.1

WoodMRespA Wood maintenance respira-

tion as a fraction of gross

photosynthesis

0.07 0.07 0.07 0.07

WoodTurnOver Annual live wood mortality

fraction

0.025 0.025 0.025 0.025

WUEConst WUE as a function of VPD

equation constant

10.9 10.9 10.9 10.9

Parameter values in bold are modified from the original values described in Aber and others (1997) and Aber and Driscoll (1997).aUnpublished data from Peter Reich.bRyu and others (2008).cDavis and others (2008).dBased on a reanalysis of the data presented in Raich and Nadelhoffer (1989) using a linear regression model through the origin, which allows foliar NPP to be reduced to zeroafter a disturbance.eRavenscroft and others (2010).

100 E. B. Peters and others

include uncertainties due to both sampling limita-

tions and experimental error, but to our knowledge

there are no systematic biases with any approach.

Additionally, no comprehensive comparison has

been made of these LAI methods, and thus, it is

unknown whether and how they contribute to

differences among stands.

Stand Disturbance Histories

Disturbance history varied considerably among the

107 forest stands. All stands at Black Hawk Island,

Cedar Creek, and most stands at the UW Arbore-

tum had experienced no stand-replacing distur-

bance events for more than 80 years, although the

Cedar Creek stands likely had some periodic fire-

wood removal and grazing until around 1940s.

Stands at Cedar Creek managed with prescribed

ground fires were not included in this study be-

cause PnET-CN cannot realistically simulate the

reduced regeneration that occurs after frequent

ground fires in savannah ecosystems. One stand at

the UW Arboretum was cleared in the 1870s, used

for agriculture until the 1930s, and then allowed to

naturally regenerate. Stands at Coulee were cleared

in 1910, used for agriculture until the 1960s, and

replanted as monoculture plantations. Stands in

northern MN experienced major logging and

intense slash fires near the turn of the twentieth

century, followed in the 1930s by implementation

of fire suppression policies and continued logging

(Heinselman 1981; White and Host 2008).

Although detailed stand-specific histories are not

known, we used these general local and

regional histories, as well as stand origin date and

Table 3. Parameters in PnET-CN Used to Simulate the Disturbance History of Forest Stands in Minnesotaand Wisconsin

Years Disturbance type Stand

mortality

fraction

Dead

wood-removal

fraction

Soil-removal

fraction

Northern MN

1880–1998 Crown fire 0.95a 0.5a 0.15

1880–1930 Clear-cut timber harvest

followed by slash fire

0.95b 0.75b 0.15

1930–1998 Clear-cut timber harvest 0.95b 0.75b 0b

NA Selective-cut timber harvest 0.2b 0.75b 0b

NA Wind throw 0.95 0 0

Cedar Creek, MN

NA NA NA NA NA

Coulee, WI

1910 Vegetation removal 0.9c 0.6c 0c

1910–1960 Agricultural period 0.05 y-1c 1c 0c

UW Arboretum, WI

1870 Vegetation removal 0.9c 0.6c 0c

1870–1930 Agricultural period 0.05 y-1c 1c 0c

Black Hawk Island, WI

NA NA NA NA NA

aBradford and others (2012).bDavis and others (2009).cOllinger and others (2002).

Table 4. Parameters in PnET-CN Used to Simu-late a Range of Disturbance Types and Intensities

Disturbance

type and

intensity

Stand

mortality

fraction

Dead wood-

removal

fraction

Soil-removal

fraction

Stand mortality event

Minor 0.10 0 0

Moderate 0.50 0 0

Severe 0.90 0 0

Wood removal event

Minor 0.95 0.10 0

Moderate 0.95 0.50 0

Severe 0.95 0.90 0

Soil removal event

Minor 0.95 0.75 0.10

Moderate 0.95 0.75 0.30

Severe 0.95 0.75 0.50

Influence of Disturbance on Temperate Forest Productivity 101

disturbance type (for example, clear-cut harvest,

crown fire) for stands in northern MN, to reconstruct

the disturbance history of each stand. Stand age

ranged from 22 to 96 years in the aspen stands, 14 to

133 years in jack pine stands, 36 to 127 years in

black spruce stands, and 59 to more than 100 years

in northern hardwood stands. In northern MN, if the

stand origin was a timber harvest before 1930, we

assumed the stand was subsequently burned in a

slash fire. If the stand origin occurred after 1930, we

assumed the stand had been previously harvested

and slash burned in 1880.

We parameterized distinct disturbance events in

PnET-CN by specifying the stand mortality, dead

wood-removal fraction, and soil-removal fraction

parameters (Table 3). On average, wild fires

have been shown to reduce whole soil (forest

floor + mineral soil) C and N storage on average by

0 to 30% (Johnson and Curtis 2001; Nave and

others 2011; Bradford and others 2012). We ran

PnET-CN for each stand from 1800 to the year

when field measurements were collected, with a

spin-up time of 1000 years.

Sensitivity of NPP to Disturbance Type,Intensity, and Frequency

To assess the sensitivity of NPP to disturbance type

and intensity, we examined how NPP responded to

a range of intensities (minor, moderate, severe) for

stand mortality, wood-removal, and soil-removal

events (Table 4). To focus on the productivity

effects of disturbance over time, rather than

changing climate or atmospheric chemistry, we ran

PnET-CN with 10-year mean climate and input

parameters set to represent aspen, northern hard-

wood, jack pine, and black spruce stands in

northern MN in 1996. Atmospheric CO2 concen-

tration (360 ppm) and N deposition rates (NO3 =

0.0190 g N m-2 month-1, NH4 = 0.0158 g N m-2

month-1) were also set to 1996 levels. To examine

the legacy effects of different disturbance intensi-

ties, we ran the model for 200 years after applying

a single disturbance event with a spin-up time of

1000 years.

To test the sensitivity of NPP to different distur-

bance frequencies, we examined how modeled NPP

responded to a range of occurrences of wind throw,

clear-cut harvest, and crown fire events. Before

extensive logging began in northern MN in the 1880s,

fires occurred on average every 50 years in jack pine

and black spruce stands, 80 years in aspen stands, and

more than 400 years in northern hardwood stands

(Heinselman 1981; White and Host 2008). To

encompass this natural range of disturbance

frequencies in northern MN, we separately ran PnET-

CN with prescribed wind throw, clear-cut harvest, or

crown fire events (Table 3) at return intervals of 400,

200, 100, and 50 years. To simulate several 400-year

disturbance cycles, we ran the model for 1500 years,

with a spin-up time of 1000 years. We used the same

input parameters and climate data (constant climate,

CO2 concentration, and N deposition rates) as

described immediately above to simulate stands in

northern MN. To compare the long-term effects of

disturbance frequencyon NPP, weaveragedNPP rates

across the 1500-year simulation period for each re-

turn interval scenario. All analyses were performed

using Matlab (version 2011b) and R (version 2.11.1).

We do not test for statistically significant differences

between model simulations, as stochasticity is not

incorporated into individual PnET-CN runs.

RESULTS

Model Corroboration

Overall, PnET-CN reasonably predicted average

aboveground NPP, foliar N concentrations, and LAI

at aspen, jack pine, northern hardwood, and black

spruce stands in MN and WI, but tended to over-

predict net N mineralization rates (Figure 1). More

specifically in aspen stands, on average PnET-CN

simulations both with and without disturbance

accurately predicted foliar NPP of 274 g m-2 y-1

and foliar N of 2.2%, but simulations with and

without disturbance over-predicted ANPP of

461 g m-2 y-1 (14 and 20%, respectively), wood

NPP of 187 g m-2 y-1 (38 and 51%, respectively),

net N mineralization of 4.0 g m-2 y-1 (86 and

103%, respectively), and LAI of 4.0 m2 m-2 (13

and 16%, respectively). In jack pine stands, on

average, the model accurately predicted ANPP

(390 g m-2 y-1), wood NPP (138 g m-2 y-1), and

foliar NPP (253 g m-2 y-1), but simulations with

and without disturbance over-predicted net N

mineralization (3.2 g m-2 y-1 by 59 and 103%,

respectively) and under-predicted foliar N (1.3% by

-14 and -11%, respectively) and LAI (3.5 m2 m-2

by -15 and -7%, respectively). In northern

hardwoods stands, on average, the model accu-

rately predicted ANPP (950 g m-2 y-1), net N

mineralization rates (9.3 g m-2 y-1), foliar N

(2.3%), and LAI (5.2 m2 m-2). In black spruce

stands, on average, the model simulations with and

without disturbance over-predicted ANPP

(209 g m-2 y-1 by 40 and 54%, respectively),

wood NPP (106 g m-2 y-1 by 38 and 52%,

respectively), foliar NPP (103 g m-2 y-1 by 41 and

56%, respectively), and LAI (2.2 m2 m-2 by 52 and

102 E. B. Peters and others

67%, respectively), and under-predicted foliar N

(0.8% by -26 and -19%, respectively). Predic-

tions differed only slightly between model simula-

tions with versus without disturbance, suggesting

disturbances played a minor role in influencing

current productivity, net N mineralization, foliar N,

and LAI at these northern forest stands. In addition,

we found no relationship between measured ANPP

and stand age for all forest types.

Model simulations generally matched the rank

order of measurements by forest type. Productivity,

net N mineralization, foliar N, and LAI werehighest at

northern hardwood stands, followed by aspen

stands, jack pine stands, and black spruce stands. As

expected, the variance in modeled output was lower

than the observed data; whereas measurements

incorporated site and climate variability among

stands, as well as measurement error, model simula-

tions were based on regionally averaged and spatially

smoothed climate, soil, and disturbance histories.

Sensitivity of NPP to Disturbance Type,Intensity, and Frequency

Disturbance type and intensity greatly influenced

the magnitude and temporal pattern of NPP, with

Figure 1. Comparison of field measurements (black bars) and the PnET-CN model without disturbance (gray bars) and

with disturbance (white bars) of A aboveground net primary productivity (ANPP) in g biomass m-2 y-1, B wood NPP in g

biomass m-2 y-1, C foliar NPP in g biomass m-2 y-1, D net nitrogen mineralization in g N m-2 y-1, E foliar nitrogen

concentration in %, and F leaf area index (LAI) in m2 m-2 for aspen (n = 30), jack pine (n = 43), northern hardwood

(n = 16), and black spruce (n = 18) forest stands in Minnesota and Wisconsin. Wood and foliar NPP were not available for

northern hardwood stands and net N mineralization data were not available for black spruce stands. Error bars ±1 standard

deviation.

Influence of Disturbance on Temperate Forest Productivity 103

modeled NPP showing greater sensitivity to soil re-

moval than stand mortality or wood removal. Gen-

erally after a disturbance, modeled NPP immediately

declined, underwent a recovery period until peak

NPP was reached, and then subsequently declined to

pre-disturbance levels (for example, aspen and black

spruce, Figure 2). Increasing the stand mortality

from 10 to 90% increased the peak NPP after a dis-

turbance, and increased the recovery time in all

forest types. However, even after a 90% mortality

event, stands recovered relatively quickly, reaching

peak NPP within 8 years in the aspen and northern

hardwood stands, 17 years in the jack pine stand,

and 28 years in the black spruce stand. Increasing

the wood removal from 10 to 90% had little effect on

post-disturbance NPP in all forest types. Increasing

the soil removal from 10 to 50% strongly reduced

NPP. For soil-removal events greater than 40%, NPP

did not recover to pre-disturbance levels even

200 years after the disturbance. Soil removal events

>50% had severe effects on N pools in the PnET-CN

model (data not shown), unrealistically forcing most

N pools to zero and foliar N concentrations to the

minimum value allowed by vegetation parameters

(Table 1). The effects of disturbance type and

intensity on NPP of jack pine and northern hard-

wood stands (data not shown) showed similar pat-

terns to the black spruce and aspen stand,

respectively (Figure 2), although they differed in

magnitude.

Figure 2. Sensitivity of

net primary productivity

(NPP), in g biomass m-2

y-1, to minor (solid line),

moderate (dashed line),

and severe (dotted line)

intensities of A, B stand

mortality, C, D dead

wood removal, and E, F

soil removal disturbance

events for aspen and

black spruce stands in

northern MN. NPP was

modeled using PnET-CN.

Disturbances were

parameterized according

to Table 4.

104 E. B. Peters and others

Disturbance frequency had a greater influence on

modeled NPP in evergreen conifer stands compared

to deciduous broadleaf stands (for example, black

spruce and aspen, Figure 3). Decreasing the return

interval of wind throw, clear-cut harvest, and

crown fire events from every 400 to 50 years in the

jack pine stand (data not shown) resulted, respec-

tively, in a 16, 18, and 25% decrease in mean

(integrated over all years) NPP. In the black spruce

stand (Figure 3), mean NPP similarly declined by

20% with decreasing return intervals for all dis-

turbance types. In the aspen (Figure 3) and north-

ern hardwood stands (data not shown), mean NPP

showed no decline with increasing disturbance

frequency, except when crown fire occurred every

50 years. The maximum age of aspen trees is about

90 years, about 70 years for jack pine, around

200 years for black spruce, and more than

250 years for red oak, white oak, and sugar maple

(Ravenscroft and others 2010). Increased mortality

at older stand ages is not accounted for in the model.

Increasing disturbance intensity in addition to

increasing disturbance frequency further reduced

mean NPP in the two evergreen forest types, but not

in the two deciduous types (for example, aspen and

black spruce, Figure 4). At long-return intervals

(200 years), increasing harvest intensity from a

selective to clear-cut resulted in 11 and 10% lower

mean NPP in black spruce (Figure 4) and jack

pine stands (data not shown), respectively. At shorter

return intervals (50 years), clear-cut harvests

reducedmeanNPP by 24and 23% in black spruceand

jack pine stands, respectively, compared to selective

cuts. In aspen (Figure 4) and northern hardwood

stands (data not shown), mean NPP showed no rela-

tionship with disturbance return interval or intensity.

For all forest types, more intense and more frequent

disturbances resulted in greater variation in NPP

across the 1500-year simulation period (Figures 3, 4).

DISCUSSION

Model Corroboration

We found that PnET-CN reasonably predicted

average ANPP, foliar N, and LAI, but tended to

over-predict net N mineralization measured at

aspen, northern hardwood, jack pine, and black

Figure 3. Mean net primary productivity (NPP), in g

biomass m-2 y-1, in relation to the return interval of

three disturbance types for aspen and black spruce stands

in northern Minnesota. NPP was modeled using PnET-

CN. Disturbance types were parameterized according to

Table 3. Means were constructed across a 1500-year

simulation period. Error bars ±1 standard deviation.

Figure 4. Mean net primary productivity (NPP), in g

biomass m-2 y-1, in relation to harvest return interval and

intensity for an A aspen and B black spruce stand in

northern MN. Clear-cut harvests (gray bars) and selective-

cut harvests (white bars) were parameterized according to

Table 3. Means were constructed across a 1500-year sim-

ulation period. Error bars ±1 standard deviation.

Influence of Disturbance on Temperate Forest Productivity 105

spruce stands. The model appeared to capture the

way in which differences in both ecosystem-scale

traits and climate would influence both ecosystem

structure and function, given that differences

among forest types involve both to a substantial

degree (Reich 2012). Model simulations that

included past disturbances modestly improved

predictions compared to simulations without dis-

turbance, suggesting that disturbances in the dis-

tant past exert minor influence over current

productivity at these northern forest stands, likely

because of the combination of long-return intervals

and moderate intensities of disturbance.

In the Mid-Atlantic and Northeast regions of the

US, prior studies have also generally found good

agreement between PnET-CN and inventory or

satellite-based estimates of NPP (Aber and others

1997; Pan and others 2006, 2009) and between

PnET-CN and measured foliar N (Aber and others

1997). Our slight over-predictions of ANPP are

consistent with known discrepancies between

modeled and measured wood NPP. Whereas our

field-measured wood NPP only represents above-

ground wood, wood allocation in PnET-CN

includes aboveground wood and coarse woody

roots. Although coarse roots can be a large biomass

pool in boreal forests, they represent a relatively

small fraction (�6%) of total NPP (Seedre and

others 2011), which is consistent with our slight

over-estimates of average ANPP in aspen, jack pine

and black spruce stands. Although LAI estimates

from PnET-CN have not been previously corrobo-

rated until this study, a different version of the

PnET model (PnET-II) was also found to reasonably

predict empirically derived LAI estimates in a

northern hardwood forest in Massachusetts, USA

(Wythers and others 2003). Because PnET-CN

assumes a horizontally uniform canopy, it cannot

accurately represent the patchy structure that is

often typical of mature black spruce stands, espe-

cially ones verging on being bogs, which largely

explains our over-predictions of LAI in this forest

type. Similar to Aber and others (1997), which

showed that PnET-CN roughly matched the rank

order of net N mineralization rates across 14 forest

stands in northeastern USA, we found PnET-CN

correctly predicted the rank order of net N miner-

alization rates across our four forest types despite

tending to over-predict mean values.

Legacy of Disturbance on Great LakesForest Productivity

Although previous studies highlight the legacy of

past disturbances in altering species composition

and forest structure in the upper Great Lakes region

(Reich and others 2001a; Friedman and Reich

2005), this study suggests that disturbances did not

significantly reduce forest productivity rates for the

sets of stands used in these studies. Consistent with

this, Reich and others (2001a), who used a subset of

the forest stands in this study, found ANPP varied

more with forest type than by stand origin (fire,

harvest) or age. Similarly, in a hardwood forest in

the Central Appalachians, Davis and others (2009)

found that carefully managed harvests, including

clear-cut, selective cut, and single tree selection cut,

only reduced short-term productivity (<15 years)

compared to an uncut reference stand. By contrast,

aspen stands in Michigan twice disturbed by harvest

and fire have been found to have lower NPP com-

pared to a reference stand because site quality was

driven lower by the disturbances (Gough and others

2007). Although the 107 stands used in this study

were not chosen in a way to mimic average dis-

turbance history, they do represent much, and

perhaps most, of the regional forests in having

occasional stand-clearing disturbances. Hence, our

results suggest that climate, soil, and species traits

have been relatively more important than distur-

bance in influencing long-term productivity rates in

northern forests. Using data from these same 107

forest stands and others, Reich (2012) showed

ANPP was strongly controlled by forest LAI, canopy

N concentrations, mean annual temperature, and

growing season length. The effects of disturbance on

total forest biomass or ecosystem carbon storage

over time may be relatively more important and

longer lasting compared to the effects on produc-

tivity (Law and others 2003; Martin and others

2005; Gough and others 2007). Our model results

suggest a decrease of between 0 and 56% in live

biomass with the disturbances we imposed in our

simulations; unfortunately, data to corroborate

these results were not available for the 107 forest

stands we simulated.

Effects of Disturbance Type, Intensity,and Frequency on Forest Productivity

The temporal patterns of simulated NPP using the

PnET-CN model were similar to those documented

in other studies. Although the overall temporal

pattern of modeled NPP after a stand-replacing

disturbance matches those observed in chronose-

quence studies of the same forest types (Wang

and others 2003; Howard and others 2004; Martin

and others 2005; Davis and others 2009; Amiro and

others 2010; Goulden and others 2011), our

simulated NPP tended to recover more quickly

106 E. B. Peters and others

(8–28 years) than has been observed in natural

forests (18–70 years). Recognizing this limitation of

the PnET-CN model, Davis and others (2009)

modified algorithms to limit foliar growth to a

percentage of the standing wood mass, preventing

a mature forest from developing until wood mass is

comparable to that of a mature forest. Although

this modified version of PnET-CN accurately pre-

dicted the magnitude and temporal pattern of

measured ANPP in a recovering West Virginia

deciduous forest, we did not apply these model

adjustments in this study because they artificially

constrain the foliar mass: wood mass ratio to a

minimum value, which is known to decrease fol-

lowing a disturbance (Wang and others 2003).

Future studies should examine how best to im-

prove PnET-CN’s recovery time following a dis-

turbance without sacrificing the accuracy of other

ecological processes. Although the fast recovery

time of modeled NPP represents a potential over-

estimate of modeled ANPP at young stand ages,

only two stands in this study were less than

20 years. In general, this study suggests that past

disturbances have a relatively short-term legacy in

influencing forest productivity in the upper Great

Lakes region.

Although fire became less common in the upper

Great Lakes region during the twentieth century

than it was previously (White and Host 2008), the

frequency and intensity of storms and fire are pre-

dicted to increase because of higher temperatures

and increased drought frequency from climate

change (Flannigan and others 2009) as well as a

build-up of fuels from fire suppression management

(Fraver and others 2011). These changes are likely

to have significant impacts on forest structure and

function (for example, Rich and others 2007; Fre-

lich and Reich 2010; Carlson and others 2011;

D’Amato and others 2011). Our findings suggest

that increased disturbance frequency will reduce

forest productivity primarily in conifer forest types

with longer recovery times. Although these studies

suggest forest productivity in the upper Great Lakes

region could be further reduced in the future, dis-

turbance intensity and frequency trends are still

difficult to predict because of the complex and

nonlinear interactions among weather, vegetation,

and people (Flannigan and others 2009).

Although few studies have assessed the relative

importance of disturbance, N deposition, elevated

CO2, and climate change factors on forest produc-

tivity in the upper Great Lakes region, Ollinger and

others (2002) found factors that enhanced tree

growth (elevated CO2 and N deposition) in the

northeastern USA were similar in magnitude to

factors that reduced growth (disturbance and

ozone). Similarly, Pan and others (2009) estimated

that NPP in the Mid-Atlantic region of the US

increased 29% over the twentieth century, and

attempted to partition this as follows: elevated CO2

contributed a +14% change in NPP, N deposition a

+17% change, climate change a +4% change, tro-

pospheric ozone a -8% change, and disturbance a

-6% change. A modeling study of boreal forests in

central Russia found that the difference in pro-

ductivity between two climate change scenarios

(11–12%) was greater than the difference between

several forest management regimes, including fire,

clear-cut, and selective harvests (Shanin and others

2011). In contrast, a modeling study of black spruce

forests in central Canada found that disturbances

(harvest and fire) had a greater impact on pro-

ductivity than climate change or CO2 fertilization

when averaged over a 150-year period (Chertov

and others 2009). The highly variable results from

these studies underscore the need for further study,

particularly in areas susceptible to increases in

disturbance frequency or intensity because of

climate change.

CONCLUSIONS

This study advances the understanding of the effects

of disturbance on forest productivity, particularly by

comparing output from the ecosystem model PnET-

CN with regional observations. The model reason-

ably predicted differences among forest types in

ANPP, foliar N, LAI, and net N mineralization.

Model simulations that included past disturbances

only modestly improved predictions of ANPP, net N

mineralization, foliar N and LAI compared to sim-

ulations without disturbance, suggesting distur-

bances in the distant past played a minor role in

influencing the current productivity of aspen, jack

pine, northern hardwood, and black spruce stands

in the upper Great Lakes region. We found that

modeled NPP is more sensitive to the intensity of

soil removal during a disturbance than the fraction

of stand mortality or wood removal. Increasing the

frequency of crown fires resulted in lower long-

term rates of NPP, specifically for conifer forest types

with longer leaf life spans and longer recovery

times. These findings illustrate the importance of

including disturbance when applying ecological

simulation models over short time periods to esti-

mate regional productivity rates and carbon bud-

gets, but that long-return interval disturbances may

be a relatively less important control than climate,

soil, and species traits over long time periods.

Influence of Disturbance on Temperate Forest Productivity 107

ACKNOWLEDGMENTS

We thank Scott Ollinger for help with PnET-CN

parameters and algorithms; Peter Snyder for advice

on climate data; Matthew Peters for providing soil

water holding capacity data; and all the researchers

whose work provided useful comparative data. This

work was funded by a Discovery Grant from the

University of Minnesota’s Institute on the Envi-

ronment, the Wilderness Research Foundation,

and the Superior National Forest. We also thank

the NSF LTER program (DEB 0620652) for support.

REFERENCES

Aber JD, Driscoll CT. 1997. Effects of land use, climate variation,

and N deposition on N cycling and C storage in northern

hardwood forests. Global Biogeochem Cycles 11(4):639–48.

Aber JD, Ollinger SV, Federer CA, Reich PB, Goulden ML,

Kicklighter DW, Melillo JM, Lathrop RG. 1995. Predicting the

effects of climate change on water yield and forest production

in the northeastern United States. Clim Res 5(3):207–22.

Aber JD, Reich PB, Goulden ML. 1996. Extrapolating leaf CO2

exchange to the canopy: a generalized model of forest pho-

tosynthesis compared with measurements by eddy correla-

tion. Oecologia 106(2):257–65.

Aber JD, Ollinger SV, Driscoll CT. 1997. Modeling nitrogen

saturation in forest ecosystems in response to land use and

atmospheric deposition. Ecol Model 101(1):61–78.

Aber JD, Ollinger SV, Driscoll CT, Likens GE, Holmes RT,

Freuder RJ, Goodale CL. 2002. Inorganic nitrogen losses from

a forested ecosystem in response to physical, chemical, biotic,

and climatic perturbations. Ecosystems 5(7):648–58.

Amiro BD, Barr AG, Barr JG, Black TA, Bracho R, Brown M,

Chen J, Clark KL, Davis KJ, Desai AR, Dore S, Engel V,

Fuentes JD, Goldstein AH, Goulden ML, Kolb TE, Lavigne

MB, Law BE, Margolis HA, Martin T, McCaughey JH, Misson

L, Montes-Helu M, Noormets A, Randerson JT, Starr G, Xiao J.

2010. Ecosystem carbon dioxide fluxes after disturbance in

forests of North America. J Geophys Res Biogeosci 115:

G00K02.

Balshi MS, McGuire AD, Zhuang Q, Melillo J, Kicklighter DW,

Kasischke E, Wirth C, Flannigan M, Harden J, Clein JS,

Burnside TJ, McAllister J, Kurz WA, Apps M, Shvidenko A.

2007. The role of historical fire disturbance in the carbon

dynamics of the pan-boreal region: a process-based analysis. J

Geophys Res Biogeosci 112(G2):ARTN G02029.

Bradford JB. 2011. Divergence in forest-type response to climate

and weather: evidence for regional links between forest-type

evenness and net primary productivity. Ecosystems 14(6):975–

86.

Bradford JB, Fraver S, Milo AM, D’Amato AW, Palik B, Shinn-

eman DJ. 2012. Effects of multiple interacting disturbances

and salvage logging on forest carbon stocks. For Ecol Manage

267:209–14.

Carlson DJ, Reich PB, Frelich LE. 2011. Fine-scale heterogeneity

in overstory composition contributes to heterogeneity of

wildfire severity in southern boreal forest. J For Res

16(3):203–14.

Chertov O, Bhatti JS, Komarov A, Mikhailov A, Bykhovets S.

2009. Influence of climate change, fire and harvest on the

carbon dynamics of black spruce in Central Canada. For Ecol

Manage 257(3):941–50.

Dale VH, Joyce LA, McNulty S, Neilson RP. 2000. The interplay

between climate change, forests, and disturbances. Sci Total

Environ 262(3):201–4.

D’Amato AW, Fraver S, Palik BJ, Bradford JB, Patty L. 2011.

Singular and interactive effects of blowdown, salvage logging,

and wildfire in sub-boreal pine systems. For Ecol Manage

262:2070–8.

Davis SC, Hessl AE, Thomas RB. 2008. A modified nitrogen

budget for temperate deciduous forests in an advanced stage

of nitrogen saturation. Glob Biogeochem Cycles 22(4):ARTN

GB4006.

Davis SC, Hessl AE, Scott CJ, Adams MB, Thomas RB. 2009.

Forest carbon sequestration changes in response to timber

harvest. For Ecol Manage 258(9):2101–9.

Ellsworth DS, Reich PB. 1993. Canopy structure and vertical

patterns of photosynthesis and related leaf traits in a decidu-

ous forest. Oecologia 96(2):169–78.

Flannigan MD, Krawchuk MA, de Groot WJ, Wotton BM,

Gowman LM. 2009. Implications of changing climate for

global wildland fire. Int J Wildland Fire 18(5):483–507.

Foster D, Swanson F, Aber J, Burke I, Brokaw N, Tilman D,

Knapp A. 2003. The importance of land-use legacies to ecol-

ogy and conservation. Bioscience 53(1):77–88.

Fownes JH. 1986. Water use and primary production of Wis-

consin hardwood forests. Dissertation, University of Wiscon-

sin, Madison, WI.

Fraver S, Jain T, Bradford JB, D’Amato AW, Kastendick D, Palik

B, Shinneman D, Stanovick J. 2011. The efficacy of salvage

logging in reducing subsequent fire severity in conifer-domi-

nated forests of Minnesota, USA. Ecol Appl 21(6):1895–901.

Frelich LE, Reich PB. 2010. Will environmental changes rein-

force the impact of global warming on the prairie-forest bor-

der of central North America? Front Ecol Environ 8:371–8.

Friedman SK, Reich PB. 2005. Regional legacies of logging:

departure from presettlement forest conditions in northern

Minnesota. Ecol Appl 15(2):726–44.

Galloway JN, Dentener FJ, Capone DG, Boyer EW, Howarth

RW, Seitzinger SP, Asner GP, Cleveland CC, Green PA, Hol-

land EA, Karl DM, Michaels AF, Porter JH, Townsend AR,

Vorosmarty CJ. 2004. Nitrogen cycles: past, present, and

future. Biogeochemistry 70(2):153–226.

Gough CM, Vogel CS, Harrold KH, George K, Curtis PS. 2007.

The legacy of harvest and fire on ecosystem carbon storage in

a north temperate forest. Glob Change Biol 13(9):1935–49.

Gough CM, Vogel CS, Schmid HP, Curtis PS. 2008. Controls on

annual forest carbon storage: lessons from the past and pre-

dictions for the future. Bioscience 58(7):609–22.

Goulden ML, McMillan AMS, Winston GC, Rocha AV, Manies

KL, Harden JW, Bond-Lamberty BP. 2011. Patterns of NPP,

GPP, respiration, and NEP during boreal forest succession.

Glob Change Biol 17(2):855–71.

Gower ST, Reich PB, Son Y. 1993. Canopy dynamics and

aboveground production of 5 tree species with different leaf

longevities. Tree Physiol 12(4):327–45.

Gower ST, McMurtrie RE, Murty D. 1996. Aboveground net

primary production decline with stand age: potential causes.

Trends Ecol Evol 11(9):378–82.

Green DS, Erickson JE, Kruger EL. 2003. Foliar morphology and

canopy nitrogen as predictors of light-use efficiency in ter-

restrial vegetation. Agric For Meteorol 115(3–4):163–71.

108 E. B. Peters and others

Heinselman ML. 1981. Fire intensity and frequency as

factors in the distribution and structure of northern ecosys-

tems. In: Mooney HH, Bonnicksen JM, Christensen NL, Lotan

JE, Reiners WA, Eds. Fire regimes and ecosystem properties.

US Forest Service General Technical Report WO-26.

pp 7–57.

Howard EA, Gower ST, Foley JA, Kucharik CJ. 2004. Effects of

logging on carbon dynamics of a jack pine forest in Sas-

katchewan, Canada. Glob Change Biol 10(8):1267–84.

Irland LC, Adams D, Alig R, Betz CJ, Chen CC, Hutchins M,

McCarl BA, Skog K, Sohngen BL. 2001. Assessing socioeco-

nomic impacts of climate change on US forests, wood-product

markets, and forest recreation. Bioscience 51(9):753–64.

Johnson DW, Curtis PS. 2001. Effects of forest management on

soil C and N storage: meta analysis. For Ecol Manage 140

(2–3):227–38.

Katsuyama M, Shibata H, Yoshioka T, Yoshida T, Ogawa A, Ohte

N. 2009. Applications of a hydro-biogeochemical model and

long-term simulations of the effects of logging in forested

watersheds. Sustain Sci 4(2):179–88.

Latty EF, Canham CD, Marks PL. 2004. The effects of land-use

history on soil properties and nutrient dynamics in northern

hardwood forests of the Adirondack Mountains. Ecosystems

7(2):193–207.

Law BE, Sun OJ, Campbell J, Van Tuyl S, Thornton PE. 2003.

Changes in carbon storage and fluxes in a chronosequence of

ponderosa pine. Glob Change Biol 9(4):510–24.

Lennon JM, Aber JD, Melillo JM. 1985. Primary production and

nitrogen allocation of field-grown sugar maples in relation to

nitrogen availability. Biogeochemistry 1(2):135–54.

Martin JL, Gower ST, Plaut J, Holmes B. 2005. Carbon pools in a

boreal mixedwood logging chronosequence. Glob Change Biol

11(11):1883–94.

Mueller KE, Hobbie SE, Oleksyn J, Reich PB, Eissenstat DM.

2012. Do evergreen and deciduous trees have different effects

on net N mineralization in soil? Ecology 93(6):1463–72.

Nadelhoffer KJ, Aber JD, Melillo JM. 1983. Leaf-litter produc-

tion and soil organic-matter dynamics along a nitrogen-

availability gradient in southern Wisconsin (USA). Can J For

Res 13(1):12–21.

Nadelhoffer KJ, Aber JD, Melillo JM. 1985. Fine roots, net pri-

mary production, and soil-nitrogen availability—a new

hypothesis. Ecology 66(4):1377–90.

Nave LE, Vance ED, Swanston CW, Curtis PS. 2011. Fire effects

on temperate forest soil C and N storage. Ecol Appl

21(4):1189–201.

Odum EP. 1969. The strategy of ecosystem development. Sci-

ence 164:262–70.

Ollinger SV, Aber JD, Reich PB, Freuder RJ. 2002. Interactive

effects of nitrogen deposition, tropospheric ozone, elevated

CO2 and land use history on the carbon dynamics of northern

hardwood forests. Glob Change Biol 8(6):545–62.

Ollinger SV, Richardson AD, Martin ME, Hollinger DY, Frolking

SE, Reich PB, Plourde LC, Katul GG, Munger JW, Oren R,

Smith ML, Paw KT, Bolstad PV, Cook BD, Day MC, Martin TA,

Monson RK, Schmid HP. 2008. Canopy nitrogen, carbon

assimilation, and albedo in temperate and boreal forests:

functional relations and potential climate feedbacks. Proc Nat

Acad Sci USA 105(49):19336–41.

Pan Y, Birdsey R, Hom J, McCullough K, Clark K. 2006. Improved

estimates of net primary productivity from MODIS satellite data

at regional and local scales. Ecol Appl 16(1):125–32.

Pan YD, Birdsey R, Hom J, McCullough K. 2009. Separating

effects of changes in atmospheric composition, climate and

land-use on carbon sequestration of US Mid-Atlantic tem-

perate forests. For Ecol Manage 259(2):151–64.

Pan YD, Birdsey RA, Fang JY, Houghton R, Kauppi PE, Kurz

WA, Phillips OL, Shvidenko A, Lewis SL, Canadell JG, Ciais P,

Jackson RB, Pacala SW, McGuire AD, Piao SL, Rautiainen A,

Sitch S, Hayes D. 2011. A large and persistent carbon sink in

the world’s forests. Science 333(6045):988–93.

Pastor J, Aber JD, McClaugherty CA, Melillo JM. 1984. Above-

ground production and N and P cycling along a nitrogen

mineralization gradient on Blackhawk Island, Wisconsin.

Ecology 65(1):256–68.

Peckham SD, Gower ST. 2011. Simulated long-term effects of

harvest and biomass residue removal on soil carbon and

nitrogen content and productivity for two Upper Great

Lakes forest ecosystems. Glob Change Biol Bioenergy 3(2):

135–47.

Potter CS, Klooster S, Brooks V. 1999. Interannual variability in

terrestrial net primary production: exploration of trends and

controls on regional to global scales. Ecosystems 2(1):36–48.

Raich JW, Nadelhoffer KJ. 1989. Belowground carbon allocation

in forest ecosystems: global trends. Ecology 70(5):1346–54.

Ravenscroft C, Scheller RM, Mladenoff DJ, White MA. 2010.

Forest restoration in a mixed-ownership landscape under

climate change. Ecol Appl 20(2):327–46.

Reich PB. 2012. Key canopy traits drive forest productivity. Proc

R Soc B 279(1736):2128–34.

Reich PB, Grigal DF, Aber JD, Gower ST. 1997. Nitrogen min-

eralization and productivity in 50 hardwood and conifer

stands on diverse soils. Ecology 78(2):335–47.

Reich PB, Ellsworth DS, Walters MB. 1998. Leaf structure

(specific leaf area) modulates photosynthesis-nitrogen rela-

tions: evidence from within and across species and functional

groups. Funct Ecol 12(6):948–58.

Reich PB, Bakken P, Carlson D, Frelich LE, Friedman SK, Grigal

DF. 2001a. Influence of logging, fire, and forest type on bio-

diversity and productivity in southern boreal forests. Ecology

82(10):2731–48.

Reich PB, Peterson DW, Wedin DA, Wrage K. 2001b. Fire and

vegetation effects on productivity and nitrogen cycling across

a forest-grassland continuum. Ecology 82(6):1703–19.

Rich RL, Frelich LE, Reich PB. 2007. Wind-throw mortality in

the southern boreal forest: effects of species, diameter and

stand age. J Ecol 95(6):1261–73.

Ryan MG, Binkley D, Fownes JH. 1997. Age-related decline

in forest productivity: pattern and process. Adv Ecol Res

27:213–62.

Ryu SR, Chen J, Noormets A, Bresee MK, Ollinger SV. 2008.

Comparisons between PnET-Day and eddy covariance based

gross ecosystem production in two Northern Wisconsin

forests. Agric For Meteorol 148(2):247–56.

Scheller RM, Hua D, Bolstad PV, Birdsey RA, Mladenoff DJ.

2011. The effects of forest harvest intensity in combination

with wind disturbance on carbon dynamics in Lake States

Mesic Forests. Ecol Model 222(1):144–53.

Seedre M, Shrestha BM, Chen HYH, Colombo S, Jogiste K. 2011.

Carbon dynamics of North American boreal forest after

stand replacing wildfire and clearcut logging. J For Res

16(3):168–83.

Shanin VN, Komarov AS, Mikhailov AV, Bykhovets SS. 2011.

Modelling carbon and nitrogen dynamics in forest ecosystems

Influence of Disturbance on Temperate Forest Productivity 109

of Central Russia under different climate change scenarios and

forest management regimes. Ecol Model 222(14):2262–75.

Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB,

Tignor M, Miller HL, Eds. 2007. The physical science basis.

Contribution of working group I to the fourth assessment

report of the intergovernmental panel on climate change

2007. Cambridge: Cambridge University Press.

Twine TE, Kucharik CJ. 2009. Climate impacts on net primary

productivity trends in natural and managed ecosystems of the

central and eastern United States. Agric For Meteorol 149(12):

2143–61.

Wang CK, Bond-Lamberty B, Gower ST. 2003. Carbon distri-

bution of a well- and poorly-drained black spruce fire

chronosequence. Glob Change Biol 9(7):1066–79.

White MA, Host GE. 2008. Forest disturbance frequency and

patch structure from pre-European settlement to present in

the Mixed Forest Province of Minnesota, USA. Can J For Res

38(8):2212–26.

Wythers KR, Reich PB, Turner DP. 2003. Predicting leaf area

index from scaling principles: corroboration and conse-

quences. Tree Physiol 23(17):1171–9.

110 E. B. Peters and others