SEARCH

SEARCH BY CITATION

Keywords:

  • carbon cycle;
  • ecosystem;
  • eddy covariance;
  • gross primary productivity (GPP);
  • leaf;
  • respiration;
  • soil;
  • wood

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  • • 
    Quantitative assessment of carbon (C) storage by forests requires an understanding of climatic controls over respiratory C loss. Ecosystem respiration can be estimated biometrically as the sum (RΣ) of soil (Rs), leaf (Rl) and wood (Rw) respiration, and meteorologically by measuring above-canopy nocturnal CO2 fluxes (Fcn).
  • • 
    Here we estimated RΣ over 5 yr in a forest in Michigan, USA, and compared RΣ and Fcn on turbulent nights. We also evaluated forest carbon-use efficiency (Ec = PNP/PGP) using biometric estimates of net primary production (PNP) and RΣ and Fcn-derived estimates of gross primary production (PGP).
  • • 
    Interannual variation in RΣ was modest (142 g C m−2 yr−1). Mean annual RΣ was 1425 g C m−2 yr−1; 71% from Rs, 18% from Rl, and 11% from Rw. Hourly RΣ was well correlated with Fcn, but 11 to 58% greater depending on the time of year. Greater RΣ compared with Fcn resulted in higher estimated annual PGP and lower annual Ec (0.42 vs 0.54) using biometric and meteorological data, respectively.
  • • 
    Our results provide one of the first multiyear estimates of RΣ in a forested ecosystem, and document the responses of component respiratory C losses to major climatic drivers. They also provide the first assessment of Ec in a deciduous forest using independent estimates of PGP.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Carbon cycling by terrestrial vegetation directly affects the chemical and biological properties of an ecosystem's solid, aqueous and gas-phase components, as well as sustaining human requirements for terrestrial sources of food, fuel and fiber. Elements of this cycle have been studied for many years, and for most widespread vegetation types the essential components of the C cycle are well understood and at least qualitatively well described (Geider et al., 2001). However, a quantitative and temporally dynamic assessment of the terrestrial C cycle is of increasing interest because of concerns over anthropogenic alterations of atmospheric CO2 concentration and the possibility of managing natural vegetation for enhanced C storage (Malhi et al., 2002). Importantly, advancements in sensor technology over the past 20 yr have enabled measurements of CO2 exchange at multiple scales and with a precision and speed that allow such an assessment (Baldocchi et al., 1996).

A prominent part of current discussions of the terrestrial C cycle is how climate and ecosystem characteristics interact to affect the potential of vegetation and associated soils to store C and help mitigate anthropogenic emissions of CO2 or, conversely, how these interactions might stimulate C loss and accelerate the rate of atmospheric CO2 buildup (Melillo et al., 2002; Pendall et al., 2004; Xiao & Moody, 2004). An ecosystem's short-term C storage or loss rate in large part represents the difference between gross primary production (PGP) and the combined release of CO2 from the respiratory metabolism of autotrophs (Ra) and heterotrophs (Rh). Ecosystem respiration (Re) is the sum of Ra and Rh (see Table 1 for a list of variables used). While the accurate measurement of either PGP or Re presents formidable challenges, quantification of Re has proven particularly difficult because of gaps in our understanding of the regulatory biochemistry of respiration, and the spatially complex and interdependent array of component sources of respiratory CO2 release, including leaves, stems, roots, soil invertebrates, fungi and bacteria (Gifford, 2003).

Table 1.  Variables used and their description
VariableDescription
AbBole basal area (m2 ha−1)
AlLeaf area index (m2 m−2)
AlmaxMaximum annual Al (m2 m−2)
CafCO2 concentration immediately above the forest floor (µl l−1)
DBole diameter at 1.3 m (cm)
EcCarbon-use efficiency (dimensionless)
EcbBiometric annual Ec (dimensionless)
EcmMeteorological annual Ec (dimensionless)
FcAbove-canopy net CO2 flux (µmol CO2 m−2 s−1)
FcdDaytime Fc (µmol CO2 m−2 s−1)
FcnNocturnal Fc (µmol CO2 m−2 s−1)
inline imageEstimated daytime Re based on measured Fcn (µmol CO2 m−2 s−1)
MawAbove-ground wood mass (g m−2)
MclFine litter C mass (g m−2)
pProportional contribution of a species to Almax
PfrAnnual fine root mass production (g m−2 yr-1)
PlAnnual leaf mass production (g m−2 yr−1)
PwAnnual above- and below-ground wood mass production (g m−2 yr−1)
PNPNet annual primary production (g C m−2 yr−1)
PGPGross annual primary production (g C m−2 yr−1)
Q10Temperature-response coefficient (dimensionless)
RaAutotrophic respiration rate (µmol CO2 m−2 s−1, g C m−2 yr−1)
ReEcosystem respiration rate (µmol CO2 m−2 s−1, g C m−2 yr−1)
RhHeterotrophic respiration rate (µmol CO2 m−2 s−1, g C m−2 yr−1)
RlLeaf respiration rate, land surface area basis (µmol CO2 m−2 s−1, g C m−2 d−1, g C m−2 yr−1)
RliMean hourly Rl (µmol CO2 m−2 s−1)
RlaLeaf respiration rate, leaf area basis (µmol CO2 m−2 s−1)
RlaiMean hourly Rla (µmol CO2 m−2 s−1)
Rl15Rla normalized to 15°C (µmol CO2 m−2 s−1)
RlgLeaf growth respiration rate (g C m−2 h−1)
RsSoil respiration rate, land surface area basis (µmol CO2 m−2 s−1, g C m−2 d−1, g C m−2 yr−1)
Rs10Rs normalized to 10°C (µmol CO2 m−2 s−1)
RsiMean hourly Rs (µmol CO2 m−2 s−1)
RwAbove-ground wood respiration rate, land surface area basis (µmol CO2 m−2 s−1, g C m−2 d−1, g C m−2 yr−1)
RwiMean hourly Rw (µmol CO2 m−2 s−1)
RwvWood respiration rate on a sapwood volume basis (µmol CO2 m−3 s−1)
RwviMean hourly Rwv (µmol CO2 m−3 s−1)
Rw15Rwv normalized to 15°C (µmol m−3 s−1)
RΣSum of Rs, Rl, and Rw (µmol CO2 m−2 s−1, g C m−2 d−1, g C m−2 yr−1)
TaAir temperature (°C)
TaminMinimum air temperature at which Rla measurements were made (°C)
TlrefLeaf reference temperature (15°C)
TsSoil temperature (°C)
TsiMean hourly Ts (°C)
TsrefSoil reference temperature (10°C)
TwWood temperature (°C)
TwiMean hourly Tw (°C)
TwminMinimum wood temperature at which Rwv measurements were made (°C)
TwrefWood reference temperature (15°C)
u*Friction velocity (m s−1)
VswSapwood volume (m3 ha−1)
σStandard deviation of daily means
σR̄Standard error of R̄
βo, β1Regression coefficients (dimensionless)
θvVolumetric soil water content (%)
θviMean hourly θv (%)

Several different approaches have been used for estimating Re: the biometric approach, in which measurements of respiratory fluxes from individual ecosystem components are scaled to a common land surface area basis and summed (RΣ); the meteorological approach, which is based on eddy covariance measurements of nocturnal CO2 fluxes (Fcn) including canopy air-layer storage fluxes; and diverse modeling approaches, which generally involve a combination of basic physiological principles and empirical relationships (Mäkeläet al., 2000). While there is now a substantial database of short-term Re estimates (e.g. Sanderman et al., 2003), only rarely have annual measurements, or intercomparison of results from different measurement methods, been made. Ryan et al. (1997) used a biometric approach in several Canadian forests and estimated that annual Ra ranged from a low of 535 g C m−2 yr−1 in Pinus banksiana forests to a high of 908 g C m−2 yr−1 in Populus tremuloides stands. However, growing-season RΣ was poorly correlated with Fcn and, on average, 36% higher (Lavigne et al., 1997). Law et al. (1999b) reported the first full annual assessment of RΣ. In the mixed-age Pinus ponderosa forest they studied, RΣ was moderately correlated with Fcn and of similar magnitude during calm nights, but poorly correlated and up to 50% greater on turbulent nights. Bolstad et al. (2004) reported comparatively high annual RΣ (up to 1469 g C m−2 yr−1) in mature P. tremuloides stands and, although RΣ also was moderately correlated with Fcn, it was up to 300% higher. Wang et al. (2004) estimated annual Re in a Finnish Pinus sylvestris forest using both modeling and meteorological approaches. Their biophysical model, parameterized with respiratory data from the same site, showed an average Re of 611 g C m−2 yr−1, which did not differ significantly from Fcn-based estimates. These results indicate continued uncertainty regarding the accuracy of Re estimates, and clearly support the suggestion of Canadell et al. (2000) that multiple approaches to estimating C-cycle components is a necessary element of climate-change research.

An ecosystem's C-storage potential is also reflected in the carbon-use efficiency (Ec) of its plant community, or the fraction of PGP converted to net primary production (PNP). That is, Ec = PNP/PGP. As PGP = PNP + Ra, Ec is inherently sensitive to factors affecting plant respiration. For forests, determination of Ec is made difficult primarily by uncertainties surrounding measurements of PGP (Mäkeläet al., 2000). Biometric, meteorological and modeling approaches have each been used, and PGP estimates of similar forest types may vary considerably. For example, Janssens et al. (2001) summarized PGP estimates from eddy covariance data above European forests and found an average uptake of 1340 g C m−2 yr−1 for less disturbed stands, including both evergreen and deciduous forests. However, the ‘Pipestem’ model of Mäkelä & Valentine (2001) predicted a minimum of ≈4000 g C m−2 yr−1 for mature Scots pine growing in northern Europe, well outside the range of values reported by Janssens et al. (2001). Modeled PGP for eastern North American forests (2000–2900 g C m−2 yr−1, White et al., 1999) was also substantially higher than meteorological estimates from forests in this region (900–1500 g C m−2 yr−1, Falge et al., 2002). However, biometric estimates of PGP from old growth Pseudotsuga menziesii (Harmon et al., 2004) were within 25% of meteorological estimates from the same site (Paw et al., 2004). Differences among PGP estimates of the order 20–50%, not unreasonable given different estimation approaches, would translate into proportional differences in estimated Ec. Independent estimations and comparisons among sites and years will be necessary to resolve these differences and improve the utility of this measure in assessing forest C-storage potential.

Our objectives were to quantify Re within an aspen-dominated mixed hardwood forest typical of the northern Great Lakes region of continental North America, and to partition this respiratory CO2 flux into its primary source components of soil respiration (Rs), leaf respiration (Rl) and above-ground live wood respiration (Rw). The hardwood forests of this region cover ≈29 × 106 ha in the USA alone (USDA, 2001) and support a diverse forest products and recreational economy, as well as providing important ecological goods and services. Among the latter, C sequestration has received increased attention, and these ecosystems may play an important role in the suspected North American C sink (Fan et al., 1998). We were interested in how different climatic factors affected these sources of respiratory CO2 and how these fluxes varied interannually. We applied both biometric and meteorological approaches to estimate Re and used these results to quantify forest Ec. Our results also contribute to the comparative database on ecosystem C-cycle dynamics that is a central objective of the multinational Fluxnet program (Baldocchi et al., 2001).

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Study site

Our study was conducted at the University of Michigan Biological Station (UMBS) in northern Michigan, USA (45°35′35.4″ N, 84°42′46.8″ W), in the transition zone between the mixed hardwood and boreal forests. The study site lies on a gently sloping high outwash plain with well drained spodosolic soils (92.9% sand, 6.5% silt, 0.6% clay, pH 4.8) derived from glacial drift and classified as entic haplorthods. Mean (1942–2003) annual temperature is 5.5°C and annual rainfall 817 mm. The presettlement forest dominated by Pinus strobus L., Pinus resinosa Aiton. and Tsuga canadensis L. was cut starting in 1880 and disturbed repeatedly by subsequent cutting and fire until 1923 (Kilburn, 1960).

The forest within the 1.1 ha study plot surrounding our meteorological tower was dominated by Populus grandidentata Michx. (42% of total basal area, Ab), P. strobus (24% of total Ab), Quercus rubra L. (14% of total Ab), Acer rubrum L. (11% of total Ab), and Betula papyrifera Marsh. (7% of total Ab) (Table 2). Understory vegetation was primarily bracken fern (Pteridium aquilinum L.) and seedlings and saplings of P. strobus and A. rubrum. We used allometric equations to estimate above-ground (bole plus branch) wood mass (Maw) from measurements of diameter at 1.3 m (D) of all individuals >3.0 cm D in the 1.1 ha plot (Curtis et al., 2002). Annual above- and below-ground wood mass production (Pw) was estimated by measuring change in D using band dendrometers. Allometric equations were developed from on-site harvests (Cooper, 1981, A. W. Cooper, personal communication; Koerper, 1977) or from general allometries for north-eastern trees (Wiant et al., 1977; Ker, 1980; Young et al., 1980; Schmitt & Grigal, 1981; Crow & Erdmann, 1983; Hocker & Early, 1983; Perala & Alban, 1994; Ter-Michaelian & Korzukhin, 1997). Annual fine root mass production (Pfr) was calculated from estimates of fine root turnover from mini-rhizotron images and fine root standing stock from soil cores, and is described in more detail by Gough et al. (2005). Whole-tree sapwood volume (Vsw) was estimated on an annual basis from species-specific equations relating D to sapwood area described by Bovard et al. (2005), Maw, and wood density measurements made on site or as reported by Perala & Alban (1994).

Table 2.  Stand characteristics of the 1.1 ha study plot detailing the abundance, mean height, mean diameter at breast height (D), bole basal area (Ab), above-ground mass (Maw), sapwood volume (Vsw) and proportional contribution to maximum leaf area (p) of the dominant canopy tree species
SpeciesStems (ha−1)Height (m)D (cm)Ab (m2 ha−1)Maw (Mg ha−1)Vsw (m3 ha−1)p
  1. All measures are from 2003 except height (1997) and p (mean across years). Standard errors for height and D are in parentheses.

Populus grandidentata 26619.0 (0.2)23.9 (0.3)12.6 701010.31
Pinus strobus1373 5.7 (0.1) 6.7 (0.1) 7.1 17 260.09
Quercus rubra 12412.1 (0.6)15.7 (1.2) 4.3 33 110.24
Acer rubrum 30011.2 (0.2)10.7 (0.3) 3.4 11 190.22
Betula papyrifera 11412.8 (0.4)13.4 (0.6) 2.0  9 160.08
Fagus grandifolia  36 7.3 (0.5) 9.4 (0.7) 0.3  2  30.06
Total2214  29.7142176 

Changes in leaf area index (Al) from leaf expansion through leaf abscision were monitored using an LAI-2000 Plant Canopy Analyzer (Li-Cor, Lincoln, NE, USA). Readings were taken every 3 m along seven transects in the 1.1 ha plot for an average of 120 samples on each of ≈12 sampling dates from May to November. Maximum Al (Almax), the proportional contribution to Almax by each tree species (p), and annual leaf mass production (Pl) was measured each year using 20 litter traps (0.179 or 0.264 m2) placed in a stratified random sample throughout the 1.1 ha plot. Pinus strobus retains its needles for 2 yr, dropping its oldest needles during the early summer of their third year (≈3 months after new needle expansion initiated), so its contribution to Almax and Pl was estimated as 2.25 times that recovered in litter traps. The contribution of P. aquilinum to Almax was estimated from a census of frond density and area in 60 1 m2 subplots distributed randomly within the 1.1 ha plot.

In addition to the 1.1 ha plot, we established 60 0.1 ha plots located at 100 m intervals along radial transects extending up to 1000 m from the center of the 1.1 ha plot. Transects were located 20° apart from 255° to 15°, the primary wind direction in this area. Thus these plots allowed periodic sampling more extensively within the meteorological tower source footprint area (Schmid, 1997). Vegetation in the 0.1 ha plots was measured as described above, and was very similar in species composition to that in the 1.1 ha plot, again dominated by P. grandidentata (37% of Ab), P. tremuloides (17% of Ab), B. papyrifera (9% of Ab), Q. rubra (9% of Ab), and A. rubrum (18% of Ab), but with relatively less P. strobus (3% of Ab). Site index (base age 50 yr) of the 1.1 ha plot and eight of the 0.1 ha plots was calculated for P. grandidentata using equations from Lundgren & Dolid (1970) where the age of dominant overstory trees was estimated from growth rings.

Soil respiration

Point measurements  Point measurements of soil respiration (Rs, µmol m−2 s−1) were made using an LI-6400 portable photosynthesis system and LI-6400-09 soil CO2 flux chamber (Li-Cor). In the absence of snow cover, the chamber was placed on 0.10 m diameter polyvinyl chloride (PVC) collars inserted ≈0.02 m into the forest floor. These collars were put in place in 1998. Within the 1.1 ha plot there were eight Rs measurement stations, and at each station there were three collars spaced 1 m apart. Stations were placed randomly within each of eight quadrats covering the entire plot (stratified random sampling). Leaf litter was left in the collars, although any woody debris was removed. During periods of snow cover the existing soil respiration collars were incrementally increased in length as snow depth increased, with interlocking PVC rings, such that soil respiration was measured through the existing snow pack. Measurement protocol followed standard operating procedure for this instrument: ambient CO2 concentration just above the forest floor (Caf) was measured and, following manual placement of the chamber on the collar, the internal chamber CO2 concentration was lowered 5–25 ppm below Caf and then allowed to rise the same amount above Caf. Recorded values of Rs represent the last of three cycles of CO2 accumulation and lowering within the chamber. Measurements were made at varying times throughout the year: during the summer Rs typically was measured twice per week, but during the winter only twice per month. On a measurement day, one measurement was taken at each station, with the specific collar used alternating at random among measurement days. At each measurement station there were thermocouples inserted at 0.02 and 0.075 m into the soil, and one 0.30 m time domain reflectometry (TDR) probe (ESI model MP-917, ESI, Victoria, British Columbia, Canada). Point measurements of soil temperature (Ts) and volumetric soil water content (θv) were recorded immediately following Rs measurements. Rs was also measured four times during the 2000 growing season in 30 0.1 ha plots.

Exponential functions of the form:

  • image(Eqn 1 )

were fitted to point measurements from the 1.1 ha plot using sigmaplot (Systat Software, Inc., Richmond, CA USA), where R̄s and s are the means (n = 8) of Rs and Ts, respectively, across measurement stations on a single day. Using this expression, the temperature coefficient, inline image. Note that for soil, leaves and wood, equation 1 was developed from temperature measurements made over the course of weeks to months. Hence estimates derived from this equation necessarily reflect long-term rather than short-term temperature responses.

Curves were fitted separately for three phenological periods each year: winter, between day 280 (approximate beginning of leaf abscision) in 1 yr and day 129 (approximate beginning of leaf expansion) the following year; early season, between day 130 and day 200 (approximate mid-growing season); and late season, between days 201 and 279. Residuals from these regressions were analyzed further as either linear or logarithmic functions (based on r2) of θv using sigmaplot. Soil respiration at a soil reference temperature (Tsref) of 10°C (Rs10) was estimated from equation 1, and its standard error, inline image, where n is the number of days Rs was measured during each phenological period and σ is the standard deviation among daily means. Statistical comparisons among temperature-normalized respiration rates were made using Tukey's test at P < 0.05.

To assess spatial variability in Rs within the eddy covariance tower footprint, we compared predicted values in the 1.1 ha plot with point measurements made in plots located up to 1000 m from the tower in the direction of the prevailing north-west winds. For this analysis, Rs, Ts and θv were measured in 30 0.1 ha plots on four dates in late summer 2000 (days 214–259). Predicted Rs values were generated using the late-season 2000 Rs model specific to the 1.1 ha plot, and Ts and θv input values from 0.1 ha plots. Predicted Rs values were compared directly with actual point measurements made in the 0.1 ha plots to evaluate the agreement between Rs in the 1.1 and 0.1 ha plots at common Ts and θv. Confidence intervals for predicted Rs values in the 1.1 ha plot were generated using the PROC NLIN procedure in sas (SAS ver. 8.2; SAS Institute; Cary, NC, USA).

Scaling  Point measurements of all respiratory components were scaled to a common soil surface area basis following the methods of Ryan et al. (1997). Mean hourly Rs (Rsi, µmol m−2 s−1 for the ith hour) throughout the year was estimated from mean hourly Ts (Tsi) and θvvi) by:

  • image(Eqn 2 )

where fvi) was the linear or logarithmic function from the residual analysis described above. The standard error of Rsi, inline image, was estimated as inline image. This ignores any effects of fvi) on inline image and hence is a conservative error estimate as fvi), where significant, increases the precision of Rsi estimates. Soil temperature was measured continuously at 0.075 m depth in three locations spaced ≈10 m apart. Soil water content was continuously measured at one location in 1999 and at four locations in all other years using a CS616 soil moisture probe (Campbell Scientific, Logan, UT, USA). Output from the CS616 probes was calibrated against the TDR probes used for point measurements. Our Ts and θv point measurements encompassed the full range of continuous Ts and θv measurements. Daily and annual Rs are the sums of estimated hourly fluxes across 24 h and 1 yr, respectively. The standard error of annual Rs was estimated as the sum of hourly inline image.

Leaf respiration

Point measurements  Point measurements of leaf dark respiration (Rla, µmol m−2 s−1 expressed on a leaf area basis) for all tree species were measured at night on fully expanded detached leaves using an LI-6400. For P. aquilinum, Rla was measured at night on attached fronds and during the day on attached, darkened fronds for ambient air temperature (Ta) > 20°C. Leaf temperature in the cuvette was maintained to within ≈0.5°C of Ta. Measurements on all species except P. strobus were corrected for overestimation of Rla caused by gas flow beneath the gaskets of the LI-6400 cuvette (Pons & Welschen, 2002; unpublished data).

Measurements were conducted in the 1.1 ha plot over multiple days in 1999 and 2001. For the four canopy-level hardwood species, Rla was typically measured in six upper-canopy and six lower-canopy leaves per night, although in some cases the sample size was less. These leaves came from the two or three trees of each species we could reach from our two canopy access towers. Understory P. strobus and P. aquilinum leaves were accessed from the ground. The number of nights that measurements were taken varied among species, ranging from one for B. papyrifera to nine for P. aquilinum. As measurements were made on fully expanded tissue, Rla was assumed to represent primarily local maintenance respiration plus some additional growth-dependent costs such as phloem loading (Amthor, 2000).

Measurements were averaged across leaves within a species and canopy position to yield mean daily point values (la and T̄a). These data were combined into three groups that showed similar absolute magnitude of la and responses to temperature: P. grandidentata and Q. rubra; A. rubrum and B. papyrifera; and P. strobus and P. aquilinum. Exponential functions as in equation 1 were fitted to these mean daily values to derive estimates for regression coefficients β0 and β1 for each group.

Scaling  Leaf respiration at a leaf reference temperature (Tlref) of 15°C (Rl15) was estimated from equation 1. Mean hourly Rla (Rlai, µmol m−2 s−1) throughout the year was estimated from Rl15 and mean hourly Ta (Tai) as in equation 2 but with no θv effects. Air temperature was measured continuously at one location 21 m above the forest floor. The minimum air temperature at which Rla measurements were made (Tamin) generally was consistent with minimum Tai during the leaf expansion period for the deciduous species, and the fitted exponential function was used for all Tai without modification. This was not true, however, for the evergreen P. strobus. For that species we assumed a linear decline in Rlai between Tai = Tamin (= 14.5°C) and Tai = 0°C, and that Rlai = 0 when Tai ≤ 0°C.

Mean hourly leaf respiration on a leaf area basis was scaled to a land surface area basis (Rli, µmol m−2 s−1) by:

  • Rli = Rlai × p × Al(Eqn 3 )

For the deciduous species, Al was assumed to increase linearly during leaf expansion and decline linearly during leaf abscision. Leaf growth respiration (Rlg) was estimated from Pl and a mass-based model that assumes 0.25 g respiratory CO2 produced per g tissue constructed (Cannell & Thornley, 2000), and this respiratory cost was evenly distributed across days during leaf expansion. Daily and annual Rl are the sums of Rli across 24 h and 1 yr, respectively, except during leaf expansion when Rlg was added. Standard errors of Rl15, Rli and annual Rl were estimated as for Rs.

Above-ground wood respiration

Point measurements  Point measurements of above-ground wood respiration expressed on a sapwood volume basis (Rwv, µmol m−3 s−1) were measured in the 1.1 ha plot using a custom cuvette attached to an LI-6400. The cuvette was similar to that described by Xu et al. (2000), fashioned from opaque PVC, and its operation was analogous to that of the LI-6400-09 soil CO2 flux chamber. Plastic collars, 0.10 m in diameter, were sealed to boles at ≈1.3 m above ground using silicone caulk, and left in place. For smaller diameter trees 0.052 m collars were used. The cuvette was attached to the collar with wire springs and respiratory CO2 was allowed to accumulate within the cuvette. Cuvette air was stirred with a small fan and circulated in a closed loop to the infrared gas analyzer of the LI-6400. The volume of the cuvette, tree collar, and associated tubing averaged 0.40 l for the large tree cuvette and 0.15 l for the small tree cuvette. Bole respiration was calculated from the rate of increase in cuvette air CO2 concentration as described above for Rs. Adjacent to each collar, a thermocouple was inserted to 0.01 m depth and wood temperature (Tw) was recorded during each Rwv measurement. Because D increases throughout the growing season, early and late-season Rwv included both growth and maintenance respiration, while winter Rwv was primarily maintenance respiration (Nelson, 1994). Respiratory CO2 deriving from above-ground dead wood (coarse woody debris), either standing or down, was not considered in this analysis.

Wood respiration was measured on five tree species over multiple days in 1999–2001. Generally, only one or two species were measured on a given day. For the majority of days, at least three individuals per species were measured although this number ranged from one to nine. Measurements across individuals within a species were averaged to yield mean values for a given day (wv and w). Exponential functions as in equation 1 were fitted to these mean daily values to derive estimates of regression coefficients β0 and β1 for each species. Curves were fitted separately for the three phenological periods in each year as described above.

Scaling  Wood respiration at a reference temperature (Twref) of 15°C (Rw15) was estimated from equation 1. Mean hourly Rwv (Rwvi, µmol m−3 s−1) throughout the year was estimated from mean hourly Tw (Twi) as in equation 2, but with no θv effects. Bole temperature was measured continuously on four trees throughout the year. The minimum bole temperature at which Rwv measurements were made (Twmin) was not below ≈6°C for any species, although winter Twi was often well below 0°C. Rather than extrapolating the fitted temperature relationship beyond Twmin, we assumed a linear decline in Rwvi between Twi = Twmin and Twi = 0°C, and that Rwvi = 0 when Twi ≤ 0°C.

Mean hourly bole respiration was scaled to a land surface area basis (Rwi, µmol m−2 s−1) by:

  • Rwi = Rwvi × Σ Vsw(Eqn 4 )

where Σ Vsw is the summed individual tree Vsw within the 1.1 ha plot expressed per m2 land area and incremented annually based on changes in D. Daily and annual Rw are the sums of Rwi across 24 h and 1 yr, respectively. Standard errors of Rw15, Rwi, and annual Rw were estimated as for Rs.

Above-canopy nocturnal CO2 flux

We used eddy covariance methods to directly measure CO2 exchanges between forest and atmosphere. Measurements were made at 46 m (approximately twice canopy height). Turbulent velocities were measured with a three-dimensional sonic anemometer (model CSAT-3, Campbell Scientific) and CO2 concentrations were measured by a closed-path infrared gas analyzer (IRGA model Li-6262, LiCor). The anemometer and IRGA data were sampled at 10 Hz for calculation of above-canopy net CO2 flux (Fc). As described by Schmid et al. (2003), hourly block averages of Fc were calculated from raw 10 Hz data from the anemometer and IRGA using Reynolds decomposition.

Nocturnal Fc (Fcn) was calculated for nights showing sustained periods of adequate turbulent mixing, defined as ≥4 h when the friction velocity (u*) > 0.35 m s−1. For nights meeting these criteria, we averaged Fc from all hours where u* > 0.35 m s−1 to yield a mean Fcn (µmol m−2 s−1). A total of 485 nights in years 1999–2001 met these criteria and were used in this analysis.

Ecosystem carbon-use efficiency

We calculated annual ecosystem Ec using biometrically and meteorologically derived estimates of PGP. In both cases, PNP was calculated as:

  • PNP = Pw + Pl + Pfr(Eqn 5 )

Biometric annual Ec (Ecb) was calculated as:

  • Ecb = PNP/(PNP + | Ra |)(Eqn 6 )

where annual autotrophic respiration, Ra = Rr + Rl + Rw. Root respiration, Rr, was estimated as 0.5 × Rs based on our analysis of root-free mineral soil and O-horizon respiration compared with total Rs (Gough et al., 2005). This partitioning of soil autotrophic and heterotrophic components matches the average value reported by Hanson et al. (2000) but ignores likely seasonal variation in root compared with soil microbial respiration.

Meteorological Ec (Ecm) was calculated as:

  • image(Eqn 7 )

where inline image is the annual sum of hourly daytime ecosystem CO2 flux (Fcd) plus the absolute value of estimated daytime ecosystem respiration for each hour based on measured nocturnal CO2 fluxes (inline image). inline image was estimated from exponential functions as in equation 1 where hourly Fcn having u* > 0.35 m s−1 was regressed against Ts measured at 0.02 m. Separate regressions were fitted for early season, late season and winter periods in each year. Gap-filling procedures for missing Fcd values were as described by Schmid et al. (2003).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Climate and phenology

Patterns of Ta and Ts across the study period were typical for the upper Great Lakes region, with daily average Ta rarely exceeding 25°C during the summer, but remaining below 0°C for extended periods during the winter (Fig. 1a). Persistent snow cover during the winter effectively insulated the soil, with soil at 0.075 m rarely freezing (Fig. 1b). The winter of 2002/03 was exceptionally cold, however, resulting in Ts < 0°C for 90 d. Low Ta during this period resulted in 2003 having the lowest mean annual Ta and Ts of the 5 yr studied. Mean growing season (day 130–279) Ts was similar in 2000 and 2003, and highest in 1999. One late-winter thaw was recorded in 2000 before leaf expansion, when high Ta and a lack of snow cover resulted in increased Ts, followed thereafter by a return to colder temperatures before a sustained warming in the spring. Patterns of θv were also typical for this region and soil type, with rapid declines in θv in the absence of rainfall during the summer, but with few periods of θv < 10% lasting longer than ≈10 d (Fig. 1c).

image

Figure 1. Major environmental variables recorded in the 1.1 ha plot across the 5 yr study period: air temperature at 21 m (Ta, a), soil temperature at 0.075 m (Ts, b), and volumetric soil water content (θv, c). Mean annual and mean growing season (italic) Ta and Ts are shown for each year.

Download figure to PowerPoint

The initiation of leaf expansion and leaf abscision was similar for years 1999–2001, which as a group were ≈15 d advanced in both measures relative to 2002–03 (Fig. 2). Maximum Al measured from litter traps or assessed optically varied ≈20% during these years, being relatively higher in 2002 and 2003, and lower in 1999 and 2001. The majority of this leaf area was contributed by P. grandidentata, Q. rubra and A. rubrum (Table 2). The understory fern P. aquilinum contributed an additional 0.5 m2 m−2 leaf area, and showed similar phenological timing to the canopy tree species.

image

Figure 2. Vegetation area index (Av), measured optically and recorded between leaf expansion and leaf abscision of the deciduous canopy species in the 1.1 ha plot. Maximum leaf area index (Almax) for each year was measured from litter traps after leaf abscision.

Download figure to PowerPoint

Soil respiration

Soil respiration was well explained by seasonal variation in Ts and θv although the magnitude of Rs responses to these climate drivers varied across years (Table 3; Fig. 3). Winter Rs was never responsive to θv and showed little interannual variation in Q10 (data not shown), so a common temperature-response function was used during winter for all years. There was considerable interannual variation in the influence of θv on Rs during the growing season, however. During 1999, θv was a significant factor (P < 0.05) both early and late in the season, in 2003 θv was never significant, and in the remaining years its significance alternated among seasons (Table 3). Q10 varied between ≈2 and 3 across years, but there was no consistent rank order among seasons.

Table 3.  Seasonal soil respiration rates (µmol m−2 s−1) normalized to 10°C (Rs10), temperature response coefficients (Q10) and the significance of soil water content [fv)] in explaining residual variation in soil respiration across 5 yr in the 1.1 ha plot
YearSeasonRs10Q10fv)n
  • Early growing season was day 130–200; late growing season, day 201–279; winter, day 280 in year x – 1 through day 129 in year x, except in 1999 when winter began on day 1, and for the second winter period in 2003 which ended on day 365.

  • Standard error of Rs10 shown in parentheses; n is the number of daily means included in the regressions.

  • Generic models based on combined values across all winters.

  • +, P < 0.1;

  • *

    , P < 0.05;

  • **

    , P < 0.01;

  • ***

    , P < 0.001; ns, P > 0.1.

  • §

    Similar superscripts within years indicate no significant difference, P < 0.05.

1999Winter2.3 (0.08)2.87ns
Early2.6 (0.10)a3.14**28
Late3.5 (0.12)b§2.12**18
2000Winter2.4 (0.10)a2.87ns18
Early3.0 (0.10)b1.96ns19
Late2.6 (0.10)ab2.85***18
2001Winter2.4 (0.09)a2.87ns19
Early2.7 (0.08)b2.11*26
Late3.2 (0.09)c2.06ns19
2002Winter2.2 (0.15)a2.87ns 8
Early2.4 (0.15)a2.38+ 8
Late2.6 (0.16)a2.66*** 7
2003Winter2.3 (0.21)a2.87ns 3
Early2.4 (0.11)a3.16ns11
Late3.5 (0.12)b2.02ns 9
Winter12.3 (0.08)2.87ns
image

Figure 3. Goodness of fit of modeled soil respiration (Rs) to observed Rs across seasons and years. Modeled values were derived from parameters shown in Table 2; observed values are daily means. Solid line, linear regression (y = 0.48 + 0.90x, r2 = 0.76, n = 211); dashed line, 1 : 1 relationship.

Download figure to PowerPoint

Soil respiration rates normalized to 10°C showed significant seasonal variation as well, but exhibited more consistent relationships across seasons (Table 3). Winter Rs10 were consistently the lowest, averaging 2.3 µmol m−2 s−1 across years. Late-season Rs10 was typically the highest, averaging 3.1 µmol m−2 s−1 across years compared with an average of 2.6 µmol m−2 s−1 early in the season. Modeled Rs, based on these Ts and θv relationships, was well correlated with observed Rs (Fig. 3), with the slope of this relationship not differing significantly from 1 (P = 0.76; two-tailed t-test). There was a tendency for modeled Rs to overestimate observed Rs below 2 µmol m−2 s−1 and the model goodness-of-fit decreased with increasing Rs.

Leaf respiration

The temperature response of Rl was best characterized in three species: P. grandidentata, A. rubrum and P. aquilinum (Fig. 4), which together accounted for >50% of Almax. More limited data were available for the remaining species, therefore each was combined with one of the first group based on similarity of response within the measured temperature range. Populus grandidentata and Q. rubra also had comparatively high leaf [N] (2.0 and 2.4%, respectively) relative to A. rubrum and B. papyrifera (1.5 and 1.8%, respectively). Pinus strobus and P. aquilinum were less similar in this regard (1.3 and 2.2% leaf [N], respectively) but both primarily grew in the understory and had similar Rl at Ta≈18°C. For P. grandidentata and A. rubrum, upper canopy leaves had higher Rl compared with lower canopy leaves at similar Ta. Leaf respiration rates normalized to 15°C reflected these groupings (Table 4). Populus grandidentata and Q. rubra had significantly higher Rl15 than all other species, A. rubrum was intermediate, followed by B. papyrifera, P. strobus and P. aquilinum. Temperature-response coefficients were fairly similar across species groups, averaging 1.75.

image

Figure 4. Response of leaf respiration (Rl) to changes in ambient air temperature (Ta) in three groups of canopy species: (a) Populus grandidentata (Pgr) and Quercus rubra (Qru); (b) Acer rubrum (Aru) and Betula papyrifera (Bpa); (c) Pteridium aquilinum (Paq) and Pinus strobus (Pst). Symbols are nightly means and ±1 SE error in each variable. Open symbols, upper canopy leaves; crossed symbols, lower canopy leaves.

Download figure to PowerPoint

Table 4.  Species-specific leaf respiration (Rl15, µmol m−2 s−1), above-ground wood respiration (Rw15, µmol m−3 s−1), and their temperature response coefficients (Q10) measured early (E) or late (L) in the growing season, or during winter (W)
 SeasonSpecies
PgrQruAruBpaPstPaq
  • Respiration rates are normalized to 15°C. Standard errors in parentheses.

  • Similar superscripts indicate no significant difference, P < 0.05. Comparisons across species for Rl15, and across seasons within a species for Rw15.

  • Abbreviations as in Fig. 4.

Leaf
Rl15E,L 0.6 (0.02)a  0.6 (0.05)a 0.4 (0.02)b 0.3 (0.01)bc 0.2 (0.01)c0.3 (0.04)c
Q10E,L 1.78  1.78 1.50 1.50 1.971.97
Bole
Rw15W19.8 (2.10)a101.8 (8.96)a17.4 (1.85)a29.5 (1.76)a29.6 (2.47)a 
E41.4 (2.86)b175.9 (10.05)b37.4 (3.71)b52.9 (4.05)b48.4 (3.13)ab 
L42.1 (2.27)b180.0 (8.94)b26.0 (2.22)b48.0 (3.28)b42.4 (1.92)b 
Q10W 1.43  1.67 1.53 1.32 1.65 
E 2.72  2.10 2.85 3.11 1.50 
L 1.66  1.59 2.11 1.90 1.71 

Wood respiration

There was considerable seasonal and interspecific variation in Rw (Fig. 5; Table 4). For all four deciduous species, Q10 and Rw15 were highest early in the growing season and lowest during the winter. Pinus strobus showed little seasonal variation in Q10 but also lower Rw15 during the winter. Note that Rwv was measured on only 2 d during the winter, but these days differed in Tw by >10°C. Among the diffuse porous deciduous species, B. papyrifera had the highest Rw15 in each season and the highest mean annual Rw15 (P < 0.05, Tukey's test), followed by P. grandidentata and A. rubrum. The high absolute Rw15 in Q. rubra was caused by the comparatively small volume of sapwood in this ring-porous species.

image

Figure 5. Seasonal responses of above-ground wood respiration (Rwv) to changing wood temperature (Tw) in five tree species: Quercus rubra (a); Populus grandidentata (b); Acer rubrum (c); Pinus strobus (d); Betula papyrifera (e).

Download figure to PowerPoint

Daily and cumulative respiratory carbon losses

Continuous measurements of Ts, Tb, Ta and θv and the coefficients presented in Tables 3 and 4 were used to estimate daily respiratory C losses across years from soil, leaves and wood in the 1.1 ha plot (Table 5). In all years, Rs was the dominant component of RΣ, contributing as much as 73% of the total flux (1999), with a 5 yr mean of 71%. Leaf respiration contributed on average 18%, and Rw 11% of RΣ. There was relatively modest interannual variation in RΣ, with 164 g C m−2, or ≈10% of the 5 yr average, separating the lowest (2003) from the highest (1999) respiratory C-loss year. The largest difference in RΣ between consecutive years was 142 g C m−2, separating 1999 and 2000.

Table 5.  Yearly variation in total respiratory carbon loss (RΣ) and its 5 yr mean in the 1.1 ha plot
YearRs%Rl%Rw%RΣ
  1. Absolute and percentage contribution of soil (Rs), leaf (Rl) and wood (Rw) respiration to RΣ are shown together with the standard error in parentheses. The standard error of RΣ was calculated as the quadratic sum of respiratory component standard errors. All units are g C m−2 yr−1.

19991116 (43)73251 (11)16172 (23)111538 (50)
2000 987 (37)71251 (11)16157 (21)101396 (44)
20011005 (37)71237 (10)17171 (23)121412 (45)
2002 946 (34)67292 (13)21165 (22)121404 (43)
2003 960 (32)70250 (11)18165 (23)121375 (41)
Mean1003 (37)71256 (11)18166 (22)111425 (45)

Within a year there was considerable variation in absolute rates of respiratory C loss, and the proportional contribution of soil, leaves and wood to that loss (Fig. 6). Considering 2001 as a typical year, Rs was >90% of RΣ for most of the winter, with Rw contributing 10–20% in early spring or late autumn during periods of relatively warm Ta but outside the period of deciduous tree leaf development. Leaf respiration from the evergreen P. strobus was a negligible component of RΣ during this period. Winter RΣ averaged 1.5 g C m−2 d−1 (Fig. 7). In 2001, leaf expansion began on day 128 with 95% full leaf expansion observed on day 151. During this period, RΣ rose dramatically and the relative contribution of Rs dropped to ≈60% (Fig. 6). The abruptness of the increase in Rl during leaf expansion reflects the combined inputs of annual growth and maintenance respiration. The relative contribution of Rs to RΣ increased gradually during the growing season as soils warmed, reaching ≈75% at the time of leaf abscision in the autumn. Consequently, late-season RΣ was typically higher than early season RΣ (5 yr means, 8.5 and 6.9 g C m−2 d−1, respectively, Fig. 7)

image

Figure 6. Daily respiratory carbon loss in 2001 from soil (Rs), leaves (Rl), boles (Rw), and their sum (RΣ) in the 1.1 ha plot (upper panel). Lower panel, percentage contribution of Rs, Rl and Rw to RΣ.

Download figure to PowerPoint

image

Figure 7. Daily mean total respiratory carbon loss (RΣ) across seasons and years in the 1.1 ha plot. Winter was day 1–129 and 280–365; early season, day 130–200; late season, day 201–279.

Download figure to PowerPoint

Comparison with eddy covariance measures

Measurement of Fcn using eddy covariance methods offers the opportunity for an independent assessment of RΣ. However, only a subset of our Fcn measurements was suitable for direct intercomparison. Of 1096 possible nights (1999–2001), 485 (44%) had ≥4 h of turbulent conditions (u* > 0.35 m s−1) from which a robust average Fcn could be calculated. Overall, Fcn and RΣ were well correlted (Fig. 8, r2 = 0.77). The relationship between the two variables departed significantly from 1 : 1, however, with RΣ being greater than Fcn on most nights. The relative magnitude of the difference was not uniform across seasons, being smallest during the winter (11% greater RΣ), intermediate early in the season (28% greater RΣ), and largest late in the season (58% greater RΣ).

image

Figure 8. Correlation between mean nighttime net ecosystem CO2 flux (Fcn) measured using eddy covariance methods and total ecosystem respiration (RΣ) estimated as the sum of soil, leaf and bole respiration. Only nights having ≥ 4 h Fcn with friction velocity (u*) > 0.35 m s−1 were used. Solid line, linear relationship between variables (y = 0.15 + 1.26x, r2 = 0.77); broken line, 1 : 1 relationship.

Download figure to PowerPoint

We examined whether a systematic difference between Rs in the 1.1 ha plot compared with that across the much larger eddy covariance footprint could help explain these differences. For this analysis, Rs, Ts and θv measurements were made on four dates in 30 0.1 ha plots located up to 1000 m from the eddy covariance tower. We then predicted Rs in the 1.1 ha plot based on the Rs10, fv) and Q10 values shown in Table 2. Most of the observed Rs measurements from the 0.1 ha plots (59) were within the 95% confidence interval of modeled Rs from the 1.1 ha plot; 55 were greater than the 95% CI of modeled values; and only six values were lower (Fig. 9). This suggests that Rs in the 1.1 ha plot was equivalent to, or less than, what would be expected across the flux tower footprint. We also found that Rs was well correlated with site index (Fig. 9, insert) and that the 1.1 ha plot site index (14.4 m) was significantly less than the mean 0.1 ha plot site index (17.6 m) (P = 0.03, one-tailed t-test).

image

Figure 9. Soil respiration (Rs) measured in the 0.1 ha permanent plots within the flux tower footprint and Rs modeled for the same conditions of soil temperature and soil water content in the 1.1 ha study plot. Solid lines are 95% confidence intervals around a modeled 1 : 1 relationship (dashed line). Insert, relationship between site index (SI) and Rs for eight 0.1 ha plots (closed symbols) and the 1.1 ha plot (open symbol).

Download figure to PowerPoint

Ecosystem carbon-use efficiency

Gross primary production estimated biometrically (PGPb) as PNP + |Ra| showed similar interannual variation as seen in RΣ, but with the highest year (1999) separated from the lowest year (2003) by only 92 g C m−2 (Table 6). Biometric Ec was quite uniform across years (coefficient of variation of 2.5%), with a 5 yr mean of 0.42. Gross primary production estimated meteorologically (PGPm) as PNP/Σ(Fcd + | inline image |) was weakly correlated with PGPb (r = 0.59), and on average 23% lower. The spread between years also was somewhat greater (196 g C m−2 separating 1999 from 2003). The lower PGPm compared with PGPb estimates resulted in correspondingly higher meteorologically based Ec estimates, averaging 0.54 over 5 yr (coefficient of variation 6.0%.

Table 6.  Ecosystem carbon-use efficiency estimated biometrically (Ecb) or meteorologically (Ecm) in the 1.1 ha plot across years
YearPNPRaPGPbEcbPGPmEcm
  1. Annual gross primary production was estimated biometrically (PGPb) as the sum of net primary production (PNP) and autotrophic respiration (Ra) or meteorologically (PGPm) from eddy covariance data. All production units are g C m−2 yr−1.

  2. †Data from Gough et al. (2005).

199965698116370.4013230.50
200067890215800.4312350.55
200170491016140.4411780.60
200261893115490.4012400.50
200365089515450.4211270.58
Mean66192415850.4212210.54

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Soil respiration

The sensitivity of Rs to Ts and θv that we observed was typical for forest soils, with our overall mean Q10 across seasons and years of 2.7 comparing well with the global mean of 2.4 estimated by Raich & Schlesinger (1992). The coarse textured, well drained soils at UMBS are susceptible to episodic drought, and θv was often an important explanatory factor in modeling Rs, as has been observed in other eastern deciduous forests (Davidson et al., 1998; Ehman et al., 2002; Bolstad et al., 2004). By incorporating both seasonal and interannual variation in sensitivity to Ts and θv, our model explained ≈75% of the variation in measured Rs across 5 yr, comparable in accuracy to other empirical models of Rs from diverse forest ecosystems (Hibbard et al., 2005).

Although the observed pattern of Rs response to climate drivers was typical, hourly and cumulative annual Rs at UMBS was high compared with some other forests of similar PNP. Raich & Nadelhoffer (1989) proposed an empirical relationship that suggested annual Rs C losses of approximately three times the mass of annual above-ground fine litterfall C (Mcl). Davidson et al. (2002a) confirmed this general relationship with an independent data set drawn only from studies using infrared CO2 detection methods. Their analysis included 1 yr (1999) of data from UMBS, which was a notable outlier showing annual Rs > 7 × Mcl. This suggested the possibility of nonsteady-state root or soil C stocks or above-average total below-ground C allocation at UMBS. Our present results, based on 5 yr of data and with an improved Rs model, show more congruence with other temperate deciduous forests, particularly those dominated by Populus. Our 5 yr mean annual Rs was 1044 g C m−2 yr−1, or 5.6 times our mean Mcl of 185 g C m−2 yr−1 (Gough et al., 2005). Mature deciduous forests in Tennessee, Wisconsin and New Zealand all showed single-year annual Rs : Mcl ratios >5 (Davidson et al., 2002a). In a further analysis of the Wisconsin site, Bolstad et al. (2004) reported a 4 yr mean annual Rs of 1116 g C m−2 yr−1 from a P. tremuloides-dominated stand (Almax ≈ 4.7), while Russell & Voroney (1998) reported a 2 yr mean annual Rs of 887 g C m−2 yr−1 from a P. tremuloides forest in Saskatchewan, Canada (Almax ≈ 3.3). We cannot rule out declining stocks of soil C below 10 cm, but neither soil C from 0 to 10 cm (Schaetzl, 1994) nor total root length density (Gough et al., 2005) appears out of steady state on a 1–5 yr time frame at our site. Both P. grandidentata and P. tremuloides are early successional, rapidly growing species, and may have higher specific root respiration rates than later successional or slower growing species (Desrochers et al., 2002; Burton & Pregitzer, 2003), perhaps contributing to relatively high Rs in Populus stands. Given the importance of Rs in determining Re, resolving the underlying mechanisms responsible for variation in Rs across forest types remains an important challenge in climate change research.

Leaf respiration

There are both methodological and conceptual issues of importance in evaluating the accuracy of annual Rl estimates. We established Rl temperature response functions by measuring nocturnal Rl at different ambient temperatures over an entire growing season, rather than by exposing leaves to short-term temperature changes within the gas-exchange cuvette. There is considerable evidence that temperature acclimation of Rl occurs in temperate tree species (Atkin et al., 2000; Bolstad et al., 2003; Gifford, 2003), resulting in a relatively rapid lowering of respiratory capacity with increasing ambient temperature. As the majority of our gas-exchange measurements were made over the course of 100 d, it is very likely that temperature acclimation occurred in the individual trees we measured. One result of such an acclimatory response would be a flattening of the temperature response function and a reduction in Q10 relative to that obtained from short-term temperature manipulations (Gifford, 2003), and a consequent reduction in estimated annual Rl. Our Q10 values were at the low end of the 1.4–4.0 range for leaves reported by Amthor (1984), although comparable with those reported by Turnbull et al. (2001) for unacclimated Q. rubra (1.78–1.93) and A. rubrum (1.46–1.53). While accounting for temperature acclimation over time, our use of Q10 values derived from seasonal changes in Ta might fail to correctly describe short-term responses to diurnal temperature fluctuations. At UMBS these temperature fluctuations average ≈9°C during the growing season. On such a day, if we assume a uniform Q10 of 2.50 across species, more typical of values from unacclimated woody plants, Rl would be ≈10% higher than estimated using the ‘acclimated’Q10s in Table 3. Thus, to a first approximation, the short- vs long-term effects on estimated annual Rl of measuring acclimated vs unacclimated leaves will tend to offset each other.

We used three different Q10 functions to describe all Rl temperature responses, pooling sun and shade leaves and aggregating species based on similarity in Rl15 and leaf [N]. While obscuring some variation present at the individual tree level, these simplifications probably had little impact on our ecosystem-level estimates. In a detailed study of 18 deciduous North American tree species, Bolstad et al. (1999) concluded that most interspecific and intracanopy variation in Rl was reflected in differences in Rlref, rather than Q10. Furthermore, whole-canopy respiration predicted using the lumped parameter model PnET-II agreed well with results obtained by aggregating individual species-response curves (Vose & Bolstad, 1999). Our Rl15 values are comparable with those for other deciduous and evergreen species (Bolstad et al., 1999; Law et al., 1999b), although we also would argue that between-study differences in estimated Rl15 of <50% are effectively within current measurement error given the inaccuracy of standard leaf cuvettes (e.g. poorly quantified gasket effects) and most commercially available infrared gas analyzers working near their differential CO2 concentration detection limits. Lastly, in calculating daily and annual Rl we assumed that dark respiration continued during the day at a rate unaffected by light. This is a common, though not universal (Bolstad et al., 2004; Harmon et al., 2004), assumption in C-cycle studies, and is supported by the results of Pinelli & Loreto (2003) who found, using isotope-sensitive infrared gas analysis, that mitochondrial respiration was unaffected by light in several herbaceous and woody species. Other evidence, however, has suggested substantial reductions in Rl during the day (Brooks & Farquhar, 1985) which, if correct, would substantially reduce estimated annual Rl.

There have been only two previous reports of annual Rl in temperate deciduous forests, both including mature, Populus-dominated ecosystems. Bolstad et al. (2004), working in a northern Wisconsin aspen forest with Almax of 4.7, mean annual Ta of 4.8°C and leaf-out period of ≈150 d, reported a 4 yr mean Rl of 110 g C m−2 yr−1, summed over nocturnal periods only (P. Bolstad, personal communication). This estimate aligns very well with ours. We estimated a 5 yr average Rl of 256 g C m−2 yr−1 summed over 24 h, or 95 g C m−2 yr−1 summed over nocturnal periods only (mean Almax 4.0 including P. aquilinum, mean annual Ta 7.3°C, leaf-out period ≈160 d). Ryan et al. (1997) reported a considerably higher single-year Rl of 464 g C m−2 yr−1, assuming 24 h foliage respiration, in a southern Canadian aspen forest with Almax 3.3, mean annual Ta−0.4°C, and leaf-out period ≈120 d. However, they did not record making gasket corrections and therefore may have overestimated base Rl rates. Clearly, an analysis of biological processes that might lead to such differences must be combined with an improved understanding of the accuracy of these estimates. Independent estimates of Rl using meteorological methods (Law et al., 1999a) or the 13C signature of different sources of respiratory CO2 (Dawson et al., 2002) may provide important comparative data in this regard.

Above-ground wood respiration

Accurate assessment of annual Rw also can be a problematic element in the biometric analysis of forest Re. There are two primary reasons for this. First, CO2 fluxes measured at the stem surface may not accurately reflect the net exchange of respiratory CO2 derived from cells lying beneath the gas-exchange cuvette itself. On the one hand, vertical transport of respiratory CO2 in the xylem sap or storage in sapwood tissues will affect the magnitude of surface fluxes, generally leading to an underestimation of Rw (McGuire & Teskey, 2004). On the other hand, failure to account for refixation of respiratory CO2 via corticular photosynthesis (Strain & Johnson, 1963) can lead to an overestimation of Rw. Second, scaling point measurements to the whole-tree or stand level introduces additional, often poorly defined errors. For example, measurements made at 1.3 m may not be representative of respiratory rates at other heights or in branches, due both to variation in the density and activity of living sapwood cells (Pruyn et al., 2002) and to potentially large radial and vertical gradients in stem temperature (Stockfors, 2000). Additionally, stand-level estimates of total sapwood volume also carry with them large uncertainties (Oren et al., 1998).

Our methods were similar to those of numerous other workers, and our calculated Rw15 and Q10 values compare well with published reports (Edwards & Hanson, 1996; Ryan et al., 1996; Ryan et al., 1997; Bolstad et al., 2004). Although we did not directly measure the effects of sap flow on Rw, an examination of meteorological data on days during the growing season when measurements were taken showed no clear relationship between Rw and vapor pressure deficit (unpublished data), the primary determinant of sap-flow velocity for the canopy species at UMBS (Bovard et al., 2005). We also have not accounted for corticular photosynthesis, which certainly is present in several of our species. However, in P. grandidentata the thick bark characteristic of the mature trees at our site may act to reduce the magnitude of this effect (Cernusak & Marshall, 2000). Although we cannot rule out these potential artifacts as sources of error, the temporal variation we observed in Rwv was consistent with expectations based on the seasonality of growth and maintenance respiration in trees at our site. Winter Rw15 and Q10 were lowest, reflecting the predominance of maintenance respiration at this time (Nelson, 1994). Bole radial growth (Gough et al., 2005) and hence growth respiration generally was greatest early in the growing season, which showed the highest Rw15 and Q10, with both radial growth and Rwv then declining after day 200.

Recognizing these potential sources of error, estimated annual Rw was, perhaps surprisingly, quite similar in the three aspen forests studied to date. In the Canadian old aspen site, with a basal area of 27 m−2 ha−1 and height ≈20 m, single-year annual Rw was 123 g C m−2 yr−1 (Ryan et al., 1997). In the Wisconsin mature aspen site of basal area 28 m−2 ha−1 and height ≈22 m, 4 yr mean Rw was 154 g C m−2 yr−1 (Bolstad et al., 2004), while our 5 yr mean Rw was 166 g C m−2 yr−1 (basal area 30 m−2 ha−1, height ≈19 m). This degree of congruence across forests of similar composition, structure and climate regime lends a measure of confidence to the accuracy of these estimates.

Summed respiratory components

At UMBS, RΣ was dominated by Rs at all times of year, varying from a high of 100% during winter to a low of ≈60% during early summer. Leaf respiration was the second greatest contributor, representing ≈30% of RΣ during leaf expansion and ≈18% on an annual basis. Above-ground wood contributed as much as 20% of RΣ during early or late winter, and ≈11% overall. This pattern of partitioning of RΣ among forest ecosystem components appears fairly typical (Lavigne et al., 1997; Law et al., 1999b; Wang et al., 2004). Hence the relatively high RΣ at UMBS compared with other deciduous forests was driven primarily by high annual Rs rather than large differences in component contributions.

The interannual variation we observed in RΣ was modest, with the highest RΣ year (1999) differing from the lowest (2003) by <15%. The largest difference between any two consecutive years was 142 g C m−2, between 1999 and 2000, which differed in growing season air and soil temperatures by ≈1°C. This difference in RΣ, while small relative to annual RΣ (10% of the 5 yr mean) is nonetheless 50–100% of annual C storage in this, and other, northern hardwood forests (Lee et al., 1999; Barford et al., 2001; Curtis et al., 2002; Schmid et al., 2003; Gough et al., 2005). This result supports the conclusions of Law et al. (1999b) that small changes in respiratory fluxes driven by small differences in temperature can have important effects on the overall magnitude of ecosystem C storage.

Biometric and meteorological comparison

Both biometric and meteorological approaches to the measurement of Re carry with them significant sources of uncertainty. One benefit of colocating research using both strategies is the possibility of intercomparison of results from methods with independent errors, and thus the potential for assessment of accuracy and constraining flux estimates (Baldocchi, 2003). The direct intercomparison of RΣ and Fcn has been reported only rarely (Lavigne et al., 1997; Law et al., 1999b; Bolstad et al., 2004) and with varying results. However, with one exception (Law et al., 1999b, when u* < 0.2), RΣ was greater than Fcn, often considerably. Our results fit this pattern as well. We found good correlation between average nightly RΣ and Fcn on nights with sustained turbulence, but also a systematic offset of between +11% in the winter and +58% late in the growing season.

One possible reason for incongruence between RΣ and Fcn is differences in the ‘footprints’ of the two methods. Perhaps the 1.1 ha plot, which surrounds our meteorological tower, had significantly higher RΣ than those landscape elements contributing to the eddy covariance signal on turbulent nights. Two lines of evidence suggest this was not the case. First, Rs measured in 30 0.1 ha plots distributed throughout the likely tower footprint was almost always similar to, or higher rather than lower than, what would be expected under similar Ts and θv conditions in the 1.1 ha plot. Second, we found a strong correlation between site index and Rs, with the 1.1 ha plot having a relatively low site index compared with plots in the tower footprint. Although we cannot rule out abnormally high Rl or Rw in the 1.1 ha plot, the much smaller contribution of these components to RΣ compared with Rs argues against footprint incongruity being the cause of the quantitative offset we observed in RΣ and Fcn.

A second possibility is a systematic positive bias in chamber measurements relative to the true respiratory flux. Davidson et al. (2002b) considered this possibility at length for Rs measurements and concluded that most identifiable sources of error in both closed and open-chamber systems would tend to produce negative biases, not positive ones. Butnor & Johnsen (2004) evaluated the accuracy of the LI-6400-09 over inert media with a known CO2 efflux, and also found small-to-moderate underestimations of the true flux. Failure to account for gasket effects in measuring Rl can lead to overestimation of Rl by as much as 50% (Pons & Welschen, 2002). While we corrected for this error, some degree of positive bias might have remained. In our system, reducing Rl by 50% results in a ≈12% reduction in RΣ and therefore cannot fully account for the offset. Finally, biases in chamber measurements of Rw appear as likely to be negative (vertical CO2 transport, sapwood temperature and specific activity variation) as positive (corticular photosynthesis). Hence we find no clear evidence of systematic, positive biases in the biometric estimate of RΣ. Failure to include other possible sources of respiratory CO2 in RΣ, such as that from coarse woody debris, would also cause us to underestimate Re.

Finally, eddy covariance measurements of Fcn may underestimate Re by several mechanisms. Based on long-term flux measurements over Harvard Forest and UMBS, respectively, Goulden et al. (1996) and Schmid et al. (2003) showed that underestimation occurs during weak turbulent mixing periods (low u*). The use of a u* filter inevitably creates more data gaps, however, leading to questions regarding the validity of various gap-filling methods and their effects on annual Fc estimates (Falge et al., 2001; Schmid et al., 2003). Additionally, based on principles of mass balance, both vertical (Lee, 1998; Finnigan, 1999) and horizontal (Finnigan et al., 2003) advection occurring under various atmospheric conditions and over nonflat terrain can lead to underestimation of Re. Resolving these issues and understanding their relative importance remain important research questions within the Fluxnet community.

Ecosystem carbon-use efficiency

Our biometric estimates of PGP are the first using modern gas-exchange and scaling methods published for a deciduous forest, and the second comparison of biometric and eddy covariance-based estimates of PGP– the first, by Harmon et al. (2004), being from an old-growth P. menziesii stand. Earlier estimates of forest PGP dating from the International Biological Program are summarized by Kira (1975) and Harris et al. (1975). As noted previously, PGP estimates from similar forest types can vary considerably, depending on the methods used. There is thus little benefit in a comparative analysis for narrowly constraining PGP estimates from our site. Both PGPb and PGPm estimates are easily accommodated within the range reported for temperate forests (Sanderman et al., 2003). Carbon-use efficiency, however, may be a more sensitive comparative index. Amthor (2000) argued for a fundamental constraint on Ec between 0.20 and 0.65, but noted there was little empirical basis for constraints within the range 0.40–0.65. Crop plants growing under controlled conditions approach Ec values above 0.55, while forests are well represented by Ec values below 0.45 (Gifford, 2003). Waring et al. (1998) argued for the constancy of forest Ec at ≈0.47 across stand types and ages (but see Mäkelä & Valentine, 2001).

Our biometric and meteorological Ec estimates span the 0.47 value of Waring et al. (1998), suggesting either a forest of below-average Ec (≈0.40), perhaps connected with age-related declines, or one of above-average efficiency, in good years approaching that of crop plants (e.g. 0.60 in 2001). Studies of forest succession at UMBS indicate a maximum age of P. grandidentata stands of ≈90 yr, after which composition shifts to a dominance by Q. rubra, Acer spp. and P. strobus (Cooper, 1981). The forest within the tower footprint is a mosaic of even-aged P. grandidentata stands, with a mean age across 12 of the 0.1 ha plots of 70 yr, but with some plots as young as 30 yr, and the 1.1 ha plot being 81 yr old. This element of the canopy clearly is mature and would suggest a relatively low Ec. Other elements of the canopy, however, are of mixed age, and active recruitment of Q. rubra, A. rubrum and P. strobus is under way across the landscape. It therefore does not appear possible strictly to favor one estimate over the other based on purely biological criteria. Rather, present uncertainty in PGP estimates from this and most other forests may make it impossible to resolve differences in Ec to better than ±0.1 units.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Respiratory C losses are important components of the forest C cycle and are sensitive to changing climatic conditions. In the aspen-dominated, mixed deciduous forest at UMBS, C losses from soils predominated, accounting for >70% of the estimated 1425 g C m−2 respired from the ecosystem each year. Maximum interannual variation in this loss (142 g C m−2 yr−1), while modest compared with total Re, was of a similar magnitude to overall annual ecosystem C storage. Our estimates of the carbon-use efficiency of this forest ranged from 0.40 based on biometric data and consistent with an aging aspen stand, to 0.60 based on meteorological data and consistent with a more productive, multi-aged forest.

Quantitative Re assessments such as ours include many poorly constrained sources of error. Independent estimates of Re from the same site and comparisons with other ecologically similar sites therefore are critical to assessing the accuracy of these Re measurements. Our meteorologically based estimates of Re provided important confirmation that our physiological measurements and scaling protocols could reproduce much of the short-term (hourly) and seasonal variation in Re evidenced in above-canopy nocturnal CO2 fluxes. They also showed a consistent positive offset between hourly biometric and meteorological estimates. The broad agreement of our multiyear Re estimates with those from two other North American aspen-dominated forests supported the general robustness of our annual sums. Further improvements in our confidence in Re estimates in this forest, as well as in others, is necessarily linked to continued research in these two areas: the intercomparison of well matched biometric and meteorological data and the development of high-quality, long-term data sets in comparable ecosystems. These goals present a substantial and continuing challenge to the international C-cycle community.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

We thank Karen Brantman, Kyle Jones, Peter Cowan, Kim Hunter, Ann Singsaas, Shane Lishawa and Chris Sutterly for assistance in data collection, Kim Pilegaard and Andreas Ibrom for comments on the manuscript, and the Plant Research Department of the Danish National Laboratory, Risø and the Danish National Bank for providing space and resources to P.S.C. during manuscript preparation. This research was supported in part by the Office of Science, US Department of Energy, through the Midwestern Regional Center of the National Institute for Global Environmental Change under Cooperative Agreement No. DE-FC03-90ER610100.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  • Amthor JS. 1984. The role of maintenance respiration in plant-growth. Plant, Cell & Environment 7: 561569.
  • Amthor JS. 2000. The McCree–de Wit–Penning de Vries–Thornley respiration paradigms: 30 years later. Annals of Botany 86: 120.
  • Atkin OK, Holly C, Ball MC. 2000. Acclimation of snow gum (Eucalyptus pauciflora) leaf respiration to seasonal and diurnal variations in temperature. the importance of changes in the capacity and temperature sensitivity of respiration. Plant, Cell & Environment 23: 1526.
  • Baldocchi DD. 2003. Assessing the eddy covariance technique for evaluating carbon dioxide exchange rates of ecosystems: past, present and future. Global Change Biology 9: 479492.
  • Baldocchi D, Valentini R, Running S, Oechel W, Dahlman R. 1996. Strategies for measuring and modelling carbon dioxide and water vapour fluxes over terrestrial ecosystems. Global Change Biology 2: 159168.
  • Baldocchi D, Falge E, Gu LH, Olson R, Hollinger D, Running S, Anthoni P, Bernhofer C, Davis K, Evans R, Fuentes J, Goldstein A, Katul G, Law B, Lee XH, Malhi Y, Meyers T, Munger W, Oechel W, Paw U KT, Pilegaard K, Schmid HP, Valentini R, Verma S, Vesala T, Wilson K, Wofsy S. 2001. FLUXNET. A new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor, and energy flux densities. Bulletin of the American Meteorological Society 82: 24152434.
  • Barford CC, Wofsy SC, Goulden ML, Munger JW, Pyle EH, Urbanski SP, Hutyra L, Saleska SR, Fitzjarrald D, Moore K. 2001. Factors controlling long- and short-term sequestration of atmospheric CO2 in a mid-latitude forest. Science 294: 16881691.
  • Bolstad PV, Mitchell K, Vose JM. 1999. Foliar temperature–respiration response functions for broad-leaved tree species in the southern Appalachians. Tree Physiology 19: 871878.
  • Bolstad PV, Reich P, Lee T. 2003. Rapid temperature acclimation of leaf respiration rates in Quercus alba and Quercus rubra. Tree Physiology 23: 969976.
  • Bolstad PV, Davis KJ, Martin J, Cook BD, Wang W. 2004. Component and whole-system respiration fluxes in northern deciduous forests. Tree Physiology 24: 493504.
  • Bovard BD, Curtis PS, Vogel CS, Su H-B, Schmid HP. 2005. Environmental controls on sap flow in a northern hardwood forest. Tree Physiology 25: 3138.
  • Brooks A, Farquhar GD. 1985. Effect of temperature on the CO2/O2 specificity of ribulose-1,5-bisphosphate carboxylase oxygenase and the rate of respiration in the light – estimates from gas-exchange measurements on spinach. Planta 165: 397406.
  • Burton AJ, Pregitzer KS. 2003. Field measurements of root respiration indicate little to no seasonal temperature acclimation for sugar maple and red pine. Tree Physiology 23: 273280.
  • Butnor JR, Johnsen KH. 2004. Calibrating soil respiration measures with a dynamic flux apparatus using artificial soil media of varying porosity. European Journal of Soil Science 55: 639647.
  • Canadell JG, Mooney HA, Baldocchi DD, Berry JA, Ehleringer JR, Field CB, Gower ST, Hollinger DY, Hunt JE, Jackson RB, Running SW, Shaver GR, Steffen W, Trumbore SE, Valentini R, Bond BY. 2000. Carbon metabolism of the terrestrial biosphere: a multitechnique approach for improved understanding. Ecosystems 3: 115130.
  • Cannell MGR, Thornley JHM. 2000. Modelling the components of plant respiration: some guiding principles. Annals of Botany 85: 4554.
  • Cernusak LA, Marshall JD. 2000. Photosynthetic refixation in branches of Western White Pine. Functional Ecology 14: 300311.
  • Cooper AW. 1981. Above-ground biomass accumulation and net primary production during the first 70 years of succession in Populus grandidentata stands on poor sites in northern lower Michigan. In: WestDC, ShugartHH, BotkinDB, eds. Forest Succession: Concepts and Application. New York, USA: Springer-Verlag.
  • Crow TR, Erdmann GG. 1983. Weight and volume equations and tables for red maple in the Lake States. US Forest Service Research Paper NC-242: 14. St Paul, MN, USA: North Central Forest Experiment Station.
  • Curtis PS, Hanson PJ, Bolstad P, Barford C, Randolph JC, Schmid HP, Wilson KB. 2002. Biometric and eddy-covariance based estimates of annual carbon storage in five eastern North American deciduous forests. Agricultural and Forest Meteorology 113: 319.
  • Davidson EA, Belk E, Boone RD. 1998. Soil water content and temperature as independent or confounded factors controlling soil respiration in a temperate mixed hardwood forest. Global Change Biology 4: 217227.
  • Davidson EA, Savage K, Bolstad P, Clark DA, Curtis PS, Ellsworth DS, Hanson PJ, Law BE, Luo Y, Pregitzer KS, Randolph JC, Zak D. 2002a. Belowground carbon allocation in forests estimated from litterfall and IRGA-based soil respiration measurements. Agricultural and Forest Meteorology 113: 3951.
  • Davidson EA, Savage K, Verchot LV, Navarro R. 2002b. Minimizing artifacts and biases in chamber-based measurements of soil respiration. Agricultural and Forest Meteorology 113: 2137.
  • Dawson TE, Mambelli S, Plamboeck AH, Templer PH, Tu KP. 2002. Stable isotopes in plant ecology. Annual Review of Ecology and Systematics 33: 507559.
  • Desrochers A, Landhausser SM, Lieffers VJ. 2002. Coarse and fine root respiration in aspen (Populus tremuloides). Tree Physiology 22: 725732.
  • Edwards NT, Hanson PJ. 1996. Stem respiration in a closed-canopy upland oak forest. Tree Physiology 16: 433439.
  • Ehman JL, Schmid HP, Grimmond CSB, Randolph JC, Hanson PJ, Wayson CA, Cropley FD. 2002. An initial intercomparison of micrometeorological and ecological inventory estimates of carbon exchange in a mid-latitude deciduous forest. Global Change Biology 8: 575589.
  • Falge E, Baldocchi D, Olson R, Anthoni P, Aubinet M, Bernhofer C, Burba G, Ceulemans R, Clement R, Dolman H, Granier A, Gross P, Grunwald T, Hollinger D, Jensen NO, Katul G, Keronen P, Kowalski A, Lai CT, Law BE, Meyers T, Moncrieff H, Moors E, Munger JW, Pilegaard K, Rannik U, Rebmann C, Suyker A, Tenhunen J, Tu K, Verma S, Vesala T, Wilson K, Wofsy S. 2001. Gap filling strategies for defensible annual sums of net ecosystem exchange. Agricultural and Forest Meteorology 107: 4369.
  • Falge E, Baldocchi D, Tenhunen J, Aubinet M, Bakwin P, Berbigier P, Bernhofer C, Burba G, Clement R, Davis KJ, Elbers JA, Goldstein AH, Grelle A, Granier A, Guomundsson J, Hollinger D, Kowalski AS, Katul G, Law BE, Malhi Y, Meyers T, Monson RK, Munger JW, Oechel W, Paw U KT, Pilegaard K, Rannik U, Rebmann C, Suyker A, Valentini R, Wilson K, Wofsy S. 2002. Seasonality of ecosystem respiration and gross primary production as derived from FLUXNET measurements. Agricultural and Forest Meteorology 113: 5374.
  • Fan S, Gloor M, Mahlman J, Pacala S, Sarmiento J, Takahashi T, Tans P. 1998. A large terrestrial carbon sink in North America implied by atmospheric and oceanic carbon dioxide data and models. Science 282: 442446.
  • Finnigan JJ. 1999. A comment on the paper by Lee (1998) on micrometeorological observations of surface-air exchange over tall vegetation. Agricultural and Forest Meteorology 97: 5564.
  • Finnigan JJ, Clements R, Malhi Y, Leuning R, Cleugh HA. 2003. A re-evaluation of long-term flux measurement techniques. Part I: averaging and coordinate rotation. Boundary-Layer Meteorology 107: 148.
  • Geider RJ, Delucia EH, Falkowski PG, Finzi AC, Grime JP, Grace J, Kana TM, La Roche J, Long SP, Osborne BA, Platt T, Prentice IC, Raven JA, Schlesinger WH, Smetacek V, Stuart V, Sathyendranath S, Thomas RB, Vogelmann TC, Williams P, Woodward FI. 2001. Primary productivity of planet earth. biological determinants and physical constraints in terrestrial and aquatic habitats. Global Change Biology 7: 849882.
  • Gifford RM. 2003. Plant respiration in productivity models: conceptualisation, representation and issues for global terrestrial carbon-cycle research. Functional Plant Biology 30: 171186.
  • Gough CM, Vogel CS, Schmid HP, Su H-B, Curtis PS. 2005. Multi-year convergence of biometric and meteorological estimates of forest carbon storage. Agricultural and Forest Meteorology. (In press.)
  • Goulden ML, Munger JW, Fan SM, Daube BC, Wofsy SC. 1996. Measurements of carbon sequestration by long-term eddy covariance: methods and a critical evaluation of accuracy. Global Change Biology 2: 169182.
  • Hanson PJ, Edwards NT, Garten CT Jr, Andrews JA. 2000. Separating root and soil microbial contributions to soil respiration: a review of methods and observations. Biogeochemistry 48: 115146.
  • Harmon ME, Bible K, Ryan MG, Shaw DC, Chen H, Klopatek J, Li X. 2004. Production, respiration, and overall carbon balance in an old-growth Pseudotsuga–Tsuga forest ecosystem. Ecosystems 7: 498512.
  • Harris WF, Sollins P, Edwards NT, Dinger BE, Shugart HH. 1975. Analysis of carbon flow and productivity in a temperate deciduous forest ecosystem. In: Productivity of World Ecosystems. Washington, DC: Special Committee for the International Biological Program, National Academy of Sciences.
  • Hibbard KA, Law BE, Reichstein M, Sulzman J, Aubinet M, Baldocchi D, Bernhofer C, Bolstad P, Bosc A, Campbell JL, Cheng Y, Yuste JC, Curtis P, Davidson EA, Epron D, Granier A, Grünwald T, Hollinger D, Janssens IA, Longdoz B, Loustau D, Martin J, Monson R, Oechel W, Pippen J, Ryel R, Savage K, Schlesinger B, Scott-Denton L, Subke J-A, Tang J, Tenhunen J, Turcu V, Vogel C. 2005. An analysis of soil respiration across northern hemisphere temperate ecosystems. Biogeochemistry. (In press.)
  • Hocker HW Jr, Early DJ. 1983. Biomass and leaf area equations for northern forest species. New Hampshire Agricultural Experiment Station University of New Hampshire Research Report 102: 27. Durham, NH, USA: New Hampshire Agricultural Experiment Station.
  • Janssens IA, Lankreijer H, Matteucci G, Kowalski AS, Buchmann N, Epron D, Pilegaard K, Kutsch W, Longdoz B, Grunwald T, Montagnani L, Dore S, Rebmann C, Moors EJ, Grelle A, Rannik U, Morgenstern K, Oltchev S, Clement R, Gudmundsson J, Minerbi S, Berbigier P, Ibrom A, Moncrieff J, Aubinet M, Bernhofer C, Jensen NO, Vesala T, Granier A, Schulze ED, Lindroth A, Dolman AJ, Jarvis PG, Ceulemans R, Valentini R. 2001. Productivity overshadows temperature in determining soil and ecosystem respiration across European forests. Global Change Biology 7: 269278.
  • Ker MF. 1980. Tree biomass equations for seven species in southwestern New Brunswick. Canadian Forest Service Maritime Forest Research Center Information Report M-X-114. Fredericton, NB, Canada: Environment Canada, Maritime Forest Research Center.
  • Kilburn PD. 1960. Effect of settlement on the vegetation of the University of Michigan Biological Station. Papers of the Michigan Academy of Science, Arts, and Letters 45: 7781.
  • Kira T. 1975. Primary productivity of forests. In: CooperJP, ed. Photosynthesis and Productivity in Different Environments. Cambridge, UK: Cambridge University Press.
  • Koerper G. 1977. The aboveground biomass and annual net production of bigtooth aspen (Populus grandidentata Michx.) on three soil types in northern lower Michigan. Ann Arbor, MI, USA: University of Michigan, MSc thesis.
  • Lavigne MB, Ryan MG, Anderson DE, Baldocchi DD, Crill PM, Fitzjarrald DR, Goulden ML, Gower ST, Massheder JM, McCaughey JH, Rayment M, Striegl RG. 1997. Comparing nocturnal eddy covariance measurements to estimates of ecosystem respiration made by scaling chamber measurements at six coniferous boreal sites. Journal of Geophysical Research – Atmospheres 102: 2897728985.
  • Law BE, Baldocchi DD, Anthoni PM. 1999a. Below-canopy and soil CO2 fluxes in a ponderosa pine forest. Agricultural and Forest Meteorology 94: 171188.
  • Law BE, Ryan MG, Anthoni PM. 1999b. Seasonal and annual respiration of a ponderosa pine ecosystem. Global Change Biology 5: 169182.
  • Lee X. 1998. On micrometeorological observations of surface-air exchange over tall vegetation. Agricultural and Forest Meteorology 91: 3949.
  • Lee XH, Fuentes JD, Staebler RM, Neumann HH. 1999. Long-term observation of the atmospheric exchange of CO2 with a temperate deciduous forest in southern Ontario, Canada. Journal of Geophysical Research-Atmospheres 104: 1597515984.
  • Lundgren AL, Dolid WA. 1970. Biological growth functions describe published site index curves for Lake States timber species. US Forest Service Research Paper. NC-36: 12. St Paul, MN, USA: North Central Forest Experiment Station.
  • Mäkelä A, Landsberg J, Ek AR, Burk TE, Ter-Mikaelian M, Ågren GI, Oliver CD, Puttonen P. 2000. Process-based models for forest ecosystem management: current state of the art and challenges for practical implementation. Tree Physiology 20: 289298.
  • Mäkelä A, Valentine HT. 2001. The ratio of NPP to GPP: evidence of change over the course of stand development. Tree Physiology 21: 10151030.
  • Malhi Y, Meir P, Brown S. 2002. Forests, carbon and global climate. Philosophical Transactions of the Royal Society of London Series A 360: 15671591.
  • McGuire MA, Teskey RO. 2004. Estimating stem respiration in trees by a mass balance approach that accounts for internal and external fluxes of CO2. Tree Physiology 24: 571578.
  • Melillo JM, Steudler PA, Aber JD, Newkirk K, Lux H, Bowles FP, Catricala C, Magill A, Ahrens T, Morrisseau S. 2002. Soil warming and carbon-cycle feedbacks to the climate system. Science 298: 21732176.
  • Nelson CJ. 1994. Apparent respiration and plant productivity. In: BooteKJ, BennettJM, SinclairTR, PaulsonGM, eds. Physiology and Determination of Crop Yield. Madison, WI, USA: American Society of Agronomy, Crop Science Society of America, Soil Society of America, 251258.
  • Oren R, Phillips N, Katul G, Ewers BE, Pataki DE. 1998. Scaling xylem sap flux and soil water balance and calculating variance: a method for partitioning water flux in forests. Annales des Sciences Forestieres 55: 191216.
  • Paw U KT, Falk M, Suchanek TH, Ustin SL, Chen JQ, Park YS, Winner WE, Thomas SC, Hsiao TC, Shaw RH, King TS, Pyles RD, Schroeder M, Matista AA. 2004. Carbon dioxide exchange between an old-growth forest and the atmosphere. Ecosystems 7: 513524.
  • Pendall E, Bridgham S, Hanson PJ, Hungate B, Kicklighter DW, Johnson DW, Law BE, Luo YQ, Megonigal JP, Olsrud M, Ryan MG, Wan SQ. 2004. Below-ground process responses to elevated CO2 and temperature. a discussion of observations, measurement methods, and models. New Phytologist 162: 311322.
  • Perala DA, Alban DH. 1994. Allometric biomass estimators for aspen-dominated ecosystems in the Upper Great Lakes. US Forest Service Research Paper. NC-314: 38. St Paul, MN, USA: North Central Forest Experiment Station.
  • Pinelli P, Loreto F. 2003. 12CO2 emission from different metabolic pathways measured in illuminated and darkened C3 and C4 leaves at low, atmospheric and elevated CO2 concentration. Journal of Experimental Botany 54: 17611769.
  • Pons TL, Welschen RAM. 2002. Overestimation of respiration rates in commercially available clamp-on leaf chambers. Complications with measurement of net photosynthesis. Plant, Cell & Environment 25: 13671372.
  • Pruyn ML, Gartner BL, Harmon ME. 2002. Within-stem variation of respiration in Pseudotsuga menziesii (Douglas-fir) trees. New Phytologist 154: 359372.
  • Raich JW, Nadelhoffer KJ. 1989. Belowground carbon allocation in forest ecosystems – global trends. Ecology 70: 61354.
  • Raich JW, Schlesinger WH. 1992. The global carbon-dioxide flux in soil respiration and its relationship to vegetation and climate. Tellus Series B 44: 8199.
  • Russell CA, Voroney RP. 1998. Carbon dioxide efflux from the floor of a boreal aspen forest. I. Relationship to environmental variables and estimates of C respired. Canadian Journal of Soil Science 78: 301310.
  • Ryan MG, Hubbard RM, Pongracic S, Raison RJ, McMurtrie RE. 1996. Foliage, fine-root, woody-tissue and stand respiration in Pinus radiata in relation to nitrogen status. Tree Physiology 16: 333343.
  • Ryan MG, Lavigne MB, Gower ST. 1997. Annual carbon cost of autotrophic respiration in boreal forest ecosystems in relation to species and climate. Journal of Geophysical Research-Atmospheres 102: 2887128883.
  • Sanderman J, Amundson RG, Baldocchi DD. 2003. Application of eddy covariance measurements to the temperature dependence of soil organic matter mean residence time. Global Biogeochemical Cycles 17: 1061.
  • Schaetzl RJ. 1994. Changes in O-horizon mass, thickness and carbon content following fire in northern hardwood forests. Vegetatio 115: 4150.
  • Schmid HP. 1997. Experimental design for flux measurements: matching scales of observations and fluxes. Agricultural and Forest Meteorology 87: 179200.
  • Schmid HP, Su HB, Vogel CS, Curtis PS. 2003. Ecosystem-atmosphere exchange of carbon dioxide over a mixed hardwood forest in northern lower Michigan. Journal of Geophysical Research – Atmospheres 108: 4417.
  • Schmitt MDC, Grigal DF. 1981. Generalized biomass estimation equations for Betula papyrifera Marsh. Canadian Journal of Forest Research 11: 837840.
  • Stockfors J. 2000. Temperature variations and distribution of living cells within tree stems. implications for stem respiration modeling and scale-up. Tree Physiology 20: 10571062.
  • Strain BR, Johnson PL. 1963. Corticular photosynthesis and growth in Populus tremuloides. Ecology 44: 581584.
  • Ter-Michaelian MT, Korzukhin MD. 1997. Biomass equations for sixty-five North American tree species. Forest Ecology and Management 97: 124.
  • Turnbull MH, Whitehead D, Tissue DT, Schuster WSF, Brown KJ, Griffin KL. 2001. Responses of leaf respiration to temperature and leaf characteristics in three deciduous tree species vary with site water availability. Tree Physiology 21: 571578.
  • USDA Forest Service. 2001. Forest Inventory Analysis (FIA). Washington, DC, USA: USDA Forest Service.
  • Vose JM, Bolstad PV. 1999. Challenges to modelling NPP in diverse eastern deciduous forests: species-level comparisons of foliar respiration responses to temperature and nitrogen. Ecological Modelling 122: 165174.
  • Wang KY, Kellomaki S, Zha T, Peltola H. 2004. Seasonal variation in energy and water fluxes in a pine forest. an analysis based on eddy covariance and an integrated model. Ecological Modelling 179: 259279.
  • Waring RH, Landsberg JJ, Williams M. 1998. Net primary production of forests: a constant fraction of gross primary production? Tree Physiology 18: 129134.
  • White MA, Running SW, Thornton PE. 1999. The impact of growing-season length variability on carbon assimilation and evapotranspiration over 88 years in the eastern US deciduous forest. International Journal of Biometeorology 42: 139145.
  • Wiant HV, Sheetz CE, Colaninno A, DeMoss JC, Castaneda F. 1977. Tables and procedures for estimating weights of some Appalacian hardwoods. West Virginia Agricultural Experiment Station Bulletin 659: 36.
  • Xiao JF, Moody A. 2004. Photosynthetic activity of US biomes: responses to the spatial variability and seasonality of precipitation and temperature. Global Change Biology 10: 437451.
  • Xu M, DeBiase TA, Qi Y. 2000. A simple technique to measure stem respiration using a horizontally oriented soil chamber. Canadian Journal of Forest Research 30: 15551560.
  • Young HE, Ribe JH, Wainwright K. 1980. Weight Tables for Tree and Shrub Species in Maine. Miscellaneous Report no. 230. Orono, ME, USA: University of Maine.