Soil Biology & Biochemistry 32 (2000) 16251635

www.elsevier.com/locate/soilbio

Biotic and abiotic factors controlling soil respiration rates in Picea abies stands
Nina Buchmann*
Max-Planck-Institute for Biogeochemistry, P.O. Box 100164, D-07701 Jena, Germany Accepted 17 May 2000

Abstract The response of soil respiration to varying environmental factors was studied in four Picea abies stands (47-, 87-, 111- and 146-year old) during the 1998 growing season. While within-site variations of soil CO2 eux (up to 1.6 mmol CO2 m2 s1) were larger than their diurnal variability (<0.25 mmol CO2 m2 s1), spatial variations within a site were smaller than seasonal changes in soil respiration rates (up to 4.4 mmol CO2 m2 s1). Highest within-site variability of soil eux was generally found during the summer months when maximum ux rates of 46 mmol CO2 m2 s1 were reached (coecient of variation 40%). Soil temperatures (in the Of and Oh layers, and Ah horizon) showed a pronounced seasonal course, in contrast to soil moisture. An exponential equation best described the relationships between soil temperature in the Of layer and soil CO2 eux r 2 between 0.75 and 0.81). However, an Arrhenius type equation always resulted in lower r 2 values (0.520.71). The Q10 values ranged between 2.39 (146-year old stand) and 3.22 (87-year old stand), averaging 2.72 for the P. abies stands within the watershed. The removal of litter and organic layers generally aected soil CO2 eux negatively. In three of the four P. abies stands (47-, 87-, 146-year old stands), soil respiration rates were reduced by 1020% after removal of the L and Of layer, and by 3040% after removal of the L and most of the Of and Oh layers. Thus, mineral soil respiration seemed to contribute a major fraction to the total soil CO2 ux (>60%). Trenching shallow ne roots during collar insertion and mechanical inhibition of root in-growth during the following months allowed ne root respiration to be separated from microbial respiration only in times of highest root growth. Microbial respiration seemed to dominate the respiratory CO2 loss from the forest oor (>70%). The comparison of the annual soil CO2 eux in the 47-year old P. abies stand (about 710 g C m2 yr1) with annual litterfall and root net primary productivity estimates supported this conclusion. 7 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Microbial respiration; Organic layer; Root respiration; Seasonality; Soil; Respiration; Q10 value

1. Introduction Soil respiration is a major CO2 ux within terrestrial ecosystems as well as between the biosphere and the atmosphere (Schlesinger, 1977; Raich and Schlesinger, 1992; IPCC, 1996). Soil CO2 uxes originate from autotrophic root respiration and heterotrophic microbial respiration in the rhizosphere and the bulk soil. Generally, between 50% and 80% of the nocturnal biosphereatmosphere CO2 exchange, measured with
* Tel.: +49-3641-64-3721; fax: +49-3641-64-3710. E-mail address: [email protected] (N. Buchmann).

eddy covariance techniques, are due to soil CO2 eux (Lavigne et al., 1997). However, during the night, these measurements of net ecosystem CO2 exchange are associated with the largest errors (Moncrie et al., 1997). Thus, detailed information on soil CO2 uxes and on factors that control these uxes are needed to constrain the ecosystem carbon budget and to decide whether or not terrestrial ecosystems are carbon sinks or sources (Fan et al., 1995; Goulden et al., 1996; Lavigne et al., 1997; Lindroth et al., 1998). Furthermore, climate and land-use change have the potential to enhance or reduce soil CO2 uxes. Changes in temperature and precipitation, and also shifts from forest to agricultural land-use or changing management prac-

0038-0717/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 0 3 8 - 0 7 1 7 ( 0 0 ) 0 0 0 7 7 - 8

1626

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635

tices will aect soil respiration uxes and therefore, the carbon budget of terrestrial ecosystems (Raich and Potter, 1995; IPCC, 1996). Typically, soil CO2 ux rates show large spatial and temporal variations, both within and among sites, which are only partly due to methodological dierences (Raich and Nadelhoer, 1989; Hanson et al., 1993; Norman et al., 1997). Since soil respiration is a combined ux from roots and microorganisms from dierent soil depths (organic surface layers and mineral horizons), several factors and their interactions aect soil respiration rates. Soil temperature and soil moisture are among the most important factors controlling the CO2 ux (Raich and Schlesinger, 1992; Raich and Potter, 1995; Davidson et al., 1998). Root nitrogen concentrations, soil texture, substrate quantity and quality have also been shown to have an eect (Grant and Rochette, 1994; Randerson et al., 1996; Boone et al., 1998; Pregitzer et al., 1998). Great debate still exists on how to model the impact of these factors on soil respiration (e.g., Lloyd and Taylor, 1994; Thierron and Laudelout, 1996), and how the large variability inuences aggregation of soil CO2 eux estimates for a watershed or a landscape (Kicklighter et al., 1994). Thus, although the variability of soil CO2 uxes and some of the underlying causes are well known, they still bear uncertainties that need to be resolved. Since autotrophic and heterotrophic respiration will react dierently to changes in environmental conditions, it is crucial to get more insight into both components of soil respiration (Kirschbaum, 1995; Boone et al., 1998). However, the separation of root (including rhizosphere microorganisms) and microbial respiration under eld conditions is still dicult. So far, many dierent approaches have been used in situ, ranging from severe disturbances of a site to very specic requirements, e.g., changes in the carbon isotopic signature of the two respiratory components (either due to a change of the photosynthetic pathway of the vegetation or manipulations with isotope tracers). Autotrophic respiration of intact roots has been measured with root cuvettes in the eld (Gansert, 1994) or with excised roots in the laboratory (Burton et al., 1998). Trenching (Fisher and Gosz, 1986; Bowden et al., 1993; Hart and Sollins, 1998; Boone et al., 1998), labeling with 14C, 13C or 18O (Horwath et al., 1994; Swinnen et al., 1994; Hogberg and Ekblad, 1996; Lin et al., 1999), inhibiting one respiratory component with specic inhibitors or herbicides (Helal and Sauerbeck, 1991; Nakane et al., 1996), and enhancing one component over the other (Bowden et al., 1993) have been used to separate root from microbial respiration. However, the ratio between the two respiration components is generally quite site-specic and varies between 1:9 and 9:1 (Hanson et al., 2000).

Natural disturbances (e.g., re, windthrow) or anthropogenic land-use changes (e.g., logging, agricultural cultivation, aorestation) often alter the soil prole, thereby changing not only carbon stocks, but maybe also carbon uxes (Schimel et al., 1997). The magnitude of change in soil CO2 eux is dependent on whether or not litter and organic layers are removed, roots are disturbed or mineral soil horizons are exposed or mixed. However, only very limited information is available about these aspects in forest ecosystems (Edwards and Sollins, 1973; Bowden et al., 1993; Mallik and Hu, 1997; Nakane et al., 1997; Boone et al., 1998; Thuille et al., 2000). Since forest management always involves some kind of site disturbance, this adds further uncertainty to modeled soil CO2 ux estimates of managed forests. In this study, spatial versus temporal variability of soil respiration was studied in four Picea abies stands during the 1998 growing season. Spatial variability was investigated at two dierent scales, within a site and within a watershed, using stands growing close to each other (less than 500 m apart). Temporal variability was addressed on a diurnal and a seasonal basis. Furthermore, the eects of ne root exclusion and organic layer removal on soil CO2 ux rates was tested in Norway spruce forests.

2. Materials and methods 2.1. Sites Four Norway spruce stands (Picea abies (L.) Karst.) were chosen within the Lehstenbach watershed, at about 770 m elevation in the Fichtelgebirge, Northeast Bavaria, Germany (50808' N, 11852' E). The four stands were located close to each other (<500 m apart) and ranged in age from 47 to 146 years (Table 1). The higher LAI of the youngest stand was associated with a higher stand density compared to the oldest spruce stand (1018 and 363 trees ha1, respectively). The managed stands had almost no woody debris and only a sparse understory with the grasses Deschampsia exuosa and Calamagrostis villosa and the ericaceous dwarf shrubs Vaccinium myrtillus and V. vitis-idea. Forest oor characteristics were quite similar, with litter layers (L) of about 1.2 cm, and with organic Of and Oh layers (Of = <70% ne organic matter, plant material still recognizable; Oh = >70% ne organic matter) of about 5.3 cm and 4.9 cm, respectively. The humus types were classied as morlike moder to mor (``Rohhumus''; Mund et al., in preparation), with a high fraction of grass roots in the forest oor of the 111-year old stand. The soil type was classied as cambic podzol for all four stands, typically

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635

1627

with sandyloamy soil texture and low pH values (Dieenbach et al., 1997). 2.2. Measurements of soil CO2 eux and organic layer parameters Soil respiration rates were measured monthly during the growing season of 1998 (AprilOctober) using a soil respiration chamber (LI-6400-09; LiCor, Lincoln, Nebraska, USA) connected to a portable photosynthesis system (LI-6400). Four to ve PVC collars (10 cm long, 10 cm inside diameter) were inserted into the soil of each stand. Insertion depth was ca. 58 cm into the forest oor. Shorter collars (35 cm long) were unstable and tended to disturb the forest oor. Although the soil disturbance during collar insertion generally results in initially high uxes just after collar installation, uxes stabilize after 1030 min (Norman et al., 1997). In accordance with most other studies, collars were installed at least 24 h prior to measurement. The soil respiration chamber was set on top of these collars, allowing an undisturbed measurement of soil CO2 ux rates. The protocol recommended by LiCor (LI6400-09 manual) was changed to ve observations of 10 mmol mol1 change per measurement. Litter and organic layer thickness (L, Of and Oh layers) as well as soil temperatures in the Of and Oh layers and the Ah horizon (at 5, 10 and 15 cm depths, respectively) were measured next to each PVC tube. Gravimetric soil water contents of the Of and Oh layers were determined with three replicates per measurement. The wet soil samples were weighed, then dried until weight constancy, and weighed again. Soil moisture is expressed as per cent dry weight. 2.3. The eect of ne root exclusion and organic layer removal on soil CO2 eux In each of the four spruce stands, some PVC collars established earlier in the growing season remained in the soil and soil CO2 eux was measured during the following months. Thus, in October for example, soil collars were available that had been established in

April through October n 15 per month since establishment). This design was used to test whether trenching shallow ne roots (growing in the organic layers) during collar insertion and mechanical inhibition of root in-growth had an eect on current soil respiration rates. At the end of the experiment, collars were taken out and checked for root in-growth into the collar from below. However, no root in-growth was observed. In July, August, and October 1998, nine additional collars were established in each stand. The litter and Of layers were carefully removed before the insertion of three collars. For three other collars, the litter and most of both the Of and Oh layers were removed from the mineral soil, and three undisturbed soil proles served as controls. Rare coarse roots (>1 mm diameter) within the organic layers were cut with a knife without disturbing the underlying mineral soil horizons. Soil CO2 eux was determined 24 h later. This design allowed the organic layer respiration to be separated from mineral soil respiration. 2.4. Statistics Analyses of variance with a posteriori tests (Tukey's honestly signicant dierence (HSD) test, a 0X05 were used to separate the means. Data were transformed when variances were not homogenous. Nonlinear regression analyses were used to test the eect of collar establishment. Soil CO2 eux and temperature data were pooled when the interaction term was not signicant at the a 0X05 level. Correlation and nonlinear regression analyses were used to test the relationship between soil temperature, soil moisture and soil respiration (Eq. (1)). The Q10 values were calculated according to Eq. (2). Residual analyses were performed to test the exponential regression (Eq. (1)) versus the Lloyd and Taylor (1994) Arrhenius type equation (Eq. (3)). y b0 eb1 T Q10 e10b1 1 2

Table 1 Stand characteristics of four P. abies stands within the Lehstenbach watersheda Site Weidenbrunnen Weidenbrunnen3 Weidenbrunnen2 Coulissenhieb
a b

Age (yr) 47 87 111 146

LAIb 10.4 NDc ND 6.6

L (cm) 1.320.3 1.520.1 1.020.1 0.920.1

Thickness of Of layer (cm) 5.120.4 4.820.7 5.220.6 5.920.2

Thickness of Oh layer (cm) 3.620.8b 4.220.7b 7.220.6a 4.420.2b

Dierent letters within a column denote signicantly dierent thickness data (Tukey HSD test, a 0X05). Leaf area index data provided by Martina Mund, Max-Planck Institute for Biogeochemistry, Jena, Germany. c ND not determined.

1628

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635

where T is the soil temperature (8C). !  1 1 y R10 exp 308X56 56X02 T 227X13

where R10 is the soil respiration rate at 108C, and T is the absolute soil temperature (K). See Lloyd and Taylor (1994) for the derivation of the equation. 3. Results 3.1. Spatial versus temporal variability of soil respiration rates Within-site spatial variations among soil collars were larger than the diurnal variability of soil respiration rates measured with the same collars during a day (Fig. 1). While within-site dierences of soil CO2 eux among replicated collars were up to 1.6 mmol CO2 m2 s1 (about 40% of maximum values, between 60% and 70% of minimum values), diurnal variations were less than 0.25 mmol CO2 m2 s1 (<10% of average

rates per collar). During the 10 h period, soil CO2 eux remained almost constant for all four collars, reecting the time course of soil temperatures in the Oh layer (10 cm depth) more closely than those in the Of layer (5 cm depth) (average change in temperature of 0.5 vs. 1.48C). In contrast, seasonal variations of soil respiration rates were greater than within-site spatial variations for the four Picea abies stands during the 1998 growing season (Table 2). Whereas seasonal uctuations of soil CO2 eux rates were as high as 4.4 mmol CO2 m2 s1 (111-year old stand), within-site spatial variability was smaller, on average about 25% of the respective rates (coecient of variation, CV; Fig. 2). Highest within-site variability of soil respiration rates (about 40%) was generally found during the summer months (with one exception) although soil temperatures varied by less than 5% within each given site (not shown). All spruce stands studied in the Lehstenbach watershed showed the same seasonal trend in their soil CO2 eux P 0X629 for the interaction term `site X month'). Soil respiration rates increased during spring and summer, and reached maximum values of 46 mmol CO2 m2 s1 in July and August P ` 0X001 for `month' as main factor; Fig. 3). During fall (SeptemberOctober), soil respiration rates declined again, reaching values close to those in spring (April). Although soil CO2 eux did not dier signicantly among the four P. abies stands within the watershed throughout the growing season P 0X196 for `site' as main factor), soil respiration rates in the 111-year old spruce stand tended to show higher rates than those in

Fig. 1. Diurnal course of soil respiration rates measured with the same four collars (C1C4) as well as soil temperatures in the organic layers at 5 and 10 cm depth. Means and standard errors are given n 5). Collars were installed 24 h prior to measurements and were not removed during the 10 h period.

Fig. 2. Seasonal course of the coecient of within-site variation for soil CO2 eux rates in four P. abies stands. Collars were inserted into the soil 24 h before the measurements.

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635 Table 2 Soil respiration rates (mmol CO2 m2 s1) in four P. abies stands in the Fichtelgebirge during the 1998 growing seasona Month April May June July August Septemberb October
a

1629

47-year old 1.2320.09d 2.9820.14b 2.6320.30bc 3.9720.39a 4.1020.42a 2.6020.17bc 1.6020.08cd

87-year old 1.1720.12d 2.5720.67bc 3.6120.30ab 4.7420.76a 4.3520.41ab 2.6320.29bc 1.6320.09c

111-year old 1.5520.17d 3.1520.18bc 3.4120.31bc 5.0721.03ab 5.9320.87a 3.2620.30bc 1.9520.23c

146-year old 1.0820.12d 2.6220.34bc 2.3020.14cd 4.6320.78a 3.9420.27ab 2.3320.25cd 1.3720.18cd

All stands 1.2420.06c 2.8220.17b 2.9920.19b 4.5120.32a 4.5820.33a 2.7120.21b 1.6420.09c

Means and standard errors n 5 are given. Dierent letters within a spruce stand denote signicantly dierent soil respiration rates (Tukey HSD test, a 0X05). b Collars had been installed in the soil one week prior to measurements.

the other stands. Soil temperatures of the Of and Oh layers and the Ah horizon (at 5, 10 and 15 cm depth) showed the same pronounced seasonal course as the soil CO2 eux r 0X93, r 0X92, and r 0X82, respectively). In contrast, soil moisture of the Of layer (Fig. 3), and also of the Oh layer (not shown) showed a dierent pattern that did not explain a signicant

fraction of the within-watershed spatial variance in soil respiration rates r 0X41 and r 0X36, respectively). 3.2. Relationships between soil temperature and soil respiration An exponential equation (Eq. (1)) best described the relationships between soil temperature at various depths (in the organic layers and the Ah horizon) and soil CO2 eux r 2 between 0.60 and 0.93 compared to r 2 between 0.53 and 0.81 for a linear t; Fig. 4). Soil temperatures in the four stands changed dierently during the growing season P ` 0X001 for the inter-

Fig. 3. Soil respiration rates, soil temperatures in the organic layers at 5 and 10 cm depth as well as soil moisture of the Of layer in four P. abies stands during the growing season of 1998. Means (2SE) are given n 3 5). Collars were installed 24 h prior to measurements (except for September, Julian Day 244, when collars were installed one week before the measurements).

Fig. 4. Relationship between soil respiration rates and soil temperature in the Of layer at 5 cm depth for all four P. abies stands during the growing season of 1998. The exponential regression equations are given in Table 3. All collars were installed 24 h prior to measurement.

1630

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635

Table 3 Relationships between soil respiration rates (mmol CO2 m2 s1) and soil temperatures (8C) measured in the Of and Oh layers at 5 and 10 cm depth and in the Ah horizon at 15 cm depth for four P. abies stands in the Fichtelgebirge during the 1998 growing season Site 47-year old 87-year old 111-year old 146-year old y b0 eb1 T y 0X98 e0X088XT5 y 1X24 e0X085XT10 y 0X99 e0X109XT15 y 0X80 e0X117XT5 y 0X73 e0X141XT10 y 0X55 e0X166XT15 y 1X03 e0X105XT5 y 1X01 e0X119XT10 y 1X05 e0X119XT15 y 0X87 e0X087XT5 y 0X85 e0X103XT10 y 0X73 e0X117XT15 SEb0 0.13 0.18 0.18 0.16 0.08 0.08 0.17 0.14 0.23 0.13 0.11 0.13 SEb1 0.011 0.014 0.018 0.016 0.013 0.014 0.014 0.014 0.023 0.012 0.012 0.016 Q10 2.41 2.34 2.34 3.22 4.11 4.11 2.87 3.27 3.27 2.39 2.82 2.82 n 20 23 16 14 24 12 16 21 14 21 30 22 F 70.2 34.7 38.2 52.0 122.8 134.7 54.2 74.2 27.5 55.4 70.59 50.8 r2 0.80 0.62 0.73 0.81 0.85 0.93 0.80 0.80 0.70 0.75 0.71 0.72 P < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001

action term `site month'), probably due to dierences in stand structure (see Table 1). This within-watershed spatial variability of soil temperatures required separate regression analyses for the four P. abies stands within the watershed (Table 3). Comparing n, r 2 and F values as well as residuals, best results were obtained with soil temperature in the Of layer (at 5 cm) as independent variable. All relationships were highly signicant P ` 0X0001 and explained between 75% and 81% of the variance in the soil respiration rates (using the T5 relationships). The Q10 values ranged between 2.39 (146-year old stand) and 3.22 (87-year old stand), averaging 2.72 for the P. abies stands within the watershed (Eq. (2)). Furthermore, these exponential regressions were compared to the Arrhenius type equation (Lloyd and Taylor, 1994) where b0 is standardized to 108C, and the eective activation energy for respiration varies inversely with soil temperature (Eq. (3)). The overall t of the Arrhenius type equation to the four datasets generally resulted in lower r 2 values: 0.71 compared to 0.80 for the 47-year old stand, 0.68 compared to 0.81 for the 87-year old stand, 0.63 compared to 0.80 for the 111-year old stand, and 0.52 compared to 0.75 for the 146-year old stand. Furthermore, residual analyses showed no dierence between the equations. Thus, the (simpler) exponential equation with soil temperature at 5 cm depth (Of layer) was preferred (Eq. (1)). 3.3. Eect of ne root exclusion on soil respiration Trenching shallow ne roots during collar insertion and subsequent mechanical inhibition of root ingrowth aected soil respiration rates in all four stands of the watershed in the same way P 0X75 for the interaction term `site X months since installation'). Therefore, the four datasets were pooled. However, the eect of ne root exclusion could not be observed at all sampling dates during the 1998 growing season

Fig. 5. Eect of ne root exclusion on soil CO2 eux. Soil respiration rates were averaged for all four stands. Standard errors n 12 15 are given. Dierent letters within a monthly dataset denote signicantly dierent soil respiration rates (Turkey HSD test, a 0X05).

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635

1631

(Fig. 5). Soil respiration rates, measured with collars inserted 24 h prior to measurement (i.e., 0 months since installation), diered signicantly from those measured with older collars (i.e., 16 months since installation) only in May and August. No eect was detectable in June, July or October. Soil CO2 eux rates measured with newly installed collars were 20% (May) to 30% (August) higher than those measured with 1-month old collars, indicating the fraction of new root growth during May and August. However, whether roots had been excluded (i.e., collars had been installed) 1 month or longer did not have any eect on soil CO2 eux, supporting the observation that no roots were growing into the collars from below. Thus, although root respiration in the upper organic layers must have been highest in May and August, microbial respiration seemed to dominate the respiratory CO2 loss from the forest oor (>70%). 3.4. Eect of organic layer removal on soil CO2 eux Generally, the removal of the litter and organic layers aected the soil CO2 eux negatively (Fig. 6). In three of the four P. abies stands, soil respiration rates were reduced by 1020% after removal of the L and Of layers, and by 3040% after removal of the L and most of the Of and Oh layers (47-, 87-, 146-year old stands). In two of the spruce stands, this reduction was signicant P 0X042 for the 47-year old stand; P 0X030 for the 146-year old stand). Due to small soil respiration rates in the 87-year old stand and therefore relatively large variations, this treatment eect was not evident in the 87-year old spruce stand P 0X28). Within-site variability of the 111-year old

stand was much larger than for all the other stands, a trend already seen in the monthly measurements in July and August (see Table 2). Both ndings could be due to the higher presence of grass roots in the organic layer of the 111-year old stand the compared to the other stands. Consequently, no treatment eect could be detected P 0X98). Since soil moisture and soil temperatures, which had been determined inside the collars just after the ux measurements, did not dier among the treatments P b 0X6), mineral soil respiration seemed to contribute a major fraction to the total soil CO2 ux (>60%). 4. Discussion 4.1. Within-site variations Within-site spatial variations in soil respiration rates were greater than the diurnal variations, but smaller than the seasonal variability within a site. Since about 80% of the variance in soil CO2 uxes was explained by changes in soil temperature, and because soil temperatures stayed fairly constant during the day, no diurnal uctuations in soil respiration rates were detected in this study. However, as soon as daytime uctuations of soil temperature increased, soil respiration rates responded (Witkamp, 1969; Ben-Asher et al., 1994; Bouma et al., 1997; diurnal changes of 10 208C). In accordance with other studies (e.g., Hanson et al., 1993; Boone et al., 1998; Davidson et al., 1998), a distinct seasonal trend was observed with highest soil CO2 ux rates as well as greatest coecients of variation during the summer months throughout the

Fig. 6. Eect of litter and organic layer removal on soil respiration rates in four adjacent P. abies stands. Dierent letters represent signicantly dierent rates within a site (Tukey HSD test, a 0X05).

1632

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635

watershed in this study. Generally, soil temperature alone was sucient to explain seasonal variations of soil and root respiration (this study; Burton et al., 1998; Gansert, 1994; Hanson et al., 1993). However, under certain conditions, soil CO2 eux could also be constrained by precipitation or soil moisture, e.g., in arid ecosystems by low soil moisture (Conant et al., 1998) or in humid ecosystems by high soil moisture (Buchmann et al., 1997, 1998). Thus, although global models typically use soil temperature as well as soil moisture for large-scale soil CO2 eux estimates (Kicklighter et al., 1994; Raich and Potter, 1995), other factors that are highly variable at a microscale such as litter quality or water availability might be responsible for large within-site variations during summer months (CV of 40%). 4.2. Temperature dependence and Q10 values There are still uncertainties associated with modeling the strong temperature dependence of soil respiration rates. While under some circumstances, linear or sinusoidal regressions give good relationships between soil CO2 eux and soil temperature (Raich and Schlesinger, 1992; Ben-Asher et al., 1994; Fan et al., 1995), most of the studies report exponential regression equations (e.g., Boone et al., 1998; Burton et al., 1998; Davidson et al., 1998; Fan et al., 1995; Raich and Potter, 1995). However, some authors strongly recommend an Arrhenius type equation that at least in one case (where the eective activation energy changed inversely with temperature) resulted in evenly distributed residual variances across the entire temperature range (Lloyd and Taylor, 1994; Thierron and Laudelout, 1996). In this study, all three functions (linear, exponential, and Arrhenius type) were compared. Whereas linear and Arrhenius type functions gave much lower r 2 values than those of the exponential function, the residual variances of exponential and Arrhenius type functions were similarly distributed along the soil temperature range from 28C to 178C for all four stands. Lloyd and Taylor (1994) argued that over a wider temperature range soil microbial populations responsible for soil organic matter decomposition might change and that, therefore, Q10 values, the indicators of temperature sensitivity, might change as well. However, their temperature range was about 408C, more than twice the range found in this study. Thus, in this study, the relatively small temperature range might be responsible for the better t and similar residual variances using the exponential function compared to the Arrhenius function. The temperature sensitivity of soil CO2 uxes within the watershed varied between 2.4 and 3.2, well within the range of 2.03.9 generally given for bulk soil respiration (Baldocchi et al., 1986; Davidson et al., 1998;

Hanson et al., 1993; Kicklighter et al., 1994; Raich and Schlesinger, 1992; Schleser, 1982). However, soil respiration is a combination of autotrophic and heterotrophic respiration, and both can exhibit dierent Q10 values. Boone et al. (1998) reported signicantly higher Q10 values for root respiration (4.6) than for bulk soil respiration (3.5). However, soil organic matter oxidation was shown to be the most responsive to increased temperatures in small `terracosms' (Lin et al., 1999). Nevertheless, under future climate scenarios with increasing temperatures, higher root respiration rates might result in a smaller soil carbon sink than anticipated (Kirschbaum, 1995; Boone et al., 1998). 4.3. Partitioning of soil CO2 eux The potential change in soil carbon uxes due to increased temperatures will strongly depend on the ratio between root (including rhizosphere) and microbial respiration. Whereas a few studies report higher contributions from root than from microbial respiration (Helal and Sauerbeck, 1992; Thierron and Laudelout, 1996), the microbial component seems to dominate the bulk soil CO2 uxes in forest soils (Hanson et al., 2000). In this study, microbial respiration dominated the total soil CO2 ux by >70%. Similar results with higher contributions from microbial than from root respiration were also observed by Kelting et al. (1998) who separated this microbial component even further (20% of total soil respiration was due to microbial respiration in the rhizosphere, 48% were due to microorganisms in the bulk soil). Killing the root system after clear-cutting followed by herbicide application (Nakane et al., 1996) as well as using stable carbon and oxygen isotopes to separate the dierent respiratory uxes (Lin et al., 1999), resulted in microbial respiration rates that were between 50% and 70% of bulk soil respiration. Manipulating litter input, root presence and soil organic matter contents, Bowden et al. (1993) estimated that only about 33% of total soil respiration was due to root respiration. Although all of these dierent experimental approaches have uncertainties, microbial respiration (not associated with the rhizosphere) appears to represent the dominant fraction of total soil CO2 eux in a wide range of terrestrial ecosystems. Thus, trenching shallow ne roots during collar insertion, subsequent mechanical inhibition of root ingrowth and monitoring soil respiration rates in those collars over time, is a useful tool to separate root and microbial respiration in the eld. One obvious drawback of this method is the detection of respiration originating from shallow coarse roots since those can not be cut with the collars. However, larger trenching plots of 3 m 3 m (Bowden et al., 1993) resulted in a very similar distribution of root versus microbial res-

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635

1633

piration (33% vs. 67%), supporting the ndings with the small collars (<30% vs. >70%). Changes in soil respiration rates after long-term exclusion of roots can be the result of several mechanisms. Soil CO2 eux might be reduced not only because the root component is lacking, but also because trenching prevents belowground carbon input for microbial decomposition. Furthermore, soil CO2 ux rates might change due to a higher soil moisture regime as a result of restricted plant water uptake (Hart and Sollins, 1998). In this study, however, soil moisture remained the same, and during most of the growing season, the duration of root exclusion did not have an eect on soil CO2 ux rates, indicating that root respiration in the organic layer was only a minor component of bulk soil respiration. The removal of litter and organic layers in this study generally resulted in a reduction in the soil respiration rates that was less than 40% (Fig. 6). Thus, the major fraction of the respired CO2 originated not from organic surface layers, but from mineral horizons, probably the Ah horizon (>60%). Similar results were reported from a temperate deciduous forest (Edwards and Sollins, 1973) (>60% of total soil respiration originated from deeper mineral horizons) and boreal Canadian and Japanese forests (Mallik and Hu, 1997; Nakane et al., 1997) (between 58% and 74% of the total soil respiration ux originated from mineral soil horizons). Thus, any natural disturbance and land-use change that aects the composition of the soil prole will strongly aect soil CO2 eux rates. According to the results of this and others studies, such an eect will be larger if mineral soil horizons were disturbed. 4.4. Annual soil CO2 ux A comparison of annual soil CO2 eux with annual litterfall and root net primary productivity estimates supported the partitioning of total soil respiration in root and microbial respiratory uxes. The annual soil CO2 eux for the 47-year old P. abies stand totaled about 710 g C m2 yr1 (using the exponential equation given in Table 3 with daily averages of soil temperatures in the Of layer at 5 cm depth; data provided by C. Rebmann, Max-Planck-Institute for Biogeochemistry, Jena, Germany). Separating this ux into its microbial and root components resulted in >500 g C m2 yr1 (>70%) and <210 g C m2 yr1 (<30%), respectively. Total annual above-ground litterfall (including wood) was about 240 g C m2 yr1 (data provided by the Bayreuth Institute for Terrestrial Ecosystem Research, Bayreuth, Germany); annual root net primary productivity of spruce was about 290 g C m2 yr1 (Mund et al., in preparation). Assuming (1) fast turnover rates of root biomass (Schlesinger, 1997),

(2) carbon losses due to soil respiration that are higher than the carbon input due to annual above-ground litterfall (Raich and Nadelhoer, 1989; Nadelhoer and Raich, 1992) and (3) negligible contributions from understory root respiration (see methods), 530 g C m2 yr1 are available for microbial respiration 240 290 g C/m2 yr1; spruce litter and root carbon inputs). Surprisingly, the soil respiration and litterfall data for this spruce stand fall perfectly within the 95% condence interval of the regression between both parameters reported by Nadelhoer and Raich (1992). Furthermore, this potential microbial respiratory carbon loss of 530 g C m2 yr1 compares well to the value of >500 g C m2 yr1 estimated in this study. Furthermore, after accounting for the microbial fraction of bulk respiration, the remaining 180 g C m2 yr1 710 530 g C m2 yr1) should represent the root component of soil respiration. Again, this value compares well to the <210 g C m2 yr1 estimated for root respiration (see earlier). 5. Conclusions The contribution to the uncertainty of regional or landscape estimates of annual carbon uxes from small-scale, spatial variations of soil CO2 uxes in a watershed seems to be of minor importance. Despite the existence of high within-site variations at any given time, all four spruce stands followed the same seasonal trend, even though they varied in age and stand structure. Variations in soil respiration could be explained using soil temperature only. However, major errors can arise in carbon models from assumptions about the relative contribution of root, microbial, organic and mineral soil CO2 uxes. These assumptions are especially important after natural or anthropogenic disturbances (e.g., re, windthrow, logging), when carbon stocks have been lost, and physical soil parameters have been altered (e.g., depth and density of organic and mineral horizons). Here, microbial respiration may exceed the 70% reported from many eld sites, but unaccounted for in most models. Consequently, CO2 ux rates, their response to increased soil temperatures as well as the future potential for carbon sequestration in those soils might be strongly underestimated in current models. Acknowledgements The author thanks the Forstamt Selb for going permission to work at the forest sites, and J.R. Brooks and M. Mund for their comments on an earlier version of this manuscript. The author was supported by the Deutsche Forschungsgemeinschaft. Data for annual

1634

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635 Goulden, M.L., Munger, J.W., Fan, S.M., Daube, B.C., Wofsy, S.C., 1996. Exchange of carbon dioxide by a deciduous forest: response to interannual climate variability. Science 271, 15761578. Grant, R.F., Rochette, P., 1994. Soil microbial respiration at dierent water potentials and temperatures: theory and mathematical modelling. Soil Science Society of America Journal 58, 1681 1690. Hanson, P.J., Wullschleger, S.D., Bohlman, S.A., Todd, D.E., 1993. Seasonal and topographic patterns of forest oor CO2 eux from an upland oak forest. Tree Physiology 13, 115. Hanson, P.J., Edwards, N.T., Garten, C.T., Andrews, J.A., 2000. Separating root and soil microbial contributions to soil respiration: a review of methods and observations. Biogeochemistry 48, 115146. Hart, S.C., Sollins, P., 1998. Soil carbon and nitrogen pools and processes in an old-growth conifer forest 13 years after trenching. Canadian Journal of Forest Research 28, 12611265. Helal, H.M., Sauerbeck, D., 1991. Short term determination of the actual respiration rate of intact plant roots. In: McMichael, B.L., Persson, H. (Eds.), Plant Roots and Their Environment. Elsevier Applied Science, London, NY, pp. 8892. Hogberg, P., Ekblad, A., 1996. Substrate-induced respiration measured in situ in a C3-plant ecosystem using additions of C4sucrose. Soil Biology and Biochemistry 28, 11311138. Horwath, W.R., Pregitzer, K.S., Paul, E.A., 1994. 14 C allocation in treesoil systems. Tree Physiology 14, 11631176. IPCC, 1996. Climate Change 1995. The Science of Climate Change. Cambridge University Press, Cambridge, p. 572. Kelting, D.L., Burger, J.A., Edwards, G.S., 1998. Estimating root respiration, microbial respiration in the rhizosphere, and root-free soil respiration in forest soils. Soil Biology and Biochemistry 30, 961968. Kicklighter, D.W., Melillo, J.M., Peterjohn, W.T., Rastetter, E.B., McGuire, A.D., Steudler, P.A., 1994. Aspects of spatial and temporal aggregation in estimating regional carbon dioxide uxes from terrestrial forest soils. Journal of Geophysical Research 99, 13031315. Kirschbaum, M.U.F., 1995. The temperature dependence of soil organic matter decomposition, and the eect of global warming on soil organic matter. Soil Biology and Biochemistry 27, 753760. Lavigne, M.B., Ryan, M.G., Anderson, D.E., Baldocchi, D.D., Crill, P.M., Fitzjarrald, D.R., Goulden, M.L., Gower, S.T., Massheder, J.M.M., McCaughey, J.H., Rayment, M.B., Striegl, R.G., 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 102, 2897728985. Lin, G., Ehleringer, J.R., Rygiewicz, P.T., Johnson, M.G., Tingey, D.T., 1999. Elevated CO2 and temperature impacts on dierent components of soil CO2 eux in Douglas-r terracosms. Global Change Biology 5, 157168. Lindroth, A., Grelle, A., Moren, A., 1998. Long-term measurements of boreal forest carbon balance reveal large temperature sensitivity. Global Change Biology 4, 443450. Lloyd, J., Taylor, J.A., 1994. On the temperature dependence of soil respiration. Functional Ecology 8, 315323. Mallik, A.U., Hu, D., 1997. Soil respiration following site preparation treatments in boreal mixed wood forest. Forest Ecology and Management 97, 265275. Moncrie, J., Valentini, R., Greco, S., Seufert, G., Ciccioli, P., 1997. Trace gas exchange over terrestrial ecosystems: methods and perspectives in micrometeorology. Journal of Experimental Botany 48, 11331142. Mund, M., Kummetz, E., Hein, M., Bauer, G., Schulze, E.D. Growth and carbon stocks of spruce forests under the inuence of atmospheric nitrogen deposition a chronosequence study. In preparation.

soil temperatures were collected by C. Rebmann, MaxPlanck Institute for Biogeochemistry, Jena, Germany. Litterfall data were provided by the Bayreuth Institute for Terrestrial Ecosystem Research (BITOK, Bayreuth, Germany).

References
Baldocchi, D.D., Verma, S.B., Matt, D.R., Anderson, D.E., 1986. Eddy-correlation measurements of carbon dioxide eux from the oor of a deciduous forest. Journal of Applied Ecology 23, 967 975. Ben-Asher, J., Cardon, G.E., Peters, D., Rolston, D.E., Biggar, J.W., Phene, C.J., Ephrath, J.E., 1994. Determining root activity distribution by measuring surface carbon dioxide uxes. Soil Science Society of America Journal 58, 926930. Boone, R.D., Nadelhoer, K.J., Canary, J.D., Kaye, J.P., 1998. Roots exert a strong inuence on the temperature sensitivity of soil respiration. Nature 396, 570572. Bouma, T.J., Nielsen, K.L., Eissenstat, D.M., Lynch, J.P., 1997. Estimating respiration of roots in soil: interactions with soil CO2, soil temperature and soil water content. Plant and Soil 195, 221 232. Bowden, R.D., Nadelhoer, K.J., Boone, R.D., Melillo, J.M., Garrison, J.B., 1993. Contributions of above-ground litter, belowground litter, and root respiration to total soil respiration in a temperate mixed hardwood forest. Canadian Journal of Forest Research 23, 14021407. Buchmann, N., Guehl, J.M., Barigah, T.S., Ehleringer, J.R., 1997. Interseasonal comparison of CO2 concentrations, isotopic composition, and carbon dynamics in an Amazonian rainforest (French Guiana). Oecologia 110, 120131. Buchmann, N., Hinckley, T.M., Ehleringer, J.R., 1998. Carbon isotope dynamics in Abies amabilis stands in the Cascades. Canadian Journal of Forest Research 28, 808819. Burton, A.J., Pregitzer, K.S., Zogg, G.P., Zak, D.R., 1998. Drought reduces root respiration in sugar maple forests. Ecological Applications 8, 771778. Conant, R.T., Klopatek, J.M., Malin, R.C., Klopatek, C.C., 1998. Carbon pools and uxes along an environmental gradient in northern Arizona. Biogeochemistry 43, 4361. Davidson, E.A., Belk, E., Boone, R.D., 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. Dieenbach, A., Gottlein, A., Matzner, E., 1997. In-situ soil solution chemistry in an acid forest soil as inuenced by growing roots of Norway spruce (Picea abies [L.] Karst.). Plant and Soil 192, 57 61. Edwards, N.T., Sollins, P., 1973. Continuous measurement of carbon dioxide evolution from partitioned forest oor components. Ecology 54, 406412. Fan, S.M., Goulden, M.L., Munger, J.W., Daube, B.C., Bakwin, P.S., Wofsy, S.C., Amthor, J.S., Fitzjarrald, D., Moore, K.E., Moore, T.R., 1995. Environmental controls on the photosynthesis and respiration of a boreal lichen woodland: a growing season of whole-ecosystem exchange measurements by eddy-correlation. Oecologia 102, 443452. Fisher, F.M., Gosz, J.R., 1986. Eects of trenching on soil processes and properties in a New Mexico mixed-conifer forest. Biology and Fertilility of Soils 2, 3542. Gansert, D., 1994. Root respiration and its importance for the carbon balance of beech seedlings (Fagus sylvatica L.) in a montane beech forest. Plant and Soil 167, 109119.

N. Buchmann / Soil Biology & Biochemistry 32 (2000) 16251635 Nadelhoer, K.J., Raich, J.W., 1992. Fine root production estimates and belowground carbon allocation in forest ecosystems. Ecology 73, 11391147. Nakane, K., Kohno, T., Horikoshi, T., 1996. Root respiration rate before and just after clear-felling in a mature, deciduous, broadleaved forest. Ecological Research 11, 111119. Nakane, K., Kohno, T., Horikoshi, T., Nakatsubo, T., 1997. Soil carbon cycling at a black spruce (Picea mariana ) forest stand in Saskatchewan, Canada. Journal of Geophysical Research 102, 2878528793. Norman, J.M., Kucharik, C.J., Gower, S.T., Baldocchi, D.D., Crill, P.M., Rayment, M.B., Savage, K., Striegl, R.G., 1997. A comparison of six methods for measuring soil-surface carbon dioxide uxes. Journal of Geophysical Research 102, 2877328777. Pregitzer, K.S., Laskowski, M.J., Burton, A.J., Lessard, V.C., Zak, D.R., 1998. Variation in sugar maple root respiration with root diameter and soil depth. Tree Physiology 18, 665670. Raich, J.W., Nadelhoer, K.J., 1989. Belowground carbon allocation in forest ecosystems: global trends. Ecology 70, 13461354. Raich, J.W., Schlesinger, W.H., 1992. The global carbon dioxide ux in soil respiration and its relationship to vegetation and climate. Tellus 44B, 8199. Raich, J.W., Potter, C.S., 1995. Global patterns of carbon dioxide emissions from soils. Global Biogeochemical Cycles 9, 2336. Randerson, J.T., Thompson, M.V., Malmstrom, C.M., Field, C.B., Fung, I.Y., 1996. Substrate limitations for heterotrophs: impli-

1635

cations for models that estimate the seasonal cycle of atmospheric CO2. Global Biogeochemical Cycles 10, 585602. Schimel, D.S., VEMAP Participants, Braswell, B.H. Continental scale variability in ecosystem processes: models, data, and the role of disturbance. Ecological Monographs. 67 (1997) 251271. Schleser, G.H., 1982. The response of CO2 evolution from soils to global temperature changes. Zeitschrift fur Naturforschung 37a, 287291. Schlesinger, W.H., 1977. Carbon balance in terrestrial detritus. Annual Review of Ecology and Systematics 8, 5181. Schlesinger, W.H., 1997. Biogeochemistry. Academic Press, London, NY, p. 563. Swinnen, J., van Veen, J.A., Merckx, R., 1994. Rhizosphere carbon uxes in eld-grown spring wheat: model calculations based on 14 C partitioning after pulse-labelling. Soil Biology and Biochemistry 26, 171182. Thierron, V., Laudelout, H., 1996. Contribution of root respiration to total CO2 eux from the soil of a deciduous forest. Canadian Journal of Forest Research 26, 11421148. Thuille, A., Buchmann, N., Schulze, E.D., 2000. Carbon pools and soil respiration rates during deforestation, grassland use and subsequent spruce aorestation in the Southern Alps, Italy: a case study related to the Kyoto Protocol. Tree Physiology, in press. Witkamp, M., 1969. Cycles of temperature and carbon dioxide evolution from litter and soil. Ecology 50, 922924.