Representative estimates of soil and ecosystem respiration in an old beech forest

Respiration has been proposed to be the main determinant of the carbon balance in European forests and is thus essential for our understanding of the carbon cycle. However, the choice of experimental design strongly affects estimates of annual respiration and of the contribution of soil respiration to total ecosystem respiration. In a detailed study of ecosystem and soil respiration fluxes in an old unmanaged deciduous forest in Central Germany over 3 years (2000–2002), we combined soil chamber and eddy covariance measurements to obtain a comprehensive picture of respiration in this forest. The closed portable chambers offered to investigate spatial variability of soil respiration and its controls while the eddy covariance system offered continuous measurements of ecosystem respiration. Over the year, both fluxes were mainly correlated with temperature. However, when soil moisture sank below 23 vol.% in the upper 6 cm, water limitations also became apparent. The temporal resolution of the eddy covariance system revealed that relatively high respiration rates occurred during budbreak due to increased metabolic activity and after leaf fall because of increased decomposition. Spatial variability in soil respiration rates was large and correlated with fine root biomass (r ² = 0.56) resulting in estimates of annual efflux varying across plots from 730 to 1,258 (mean 898) g C m⁻² year⁻¹. Power function calculations showed that achieving a precision in the soil respiration estimate of 20% of the full population mean at a confidence level of 95%, requires about eight sampling locations. Our results can be used as guidelines to improve the representativeness of soil respiration measurements by nested sampling designs, being applied in long-term and large-scale carbon sequestration projects such as FLUXNET and CarboEurope.     

Response of soil respiration to simulated N deposition in a disturbed and a rehabilitated tropical forest in southern China

Responses of soil respiration (CO₂ emission) to simulated N deposition were studied in a disturbed (reforested forest with previous understory and litter harvesting) and a rehabilitated (reforested forest with no understory and litter harvesting) tropical forest in southern China from October 2005 to September 2006. The objectives of the study were to test the following hypotheses: (1) soil respiration is higher in rehabilitated forest than in disturbed forest; (2) soil respiration in both rehabilitated and disturbed tropical forests is stimulated by N additions; and (3) soil respiration is more sensitive to N addition in disturbed forest than in rehabilitated forest due to relatively low soil nutrient status in the former, resulting from different previous human disturbance. Static chamber and gas chromatography techniques were employed to quantify the soil respiration, following different N treatments (Control, no N addition; Low-N, 5 g N m⁻² year⁻¹; Medium-N, 10 g N m⁻² year⁻¹), which had been applied continuously for 26 months before the respiration measurement. Results showed that soil respiration exhibited a strong seasonal pattern, with the highest rates observed in the hot and wet growing season (April–September) and the lowest rates in winter (December–February) in both rehabilitated and disturbed forests. Soil respiration rates exhibited significant positive exponential relationship with soil temperature and significant positive linear relationship with soil moisture. Soil respiration was also significantly higher in the rehabilitated forest than in the disturbed forest. Annual mean soil respiration rate in the rehabilitated forest was 20% lower in low-N plots (71 ± 4 mg CO₂-C m⁻² h⁻¹) and 10% lower in medium-N plots (80 ± 4 mg CO₂-C m⁻² h⁻¹) than in the control plots (89 ± 5 mg CO₂-C m⁻² h⁻¹), and the differences between the control and low-N or medium-N treatments were statistically significant. In disturbed forest, annual mean soil respiration rate was 5% lower in low-N plots (63 ± 3 mg CO₂-C m⁻² h⁻¹) and 8% lower in medium-N plots (61 ± 3 mg CO₂-C m⁻² h⁻¹) than in the control plots (66 ± 4 mg CO₂-C m⁻² h⁻¹), but the differences among treatments were not significant. The depressed effects of experimental N deposition occurred mostly in the hot and wet growing season. Our results suggest that response of soil respiration to elevated N deposition in the reforested tropical forests may vary depending on the status of human disturbance. Responsible Editor: Hans Lambers.     

Characteristics of soil CO2 efflux variability in an aseasonal tropical rainforest in Borneo Island

Although soil carbon dioxide (CO₂) efflux from tropical forests may play an important role in global carbon (C) balance, our knowledge of the fluctuations and factors controlling soil CO₂ efflux in the Asian tropics is still poor. This study characterizes the temporal and spatial variability in soil CO₂ efflux in relation to temperature/moisture content and estimates annual efflux from the forest floor in an aseasonal intact tropical rainforest in Sarawak, Malaysia. Soil CO₂ efflux varied widely in space; the range of variation averaged 17.4 μmol m⁻² s⁻¹ in total. While most CO₂ flux rates were under 10 μmol m⁻² s⁻¹, exceptionally high fluxes were observed sporadically at several sampling points. Semivariogram analysis revealed little spatial dependence in soil CO₂ efflux. Temperature explained nearly half of the spatial heterogeneity, but the effect varied with time. Seasonal variation in CO₂ efflux had no fixed pattern, but was significantly correlated with soil moisture content. The correlation coefficient with soil moisture content (SMC) at 30 and 60 cm depth was higher than at 10 cm depths. The annual soil CO₂ efflux, estimated from the relationship between CO₂ efflux and SMC at 30 cm depth, was 165 mol m⁻² year⁻¹ (1,986 g C m⁻² year⁻¹). As this area is known to suffer severe drought every 4–5 years caused by the El Nino-Southern Oscillation, the results suggest that an unpredictable dry period might affect soil CO₂ efflux, leading an annual variation in soil C balance.     

Spatial and temporal patterns of soil respiration over the Japanese Archipelago: a model intercomparison study

We used terrestrial ecosystem models to estimate spatial and temporal variability in and uncertainty of estimated soil carbon dioxide (CO₂) efflux, or soil respiration, over the Japanese Archipelago. We compared five carbon-cycle models to assess inter-model variability: Biome-BGC, CASA, LPJ, SEIB, and VISIT. These models differ in approaches to soil carbon dynamics, root respiration estimation, and relationships between decomposition and environmental factors. We simulated the carbon budget of natural ecosystems over the archipelago for 2001–2006 at 1-day time steps and 2-min (latitude and longitude) spatial resolution. The models were calibrated using measured flux data to accurately represent net ecosystem CO₂ exchange. Each model successfully reproduced seasonal changes and latitudinal gradients in soil respiration. The five-model average of estimated total soil respiration of Japanese ecosystems was 295 Tg C year⁻¹, with individual model estimates ranging from 210 to 396 Tg C year⁻¹ (1 Tg = 10¹² g). The differences between modeled estimates were more evident in summer and in warmer years, implying that they were mainly attributable to differences in modeling the temperature dependence of soil respiration. There was a large discrepancy between models in the estimated contribution of roots to total soil respiration, ranging from 3.9 to 48.4%. Although model calibration reduced the uncertainty of flux estimates, substantial uncertainties still remained in estimates of underground processes from these terrestrial carbon-cycle models.     

Carbon stock and cycling in a bambooPhyllostachys bambusoides stand

Gross production and carbon cycling in aPhyllostachys bambusoides stand in Kyoto Prefecture, central Japan, were determined, and then a compartment model showing the carbon stock and cycling within the ecosystem was developed. Aboveground carbon stock was 52.3 tC ha⁻¹, increasing at a rate of 3.6 tC ha⁻¹ year⁻¹. Belowground carbon stock was 20.8 tC ha⁻¹ in the root system and 92.0 tC ha⁻¹ in the soil. Aboveground net production was 11.2 tC ha⁻¹ year⁻¹. Belowground net production was crudely estimated at 4.5 tC ha⁻¹ year⁻¹. The gross production was estimated at 41.8 tC ha⁻¹ year⁻¹ by summing the amount of outflow to the environment and the increment in biomass. Leaves consumed 13.7 tC ha⁻¹ year⁻¹ by respiration; the rest (41.8−13.7=28.1 tC ha⁻¹ year⁻¹) was surplus production of the leaves and flowed into the other compartments. The amounts of construction and maintenance respiration of the aboveground compartments were 3.4 and 18.5 tC ha⁻¹ year⁻¹, respectively. The annual amount of soil respiration was 11.2 tC ha⁻¹ year⁻¹. Soil respiration levels of 4.3 and 3.1 tC ha⁻¹ year⁻¹ were estimated for the flow of root respiration and root detritus. The proportion of net to gross production was 37%, which fell within the range of young and mature forests. A shorter life span of culms, compared to tree trunks, resulted in smaller biomass accumulation ratio (biomass/net production) in the ecosystem, of 4.66.     

Estimation of autotrophic and heterotrophic components of soil respiration by trenching is sensitive to corrections for root decomposition and changes in soil water content

This study aims to assess the effects of corrections for disturbances such as an increased amount of dead roots and an increase in volumetric soil water content on the calculation of soil CO₂ efflux partitioning. Soil CO₂ efflux, soil temperature and superficial soil water content were monitored in two young beech sites (H1 and H2) during a trenching experiment. Trenching induced a significant input of dead root mass that participated in soil CO₂ efflux and reduced the soil dissolved organic carbon content, while it increased superficial soil water content within the trenched plot. Annual soil CO₂ efflux in control plots was 528 g C m⁻² year⁻¹ at H1 and 527 g C m⁻² year⁻¹ at H2. The annual soil CO₂ efflux in trenched plots was 353 g C m⁻² year⁻¹ at H1 and 425 g C m⁻² year⁻¹ at H2. By taking into account annual CO₂ efflux from decaying trenched roots, the autotrophic contribution to total soil CO₂ efflux reached 69% at H1 and 54% at H2. The partitioning calculation was highly sensitive to the initial root mass estimated within the trenched plots. Uncertainties in the remaining root mass, the fraction of root C that is incorporated into soil organic matter during root decomposition, and the root decomposition rate constant had a limited impact on the partitioning calculation. Corrections for differences in superficial soil water content had a significant impact on annual respired CO₂ despite a limited effect on partitioning.     

Biometric-based estimation of net ecosystem production in a mature Japanese cedar (<Emphasis Type="Italic">Cryptomeria japonica</Emphasis>) plantation beneath a flux tower

Quantification of carbon budgets and cycling in Japanese cedar (Cryptomeria japonica D. Don) plantations is essential for understanding forest functions in Japan because these plantations occupy about 20% of the total forested area. We conducted a biometric estimate of net ecosystem production (NEP) in a mature Japanese cedar plantation beneath a flux tower over a 4-year period. Net primary production (NPP) was 7.9 Mg C ha⁻¹ year⁻¹ and consisted mainly of tree biomass increment and aboveground litter production. Respiration was calculated as 6.8 (soil) and 3.3 (root) Mg C ha⁻¹ year⁻¹. Thus, NEP in the plantation was 4.3 Mg C ha⁻¹ year⁻¹. In agreement with the tower-based flux findings, this result suggests that the Japanese cedar plantation was a strong carbon sink. The biometric-based NEP was higher among most other types of Japanese forests studied. Carbon sequestration in the mature plantation was characterized by a larger increment in tree biomass and lower mortality than in natural forests. Land-use change from natural forest to Japanese cedar plantation might, therefore, stimulate carbon sequestration and change the carbon allocation of NPP from an increment in coarse woody debris to an increase in tree biomass.     

Litterfall dynamics in carbonate and deltaic mangrove ecosystems in the Gulf of Mexico

From 1996 to 2002, we measured litterfall, standing litter crop, and litter turnover rates in scrub, basin, fringe and riverine forests in two contrasting mangrove ecosystems: a carbonate-dominated system in the Southeastern Everglades and a terrigenous-dominated system in Laguna de Terminos (LT), Mexico. We hypothesized that litter dynamics is driven by latitude, geomorphology, hydrology, soil fertility and soil salinity stress. There were significant temporal patterns in LT with litterfall rates higher during the rainy season (2.4 g m⁻² day⁻¹) than during the dry season (1.8 g m⁻² day⁻¹). Total annual litterfall was significantly higher in the riverine forest (12.8 Mg ha⁻² year⁻¹) than in the fringe and basin forests (9.7 and 5.2 Mg ha⁻² year⁻¹, respectively). In Southeastern Everglades, total annual litterfall was also significantly higher during the rainy season than during the dry season. Spatially, the scrub forest had the lowest annual litterfall (2.5 Mg ha⁻² year⁻¹), while the fringe and basin had the highest (9.1 and 6.5 Mg ha⁻² year⁻¹, respectively). In LT, annual standing litter crop was 3.3 Mg ha⁻¹ in the fringe and 2.2 Mg ha⁻¹ in the basin. Litter turnover rates were significantly higher in the fringe mangrove forest (4.1 year⁻¹) relative to the basin forests (2.2 year⁻¹). At Southeastern Everglades there were significant differences in annual standing litter crop: 1.9, 3.3 and 4.5 Mg ha⁻¹ at scrub, basin and fringe mangrove sites, respectively. Furthermore, turnover rates were similar at both basin and fringe mangrove types (2.1 and 2.0 year⁻¹, respectively) but significantly higher than scrub mangrove forest (1.3 year⁻¹). These findings suggest that litter export is important in regulating litter turnover rates in frequently flooded riverine and fringe forests, while in infrequently flooded basin forests, in situ litter decomposition controls litter turnover rates.     

Estimation of carbon storage, carbon inputs, and soil CO2 efflux of alder plantations on granite soil in central Korea: comparison with Japanese larch plantation

Alder is a typical species used for forest rehabilitation after disturbances because of its N₂-fixing activities through microbes. To investigate forest dynamics of the carbon budget, we determined the aboveground and soil carbon content, carbon input by litterfall to belowground, and soil CO₂ efflux over 2 years in 38-year-old alder plantations in central Korea. The estimated aboveground carbon storage and increment were 47.39 Mg C ha⁻¹ and 2.17 Mg C ha⁻¹ year⁻¹. Carbon storage in the organic layer and in mineral soil in the topsoil to 30 cm depth were, respectively, 3.21 and 66.85 Mg C ha⁻¹. Annual carbon input by leaves and total litter in the study stand were, respectively, 1.78 and 2.68 Mg C ha⁻¹ year⁻¹. The aboveground carbon increment at this stand was similar to the annual carbon inputs by total litterfall. The diurnal pattern of soil CO₂ efflux was significantly different in May, August, and October, typically varying approximately twofold throughout the course of a day. In the seasonally observed pattern, soil CO₂ efflux varied strongly with soil temperature; increasing trends were evident during the early growing season, with sustained high rates from mid May through late October. Soil CO₂ efflux was related exponentially to soil temperature (R ² = 0.85, P < 0.0001), but not to soil water content. The Q ₁₀ value for this plantation was 3.8, and annual soil respiration was estimated at 10.2 Mg C ha⁻¹ year⁻¹. An erratum to this article can be found at     

Annual sediment respiration in estuarine sandy intertidal flats in the Seto Inland Sea, Japan

To quantify organic matter mineralization at estuarine intertidal flats, we measured in situ sediment respiration rates using an infrared gas analyzer in estuarine sandy intertidal flats located in the northwestern Seto Inland Sea, Japan. In situ sediment respiration rates showed spatial and seasonal variations, and the mean of the rates is 38.8 mg CO₂-C m⁻² h⁻¹ in summer. In situ sediment respiration rates changed significantly with sediment temperature at the study sites (r ² = 0.70, p < 0.05), although we did not detect any significant correlations between the rates and sediment characteristics. We prepared a model for estimating the annual sediment respiration based on the in situ sediment respiration rates and their temperature coefficient (Q ₁₀ = 1.8). The annual sediment respiration was estimated to be 92 g CO₂-C m⁻² year⁻¹. The total amount of organic carbon mineralization for the entire estuarine intertidal flats through sediment respiration (43 t C year⁻¹) is equivalent to approximately 25% of the annual organic carbon load supplied from the river basin of the estuary.     

Effects of environmental factors upon variation in soil respiration of a Zoysia japonica grassland, central Japan

The effects of environmental factors on seasonal and annual variations in soil respiration were examined in the cool temperate Zoysia japonica grassland of Japan. Field measurements of soil respiration were conducted using a closed chamber method with an infrared gas analyzer at monthly intervals in the snow-free seasons from May 2007 to December 2009. There was an exponential relationship between soil respiration and soil temperature, and the soil temperature accounted for 85–86% of seasonal soil respiration variability. Moreover, a positive linear relationship between soil respiration and soil water content was detected in summer (R ² = 0.55, p < 0.001), but not in spring or autumn. Annual soil respiration was estimated at 755, 719, and 1,037 g C m⁻² year⁻¹ in 2007, 2008, and 2009, respectively. These interannual variations in soil respiration might be influenced by the strength of precipitation during rainy seasons and the timing of each snow-melt. Our results suggest that the effects of rainfall and snow-melt events on soil respiration might be important factors to understand carbon dynamics in grassland ecosystem in Japan.     

Soil carbon storage, litterfall and CO2 efflux in fertilized and unfertilized larch (Larix leptolepis) plantations

This study evaluated the effects of forest fertilization on the forest carbon (C) dynamics in a 36-year-old larch (Larix leptolepis) plantation in Korea. Above- and below-ground C storage, litterfall, root decomposition and soil CO₂ efflux rates after fertilization were measured for 2 years. Fertilizers were applied to the forest floor at rates of 112 kg N ha⁻¹ year⁻¹, 75 kg P ha⁻¹ year⁻¹ and 37 kg K ha⁻¹ year⁻¹ for 2 years (May 2002, 2003). There was no significant difference in the above-ground C storage between fertilized (41.20 Mg C ha⁻¹) and unfertilized (42.25 Mg C ha⁻¹) plots, and the C increment was similar between the fertilized (1.65 Mg C ha⁻¹ year⁻¹) and unfertilized (1.52 Mg C ha⁻¹ year⁻¹) plots. There was no significant difference in the soil C storage between the fertilized and unfertilized plots at each soil depth (0–15, 15–30 and 30–50 cm). The organic C inputs due to litterfall ranged from 1.57 Mg C ha⁻¹ year⁻¹ for fertilized to 1.68 Mg C ha⁻¹ year⁻¹ for unfertilized plots. There was no significant difference in the needle litter decomposition rates between the fertilized and unfertilized plots, while the decomposition of roots with 1–2 mm diameters increased significantly with the fertilization relative to the unfertilized plots. The mean annual soil CO₂ efflux rates for the 2 years were similar between the fertilized (0.38 g CO₂ m⁻² h⁻¹) and unfertilized (0.40 g CO₂ m⁻² h⁻¹) plots, which corresponded with the similar fluctuation in the organic carbon (litterfall, needle and root decomposition) and soil environmental parameters (soil temperature and soil water content). These results indicate that little effect on the C dynamics of the larch plantation could be attributed to the 2-year short-term fertilization trials and/or the soil fertility in the mature coniferous plantation used in this study.     

Overgrazing and soil carbon dynamics in eastern Inner Mongolia of China

Eastern Inner Mongolia of China is a typical ecotone between sandy forests and steppe. Little is known about the effect of overgrazing on carbon loss from soil in semiarid steppe and sandy forests of the north of China. The soil carbon parameters were measured in a 10,000 ha natural reserve in eastern Inner Mongolia of China (43°30′–43°36′N, 117°06′–117°16′E). Three situations were compared: primary protected (PP), moderately protected (MP) and highly degraded (HD). Soil and litter samples were recovered in spring and summer. Soil organic carbon (SOC) and CO₂–C values decreased from the PP (9.23 kg m⁻² and 157 g m⁻²) to the HD (1.69 kg m⁻² and 57 g m⁻²) sites whereas the C mineralization rate increased toward the less restored sites (1.06–2.37). Surface-litter C was different in both sites under protection (PP 648 and MP 408 g m⁻²), an was low at the HD site (17 g m⁻²). Leaves from woody species dominated the surface litter at the PP site, whereas grass material was predominant at the MP site. During summer, both CO₂–C and SOC decreased, whereas the C mineralization rate increased. We calculated that C loss since the introduction of cattle into the forest was 77 M g ha⁻¹, reaching a total of 1.1×10¹⁵ g for eastern Inner Mongolia. These values are higher than those caused by the conversion of steppe and other ecosystems into agriculture or cultivated pastures. The amount of C fixed at the PP site (650 g ha⁻¹year⁻¹) indicates that the sandy soils have a significant potential as atmospheric carbon sinks. Foundation item: The National Natural Science Foundation of China (No. 39900019, 30070129).     

Model simulation of soil carbon dynamics with stand development in Japanese cedar (<Emphasis Type="Italic">Cryptomeria japonica</Emphasis>) plantations

Soil carbon cycling in Japanese cedar plantations after clear-cutting over time was calculated with and without consideration of plantation management (pruning and thinning) using a mathematical model. The model employed a daily time step and used daily air temperature and observed precipitation near the plantations. The results of these calculations of above-ground biomass, carbon flow and accumulation were compared with the field observations reported by Shutou and Nakane (Ecol Res 19:233–244, 2004) observed in the Hiroshima prefecture. For example, the carbon accumulation and total soil respiration rate calculated with the inclusion of plantation management for a 60-year-old stand were 104 and 4.8 t C ha⁻¹year⁻¹, compared with a field observation of 110 and 4.5 t C ha⁻¹year⁻¹, respectively. The results calculated for carbon flow and accumulation were not significantly different from the field observations, assessed using one-way analysis of variance (P > 0.05), and the mean relative errors were very small (e.g., the maximum was 0.05). Consequently, it is suggested that the dynamics of carbon cycling following clear-cutting of a Japanese cedar plantation can be simulated realistically using these calculations with and without consideration of plantation management. Moreover, it is possible that the calculation without consideration of plantation management can yield the change of carbon cycling over time, even if the history of forest management in the plantations is unknown. It is expected that this study will contribute to the development of a more versatile model.     

Pools and fluxes of carbon in three Norway spruce ecosystems along a climatic gradient in Sweden

This paper presents an integrated analysis of organic carbon (C) pools in soils and vegetation, within-ecosystem fluxes and net ecosystem exchange (NEE) in three 40-year old Norway spruce stands along a north-south climatic gradient in Sweden, measured 2001–2004. A process-orientated ecosystem model (CoupModel), previously parameterised on a regional dataset, was used for the analysis. Pools of soil organic carbon (SOC) and tree growth rates were highest at the southernmost site (1.6 and 2.0-fold, respectively). Tree litter production (litterfall and root litter) was also highest in the south, with about half coming from fine roots (<1 mm) at all sites. However, when the litter input from the forest floor vegetation was included, the difference in total litter input rate between the sites almost disappeared (190–233 g C m⁻² year⁻¹). We propose that a higher N deposition and N availability in the south result in a slower turnover of soil organic matter than in the north. This effect seems to overshadow the effect of temperature. At the southern site, 19% of the total litter input to the O horizon was leached to the mineral soil as dissolved organic carbon, while at the two northern sites the corresponding figure was approx. 9%. The CoupModel accurately described general C cycling behaviour in these ecosystems, reproducing the differences between north and south. The simulated changes in SOC pools during the measurement period were small, ranging from −8 g C m⁻² year⁻¹ in the north to +9 g C m⁻² year⁻¹ in the south. In contrast, NEE and tree growth measurements at the northernmost site suggest that the soil lost about 90 g C m⁻² year⁻¹. An erratum to this article can be found at     

Carbon dynamics and budget in a Zoysia japonica grassland, central Japan

The ecosystem carbon budget was estimated in a Japanese Zoysia japonica grassland. The green biomass started to grow in May and peaked from mid-July to September. Seasonal variations in soil CO₂ flux and root respiration were mediated by changes in soil temperature. Annual soil CO₂ flux was 1,121.4 and 1,213.6 g C m⁻² and root respiration was 471.0 and 544.3 g C m⁻² in 2007 and 2008, respectively. The root respiration contribution to soil CO₂ flux ranged from 33% to 71%. During the growing season, net primary production (NPP) was 747.5 and 770.1 g C m⁻² in 2007 and 2008, respectively. The biomass removed by livestock grazing (GL) was 122.1 and 102.7 g C m⁻², and the livestock returned 28.2 and 25.6 g C m⁻² as fecal input (FI) in 2007 and 2008, respectively. The decomposition of FI (DL, the dry weight loss due to decomposition) was very low, 1.5 and 1.4 g C m⁻², in 2007 and 2008. Based on the values of annual NPP, soil CO₂ flux, root respiration, GL, FI, and DL, the estimated carbon budget of the grassland was 1.7 and 22.3 g C m⁻² in 2007 and 2008, respectively. Thus, the carbon budget of this Z. japonica grassland ecosystem remained in equilibrium with the atmosphere under current grazing conditions over the 2 years of the study.     

Scaling-up technique for net ecosystem productivity of deciduous broadleaved forests in Japan using MODIS data

We attempted to obtain carbon sequestration maps of deciduous forests in Japan using detectable parameters from the Moderate Resolution Imaging Spectrometer (MODIS) sensor and to determine how the spatial pattern of carbon sequestration differs within the same forest ecosystem type. For this investigation, we firstly parameterized the MODIS algorithm at one flux tower site, Takayama, for the years 2002–2003. The MODIS algorithm could link flux-based net ecosystem productivity (NEP) with simple functions controlled by a thermal infrared band and a vegetation index. Second, the performance of the MODIS algorithm was validated through comparisons with the flux-based NEP at another flux tower site, Hitsujigaoka. The MODIS-based NEP at Hitsujigaoka was also within an accuracy of a flux-based NEP with R ² of 0.879 and root mean square error of 1.64 gC m⁻² day⁻¹, regardless of canopy structure and age. The MODIS algorithm was noteworthy for its general applicability in different locations. Finally, we used the MODIS algorithm for the same forest ecosystem type in Japan for regional extrapolation of NEP. The MODIS-based NEP of deciduous forests in Japan showed great variance with 347 ± 288 gC m⁻² year⁻¹ in 2002, according to the stand structure and climatic condition of the year. Studies for quantification of ecosystem carbon balance need to consider variance, frequency and spatial distributions of NEP. Satellite remote sensing demonstrated the potential for the large-scale mapping of NEP.     

Dynamics of soil respiration in sparse <Emphasis Type="Italic">Ulmus pumila</Emphasis> woodland under semi-arid climate

Sparse Ulmus pumila woodlands play an important role in contributing to ecosystem function in semi-arid grassland of northern China. To understand the key attributes of soil carbon cycling in U. pumila woodland, we studied dynamics of soil respiration in the canopy field (i.e., the projected crown cover area) and the open field at locations differing in distance (i.e., at 1–1.5, 3–4, 10, and >15 m) to tree stems from July through September of 2005, and measured soil biotic factors (e.g., fine root mass, soil microbial biomass, and activity) and abiotic factors [e.g., soil water content (SWC) and organic carbon] in mid-August. Soil respiration was further separated into root component and microbial component at the end of the field measurement in September. Results showed that soil respiration had a significant exponent relationship with soil temperature at 10-cm depth. The temperature sensitivity index of soil respiration, Q ₁₀, was lower than the global average of 2.0, and declined significantly (P < 0.05) with distance. The rate of soil respiration was generally greater in the canopy field than in the open field; monthly mean of soil respiration was 305.5–730.8 mg CO₂ m⁻² h⁻¹ in the canopy field and 299.6–443.1 mg CO₂ m⁻² h⁻¹ in the open field from July through September; basal soil respiration at 10°C declined with distance, and varied from ~250 mg CO₂ m⁻² h⁻¹ near tree stems to <200 mg CO₂ m⁻² h⁻¹ in the open field. Variations in soil respiration with distance were consistent with patterns of SWC, fine root mass, microbial biomass and activities. Regression analysis indicated that soil respiration was tightly coupled with microbial respiration and only weakly related to root respiration. Overall, variations in SWC, soil nutrients, microbial biomass, and microbial activity are largely responsible for the spatial heterogeneity of soil respiration in this semi-arid U. pumila woodland.     

Different patterns of ecosystem carbon accumulation between a young and an old-growth subtropical forest in Southern China

Using long-term (22 years) measurements from a young and an old-growth subtropical forest in southern China, we found that both forests accumulated carbon from 1982 to 2004, with the mean carbon accumulation rate at 227 ± 59 g C m⁻² year⁻¹ for young forest and 115 ± 89 g C m⁻² year⁻¹ for the old-growth forest. Allocation of the accumulated carbon was quite different between these two forests: the young forest accumulated a significant amount of carbon in plant live biomass, whereas the old-growth forest accumulated a significant amount of carbon in the soil. From 1982 to 2004, net primary productivity (NPP) increased for the young forest, and did not change significantly for the old-growth forest. The increase in NPP of the young forest resulted from recruitment of some dominant tree species characteristic of the subtropical mature forest in the region and an increase in tree density; decline of NPP of the old-growth forest was caused by increased mortality of the dominant trees.     

Regulatory effects of soil properties on background N2O emissions from agricultural soils in China

The background nitrous oxide (N₂O) emission (BNE) from agricultural soils originates from microbial nitrification and denitrification processes of soil nitrogen supplies, excluding emissions from nitrogen fertilizers applied in the current year. It is of great necessity to quantify BNE accurately at various spatial scales since BNE contributes considerably to the overall N₂O emissions from croplands. Annual BNE rates across various soil/climate regions and major cropping systems of China were determined by network observations during 2002–2006 using the static chamber technique. The observations show BNE rates ranging from 0.1 to 3.67 kg N ha⁻¹ year⁻¹, with a mean of 1.35 kg N ha⁻¹ year⁻¹. Empirical functions are derived for cultivated mineral soils and describe the dependences of annual BNE rates upon soil total nitrogen (TN) content, soil organic carbon (SOC) content, bulk density (BD) and clay fraction (CF), separately or collectively. These empirical functions provide simple approaches to scale up estimated national/regional BNE inventories using available database of soil properties surveys and cropland area statistics. The national BNE of China is estimated to be 0.114–0.184 Tg (1 Tg = 10¹² g) N year⁻¹ in 2000, with the range being due to the use of different approaches. However, the available observations of annual BNE rates do not cover the entire range of soil properties on a national scale. Further work is needed to verify the empirical models for a complete range of soil types. In addition, a predictive empirical relationship between annual BNE rates and TN or SOC is established for cultivated mineral soil at the global scale. However, the empirical models could not accurately predict the BNE rates of cultivated organic soils. Further studies are required to understand the regulatory effects of soil properties on annual BNE rates of cultivated organic soils.