Resolving systematic errors in estimates of net ecosystem exchange of CO2 and ecosystem respiration in a tropical forest biome

The controls on uptake and release of CO₂ by tropical rainforests, and the responses to a changing climate, are major uncertainties in global climate change models. Eddy-covariance measurements potentially provide detailed data on CO₂ exchange and responses to the environment in these forests, but accurate estimates of the net ecosystem exchange of CO₂ (NEE) and ecosystem respiration (R_eco) require careful analysis of data representativity, treatment of data gaps, and correction for systematic errors. This study uses the comprehensive data from our study site in an old-growth tropical rainforest near Santarem, Brazil, to examine the biases in NEE and R_eco potentially associated with the two most important sources of systematic error in Eddy-covariance data: lost nighttime flux and missing canopy storage measurements. We present multiple estimates for the net carbon balance and R_eco at our site, including the conventional “u* filter”, a detailed bottom-up budget for respiration, estimates by similarity with ²²²Rn, and an independent estimate of respiration by extrapolation of daytime Eddy flux data to zero light. Eddy-covariance measurements between 2002 and 2006 showed a mean net ecosystem carbon loss of 0.25 ± 0.04 μmol m⁻² s⁻¹, with a mean respiration rate of 8.60 ± 0.11 μmol m⁻² s⁻¹ at our site. We found that lost nocturnal flux can potentially introduce significant bias into these results. We develop robust approaches to correct for these biases, showing that, where appropriate, a site-specific u* threshold can be used to avoid systematic bias in estimates of carbon exchange. Because of the presence of gaps in the data and the day–night asymmetry between storage and turbulence, inclusion of canopy storage is essential to accurate assessments of NEE. We found that short-term measurements of storage may be adequate to accurately model storage for use in obtaining ecosystem carbon balance, at sites where storage is not routinely measured. The analytical framework utilized in this study can be applied to other Eddy-covariance sites to help correct and validate measurements of the carbon cycle and its components.     

Environmental control of net ecosystem CO2 exchange in a treed, moderately rich fen in northern Alberta

Peatlands cover about 21% of the landscape and contain about 80% of the soil carbon stock in western Canada. However, the current rates of carbon accumulation and the environmental controls on ecosystem photosynthesis and respiration in peatland ecosystems are poorly understood. As part of Fluxnet-Canada, we continuously measured net ecosystem carbon dioxide exchange (NEE) using the eddy covariance technique in a treed fen dominated by stunted Picea mariana and Larix laricina trees during August 2003–December 2004. The total carbon stock in the ecosystem was approximately 51,000 g C m⁻², with only 540 g C m⁻² contributed by live above ground vegetation. The NEE measurements were used to parameterize simple physiological models to assess temporal variation in maximum ecosystem photosynthesis (A_max) and ecosystem respiration rate at 10 °C (R₁₀). During mid-summer the ecosystem had a relatively high A_max (approx. 30 μmol m⁻² s⁻¹) with relatively low R₁₀ (approx. 4 μmol m⁻² s⁻¹). The peak mid-day NEE uptake rate during July and August was 10 μmol m⁻² s⁻¹. The ecosystem showed large seasonal variation in photosynthetic and respiratory activity that was correlated with shifts in temperature, with both spring increases and fall decreases in A_max well predicted by the mean daily air temperature averaged over the preceding 21 days. Leaf-level gas exchange and spectral reflectance measurements also suggested that seasonal changes in photosynthetic activity were primarily controlled by shifts in temperature. Ecosystem respiration was strongly correlated with changes in ecosystem photosynthesis during the growing season, suggesting important links between plant activity and mycorrhizae and microbial activity in the shallow layers of the peat. Only very low rates of respiration were observed during the winter months. During 2004, the peatland recorded a net annual gain of 144 g C m⁻² year⁻¹, the result of a difference between gross photosynthesis of 713 and total ecosystem respiration of 569 g C m⁻² year⁻¹.     

Temporal and spatial variability of soil respiration in a boreal mixedwood forest

Temporal and spatial variability of soil respiration (R_s) was measured and analyzed in a 74-year-old, mixedwood, boreal forest in Ontario, Canada, over a period of 2 years (August 2003–July 2005). The ranges of R_s measured during the two study years were 0.5–6.9 μmol CO₂ m⁻² s⁻¹ for 2003–2004 (Year 1) and 0.4–6.8 μmol CO₂ m⁻² s⁻¹ for 2004–2005 (Year 2). Mean annual R_s for the stand was the same for both years, 2.7 μmol CO₂ m⁻² s⁻¹. Temporal variability of R_s was controlled mainly by soil temperature (T_s), but soil moisture had a confounding effect on T_s. Annual estimates of total soil CO₂ emissions at the site, calculated using a simple empirical R_s–T_s relationship, showed that R_s can account for about 88 ± 27% of total annual ecosystem respiration at the site. The majority of soil CO₂ emissions came from the upper 12 to 20 cm organic LFH (litter–fibric–humic) soil layer. The degree of spatial variability in R_s, along the measured transect, was seasonal and followed the seasonal trend of mean R_s: increasing through the growing season and converging to a minimum in winter (coefficient of variation (CV) ranged from 4 to 74% in Year 1 and 4 to 62% in Year 2). Spatial variability in R_s was found to be negatively related to spatial variability in the C:N ratio of the LHF layer at the site. Spatial variability in R_s was also found to depend on forest tree species composition within the stand. R_s was about 15% higher in a broadleaf deciduous tree patch compared to evergreen coniferous area. However, the difference was not always significant (at 95% CI). In general, R_s in the mixedwood patch, having both deciduous and coniferous species, was dominated by broadleaf trees, reflecting changing physiological controls on R_s with seasons. Our results highlight the importance of discerning soil CO₂ emissions at a variety of spatial and temporal scales. They also suggest including the LFH soil layer and allowing for seasonal variability in CO₂ production within that layer, when modeling soil respiration in forest ecosystems.     

Spatial and temporal variability of soil CO2 emission in a sugarcane area under green and slash-and-burn managements

Soil management causes changes in physical, chemical, and biological properties that consequently affect soil CO₂ emission (FCO2). Here, we studied the soil carbon dynamics in areas with sugarcane production in southern Brazil under two different sugarcane management systems: green (G), consisting of mechanized harvesting that produces a large amount of crop residues left on the soil surface, and slash-and-burn (SB), in which the residues are burned before manual harvest, leaving no residues on the soil surface. The study was conducted during the period after harvest in two side-by-side grids installed in adjacent areas, having 60 points each. The aim was to characterize the temporal and spatial variability of FCO2, and its relation to soil temperature and soil moisture, in a red latosol (Oxisol) where G and SB management systems have been recently used. Mean FCO2 emission was 39% higher in the SB plot (2.87 μmol m⁻² s⁻¹) when compared to the G plot (2.06 μmol m⁻² s⁻¹) throughout the 70-day period after harvest. A quadratic equation of emissions versus soil moisture was able to explain 73% and 50% of temporal variability of FCO2 in SB and G, respectively. This seems to relate to the sensitivity of FCO2 to precipitation events, which caused a significant increase in SB emissions but not in G-managed area emissions. FCO2 semivariogram models were mostly exponential in both areas, ranging from 72.6 to 73.8 m and 63.0 to 64.7 m for G and SB, respectively. These results indicate that the G management system results in more homogeneous FCO2 when spatial and temporal variability are considered. The spatial variability analysis of soil temperature and soil moisture indicates that those parameters do not adequately explain the changes in spatial variability of FCO2, but emission maps are clearly more homogeneous after a drought period when no rain has occurred, in both sites.     

Does variability in litter quality determine soil microbial respiration in an Amazonian rainforest?

Tree species-rich tropical rainforests are characterized by a highly variable quality of leaf litter input to the soil at small spatial scales. This diverse plant litter is a major source of energy and nutrients for soil microorganisms, particularly in rainforests developed on old and nutrient-impoverished soils. Here we tested the hypothesis that the variability in leaf litter quality produced by a highly diverse tree community determines the spatial variability of the microbial respiration process in the underlying soil. We analyzed a total of 225 litter-soil pairs from an undisturbed Amazonian rainforest in French Guiana using a hierarchical sampling design. The microbial respiration process was assessed using substrate-induced respiration (SIR) and compared to a wide range of quality parameters of the associated litter layer (litter nutrients, carbon forms, stoichiometry, litter mass and pH). The results show that the variability of both litter quality and SIR rates was more important at large than at small scales. SIR rates varied between 1.1 and 4.0 μg g⁻¹ h⁻¹ and were significantly correlated with litter layer quality (up to 50% of the variability explained by the best mixed linear model). Total litter P content was the individual most important factor explaining the observed spatial variation in soil SIR, with higher rates associated to high litter P. SIR rates also correlated positively with total litter N content and with increasing proportions of labile C compounds. However, contrary to our expectation, SIR rates were not related to litter stoichiometry. These data suggest that in the studied Amazonian rainforest, tree canopy composition is an important driver of the microbial respiration process via leaf litter fall, resulting in potentially strong plant-soil feedbacks.     

Carbon dioxide fluxes in coastal Douglas-fir stands at different stages of development after clearcut harvesting

Forests play a significant role in the global carbon (C) cycle. Variability in weather, species, stand age, and current and past disturbances are some of the factors that control stand-level C dynamics. This study examines the relative roles of stand age and associated structural characteristics and weather variability on the exchange of carbon dioxide between the atmosphere and three different coastal Douglas-fir stands at different stages of development after clearcut harvesting. The eddy covariance technique was used to measure carbon dioxide fluxes and a portable soil chamber system was used to measure soil respiration in the three stands located within 50 km of each other on the east coast of Vancouver Island, British Columbia, Canada. In 2002, the recently clearcut harvested stand (HDF00) was a large C source, the pole/sapling aged stand (HDF88) was a moderate C source, and the rotation-aged stand (DF49) was a moderate C sink (net ecosystem production of −606, −133, and 254 g C m⁻² year⁻¹, respectively). Annual gross ecosystem production and ecosystem respiration also increased with increasing stand age. Differences in stand structural characteristics such as species composition and phenology were important in determining the timing and magnitude of maximum gross ecosystem production and net ecosystem production through the year. Both soil and ecosystem respiration were exponentially related to soil temperature in each stand with total ecosystem respiration differing more among stands than soil respiration. Between 1998 and 2003, annual net ecosystem production ranged from 254 to 424 g C m⁻² year⁻¹ over 6 years for DF49, from −623 to −564 g C m⁻² year⁻¹ over 3 years for HDF00, and from −154 to −133 g C m⁻² year⁻¹ over 2 years for HDF88. Interannual variations in C exchange of the oldest, most structurally stable stand (DF49) were related to variations in spring weather while the rapid growth of understory and pioneer species influenced variations in HDF00. The differences in net ecosystem production among stands (maximum of 1000 g C m⁻² year⁻¹ between the oldest and youngest stands) were an order of magnitude greater than the differences among years within a stand and emphasized the importance of age-related differences in stand structure on C exchange processes.     

Influence of temperature and drought on seasonal and interannual variations of soil, bole and ecosystem respiration in a boreal aspen stand

Continuous half-hourly measurements of soil (R_s) and bole respiration (R_b), as well as whole-ecosystem CO₂ exchange, were made with a non steady-state automated chamber system and with the eddy covariance (EC) technique, respectively, in a mature trembling aspen stand between January 2001 and December 2003. Our main objective was to investigate the influence of long-term variations of environmental and biological variables on component-specific and whole-ecosystem respiration (R_e) processes. During the study period, the stand was exposed to severe drought conditions that affected much of the western plains of North America. Over the 3 years, daily mean R_s varied from a minimum of 0.1 μmol m⁻² s⁻¹ during winter to a maximum of 9.2 μmol m⁻² s⁻¹ in mid-summer. Seasonal variations of R_s were highly correlated with variations of soil temperature (T_s) and water content (θ) in the surface soil layers. Both variables explained 96, 95 and 90% of the variance in daily mean R_s from 2001 to 2003. Aspen daily mean R_b varied from negligible during winter to a maximum of 2.5 μmol m⁻² bark s⁻¹ (2.2 μmol m⁻² ground s⁻¹) during the growing season. Maximum R_b occurred at the end of the aspen radial growth increment and leaf emergence period during each year. This was 2 months before the peak in bole temperature (T_b) in 2001 and 2003. Nonetheless, R_b was highly correlated with T_b and this variable explained 77, 87 and 62% of the variance in R_b in the respective years. Partitioning of R_b between its maintenance (R_bm) and growth (R_bg) components using the mature tissue method showed that daily mean R_bg occurred at the same time as aspen radial growth increment during each growing season. This method led, however, to systematic over- and underestimations of R_bm and R_bg, respectively, during each year. Annual totals of R_s, R_b and estimated foliage respiration (R_f) from hazelnut and aspen trees were, on average, 829, 159 and 202 g C m⁻² year⁻¹, respectively, over the 3 years. These totals corresponded to 70, 14 and 16%, respectively, of scaled-up respiration estimates of R_e from chamber measurements. Scaled R_e estimates were 25% higher (1190 g C m⁻² year⁻¹) than the annual totals of R_e obtained from EC (949 g C m⁻² year⁻¹). The independent effects of temperature and drought on annual totals of R_e and its components were difficult to separate because the two variables co-varied during the 3 years. However, recalculation of annual totals of R_s to remove the limitations imposed by low θ, suggests that drought played a more important role than temperature in explaining interannual variations of R_s and R_e.     

Spatial variability in soil heat flux at three Inner Mongolia steppe ecosystems

Closing the energy budget at flux measurement sites is problematic, even when the fetch extends over flat, homogeneous surfaces with low vegetation cover. We used the residual energy balance and ordinary least square (OLS) linear regression methods to quantify spatial variability in soil heat flux contributing to energy balance closure (EBC), by deploying a mobile energy system within the footprints of three Eddy-covariance towers located in the steppe of Inner Mongolia, China. The EBC at the study sites had a daily average residual of 8–19 W m⁻² with OLS slopes of 0.83–0.96. The EBC was better achieved at the wet site than at the dry site. The spatial variability in soil heat flux was 48 W m⁻² (13% of R_n) during the day and 15 W m⁻² (34%) at night, with an average of 29 W m⁻² (24%) across the three sites. A 9% OLS slope difference due to this variability was recorded from our eight plot measurements. A large amount of missing energy (110 W m⁻² at peak) could occur with decreasing OLS slope of 23% across the three grassland sites when soil heat flux is not taken into account. In particular, heat storage in the top soil layer not only influenced the magnitude of EBC, but also adjusted soil heat flux to match the ‘truth schedule’. Heat storage in the top soil layer comprised half of the soil heat flux when the heat flux plate was deployed at a depth of 30 mm. If this part of heat storage was neglected, the residual of EBC would increase as large as 60 W m⁻² with OLS slope decreasing 9%. Comparing them with the multiple-location soil heat flux measurements, the single-location measurements from near the Eddy-covariance towers obtained a slightly better EBC with the OLS slope increasing by 4%.     

Spatial patterns of soil biological and physical properties in a ridge tilled and a ploughed Luvisol

The present study was conducted to determine the spatial heterogeneity of bulk density, soil moisture, inorganic N, microbial biomass C, and microbial biomass N in the ridge tillage system of Turiel compared to conventional mouldboard ploughing on three sampling dates in May, July, and August. The soil sampling was carried out under vegetation representing the ridge in a high spatial resolution down the soil profile. Bulk density increased with depth and ranged from 1.3 g cm⁻³ at 10 cm depth to 1.6 g cm⁻³ at 35 cm in ploughed plots and from 1.0 g m⁻³ at 5 cm to 1.4 g m⁻³ at 35 cm in the ridges. In the ploughed plots, the contents of microbial biomass C and microbial biomass N remained roughly constant at 215 and 33 μg g⁻¹ soil, respectively, throughout the experimental period. The microbial biomass C/N ratio varied in a small range around 6.4. In the ridged plots, the contents of microbial biomass C and microbial biomass N were 5% and 6% higher compared to the ploughed plots. Highest microbial biomass C contents of roughly 300 μg g⁻¹ soil were always measured in the crowns in July. The lowest contents of microbial biomass C of 85–137 μg g⁻¹ soil were measured in the furrows. The ridges showed strong spatial heterogeneity in bulk density, soil water content, inorganic nitrogen and microbial biomass.     

Temporal change in spatial variability of soil respiration on a slope of Japanese cedar (Cryptomeria japonica D. Don) forest

Although information regarding the spatial variability of soil respiration is important for understanding carbon cycling and developing a suitable sampling design for estimating average soil respiration, it remains relatively understudied compared to temporal changes. In this study, soil respiration was measured at 35 locations by season on a slope of Japanese cedar forest in order to examine temporal changes in the spatial distribution of soil respiration. Spatial variability of soil respiration varied between seasons, with the highest coefficient variation in winter (42%) and lowest in summer (26%). Semivariogram analysis and kriged maps revealed different patterns of spatial distribution in each season. Factors affecting the spatial variability were relief index (autumn), soil hardness of the A layer (winter), soil hardness at 50 cm depth (spring) and the altitude and relief index (summer). Annual soil respiration (average: 39 mol m⁻² y⁻¹) varied from 26 mol m⁻² y⁻¹ to 55 mol m⁻² y⁻¹ between the 35 locations and was higher in the upper part of the slope and lower in the lower part. The average Q₁₀ value was 2.3, varying from 1.3 to 3.0 among the locations. These findings suggest that insufficient information on the spatial variability of soil respiration and imbalanced sampling could bias estimates of current and future carbon budgets.     

Enzymatic activities and microbial communities in an Antarctic dry valley soil: Responses to C and N supplementation

The soils of the Antarctic dry valleys are exposed to extremely dry and cold conditions. Nevertheless, they contain small communities of micro-organisms that contribute to the biogeochemical transformations of the bioelements, albeit at slow rates. We have determined the dehydrogenase, β-glucosidase, acid and alkaline phosphatase and arylsulphatase activities and the rates of respiration (CO₂ production) in laboratory assays of soils collected from a field experiment in an Antarctic dry valley. The objective of the field experiment was to test the responses of the soil microbial community to additions of C and N in simple (glucose and NH₄Cl) and complex forms (glycine and lacustrine detritus from the adjacent lake comprising principally cyanobacterial necromass). The soil samples were taken 3 years after the experimental treatments had been applied. In unamended soil, all enzyme activities and respiration were detected indicating that the enzymatic capacity to mineralize organic C, P and S compounds existed in the soil, despite the very low organic matter content. Relative to the control (unamended soil), respiration was significantly increased by all the experimental additions of C and N except the smallest NH₄Cl addition (1 mg N g⁻¹ soil) and the smallest detritus addition (1.5 mg C g⁻¹ soil and 0.13 mg N g⁻¹ soil). The activities of all enzymes except dehydrogenase were increased by C and combined large C (10 mg C g⁻¹ soil) and N additions, but either unchanged or diminished by addition of either N only or N (up to 10 mg N g⁻¹ soil) with only small C (1 mg C g⁻¹ soil) additions in the form of glucose and NH₄Cl. This suggests that in the presence of a large amount of N, the C supply for enzyme biosynthesis was limited. When normalized with respect to soil respiration, only arylsulphatase per unit of respiration showed a significant increase with C and N additions as glucose and NH₄Cl, consistent with S limitation when C and N limitations have been alleviated. Based on the positive responses of enzyme activity, detritus appeared to provide either conditions or resources which led to a larger biological response than a similar amount of C and more N added in the form of defined compounds (glucose, NH₄Cl or glycine). Assessment of the soil microbial community by ester-linked fatty acid (ELFA) analysis provided no evidence of changes in the community structure as a result of the C and N supplementation treatments. Thus the respiration and enzyme activity responses to supplementation occurred in an apparently structurally stable or unresponsive microbial community.     

Regional surface flux of CO2 inferred from changes in the advected CO2 column density

A Lagrangian experiment was conducted over Iowa during the daytime (9:00–17:30 LT) on June 19, 2007 as part of the North American Carbon Program's Mid-Continent Intensive using a light-weight and operationally flexible aircraft to measure a net drawdown of CO₂ concentration within the boundary layer. The drawdown can be related to net ecosystem exchange when anthropogenic emissions are estimated using a combination of the Vulcan fossil fuel emissions inventory coupled with a source contribution analysis using HYSPLIT. Results show a temporally and spatially averaged net CO₂ flux of −9.0 ± 2.4 μmol m⁻² s⁻¹ measured from the aircraft data. The average flux from anthropogenic emissions over the measurement area was 0.3 ± 0.1 μmol CO₂ m⁻² s⁻¹. Large-scale subsidence occurred during the experiment, entraining 1.0 ± 0.2 μmol CO₂ m⁻² s⁻¹ into the boundary layer. Thus, the CO₂ flux attributable to the vegetation and soils is −10.3 ± 2.4 μmol m⁻² s⁻¹. The magnitude of the calculated daytime biospheric flux is consistent with tower-based eddy covariance fluxes over corn and soybeans given existing land-use estimates for this agricultural region. Flux values are relatively insensitive to the choice of integration height above the boundary layer and emission footprint area. Flux uncertainties are relatively small compared to the biospheric fluxes, though the measurements were conducted at the height of the growing season.     

Interpreting the dependence of soil respiration on soil temperature and water content in a boreal aspen stand

Continuous half-hourly measurements of soil CO₂ efflux made between January and December 2001 in a mature trembling aspen stand located at the southern edge of the boreal forest in Canada were used to investigate the seasonal and diurnal dependence of soil respiration (R_s) on soil temperature (T_s) and water content (θ). Daily mean R_s varied from a minimum of 0.1 μmol m⁻² s⁻¹ in February to a maximum of 9.2 μmol m⁻² s⁻¹ in mid-July. Daily mean T_s at the 2-cm depth was the primary variable accounting for the temporal variation of R_s and no differences between Arrhenius and Q₁₀ response functions were found to describe the seasonal relationship. R_s at 10 °C (R_s10) and the temperature sensitivity of R_s (Q_10Rs) calculated at the seasonal time scale were 3.8 μmol m⁻² s⁻¹ and 3.8, respectively. Temperature normalization of daily mean R_s (R_sN) revealed that θ in the 0–15 cm soil layer was the secondary variable accounting for the temporal variation of R_s during the growing season. Daily R_sN showed two distinctive phases with respect to soil water field capacity in the 0–15 cm layer (θ_fc, 0.30 m³ m⁻³): (1) R_sN was strongly reduced when θ decreased below θ_fc, which reflected a reduction in microbial decomposition, and (2) R_sN slightly decreased when θ increased above θ_fc, which reflected a restriction of CO₂ or O₂ transport in the soil profile.Diurnal variations of half-hourly R_s were usually out of phase with T_s at the 2-cm depth, which resulted in strong diurnal hysteresis between the two variables. Daily nighttime R_s10 and Q_10Rs parameters calculated from half-hourly nighttime measurements of R_s and T_s at the 2-cm depth (when there was steady cooling of the soil) varied greatly during the growing season and ranged from 6.8 to 1.6 μmol m⁻² s⁻¹ and 5.5 to 1.3, respectively. On average, daily nighttime R_s10 (4.5 μmol m⁻² s⁻¹) and Q_10Rs (2.8) were higher and lower, respectively, than the values obtained from the seasonal relationship. Seasonal variations of these daily parameters were highly correlated with variations of θ in the 0–15 cm soil layer, with a tendency of low R_s10 and Q_10Rs values at low θ. Overall, the use of seasonal R_s10 and Q_10Rs parameters led to an overestimation of daily ranges of half-hourly R_s (ΔR_s) during drought conditions, which supported findings that the short-term temperature sensitivity of R_s was lower during periods of low θ. The use of daily nighttime R_s10 and Q_10Rs parameters greatly helped at simulating ΔR_s during these periods but did not improve the estimation of half-hourly R_s throughout the year as it could not account for the diurnal hysteresis effect.     

Modelling multi-year coupled carbon and water fluxes in a boreal aspen forest

In order to test two hypotheses: (i) that carbon (C) and energy exchanges between terrestrial ecosystems and the atmosphere are closely constrained by soil water availability, and (ii) that vegetation is able to optimize soil water uptake from different soil layers; two model simulations were conducted. The Boreal Ecosystem Productivity Simulator (BEPS) model was run to simulate an aspen forest in Saskatchewan, Canada during the period 1997–2004. In Simulation 1, the effect of soil water availability in different soil layers on stomatal conductance was weighted only by root fraction. In Simulation 2, the influence of soil water availability in different soil layers on stomatal conductance was weighted according to both the root fraction and soil water availability, in order to allow easier access of roots to soil layers containing more water.Comparison against measured fluxes showed that Simulation 2 was an improvement over Simulation 1 in predicting C, water and energy fluxes at different time scales in dry years. In Simulation 1, the daytime C and water fluxes were underestimated during the transition from adequate to insufficient soil water content in the upper layers. In this run, the model captured 92, 79 and 91% of the daily variances in gross primary productivity (GPP), net ecosystem productivity (NEP), and ecosystem respiration (R_e) during 1997–2004. In Simulation 2, the daily variances of GPP, NEP, and R_e explained by the model increased to 93, 82 and 92%, respectively. In Simulation 1, the annual NEP was considerably underestimated in the dry years and years with dry periods, with a root mean square error (RMSE) of 45 g C m⁻² year⁻¹ (n = 8) from 1997 to 2004. In Simulation 2, the RMSE value of simulated annual NEP was reduced to 14 g C m⁻² year⁻¹, a relatively small value compared with the average NEP of 157 g C m⁻² year⁻¹ during 1997–2004. This suggested that the ability of plant roots to extract water from deep soil layers is critical for the forest to maintain growth when surface layers dried out. Our model results showed that NEP was very sensitive to water conditions at this site. In wet years, heterotrophic respiration was enhanced and NEP was reduced.     

Above- and belowground ecosystem biomass and carbon pools in an age-sequence of temperate pine plantation forests

We assessed the successional development of above- and belowground ecosystem biomass and carbon (C) pools in an age-sequence of four White pine (Pinus strobus L.) plantation stands (2-, 15-, 30-, and 65-years-old) in Southern Ontario, Canada. Biomass and C stocks of above- and belowground live and dead tree biomass, understorey and forest ground vegetation, forest floor C (LFH-layer), and woody debris were determined from plot-level inventories and destructive tree sampling. Small root biomass (<5 mm) and mineral soil C stocks were estimated from soil cores. Aboveground tree biomass became the major ecosystem C pool with increasing age, reaching 0.5, 66, 92, and 176 t ha⁻¹ in the 2-, 15-, 30-, and 65-year-old stands, respectively. Tree root biomass increased from 0.1 to 10, 18, 38 t ha⁻¹ in the 2-, 15-, 30-, and 65-year-old stands, respectively, contributing considerably to the total ecosystem C in the three older stands. Forest floor C was 0.8, 7.5, 5.4, and 12.1 t C ha⁻¹ in the 2-, 15-, 30-, and 65-year-old stands, respectively, indicating an increase during the first two decades, but no further age-effect during the later growth phase. Mineral soil C was age-independent with 37.2, 33.9, 39.1, and 36.7 t C ha⁻¹ in the 2-, 15-, 30-, and 65-year-old stands, respectively. Aboveground ecosystem C increased with age from 3 to 40, 52, and 100 t C ha⁻¹ in the 2-, 15-, 30-, and 65-year-old stands, respectively, due to an increase in aboveground tree biomass. Belowground ecosystem C remained similiar in the early decades after establishment with 37, 39, and 39 t C ha⁻¹ in the 2-, 15-, and 30-year-old stands, but increased to 56 t C ha⁻¹ in the 65-year-old stand due to an increase in root biomass. The difference in total ecosystem C between the 2- and 65-year-old stand was 116 t C ha⁻¹. Our results highlight the importance of considering the successional development of forest ecosystem C pools, when estimating C sink potentials over their complete life cycle.     

Comparison of net ecosystem CO2 exchange in two peatlands in western Canada with contrasting dominant vegetation, Sphagnum and Carex

Net ecosystem carbon dioxide exchange was measured in two contrasting peatlands in northern Alberta, Canada using the eddy covariance technique during the growing season (May–October). Sphagnum spp. made up approximately 66% of the total LAI (1.52 m² m⁻²) at the poor fen and the total N content of Sphagnum capitula was 7.8 mg g⁻¹ at the peak of the growing season. In contrast, the dominant plant species at the extreme-rich fen site, the perennial sedge, Carex lasiocarpa, accounted for approximately 60% of the total LAI (1.09 m² m⁻²), and had leaf total N content of 19.3 mg g⁻¹ at peak biomass. In addition, the peak aboveground biomass was higher at the poor fen (230.9 g m⁻²) than at the extreme-rich fen (157.1 g m⁻²). Both sites had maximum daily rates of net CO₂ uptake of approximately 5 μmol m⁻² s⁻¹, and typical nighttime rates of CO₂ loss of approximately 2 μmol m⁻² s⁻¹ during the peak of the growing season. Calculations of maximum photosynthetic and respiratory capacity were consistently higher at the extreme-rich fen. The poor fen was a net sink for CO₂ during 4 of the 6 months (peaking at 44 g C m⁻² in July), while only slight net losses of CO₂ (3 g C m⁻²) occurred in May and September. In contrast, the extreme-rich fen was calculated to be a significant net sink for CO₂ only during 2 months of the growing season (peaking at 30 g C m⁻² in August), while significant net losses of CO₂ occurred in May (8 g C m⁻²) and in October (13 g C m⁻²). The plant species at the poor fen site were active earlier and later in the growing season, while it took longer for C. lasiocarpa to develop leaf tissue, and leaf senescence and reduction in photosynthetic activity occurred earlier in the fall at the extreme-rich fen. When integrated over the 6-month growing season, the poor fen was a net sink (90 g C m⁻²) that was three times larger than the extreme-rich fen (31 g C m⁻²). The ratio of cumulative total ecosystem respiration to gross primary production was 0.7 at the poor fen and 0.9 at the extreme-rich fen.     

Influence of different land uses on soil nitrogen transformations after conversion from an Indian dry tropical forest

The effects of alternate land uses, such as grassland, cropland and mine spoil on mineral nitrogen (N), N-transformation rate and microbial biomass N (MBN) in dry tropical forest soils of India were studied. The mean annual mineral N in the forest, grassland, cropland and mine spoil ecosystems, respectively ranged from 15.24 to 19.58, 17.8 to 18.56, 16.49 to 19.85 and 10.52 to 13.44 µg g^− 1, net nitrification rate from 14.15 to 23.4, 10.11 to 11.38, 8.07 to 9.16, 10.52 to 13.44 µg g^− 1mo^− 1; net N-mineralization rate from 17.38 to 26.36, 13.99 to 15.41, 10.99 to 12.5, 5.43 to 7.68 µg g^− 1mo^− 1and and microbial biomass N from 41.25 to 58.87, 34.47 to 47.95, 27.88 to 30.43 and 22.95 to 25.26 µg g^− 1, respectively. The values were within the range reported by previous studies in different tropical environments. The mean annual net nitrification rates declined after conversion into grassland, cropland and mine spoil by 43, 54 and 78%, respectively, net N mineralization by 33, 46 and 70%, and microbial biomass N by 29%, 42% and 52%, respectively.The MBN was positively related to root biomass and total plant biomass, while microbial-N and inorganic N are reciprocally, while nitrification and N-mineralization are directly related to seasonal soil moisture and temperature. The microbial biomass N, nitrification and N-mineralization are negatively related to smaller fraction (< 0.1 mm) of the soil. Above- and below-ground biomass also have had their impact on microbial biomass N, and thereby N-mineralization. Thus, in dry tropical forests, land-use change affects remarkably the nitrogen transformation process in soil.     

Incorporation of spectral data into multivariate geostatistical models to map soil phosphorus variability in a Florida wetland

Hybrid geostatistical prediction methods incorporate (i) spatially-explicit soil observations and exhaustive grids of ancillary environmental variables (e.g. derived from remote sensing), (ii) spatial autocorrelation, (iii) spatial covariation, and/or (iv) combinations of the above. In numerous studies of terrestrial soils it has been shown that hybrid geostatistical methods outperform univariate spatial and regression (aspatial) methods. However, hybrid methods have rarely been employed to predict soil properties in wetlands. In this study we used spectral data and derived indices from two remote sensors (Landsat ETM+ and ASTER), with different spatial resolutions, from different seasons, but with similar spectral range, ancillary environmental data, as well as floc and soil total phosphorus (TP) observations from 111 sites. The specific objective of our study was to evaluate the performance of aspatial methods (multivariate regressions — REG), univariate spatial (Ordinary Kriging — OK) and hybrid/multivariate geostatistical methods (Regression Kriging — RK and Co-kriging — CK) in predicting the spatial variability and distribution of floc and soil TP in a subtropical wetland, WCA-2A, in the Florida Everglades. Measured floc TP ranged from 194 to 1865 mg kg^− 1 with a median of 751 mg kg^− 1 and standard deviation (SD) of 381 mg kg^− 1. According to cross-validation, predictions for floc TP based on the root mean square prediction error (RMSE) were best in the following order: RK_quadratic (134.9) > RK_multivariate (201.1) > OK (206.1) > CK (212.1) > REG_multivariate (218.3) > REG_quadratic (220.3) > REG_linear (264.4); and based on the mean prediction error (ME) followed the order RK_multivariate (0.9)  RK_quadratic (1.1) > CK (− 6.7) > REG_multivariate (18.2) > REG_linear (25.1) > OK (− 27.3) > REG_quadratic (27.3). The Normalized Difference Vegetation Index (NDVI)-green derived from Landsat ETM+ showed the largest predictive power for floc TP. Measured soil TP ranged from 155 to 1702 mg kg^− 1 with a median of 433 mg kg^− 1 and standard deviation of 316 mg kg^− 1. Predictions for soil TP based on RMSE were best in the following order: RK_ASTER (200.1) > CK_ASTER (238.2)  CK_ETM (239.0) > OK (258.0) > RK_ETM (279.2) > REG_ASTER (281.8) > REG_ETM (356.1); and based on ME followed the order: CK_ASTER (0.1)  CK_ETM (0.2) > RK_ASTER (− 5.2) > RK_ETM (− 31.5) > OK (− 41.8) > REG_ASTER (94.4) > REG_ETM (133.7). The NDVI showed the largest predictive power for soil TP. This comparative study in a subtropical wetland demonstrated the benefits of incorporating remote sensing data into floc and soil TP prediction models. Overall, hybrid geostatistical methods (CK and RK) performed better than regressions and spatial univariate models (OK) in the prediction of floc and soil TP. Depending on the strength of the spatial covariance between primary and secondary variables (CK) and the ability of the regression model in RK to explain the variability of a target variable (e.g., floc or soil TP), either CK or RK performed best. Our findings in this wetland confirmed results from earlier studies on terrestrial soils indicating the superior performance of hybrid geostatistical methods in predicting soil properties.     

Soil microbial community as bioindicator of the recovery of soil functioning derived from metal phytoextraction with sorghum

A three-month microcosm study was carried out in order to evaluate: (i) the capacity of sorghum plants to phytoextract Cd (50 mg kg⁻¹) and Zn (1000 mg kg⁻¹) from artificially polluted soil and (ii) the possibility of biomonitoring the efficiency of phytoremediation using parameters related to the size, activity and functional diversity of the soil microbial community. Apart from plant and soil (total and bioavailable) metal concentrations, the following parameters were determined: soil physicochemical properties (pH, OM content, electrical conductivity, total N, and extractable P and K), dehydrogenase activity, basal- and substrate-induced respiration (with glucose and a model rhizodeposit solution, both adjusted to 800 mg C kg⁻¹ DW soil and 45.2 mg N kg⁻¹ DW soil), microbial respiration quotient, functional diversity through community level physiological profiles and, finally, seed germination toxicity tests with Lepidium sativum. Sorghum plants were highly tolerant to metal pollution and capable of reaching high biomass values in the presence of metals. In the first two harvests, values of shoot Cd concentrations were higher than 100 mg Cd kg⁻¹ DW, the threshold value for hyperaccumulators. Nonetheless, in the third harvest, the bioconcentration factor was 1.34 and 0.35 for Cd and Zn, respectively, well below the threshold value of 10 considered for a phytoextraction process to be feasible. In general, microbial parameters showed lower values in metal polluted than in control non-polluted soils, and higher values in planted than in control unplanted pots. As a result of the phytoextraction process, which includes both plant growth and metal phytoextraction, the functioning of the phytoremediated soil, as reflected by the values of the different microbial parameters here determined, was restored. Most importantly, although the phytoextracted soil recovered its function, it was still more phytotoxic than the control non-polluted soil.     

Carbon dioxide and energy flux partitioning between the understorey and the overstorey of a maritime pine forest during a year with reduced soil water availability

Carbon dioxide, water vapour and energy fluxes were measured above and within a maritime pine forest during an atypical year with long-lasting reduced soil water availability. Energy balance closure was adequately good at both levels. As compared with what is usually observed at this site the ecosystem dissipated less energy via latent heat flux and more via sensible heat flux. The understorey canopy was responsible for a variable, significant component of the whole canopy fluxes of water vapour and carbon dioxide. The annual contribution of the understorey was 38% (154 mm) of the overall evaporation (399 mm) and 32% (89 mm) of the overall sensible heat flux (274 mm). The participation of the understorey reached 45% of the overall evaporation and 30% of the daytime overall assimilation during significant soil water deficit periods in summertime. Even during winter, understorey photosynthesis was consistent as it compensated soil and understorey respiration. The ecosystem behaved as a sink of carbon, with a negative annual carbon budget (−57 g C m⁻²). However, due to high soil water deficit, the annual ecosystem GPP was 40% less than usually observed at this site. This budget resulted from a sink of −131 g C m⁻² for the overstorey and a source of +74 g C m⁻² for the understorey. Moreover, on an annual basis the overstorey layer contributed to almost two-thirds of the ecosystem respiration. Finally, the effect of long-lasting soil water deficit on the maritime pine forest was found more important than the effect of the heat wave and drought of summer 2003.