Winter CO2 CH4 and N2O fluxes on some natural and drained boreal peatlands

CO₂ and CH₄ fluxes during the winter were measured at natural and drained bog and fen sites in eastern Finland using both the closed chamber method and calculations of gas diffusion along a concentration gradient through the snowpack. The snow diffusion results were compared with those obtained by chamber, but the winter flux estimates were derived from chamber data only. CH₄ emissions from a poor bog were lower than those from an oligotrophic fen, while both CO₂ and CH₄ fluxes were higher in theCarex rostrata- occupied marginal (lagg) area of the fen than in the slightly less fertile centre. Average estimated winter CO₂-C losses from virgin and drained forested peatlands were 41 and 68 g CO₂-C m^–2, respectively, accounting for 23 and 21% of the annual total CO₂ release from the peat. The mean release of CH₄-C was 1.0 g in natural bogs and 3.4 g m^–2 in fens, giving rise to winter emissions averaging to 22% of the annual emission from the bogs and 10% of that from the fens. These wintertime carbon gas losses in Finnish natural peatlands were even greater than reported average long-term annual C accumulation values (less than 25g C m^–2). The narrow range of 10–30% of the proportion of winter CO₂ and CH₄ emissions from annual emissions found in Finnish peatlands suggest that a wider generalization in the boreal zone is possible. Drained forested bogs emitted 0.3 g CH₄-C m^–2 on the average, while the effectively drained fens consumed an average of 0.01 g CH₄-C m^–2. Reason for the low CH₄. efflux or net oxidation in drained peatlands probably lies in low substrate supply and thus low CH₄ production in the anoxic deep peat layers. N₂O release from a fertilized grassland site in November–May was 0.7 g N₂O m^–2, accounting for 38% of the total annual emission, while a forested bog released none and two efficiently drained forested fens 0.09 (28% of annual release) and 0.04 g N₂O m^–2 (27%) during the winter, respectively.     

Diurnal and seasonal variation of carbon dioxide exchange from a former true raised bog

Carbon dioxide exchange was measured, using the eddy covariance technique, during a one and a half year period in 1994 and 1995. The measurements took place over a former true raised bog, characterized by a shallow peat layer and a vegetation dominated by Molinia caerulea. The growing season extended from May until late October, with a maximum LAI in August of 1.7. The carbon balance shows a net release of 97 g C m^–2 y^–1 (265 kg C ha^–1 y^–1) from the peat bog ecosystem to the atmosphere. During June, July and August there is net consumption of CO₂, while during the rest of the year there is net production of CO₂. The average daytime assimilation rates ranged between – 0.2 and – 0.5 mg CO₂ m^–2 s^–1 (– 45 and –11.3 μmol CO₂ m^–2 s^–1), in a period where the LAI ranged between 1 and 1.7. A high vapour pressure deficit (> 15 hPa) corresponding with high temperatures was found to reduce the assimilation rate by on average 50%. Apart from these factors, LAI and the soil temperature codetermine the net exchange of CO₂. The total nocturnal respiration during the growing season lies within the same order as the average daytime net assimilation rate. Temperature was found to be the main factor controlling soil respiration, with a Q₁₀ of 4.8.     

Reconstruction of the carbon balance for microsites in a boreal oligotrophic pine fen, Finland

 Carbon dioxide (CO₂) exchange was studied at flark (minerotrophic hollow), lawn and hummock microsites in an oligotrophic boreal pine fen. Statistical response functions were constructed for the microsites in order to reconstruct the annual CO₂ exchange balance from climate data. Carbon accumulation was estimated from the annual net CO₂ exchange, methane (CH₄) emissions and leaching of carbon. Due to high water tables in the year 1993, the average carbon accumulation at the flark, Eriophorum lawn, Carex lawn and hummock microsites was high, 2.91, 6.08, 2.83 and 2.66 mol C m^–2, respectively, and for the whole peatland it was 5.66 mol m^–2 year^–1. During the maximum primary production period in midsummer, hummocks with low water tables emitted less methane than predicted from the average net ecosystem exchange (NEE), while the Carex lawns emitted slightly more. CH₄ release during that period corresponded to 16% of the contemporary NEE. Annual C accumulation rate did not correlate with annual CH₄ release in the microsites studied, but the total community CO₂ release seemed to be related to CH₄ emissions in the wet microsites, again excluding the hummocks. The dependence of CO₂ exchange dynamics on weather events suggests that daily balances in C accumulation are labile and can change from net carbon uptake to net release, primarily in high hummocks on fens under warmer, drier climatic conditions. Received: 16 August 1996 / Accepted: 30 November 1996     

Carbon storage and fluxes in ponderosa pine forests at different developmental stages

We compared carbon storage and fluxes in young and old ponderosa pine stands in Oregon, including plant and soil storage, net primary productivity, respiration fluxes, eddy flux estimates of net ecosystem exchange (NEE), and Biome‐BGC simulations of fluxes. The young forest (Y site) was previously an old‐growth ponderosa pine forest that had been clearcut in 1978, and the old forest (O site), which has never been logged, consists of two primary age classes (50 and 250 years old). Total ecosystem carbon content (vegetation, detritus and soil) of the O forest was about twice that of the Y site (21 vs. 10 kg C m^?2 ground), and significantly more of the total is stored in living vegetation at the O site (61% vs. 15%). Ecosystem respiration (R_e) was higher at the O site (1014 vs. 835 g C m^?2 year^?1), and it was largely from soils at both sites (77% of R_e). The biological data show that above‐ground net primary productivity (ANPP), NPP and net ecosystem production (NEP) were greater at the O site than the Y site. Monte Carlo estimates of NEP show that the young site is a source of CO₂ to the atmosphere, and is significantly lower than NEP(O) by c. 100 g C m^?2 year^?1. Eddy covariance measurements also show that the O site was a stronger sink for CO₂ than the Y site. Across a 15‐km swath in the region, ANPP ranged from 76 g C m^?2 year^?1 at the Y site to 236 g C m^?2 year^?1 (overall mean 158 ± 14 g C m^?2 year^?1). The lowest ANPP values were for the youngest and oldest stands, but there was a large range of ANPP for mature stands. Carbon, water and nitrogen cycle simulations with the Biome‐BGC model suggest that disturbance type and frequency, time since disturbance, age‐dependent changes in below‐ground allocation, and increasing atmospheric concentration of CO₂ all exert significant control on the net ecosystem exchange of carbon at the two sites. Model estimates of major carbon flux components agree with budget‐based observations to within ± 20%, with larger differences for NEP and for several storage terms. Simulations showed the period of regrowth required to replace carbon lost during and after a stand‐replacing fire (O) or a clearcut (Y) to be between 50 and 100 years. In both cases, simulations showed a shift from net carbon source to net sink (on an annual basis) 10–20 years after disturbance. These results suggest that the net ecosystem production of young stands may be low because heterotrophic respiration, particularly from soils, is higher than the NPP of the regrowth. The amount of carbon stored in long‐term pools (biomass and soils) in addition to short‐term fluxes has important implications for management of forests in the Pacific North‐west for carbon sequestration.     

Spring photosynthesis in a cool temperate bog

In northern ecosystems, the onset and growth of spring photosynthesis may have an important influence on the annual carbon (C) budget, yet the controls have not been clearly identified, especially for peatlands. We used a 5‐year set of daily carbon dioxide (CO₂) exchange measurements derived from an eddy covariance tower located at Mer Bleue, an ombrotrophic bog near Ottawa, Canada, from March to May [day‐of‐year (DOY) 60–150], 1999–2003. We used half‐hourly measured net ecosystem exchange minus modelled ecosystem respiration to estimate daily photosynthesis, as gross ecosystem production (GEP). The onset of GEP in each year was closely related to the thinning and disappearance of the snow cover, occurring between DOY 86 and 101. GEP increased during the spring, reaching 10‐day average values of between 5 and 9 g CO₂ m^?2 day^?1 by the end of May. This increase was initially associated with moss activity (Sphagnum and Polytrichum), followed by the evergreen shrubs. Peat temperatures in the rooting zone (10–20 cm depth) and increases in shrub leaf nitrogen and chlorophyll a concentrations contributed to this rapid increase in GEP. Examination of moderate‐resolution imaging spectroradiometer (MODIS) images over several years revealed that the temporal resolution (16‐day composites) was inadequate to capture the onset of GEP but estimates of gross primary productivity and photosynthesis from MODIS 8‐day composites for the most part followed the pattern and magnitude of CO₂ exchange observed at the tower.     

CO2 exchange in an organic field growing barley or grass in eastern Finland

The CO₂ dynamics were measured in an organic soil in eastern Finland during the growing season and wintertime, and the annual CO₂ balance was calculated for plots where barley or grass was grown. During the summer, the CO₂ dynamics were measured by transparent and opaque chambers using a portable infrared gas analyser for the CO₂ analyses. During the winter, the CO₂ release was measured by opaque chambers analysing the samples in the laboratory with a gas chromatograph. Statistical response functions for CO₂ dynamics were constructed to evaluate the annual CO₂ exchange from the climatic data. The net CO₂ exchange was calculated for every hour in the snow‐free season. The carbon balance varied extensively depending on the weather conditions, and type and phenology of vegetation. During the growing season, the grassland was a net source while the barley field was a net sink for CO₂. However, both soils were net sources for CO₂ when autumn, winter and spring were included also. The annual CO₂ emissions from the grassland and barley soil were 750 g CO₂‐C m^?2 and 400 g CO₂‐C m^?2, respectively. The carbon accumulated in root and shoot biomass during the growing season was 330 g m^?2 for grass and 520 g m^?2 for barley. The C in the aboveground plant biomass ranged from 43 to 47% of the carbon fixed in photosynthesis (P_G) and the proportion of C in the root biomass was 10% of the carbon fixed in photosynthesis. The bare soils had 10–60% higher net CO₂ emission than the vegetated soils. These results indicate that the carbon balance of organic soils is affected by the characteristics of the prevailing plant cover. The dry summer of 1997 may have limited the growth of grass in the late summer thus reducing photosynthesis, which could be one reason for the high CO₂ release from this grass field.     

Seasonal Net Ecosystem Carbon Exchange of a Regenerating Cutaway Bog: How Long Does it Take to Restore the C‐Sequestration Function?

We measured the net ecosystem exchange (NEE) and respiration rates and modeled the photosynthesis and respiration dynamics in a cutover bog in the Swiss Jura Mountains during one growing season at three stages of regeneration (29, 42, and 51 years after peat cutting; coded sites A, B, and C) to determine if reestablishment of Sphagnum suffices to restore the C‐sequestration function. From the younger to the older stage Sphagnum cover increased, while net primary Sphagnum production over the growing season decreased (139, 82, and, 67 g m^?2 y^?1 for A, B, and C respectively), and fen plant species were replaced by bog species. According to our NEE estimations, over the vegetation period site A was a net CO₂‐C source emitting 40 g CO₂‐C/m² while sites B and C were accumulating CO₂‐C, on average 222 and 209 g CO₂‐C/m², respectively. These differences are due to the higher respiration in site A during the summer, suggesting that early regeneration stages may be more sensitive to a warmer climate. Methane fluxes increased from site A to C in parallel with Eriophorum vaginatum cover and vascular plant leaf area. Our results show that reestablishing a Sphagnum cover is not sufficient to restore a CO₂‐sequestrating function but that after circa 50 years the ecosystem may naturally regain this function over the growing season.     

The contribution of bryophytes to the carbon exchange for a temperate rainforest

Bryophytes blanket the floor of temperate rainforests in New Zealand and may influence a number of important ecosystem processes, including carbon cycling. Their contribution to forest floor carbon exchange was determined in a mature, undisturbed podocarp‐broadleaved forest in New Zealand, dominated by 100–400‐year‐old rimu (Dacrydium cupressimum) trees. Eight species of mosses and 13 species of liverworts contributed to the 62% cover of the diverse forest floor community. The bryophyte community developed a relatively thin (depth <30 mm), but dense, canopy that experienced elevated CO₂ partial pressures (median 46.6 Pa immediately below the bryophyte canopy) relative to the surrounding air (median 37.6 Pa at 100 mm above the canopy). Light‐saturated rates of net CO₂ exchange from 14 microcosms collected from the forest floor were highly variable; the maximum rate of net uptake (bryophyte photosynthesis – whole‐plant respiration) per unit ground area at saturating irradiance was 1.9 μmol m^?2 s^?1 and in one microcosm, the net rate of CO₂ exchange was negative (respiration). CO₂ exchange for all microcosms was strongly dependent on water content. The average water content in the microcosms ranged from 1375% when fully saturated to 250% when air‐dried. Reduction in water content across this range resulted in an average decrease of 85% in net CO₂ uptake per unit ground area. The results from the microcosms were used in a model to estimate annual carbon exchange for the forest floor. This model incorporated hourly variability in average irradiance reaching the forest floor, water content of the bryophyte layer, and air and soil temperature. The annual net carbon uptake by forest floor bryophytes was 103 g m^?2, compared to annual carbon efflux from the forest floor (bryophyte and soil respiration) of ?1010 g m^?2. To put this in perspective of the magnitude of the components of CO₂ exchange for the forest floor, the bryophyte layer reclaimed an amount of CO₂ equivalent to only about 10% of forest floor respiration (bryophyte plus soil) or ～11% of soil respiration. The contribution of forest floor bryophytes to productivity in this temperate rainforest was much smaller than in boreal forests, possibly because of differences in species composition and environmental limitations to photosynthesis. Because of their close dependence on water table depth, the contribution of the bryophyte community to ecosystem CO₂ exchange may be highly responsive to rapid changes in climate.     

Carbon exchange of grazed pasture on a drained peat soil

Land‐use changes have contributed to increased atmospheric CO₂ concentrations. Conversion from natural peatlands to agricultural land has led to widespread subsidence of the peat surface caused by soil compaction and mineralization. To study the net ecosystem exchange of carbon (C) and the contribution of respiration to peat subsidence, eddy covariance measurements were made over pasture on a well‐developed, drained peat soil from 22 May 2002 to 21 May 2003. The depth to the water table fluctuated between 0.02 m in winter 2002 to 0.75 m during late summer and early autumn 2003. Peat soil moisture content varied between 0.6 and 0.7 m³ m^?3 until the water table dropped below 0.5 m, when moisture content reached 0.38 m³ m^?3. Neither depth to water table nor soil moisture was found to have an effect on the rate of night‐time respiration (ranging from 0.4–8.0 μmol CO₂ m^?2 s^?1 in winter and summer, respectively). Most of the variance in night‐time respiration was explained by changes in the 0.1 m soil temperature (r²=0.93). The highest values for daytime net ecosystem exchange were measured in September 2002, with a maximum of ?17.2 μmol CO₂ m^?2 s^?1. Grazing events and soil moisture deficiencies during a short period in summer reduced net CO₂ exchange. To establish an annual C balance for this ecosystem, non‐linear regression was used to model missing data. Annually integrated (CO₂) C exchange for this peat–pasture ecosystem was 45±500 kg C ha^?1 yr^?1. After including other C exchanges (methane emissions from cows and production of milk), the net annual C loss was 1061±500 kg C ha^?1 yr^?1.     

An initial intercomparison of micrometeorological and ecological inventory estimates of carbon exchange in a mid-latitude deciduous forest

The role of mid‐latitude forests in the sequestration of carbon (C) is of interest to an increasing number of scientists and policy‐makers alike. Net CO₂ exchange can be estimated on an annual basis, using eddy‐covariance techniques or from ecological inventories of C fluxes to and from a forest. Here we present an intercomparison of annual estimates of C exchange in a mixed hardwood forest in the Morgan‐Monroe State Forest, Indiana, USA for two years, 1998 and 1999. Based on eddy‐covariance measurements made at 1.8 times canopy height from a tower, C uptake by the forest was 237 and 287 g C m^?2 y^?1 for 1998 and 1999, respectively. For the same time period, biometric and ecophysiological measures and modelled estimates of all significant carbon fluxes within deciduous forests were made, including: change in living biomass, aboveground and belowground detritus production, foliage consumption, and forest floor and soil respiration. Using this ecological inventory method for these same two time periods, C uptake was estimated to be 271 and 377 g C m^?2 y^?1, which are 14.3% and 31.4% larger, respectively, than the tower‐based values. The relative change between this method's annual estimates is consistent with that of the eddy‐covariance based values. Our results indicate that the difference in annual C exchange rates was due to reduced heterotrophic soil respiration in 1999.     

Contemporary carbon accumulation in a boreal oligotrophic minerogenic mire – a significant sink after accounting for all C-fluxes

Based on theories of mire development and responses to a changing climate, the current role of mires as a net carbon sink has been questioned. A rigorous evaluation of the current net C-exchange in mires requires measurements of all relevant fluxes. Estimates of annual total carbon budgets in mires are still very limited. Here, we present a full carbon budget over 2 years for a boreal minerogenic oligotrophic mire in northern Sweden (64°11′N, 19°33′E). Data on the following fluxes were collected: land–atmosphere CO₂ exchange (continuous Eddy covariance measurements) and CH₄ exchange (static chambers during the snow free period); TOC (total organic carbon) in precipitation; loss of TOC, dissolved inorganic carbon (DIC) and CH₄ through stream water runoff (continuous discharge measurements and regular C-concentration measurements). The mire constituted a net sink of 27±3.4 (±SD) g C m⁻² yr⁻¹ during 2004 and 20±3.4 g C m⁻² yr⁻¹ during 2005. This could be partitioned into an annual surface–atmosphere CO₂ net uptake of 55±1.9 g C m⁻² yr⁻¹ during 2004 and 48±1.6 g C m⁻² yr⁻¹ during 2005. The annual NEE was further separated into a net uptake season, with an uptake of 92 g C m⁻² yr⁻¹ during 2004 and 86 g C m⁻² yr⁻¹ during 2005, and a net loss season with a loss of 37 g C m⁻² yr⁻¹ during 2004 and 38 g C m⁻² yr⁻¹ during 2005. Of the annual net CO₂-C uptake, 37% and 31% was lost through runoff (with runoff TOC>DIC≫CH₄) and 16% and 29% through methane emission during 2004 and 2005, respectively. This mire is still a significant C-sink, with carbon accumulation rates comparable to the long-term Holocene C-accumulation, and higher than the C-accumulation during the late Holocene in the region.     

Modelling night‐time ecosystem respiration by a constrained source optimization method

One of the main challenges to quantifying ecosystem carbon budgets is properly quantifying the magnitude of night‐time ecosystem respiration. Inverse Lagrangian dispersion analysis provides a promising approach to addressing such a problem when measured mean CO₂ concentration profiles and nocturnal velocity statistics are available. An inverse method, termed ‘Constrained Source Optimization’ or CSO, which couples a localized near‐field theory (LNF) of turbulent dispersion to respiratory sources, is developed to estimate seasonal and annual components of ecosystem respiration. A key advantage to the proposed method is that the effects of variable leaf area density on flow statistics are explicitly resolved via higher‐order closure principles. In CSO, the source distribution was computed after optimizing key physiological parameters to recover the measured mean concentration profile in a least‐square fashion. The proposed method was field‐tested using 1 year of 30‐min mean CO₂ concentration and CO₂ flux measurements collected within a 17‐year‐old (in 1999) even‐aged loblolly pine (Pinus taeda L.) stand in central North Carolina. Eddy‐covariance flux measurements conditioned on large friction velocity, leaf‐level porometry and forest‐floor respiration chamber measurements were used to assess the performance of the CSO model. The CSO approach produced reasonable estimates of ecosystem respiration, which permits estimation of ecosystem gross primary production when combined with daytime net ecosystem exchange (NEE) measurements. We employed the CSO approach in modelling annual respiration of above‐ground plant components (c. 214 g C m^?2 year^?1) and forest floor (c. 989 g C m^?2 year^?1) for estimating gross primary production (c. 1800 g C m^?2 year^?1) with a NEE of c. 605 g C m^?2 year^?1 for this pine forest ecosystem. We conclude that the CSO approach can utilise routine CO₂ concentration profile measurements to corroborate forest carbon balance estimates from eddy‐covariance NEE and chamber‐based component flux measurements.     

Rain events decrease boreal peatland net CO2 uptake through reduced light availability

Boreal peatlands store large amounts of carbon, reflecting their important role in the global carbon cycle. The short‐term exchange and the long‐term storage of atmospheric carbon dioxide (CO₂) in these ecosystems are closely associated with the permanently wet surface conditions and are susceptible to drought. Especially, the single most important peat forming plant genus, Sphagnum, depends heavily on surface wetness for its primary production. Changes in rainfall patterns are expected to affect surface wetness, but how this transient rewetting affects net ecosystem exchange of CO₂ (NEE) remains unknown. This study explores how the timing and characteristics of rain events during photosynthetic active periods, that is daytime, affect peatland NEE and whether rain event associated changes in environmental conditions modify this response (e.g. water table, radiation, vapour pressure deficit, temperature). We analysed an 11‐year time series of half‐hourly eddy covariance and meteorological measurements from Degerö Stormyr, a boreal peatland in northern Sweden. Our results show that daytime rain events systematically decreased the sink strength of peatlands for atmospheric CO₂. The decrease was best explained by rain associated reduction in light, rather than by rain characteristics or drought length. An average daytime growing season rain event reduced net ecosystem CO₂ uptake by 0.23–0.54 gC m^?2. On an annual basis, this reduction of net CO₂ uptake corresponds to 24% of the annual net CO₂ uptake (NEE) of the study site, equivalent to a 4.4% reduction of gross primary production (GPP) during the growing season. We conclude that reduced light availability associated with rain events is more important in explaining the NEE response to rain events than rain characteristics and changes in water availability. This suggests that peatland CO₂ uptake is highly sensitive to changes in cloud cover formation and to altered rainfall regimes, a process hitherto largely ignored.     

Effect of the replacement of a native savanna by an African Brachiaria decumbens pasture on the CO2 exchange in the Orinoco lowlands,Venezuela

In the Orinoco lowlands, savannas have been often replaced by pastures composed of the C₄ grass, Brachiaria decumbens Stapf. We addressed following questions: (1) How does the replacement of the native vegetation affect CO₂ exchange on seasonal and annual scales? (2) How do biophysical constraints change when the landscape is transformed? To assess how these changes affect carbon exchange, we determined simultaneously the CO₂ fluxes by eddy covariance, and the soil CO₂ efflux by a chamber-based system in B. decumbens and herbaceous savanna stands. Measurements covered a one-year period from the beginning of the dry season (November 2008) to the end of the wet season (November 2009). During the wet season, the net ecosystem CO₂ exchange reached maximum values of 23 and 10 μmol(CO₂) m^?2 s^?1 in the B. decumbens field and in the herbaceous savanna stand, respectively. The soil CO₂ efflux for both stands followed a temperature variation during the dry and wet seasons, when the soil water content (SWC) increased above 0.087 m³ m^?3 in the latter case. Bursts of CO₂ emissions were evident when the dry soil experienced rehydration. The carbon source/sink dynamics over the two canopies differed markedly. Annual measurements of the net ecosystem production indicated that the B. decumbens field constituted a strong carbon sink of 216 g(C) m^?2 y^?1. By contrast, the herbaceous savanna stand was found to be only a weak sink [36 g(C) m^?2 y^?1]. About 53% of the gross primary production was lost as the ecosystem respiration. Carbon uptake was limited by SWC in the herbaceous savanna stand as evident from the pattern of water-use efficiency (WUE). At the B. decumbens stand, WUE was relatively insensitive to SWC. Although these results were specific to the studied site, the effect of land use changes and the physiological response of the studied stands might be applicable to other savannas.     

Long-term carbon exchange in a sparse, seasonally dry tussock grassland

Rainfall and its seasonal distribution can alter carbon dioxide (CO₂) exchange and the sustainability of grassland ecosystems. Using eddy covariance, CO₂ exchange between the atmosphere and a sparse grassland was measured for 2 years at Twizel, New Zealand. The years had contrasting distributions of rain and falls (446 mm followed by 933 mm; long‐term mean=646 mm). The vegetation was sparse with total above‐ground biomass of only 1410 g m^?2. During the dry year, leaf area index peaked in spring (November) at 0.7, but it was <0.2 by early summer. The maximum daily net CO₂ uptake rate was only 1.5 g C m^?2 day^?1, and it occurred before mid‐summer in both years. On an annual basis, for the dry year, 9 g C m^?2 was lost to the atmosphere. During the wet year, 41 g C m^?2 was sequestered from the atmosphere. The net exchange rates were determined mostly by the timing and intensity of spring rainfall. The components of ecosystem respiration were measured using chambers. Combining scaled‐up measurements with the eddy CO₂ effluxes, it was estimated that 85% of ecosystem respiration emanated from the soil surface. Under well‐watered conditions, 26% of the soil surface CO₂ efflux came from soil microbial activity. Rates of soil microbial CO₂ production and net mineral‐N production were low and indicative of substrate limitation. Soil respiration declined by a factor of four as the soil water content declined from field capacity (0.21 m³ m^?3) to the driest value obtained (0.04 m³ m^?3). Rainfall after periods of drought resulted in large, but short‐lived, respiration pulses that were curvilinearly related to the increase in root‐zone water content. Coupled with the low leaf area and high root : shoot ratio, this sparse grassland had a limited capacity to sequester and store carbon. Assuming a proportionality between carbon gain and rainfall during the summer, rainfall distribution statistics suggest that the ecosystem is sustainable in the long term.     

Seasonal variation in leaf properties and ecosystem carbon budget in a cool-temperate deciduous broad-leaved forest: simulation analysis at Takayama site, Japan

Seasonal changes in gross primary production (GPP) and net ecosystem production (NEP) in temperate deciduous forests are mostly driven by environmental conditions and the phenology of leaf demography. This study addresses another factor, temporal changes in leaf properties, i.e., leaf aging from emergence to senescence. A process-based model was used to link the ecosystem-scale carbon budget with leaf-level properties on the basis of field observation and scaling procedures; temporal variations in leaf thickness (leaf mass per area, LMA), photosynthetic rubisco (V_cmax) and electron-transport (J_max) capacity, and dark respiration (R_d) were empirically parameterized. The model was applied to a cool-temperate deciduous broad-leaved forest at Takayama, in central Japan, and validated with data of net ecosystem CO₂ exchange (NEE=–NEP) measured using the eddy-covariance method. NEP of the Takayama site varied seasonally from 3 g C m^–2 day^–1 net source in late winter to 5 g C m^–2 day^–1 net sink in early to mid-summer. A sensitivity experiment showed that removing the leaf-aging effect changed the seasonal CO₂ exchange pattern, and led to overestimation of annual GPP by 6% and annual NEP by 38%. We found that seasonal variation in V_cmax affected the seasonal pattern and annual budget of CO₂ exchange most strongly; LMA and R_d had moderate influences. The rapid change in V_cmax and R_d during leaf emergence and senescence was important in evaluating GPP and NEP of the temperate deciduous forest.     

Seasonal variations of CO2 and water vapour exchange rates over a temperate deciduous forest

Long-term and direct measurements of CO₂ and water vapour exchange are needed over forested ecosystems to determine their net annual fluxes of carbon dioxide and water. Such measurements are also needed to parameterize and test biogeochemical, ecological and hydrological assessment models. Responding to this need, eddy covariance measurements of CO₂ and water vapour were made ever a deciduous forest growing near Oak Ridge, TN, between April 1993 and April 1994. Periodic measurements were made of leaf area index, stomatal resistance, soil moisture and pre-dawn leaf water potential to characterize the gas exchange capacity of the canopy. Four factors had a disproportionate influence on the seasonal variation of CO₂ flux densities. These factors were photon flux densities (during the growing season), temperature (during the dormant season), leaf area index and the occurrence of drought The drought period occurred during the peak of the growing season and caused a significant decline in daily and hourly CO₂ flux densities, relative to observations over the stand when soil moisture was plentiful. The annual net uptake of carbon was calculated by integrating flux measurements and filling missing and spurious data with the relations obtained between measured CO₂ fluxes and environmental forcing variables. The net flux of carbon for the period between April 1993 and April 1994 was -525 g C m^?2 y^?1. This value represents a net flux of carbon from the atmosphere and into the forest. The net annual carbon exchange of this southern temperate broadleaved forest exceeded values measured over a northern temperate forest (which experiences a shorter growing season and has less leaf area) by 200 g C m^?2 y^?1 (cf. Wofsy et al 1993). The seasonal variation of canopy evaporation (latent heat flux) was controlled mostly by changes in leaf area and net radiation. A strong depression in evaporation rates was not observed during the drought Over a broadleaved forest large vapour pressure deficits promote evaporation and trees in a mixed stand are able to tap a variety of deep and shallow water sources.     

CO2 exchange in three Canadian High Arctic ecosystems: response to long-term experimental warming

Carbon dioxide exchange, soil C and N, leaf mineral nutrition and leaf carbon isotope discrimination (LCID‐Δ) were measured in three High Arctic tundra ecosystems over 2 years under ambient and long‐term (9 years) warmed (～2°C) conditions. These ecosystems are located at Alexandra Fiord (79°N) on Ellesmere Island, Nunavut, and span a soil water gradient; dry, mesic, and wet tundra. Growing season CO₂ fluxes (i.e., net ecosystem exchange (NEE), gross ecosystem photosynthesis (GEP), and ecosystem respiration (R_e)) were measured using an infrared gas analyzer and winter C losses were estimated by chemical absorption. All three tundra ecosystems lost CO₂ to the atmosphere during the winter, ranging from 7 to 12 g CO₂‐C m^?2 season^?1 being highest in the wet tundra. The period during the growing season when mesic tundra switch from being a CO₂ source to a CO₂ sink was increased by 2 weeks because of warming and increases in GEP. Warming during the summer stimulated dry tundra GEP more than R_e and thus, NEE was consistently greater under warmed as opposed to ambient temperatures. In mesic tundra, warming stimulated GEP with no effect on R_e increasing NEE by ～10%, especially in the first half of the summer. During the ～70 days growing season (mid‐June–mid‐August), the dry and wet tundra ecosystems were net CO₂‐C sinks (30 and 67 g C m^?2 season^?1, respectively) and the mesic ecosystem was a net C source (58 g C m^?2 season^?1) to the atmosphere under ambient temperature conditions, due in part to unusual glacier melt water flooding that occurred in the mesic tundra. Experimental warming during the growing season increased net C uptake by ～12% in dry tundra, but reduced net C uptake by ～20% in wet tundra primarily because of greater rates of R_e as opposed to lower rates of GEP. Mesic tundra responded to long‐term warming with ～30% increase in GEP with almost no change in R_e reducing this tundra type to a slight C source (17 g C m^?2 season^?1). Warming caused LCID of Dryas integrafolia plants to be higher in dry tundra and lower in Salix arctic plants in mesic and wet tundra. Our findings indicate that: (1) High Arctic ecosystems, which occur in similar mesoclimates, have different net CO₂ exchange rates with the atmosphere; (2) long‐term warming can increase the net CO₂ exchange of High Arctic tundra by stimulating GEP, but it can also reduce net CO₂ exchange in some tundra types during the summer by stimulating R_e to a greater degree than stimulating GEP; (3) after 9 years of experimental warming, increases in soil carbon and nitrogen are detectable, in part, because of increases in deciduous shrub cover, biomass, and leaf litter inputs; (4) dry tundra increases in GEP, in response to long‐term warming, is reflected in D. integrifolia LCID; and (5) the differential carbon exchange responses of dry, mesic, and wet tundra to similar warming magnitudes appear to depend, in part, on the hydrologic (soil water) conditions. Annual net ecosystem CO₂‐C exchange rates ranged from losses of 64 g C m^?2 yr^?1 to gains of 55 g C m^?2 yr^?1. These magnitudes of positive NEE are close to the estimates of NPP for these tundra types in Alexandra Fiord and in other High Arctic locations based on destructive harvests.     

Contemporary carbon balance and late Holocene carbon accumulation in a northern peatland

Northern peatlands contain up to 25% of the world's soil carbon (C) and have an estimated annual exchange of CO₂‐C with the atmosphere of 0.1–0.5 Pg yr⁻¹ and of CH₄‐C of 10–25 Tg yr⁻¹. Despite this overall importance to the global C cycle, there have been few, if any, complete multiyear annual C balances for these ecosystems. We report a 6‐year balance computed from continuous net ecosystem CO₂ exchange (NEE), regular instantaneous measurements of methane (CH₄) emissions, and export of dissolved organic C (DOC) from a northern ombrotrophic bog. From these observations, we have constructed complete seasonal and annual C balances, examined their seasonal and interannual variability, and compared the mean 6‐year contemporary C exchange with the apparent C accumulation for the last 3000 years obtained from C density and age‐depth profiles from two peat cores. The 6‐year mean NEE‐C and CH₄‐C exchange, and net DOC loss are −40.2±40.5 (±1 SD), 3.7±0.5, and 14.9±3.1 g m⁻² yr⁻¹, giving a 6‐year mean balance of −21.5±39.0 g m⁻² yr⁻¹ (where positive exchange is a loss of C from the ecosystem). NEE had the largest magnitude and variability of the components of the C balance, but DOC and CH₄ had similar proportional variabilities and their inclusion is essential to resolve the C balance. There are large interseasonal and interannual ranges to the exchanges due to variations in climatic conditions. We estimate from the largest and smallest seasonal exchanges, quasi‐maximum limits of the annual C balance between 50 and −105 g m⁻² yr⁻¹. The net C accumulation rate obtained from the two peatland cores for the interval 400–3000 bp (samples from the anoxic layer only) were 21.9±2.8 and 14.0±37.6 g m⁻² yr⁻¹, which are not significantly different from the 6‐year mean contemporary exchange.     

Photosynthesis and carbon balance of a Sahelian fallow savanna

Eddy-covariance measurements of CO₂ exchange above a Sahelian savanna consisting of small shrubs over a near-continuous herb layer were made during the HAPEX-Sahel experiment in Niger, West Africa. The measurements were made near-continuously during an 8-week period, covering the main part of the rainy season and three weeks at the beginning of the dry season. The measurements were corrected for in-canopy storage of CO₂ and the night-time measurements used to derive respiration functions for the soil, roots and above-ground plant material. Photosynthetic CO₂ uptake was estimated and compared to simulations using a biochemical photosynthesis model in a simple, ‘big-leaf’, implementation. The model satisfactorily reproduced the measurements (coefficient of determination 0.80) using parameters defined from the literature and based on soil nutrient concentrations. When the quantum yield (α) and rubisco capacity (V_mr) were fitted to the data with allowance for physiological changes through the season, an excellent agreement between model and measurements was obtained (coefficient of determination 0.93, RMS error 1.46 μmol m^–2 s^–1). The fitted photosynthesis and respiration model was used to estimate the carbon balance of the savanna site during the growing season of 1992 and for the complete calendar year. Harvest estimates of net plant biomass accumulation during the growing season and annual wood accumulation agreed well with modelled net photosynthesis and annual net carbon accumulation, respectively. Peak instantaneous ecosystem CO₂ uptake was comparable to peak values observed in other biomes, but annual photosynthesis and carbon sequestration were considerably lower than observed elsewhere.