Ernst Linder

N2O emissions from humid tropical agricultural soils: effects of soil moisture, texture and nitrogen availability

We studied soil moisture dynamics and nitrous oxide (N2O) fluxes from agricultural soils in the humid tropics of Costa Rica. Using a split-plot design on two soils (clay, loam) we compared two crop types (annual, perennial) each unfertilized and fertilized. Both soils are of andic origin. Their properties include relatively low bulk density and high organic matter content, water retention capacity, and hydraulic conductivity. The top 2–3xa0cm of the soils consists of distinct small aggregates (dia. <0.5xa0cm). We measured a strong gradient of bulk density and moisture within the top 7xa0cm of the clay soil. Using automated sampling and analysis systems we measured N2O emissions at 4.6xa0h intervals, meteorological variables, soil moisture, and temperature at 0.5xa0h intervals. Mean daily soil moisture content at 5xa0cm depth ranged from 46% water filled pore space (WFPS) on clay in April 1995 to near saturation on loam during a wet period in February 1996. On both soils the aggregated surface layer always remained unsaturated. Soils emitted N2O throughout the year. Mean N2O fluxes were 1.04±0.72xa0ng N2O-N cm−2xa0h−1 (mean±standard deviation) from unfertilized loam under annual crops compared to 3.54±4.31xa0ng N2O-N cm−2xa0h−1 from the fertilized plot (351 days measurement). Fertilization dominated the temporal variation of N2O emissions. Generally fluxes peaked shortly after fertilization and were increased for up to 6 weeks (‘post fertilization flux’). Emissions continued at a lower rate (‘background flux’) after fertilization effects faded. Mean post-fertilization fluxes were 6.3±6.5xa0ng N2O-N cm−2xa0h−1 while the background flux rate was 2.2±1.8xa0ng N2O-N cm−2xa0h−1. Soil moisture dynamics affected N2O emissions. Post fertilization fluxes were highest from wet soils; fluxes from relatively dry soils increased only after rain events. N2O emissions were weakly affected by soil moisture during phases of low N availability. Statistical modeling confirmed N availability and soil moisture as the major controls on N2O flux. Our data suggest that small-scale differences in soil structure and moisture content cause very different biogeochemical environments within the top 7xa0cm of soils, which is important for net N2O fluxes from soils.

Peatland responses to varying interannual moisture conditions as measured by automatic CO2 chambers

[1]xa0Net ecosystem CO2 exchange (NEE) was measured from June 2000 through October 2001 by 10 automatic chambers at a peatland in southeastern New Hampshire. The high temporal frequency of this sampling method permitted detailed examination of NEE as it varied daily and seasonally. Summer of 2001 was significantly drier than the 30-year average, while summer of 2000 was wetter than normal. Although NEE varied spatially across the peatland with differences in plant species composition and biomass, maximum CO2 uptake was 30–40% larger in the drier summer in evergreen and deciduous shrub communities but the same or lower in sedge sites. Ecosystem respiration rates were 13–84% larger in the drier summer depending on plant growth form with water table and temperature as strong predictors. Ecosystem respiration was also correlated with maximum ecosystem productivity and foliar biomass suggesting that plant processes, water table, and temperature are tightly linked in their control of respiratory losses. The ratio between maximum productivity and respiration declined for evergreen shrub and sedge sites between the wet and dry summer, but increased in deciduous shrub sites. A drier climate may reduce the CO2 sink function of peatlands for some growth forms and increase it for others, suggesting that ecosystem carbon and climate models should account for differences in growth form responses to climate change. It also implies that plant functional types respond on short timescales to changes in moisture, and that the transition from sedges to shrubs could occur rapidly in peatlands under a drier and warmer climate.

Global N removal by freshwater aquatic systems using a spatially distributed, within-basin approach

2.6-1000 km 2 ), large rivers, lakes, and reservoirs, using a 30 0 latitudelongitude river network to route and process material from continental source areas to the coastal zone. Mean annual aquatic TN removal (for the mid-1990s time period) is determined by the distributions of aquatic TN inputs, mean annual hydrological characteristics, and biological activity. Model-predicted TN concentrations at basin mouths corresponded wellwithobservations(medianrelativeerror= � 12%,interquartile rangeofrelativeerror= 85%), an improvement over assumptions of uniform aquatic removal across basins. Removal by aquatic systems globally accounted for 14% of total N inputs to continental surfaces, but represented 53% of inputs to aquatic systems. Integrated aquatic removal was similar in small rivers (16.5% of inputs), large rivers (13.6%), and lakes (15.2%), while large reservoirs were less important (5.2%). Bias related to runoff suggests improvements are needed in nonpoint N input estimates and/or aquatic biological activity. The within-basin approach represented by FrAMES-N will improve understanding of the freshwater nutrient flux response to anthropogenic change at global scales.

Net ecosystem productivity and its uncertainty in a diverse boreal peatland

Net ecosystem exchange (NEE) of CO2 was measured in four peatlands along plant community, hydrologic, and water chemistry gradients from bog to rich fen in a diverse peatland complex near Thompson, Manitoba, as part of the Boreal Ecosystem-Atmosphere Study (BOREAS). A simple model for estimating growing season net ecosystem productivity (NEP) using continuous measurements of photosynthetically active radiation (PAR), and peat temperature was constructed with weekly chamber measurements of NEE from May to October 1996. The model explained 79–83% of the variation in NEE across the four sites. Model estimation and parameter uncertainty calculations were performed using nonlinear regression analyses with a maximum likelihood objective function. The model underestimated maximum NEE and respiration during the midseason when the standard errors for each parameter were greatest. On a daily basis, uncertainty in the midday NEE simulation was higher than at night. Although the magnitude of both photosynthesis and respiration rates followed the trophic gradient bog less than poor fen less than intermediate fen less than rich fen, NEP did not follow the same pattern. NEP in the bog and rich fen was close to zero, while the poor and intermediate fens had higher NEP due to a greater imbalance between uptake and release of CO2. Although all sites had a positive growing season NEP, upper and lower 95% confidence limits showed that the bog and rich fen were either a source or sink of CO2 to the atmosphere, while the sedge-dominated poor and intermediate fens were accumulating approximately 20–100 g CO2 C m−2 over the 5 month period in 1996. Peatland ecosystems may switch from a net sink to a source of carbon on short timescales with small changes in soil temperature or water table position. Since the difference between production and decomposition is small, it is important to understand and quantify the magnitude of uncertainty in these measurements in order to predict the effect of climatic change on peatland carbon exchange.

Statistical uncertainty of eddy flux–based estimates of gross ecosystem carbon exchange at Howland Forest, Maine

[1]xa0We present an uncertainty analysis of gross ecosystem carbon exchange (GEE) estimates derived from 7 years of continuous eddy covariance measurements of forest-atmosphere CO2 fluxes at Howland Forest, Maine, USA. These data, which have high temporal resolution, can be used to validate process modeling analyses, remote sensing assessments, and field surveys. However, separation of tower-based net ecosystem exchange (NEE) into its components (respiration losses and photosynthetic uptake) requires at least one application of a model, which is usually a regression model fitted to nighttime data and extrapolated for all daytime intervals. In addition, the existence of a significant amount of missing data in eddy flux time series requires a model for daytime NEE as well. Statistical approaches for analytically specifying prediction intervals associated with a regression require, among other things, constant variance of the data, normally distributed residuals, and linearizable regression models. Because the NEE data do not conform to these criteria, we used a Monte Carlo approach (bootstrapping) to quantify the statistical uncertainty of GEE estimates and present this uncertainty in the form of 90% prediction limits. We explore two examples of regression models for modeling respiration and daytime NEE: (1) a simple, physiologically based model from the literature and (2) a nonlinear regression model based on an artificial neural network. We find that uncertainty at the half-hourly timescale is generally on the order of the observations themselves (i.e., ∼100%) but is much less at annual timescales (∼10%). On the other hand, this small absolute uncertainty is commensurate with the interannual variability in estimated GEE. The largest uncertainty is associated with choice of model type, which raises basic questions about the relative roles of models and data.

CONTROLS OF SPATIAL VARIATION IN THE PREVALENCE OF TREMATODE PARASITES INFECTING A MARINE SNAIL

Geographic variability in abundance can be driven by multiple physical and biological factors operating at multiple scales. To understand the determinants of larval trematode prevalence within populations of the marine snail host Littorina littorea, we quantified many physical and biological variables at 28 New England intertidal sites. A hierarchical, mixed-effects model identified the abundance of gulls (the final hosts and dispersive agents of infective trematode stages) and snail size (a proxy for time of exposure) as the primary factors associated with trematode prevalence. The predominant influence of these variables coupled with routinely low infection rates (21 of the 28 populations exhibited prevalence <12%) suggest broad-scale recruitment limitation of trematodes. Although infection rates were spatially variable, formal analyses detected no regional spatial gradients in either trematode prevalence or independent environmental variables. Trematode prevalence appears to be predominantly determined by local site characteristics favoring high gull abundance.

Bayesian spatial prediction

This paper presents a complete Bayesian methodology for analyzing spatial data, one which employs proper priors and features diagnostic methods in the Bayesian spatial setting. The spatial covariance structure is modeled using a rich class of covariance functions for Gaussian random fields. A general class of priors for trend, scale, and structural covariance parameters is considered. In particular, we obtain analytic results that allow easy computation of the predictive distribution for an arbitrary prior on the parameters of the covariance function using importance sampling. The computations, as well as model diagnostics and sensitivity analysis, are illustrated with a set of precipitation data.

Evaluation of trends in derived snowfall and rainfall across Eurasia and linkages with discharge to the Arctic Ocean

[1]xa0To more fully understand the role of precipitation in observed increases in freshwater discharge to the Arctic Ocean, data from a new archive of bias-adjusted precipitation records for the former USSR (TD9813), along with the CRU and Willmott-Matsuura data sets, were examined for the period 1936–1999. Across the six largest Eurasian river basins, snowfall derived from TD9813 exhibits a strongly significant increase until the late 1950s and a moderately significant decrease thereafter. A strongly significant decline in derived rainfall is also noted. Spatially, snowfall increases are found primarily across north-central Eurasia, an area where the rainfall decreases are most prominent. Although no significant change is determined in Eurasian-basin snowfall over the entire 64 year period, we note that interpolation from early, uneven station networks causes an overestimation of spatial precipitation, and that the local snowfall trends determined from gridded TD9813 data are likely underestimated. Yet, numerous uncertainties in historical Arctic climate data and the sparse, irregular nature of Arctic station networks preclude a confident assessment of precipitation-discharge linkages during the period of reported discharge trends.

Global perspective of nitrate flux in ice cores

The relationships between the concentration and the flux of chemical species (Cl{sup {minus}}, NO{sub 3}{sup {minus}}, SO{sub 4}{sup 2{minus}}, Na{sup +}, K{sup +}, NH{sub 4}{sup +}, Mg{sub 2+}, Ca{sup 2+}) versus snow accumulation rate were examined at GISP2 and 20D in Greenland, Mount Logan from the St. Elias Range, Yukon Territory, Canada, and Sentik Glacier from the northwest end of the Zanskar Range in the Indian Himalayas. At all sites, only nitrate flux is significantly ({alpha}=0.05) related to snow accumulation rate. Of all the chemical series, only nitrate concentration data are normally distributed. Therefore the authors suggest that nitrate concentration in snow is affected by postdepositional exchange with the atmosphere over a broad range of environmental conditions. The persistant summer maxima in nitrate observed in Greenland snow over the entire range of record studied (the last 800 years) may be mainly due to NO{sub x} released from peroxyacetyl nitrate by thermal decomposition in the presence of higher OH concentrations in summer. The late winter/early spring nitrate peak observed in modern Greenland snow may be related to the buildup of anthropogenically derived NO{sub y} in the Arctic troposphere during the long polar winter. 58 refs., 3 figs., 4 tabs.

Spatial and temporal variability of nitrogen oxide and methane fluxes from a fertilized tree plantation in Costa Rica

Nitric oxide (NO), nitrous oxide (N2O), and methane (CH4) are naturally produced and consumed by soil biogeochemical processes. Naturally high variation between trace gas fluxes may temporarily increase due to agricultural management. We studied spatial and temporal variability of fluxes in the context of a 3-year field experiment established to identify and quantify N2O fluxes and controlling factors using automated field measurements. We measured trace gas fluxes, soil temperature, and moisture from fertilized and unfertilized balsa (Ochroma lagopus) plantations. Combining spatial and temporal sampling we evaluate if automatically measured time series of N2O emissions are representative of overall mean fluxes from fertilized loam under balsa. Soil trace gas fluxes were measured manually at 36 randomly distributed sampling locations per plot. Mean plot emissions were evaluated against fluxes measured by seven chambers commonly used for routine bimonthly manual measurements and against N2O emissions measured by two automated chambers at 4.6-hour sampling intervals. Trace gas fluxes were highly variable over 40 × 40 m plots. Nitrogen oxide fluxes were mainly spatially independent. Fertilization increased nitrogen oxide emissions but did not introduce spatial dependency of flux data. Within about 6 weeks fluxes approached pre-fertilization level again. Given high spatial variation of nitrogen oxide fluxes we find that automatically measured N2O fluxes represent the nature of the flux response well and are in the range of fluxes indicated by spatial sampling. When soils were relatively dry fertilization inhibited CH4 uptake.