Process-based Biogeochemical Model Research Articles

Twentieth-Century Droughts and Their Impacts on Terrestrial Carbon Cycling in China

Abstract Midlatitude regions experienced frequent droughts during the twentieth century, but their impacts on terrestrial carbon balance are unclear. This paper presents a century-scale study of drought effects on the carbon balance of terrestrial ecosystems in China. The authors first characterized the severe extended droughts over the period 1901–2002 using the Palmer drought severity index and then examined how these droughts affected the terrestrial carbon dynamics using tree-ring width chronologies and a process-based biogeochemistry model, the Terrestrial Ecosystem Model (TEM). It is found that China suffered from a series of severe extended droughts during the twentieth century. The major drought periods included 1920–30, 1939–47, 1956–58, 1960–63, 1965–68, 1978–80, and 1999–2002. Most droughts generally reduced net primary productivity (NPP) and net ecosystem productivity (NEP) in large parts of drought-affected areas. Moreover, some of the droughts substantially reduced the countrywide annual NPP and NEP. Out of the seven droughts, three (1920–30, 1965–68, and 1978–80) caused the countrywide terrestrial ecosystems to switch from a carbon sink to a source, and one (1960–63) substantially reduced the magnitude of the countrywide terrestrial carbon sink. Strong decreases in NPP were mainly responsible for the anomalies in annual NEP during these drought periods. Changes in heterotrophic respiration happened in the same direction, but mostly with smaller magnitude. The results show that severe extended droughts had significant effects on terrestrial carbon cycling in China, although future studies should consider other important processes such as drought-induced mortality and regrowth, land-use change, disturbances (e.g., fire), human management (e.g., fertilization and irrigation), and environmental pollution (e.g., ozone pollution, nitrogen deposition). These drought effects are of particular importance in light of projected widespread summer drying in midlatitude regions during the twenty-first century. Future droughts could lead to a reduced terrestrial carbon sink or even a source and exert a positive feedback to the global climate system.

Can we reconcile differences in estimates of carbon fluxes from land-use change and forestry for the 1990s?

Abstract. The effect of Land Use Change and Forestry (LUCF) on terrestrial carbon fluxes can be regarded as a carbon credit or debit under the UNFCCC, but scientific uncertainty in the estimates for LUCF remains large. Here, we assess the LUCF estimates by examining a variety of models of different types with different land cover change maps in the 1990s. Annual carbon pools and their changes are separated into different components for separate geographical regions, while annual land cover change areas and carbon fluxes are disaggregated into different LUCF activities and the biospheric response due to CO2 fertilization and climate change. We developed a consolidated estimate of the terrestrial carbon fluxes that combines book-keeping models with process-based biogeochemical models and inventory estimates and yields an estimate of the global terrestrial carbon flux that is within the uncertainty range developed in the IPCC 4th Assessment Report. We examined the USA and Brazil as case studies in order to assess the cause of differences from the UNFCCC reported carbon fluxes. Major differences in the litter and soil organic matter components are found for the USA. Differences in Brazil result from assumptions about the LUC for agricultural purposes. The effects of CO2 fertilization and climate change also vary significantly in Brazil. Our consolidated estimate shows that the small sink in Latin America is within the uncertainty range from inverse models, but that the sink in the USA is significantly smaller than the inverse models estimates. Because there are different sources of errors at the country level, there is no easy reconciliation of different estimates of carbon fluxes at the global level. Clearly, further work is required to develop data sets for historical land cover change areas and models of biogeochemical changes for an accurate representation of carbon uptake or emissions due to LUC.

Ten steps applied to development and evaluation of process-based biogeochemical models of estuaries

The procedures involved in model development may be set out as a ten step process, beginning with defining the purpose of the model and ending with evaluation of the appropriateness and utility of the completed model. This process, recently outlined by Jakeman et al. [Jakeman, A.J., Letcher, R.A., Norton, J.P., 2006. Ten iterative steps in development and evaluation of environmental models. Environmental Modelling and Software 21, 602–614], is often iterative as model development is a continuous process that refines and improves the intended capacity of the model. Here, the ten steps of model development are critiqued and applied using a process-based biogeochemical model of aquatic systems, with examples from two case studies: a model of phytoplankton succession and nutrient concentrations in the Swan-Canning Estuary (Western Australia) and a model of sediment and nutrient transport and transformation in the Fitzroy Estuary and Keppel Bay (Queensland).

Mapping and modelling of greenhouse gas emissions from rice paddies with satellite radar observations and the DNDC biogeochemical model

1. Rice is an important agricultural production system with more than 80 million ha of irrigated rice paddies in annual production globally. As water resources become scarcer, the competition between urban development and agriculture for available water will intensify. Paddy rice cropland distribution and management intensity will need to evolve over the coming decades to accommodate increased production demand with decreasing land and water resources. 2. While process-based biogeochemical models can provide important insights into how agricultural management of rice paddies influences water resources, yields and greenhouse gas emissions, they require accurate spatial estimates of the extent of paddy rice cropland and cropping systems. Satellite remote sensing data can provide such spatially explicit information. 3. Data from Synthetic Aperture Radar (SAR) are ideal for mapping rice paddies owing to its nearly all-weather imaging capabilities and sensitivity to flooded vegetation. 4. This paper presents a framework for combining routine SAR observations, GIS databases and a process-based biogeochemical model for a decision-support system for mapping and monitoring rice paddies. This framework is demonstrated for a site in India under a range of alternative water management strategies. Copyright © 2007 John Wiley & Sons, Ltd.

Modeling response of soil productivity to biogeochemical cycling and atmospheric acid deposition in the Hayden Brook watershed (Canada) using the ForNBM model

A process-based biogeochemistry model, the ForNBM, was calibrated to simulate stream flows, nutrient concentrations and loadings on a monthly time scale using the field data collected from 1973 to 1985 at the Hayden Brook watershed, a natural forested mixed-hardwood watershed, in central New Brunswick, Canada. The model was then used to assess the major biogeochemical processes that influence soil productivity and the sensitivity of soil acidity and nutrient leaching to different scenarios of atmospheric acid deposition in the Hayden Brook watershed. The results of the model calibration showed that the ForNBM model reproduced the field data well with the Nash–Sutcliffe coefficient of efficiency ranking from 0.63 to 0.99 for the simulations of stream flows, from −0.11 to 0.14 for the stream nutrient concentrations, from 0.03 to 0.63 for the monthly stream nutrient loadings, and from 0.94 to 0.99 for the cumulative stream nutrient loadings. Field measurements and model predictions indicated that 90% of atmospheric deposited nitrogen (N) was retained in the soils. Atmospheric N deposition may be the most important source of N for the long-term watershed soil N accumulation in the Hayden Brook watershed. Biomass had low nutrient retention rates and high nutrient return rates, which demonstrated that the biogeochemical cycling was fast in the mixed-hardwood forest. The results also showed that annual N supply for biomass uptake was mainly from soil organic matter mineralization, and base cations (Ca 2+, Mg 2+, and K +) from both soil organic matter mineralization and mineral weathering. Results of sensitivity analysis indicated that the soil pH and base cation leaching were more sensitive to atmospheric acid deposition than N. A 100% increase of current atmospheric acid deposition may cause a 0.6 unit decrease of soil pH, 44%, 39%, and 30% increase of soil Ca 2+, Mg 2+, and K + leaching, respectively. Long-term soil productivity at the Hayden Brook watershed may also be declined by soil acidity and depletion of base cations due to increased leaching in response to increased atmospheric acid deposition.

DAYCENT National‐Scale Simulations of Nitrous Oxide Emissions from Cropped Soils in the United States

Until recently, Intergovernmental Panel on Climate Change (IPCC) emission factor methodology, based on simple empirical relationships, has been used to estimate carbon (C) and nitrogen (N) fluxes for regional and national inventories. However, the 2005 USEPA greenhouse gas inventory includes estimates of N2O emissions from cultivated soils derived from simulations using DAYCENT, a process-based biogeochemical model. DAYCENT simulated major U.S. crops at county-level resolution and IPCC emission factor methodology was used to estimate emissions for the approximately 14% of cropped land not simulated by DAYCENT. The methodology used to combine DAYCENT simulations and IPCC methodology to estimate direct and indirect N2O emissions is described in detail. Nitrous oxide emissions from simulations of presettlement native vegetation were subtracted from cropped soil N2O to isolate anthropogenic emissions. Meteorological data required to drive DAYCENT were acquired from DAYMET, an algorithm that uses weather station data and accounts for topography to predict daily temperature and precipitation at 1-km2 resolution. Soils data were acquired from the State Soil Geographic Database (STATSGO). Weather data and dominant soil texture class that lie closest to the geographical center of the largest cluster of cropped land in each county were used to drive DAYCENT. Land management information was implemented at the agricultural-economic region level, as defined by the Agricultural Sector Model. Maps of model-simulated county-level crop yields were compared with yields estimated by the USDA for quality control. Combining results from DAYCENT simulations of major crops and IPCC methodology for remaining cropland yielded estimates of approximately 109 and approximately 70 Tg CO2 equivalents for direct and indirect, respectively, mean annual anthropogenic N2O emissions for 1990-2003.

Effect of Land‐Cover Change on Terrestrial Carbon Dynamics in the Southern United States

Land-cover change has significant influence on carbon storage and fluxes in terrestrial ecosystems. The southern United States is thought to be the largest carbon sink across the conterminous United States. However, the spatial and temporary variability of carbon storage and fluxes due to land-cover change in the southern United States remains unclear. In this study, we first reconstructed the annual data set of land-cover of the southern United States from 1860 to 2003 with a spatial resolution of 8 km. Then we used a spatially explicit process-based biogeochemical model (Terrestrial Ecosystem Model [TEM] 4.3) to simulate the effects of cropland expansion and forest regrowth on the carbon dynamics in this region. The pattern of land-cover change in the southern United States was primarily driven by the change of cropland, including cropland expansion and forest regrowth on abandoned cropland. The TEM simulation estimated that total carbon storage in the southern United States in 1860 was 36.8 Pg C, which likely was overestimated, including 10.8 Pg C in the southeast and 26 Pg C in the south-central. During 1860-2003, a total of 9.4 Pg C, including 6.5 Pg C of vegetation and 2.9 Pg C of soil C pool, was released to the atmosphere in the southern United States. The net carbon flux due to cropland expansion and forest regrowth on abandoned cropland was approximately zero in the entire southern region between 1980 and 2003. The temporal and spatial variability of regional net carbon exchange was influenced by land-cover pattern, especially the distribution of cropland. The land-use analysis in this study is incomplete and preliminary. Finally, the limitations, improvements, and future research needs of this study were discussed.

Analyzing the carbon dynamics of central European forests: comparison of Biome-BGC simulations with measurements

Biogeochemical models are often used for making projections of future carbon dynamics under scenarios of global change. The aim of this study was to assess the accuracy of the process-based biogeochemical model Biome-BGC for application in central European forests from the lowlands to upper treeline as a pre-requisite for environmental impact assessments. We analyzed model behavior along an altitudinal gradient across the alpine treeline, which provided insights on the sensitivity of simulated average carbon pools to changes in environmental factors. A second set of tests included medium-term (30 years) simulations of carbon fluxes, and a third set of tests focused on daily carbon and water fluxes. Model results were compared to aboveground biomass measurements, leaf area index recordings as well as net ecosystem exchange (NEE) and actual evapotranspiration (AET) measurements. The simulated medium-term forest growth agreed well with measured data. Also daily NEE fluxes were simulated adequately in most cases. Problems were detected when simulating ecosystems close to the upper timberline (overestimation of measured growth and pool sizes), and when simulating daily AET fluxes (overestimation of measured fluxes). The results showed that future applications of Biome-BGC could benefit much from an improvement of model algorithms (e.g., the Q10 model for respiration) as well as from a detailed analysis of the ecological significance of crucial parameters (e.g., the canopy water interception coefficient).

Influence of climate change factors on carbon dynamics in northern forested peatlands

Peatlands are carbon-accumulating wetland ecosystems, developed through an imbalance among organic matter production and decomposition processes. Soil saturation is the principal cause of anoxic conditions that constrain organic matter decay. Accordingly, changes in the hydrologic regime will affect the carbon (C) dynamics in forested peatlands. Our objective is to review ecological studies and experiments on managed peatlands that provide a basis for assessing the effects of an altered hydrology on C dynamics. We conclude that climate change influences will be mediated primarily through the hydrologic cycle. A lower water table resulting from altered precipitation patterns and increased atmospheric temperature may be expected to decrease soil CH4 and increase CO2 emissions from the peat surface. Correspondingly, the C balance in forested peatlands is also sensitive to management and restoration prescriptions. Increases in soil CO2 efflux do not necessarily equate with net losses from the soil C pool. While the fundamentals of the C balance in peatlands are well-established, the combined affects of global change stressors and management practices are best considered using process-based biogeochemical models. Long-term studies are needed both for validation and to provide a framework for longitudinal assessments of the peatland C cycle. Key words: Peatland, carbon cycle, methane, forest, wetland.

Combining lidar estimates of aboveground biomass and Landsat estimates of stand age for spatially extensive validation of modeled forest productivity

Extensive estimates of forest productivity are required to understand the relationships between shifting land use, changing climate and carbon storage and fluxes. Aboveground net primary production of wood (NPP Aw) is a major component of total NPP and of net ecosystem production (NEP). Remote sensing of NPP and NPP Aw is generally based on light use efficiency or process-based biogeochemistry models. However, validating these large area flux estimates remains a major challenge. In this study we develop an independent approach to estimating NPP Aw, based on stand age and biomass, that could be implemented over a large area and used in validation efforts. Stand age is first mapped by iterative unsupervised classification of a multi-temporal sequence of images from a passive optical sensor (e.g. Landsat TM). Stand age is then cross-tabulated with estimates of stand height and aboveground biomass from lidar remote sensing. NPP Aw is then calculated as the average increment in lidar-estimated biomass over the time period determined using change detection. In western Oregon, productivity estimates made using this method compared well with forest inventory estimates and were significantly different than estimates from a spatially distributed biogeochemistry model. The generality of the relationship between lidar-based canopy characteristics and stand biomass means that this approach could potentially be widely applicable to landscapes with stand replacing disturbance regimes, notably in regions where forest inventories are not routinely maintained.

Modeling stream water nutrient concentrations and loadings in response to weather condition and forest harvesting

A process-based biogeochemistry model, ForNBM, was used to simulate stream water nutrient concentrations and monthly and cumulative stream nutrient loadings according to the monthly air temperature and precipitation records at the Nashwaak Experimental Watershed in central New Brunswick, Canada. Compared with the field measurement data from 1973 to 1985, the simulation results show that the best simulation of stream water nutrient concentrations is for Mg 2+ with r 2 = 0.59, and the poorest is for NH 4 +-N with r 2 = 0.48. The best simulation of the monthly stream nutrient loadings is for Ca 2+ with r 2 = 0.66, and the poorest is for NO 3 −-N with r 2 = 0.53. All simulations of the cumulative stream nutrient loadings have r 2 values more than 0.90. Seasonal precipitation and air temperature played the most important role in determining the patterns of the stream nutrient concentrations and loadings. The stream nutrient concentrations were high during the winter and low during the spring. The peak values of the stream nutrient loadings appeared in the springs and the low values appeared in the winters. Although snowmelt diluted the stream water nutrient concentrations during the springs, it resulted in high stream nutrient loadings due to stream flooding. According to the assessments of the model simulations and the field measurements, the increased watershed soil nutrient losses through the stream exports after forest harvesting may not cause the decline of the watershed soil productivity because (1) the annual increments of the stream nutrient exports were small, with 2.34 kg ha −1 year −1 for NO 3 −-N, 0.93 for NH 4 +-N, 3.34 for Ca 2+, 0.30 for Mg 2+, and 5.21 for K + and (2) the increased watershed soil nutrient losses occurred during the stage of forest regeneration when vegetation growth requires low soil nutrient supply.

An Approach to Spatially Distributed Modeling of Net Primary Production (NPP) at the Landscape Scale and Its Application in Validation of EOS NPP Products

Substantial research seeks to improve estimates of ecosystem processes and fluxes at a range of scales, notably from the stand scale (<1 km 2) using ecosystem physiology and eddy covariance techniques, to the landscape (∼10 2 km 2) and global (10 8 km 2) scales using a variety of modeling and data acquisition approaches. One approach uses remotely sensed ecosystem properties in the scaling process. This approach combines digital maps of key ecosystem properties such as land cover type, leaf area index, and/or canopy chemistry with quantitative models of biological processes based on these ecosystem properties. Constraints on parametrizing models for global scale applications mean that relatively simple algorithms must be used which are based almost exclusively on satellite-derived inputs, for example, the planned Earth Observation System (EOS)-MODIS Land Science Team model output. Presently, there are limited ways of validating these outputs. At the landscape scale, the opportunity exists to combine remote sensing data with spatially distributed, process-based biogeochemistry models to examine variation in ecosystem processes such as NPP as a function of land cover type, canopy attributes, and/or location along environmental gradients. These process models can be validated against direct measurements made with eddy covariance flux towers and ground-based NPP sampling. Once run and validated over local landscapes, these fine scale models may provide our best opportunity to provide meaningful evaluation (or “validation” in some sense) of simpler, globally applied models. In this article, we 1) provide a biological framework that links ecosystem attributes and ecosystem carbon flux processes at a variety of scales, and summarizes the state of knowledge and models in these areas, 2) describe the need for developing NPP surfaces at a local landscape scale as a means of validating global models, in particular the MODIS NPP product, 3) describe the approach of the BigFoot project to performing such a validation exercise for a series of sites in North America, and 4) present an example using one such model (PnET-II) across diverse vegetation types in a heterogeneous landscape in central North America.

Effect of interannual climate variability on carbon storage in Amazonian ecosystems

The Amazon Basin contains almost one-half of the world's undisturbed tropical evergreen forest as well as large areas of tropical savanna1,2. The forests account for about 10 per cent of the world's terrestrial primary productivity and for a similar fraction of the carbon stored in land ecosystems2,3, and short-term field measurements4 suggest that these ecosystems are globally important carbon sinks. But tropical land ecosystems have experienced substantial interannual climate variability owing to frequent El Nino episodes in recent decades5. Of particular importance to climate change policy is how such climate variations, coupled with increases in atmospheric CO2 concentration, affect terrestrial carbon storage6,7,8. Previous model analyses have demonstrated the importance of temperature in controlling carbon storage9,10. Here we use a transient process-based biogeochemical model of terrestrial ecosystems3,11 to investigate interannual variations of carbon storage in undisturbed Amazonian ecosystems in response to climate variability and increasing atmospheric CO2 concentration during the period 1980 to 1994. In El Nino years, which bring hot, dry weather to much of the Amazon region, the ecosystems act as a source of carbon to the atmosphere (up to 0.2 petagrams of carbon in 1987 and 1992). In other years, these ecosystems act as a carbon sink (up to 0.7 Pg C in 1981 and 1993). These fluxes are large; they compare to a 0.3 Pg C per year source to the atmosphere associated with deforestation inthe Amazon Basin in the early 1990s12. Soil moisture, which is affected by both precipitation and temperature, and which affects both plant and soil processes, appears to be an important control on carbon storage.

TROPICAL DEFORESTATION AND THE GLOBAL CARBON BUDGET

▪ Abstract The CO2 concentration of the atmosphere has increased by almost 30% since 1800. This increase is due largely to two factors: the combustion of fossil fuel and deforestation to create croplands and pastures. Deforestation results in a net flux of carbon to the atmosphere because forests contain 20–50 times more carbon per unit area than agricultural lands. In recent decades, the tropics have been the primary region of deforestation. The annual rate of CO2 released due to tropical deforestation during the early 1990s has been estimated at between 1.2 and 2.3 gigatons C. The range represents uncertainties about both the rates of deforestation and the amounts of carbon stored in different types of tropical forests at the time of cutting. An evaluation of the role of tropical regions in the global carbon budget must include both the carbon flux to the atmosphere due to deforestation and carbon accumulation, if any, in intact forests. In the early 1990s, the release of CO2 from tropical deforestation appears to have been mostly offset by CO2 uptake occurring elsewhere in the tropics, according to an analysis of recent trends in the atmospheric concentrations of O2 and N2. Interannual variations in climate and/or CO2 fertilization may have been responsible for the CO2 uptake in intact forests. These mechanisms are consistent with site-specific measurements of net carbon fluxes between tropical forests and the atmosphere, and with regional and global simulations using process-based biogeochemistry models.