DNDC: A process-based model of greenhouse gas fluxes from agricultural soils
Introduction
Agricultural soils can act as a source or a sink for the three greenhouse gases, nitrous oxide (N2O), carbon dioxide (CO2) and methane (CH4). The fluxes of these gases derive from biological processes and depend on many factors that sometimes have complex feedbacks and interactions. Understanding the impacts of human activities on greenhouse gas emissions from productive soils is vital for mitigating negative effects on climate change while continuing to feed the Earth's increasing population.
As greenhouse gas emissions from soils are the result of microbial processes, the emissions exhibit a high degree of temporal and spatial variability. Direct measurement of greenhouse gas emissions for inventory purposes is impractical as it would require many measurements to be made over large areas and for long periods of time. Many countries use the IPCC default methodology for calculating N2O emissions from agricultural soils for their national inventories. This method simply assumes a fixed proportion (the “emission factor”) of the applied N is emitted as N2O. The emission factor was deduced from a limited number of observations but represents an average value over all soil types, climate conditions and management practices. As N2O emissions are highly sensitive to all these factors there is a high degree of uncertainty associated with the emission factor. In addition, the emission factor method does not account for many of the management practices that could potentially reduce N2O emissions (e.g., fertiliser timing, splitting fertiliser applications, use of nitrification inhibitors, depth of application). For these reasons the development of a more process-based approach is desirable.
The development of a process-based model not only allows the simulation of agricultural greenhouse gas emissions at a range of scales up to national or global level, but also the exploration of potential mitigation strategies. In addition, because the DNDC model simulates the interactions between the different soil processes, it is possible to determine how strategies that reduce the emission of one gas will affect emissions of the other gases, and whether there may be other adverse consequences (e.g., reduced production or increased nitrate leaching).
The DNDC model was originally developed to simulate N2O emissions from cropped soils in the US (Li et al., 1992a, US EPA, 1995). It has since been used and expanded by many research groups covering a range of countries and production systems. In this paper we describe the DNDC model and how it has been developed, validated and used, including regional and national scale simulations, sensitivity analysis and scenario assessment.
Section snippets
Model description
As discussed in the introduction DNDC was first used to model N2O emissions from agricultural soils in the US (US EPA, 1995). Since its initial development (Li et al., 1992a), other researchers have modified the model to adapt it to other production systems and many of these modifications have been incorporated into later versions of the DNDC model. DNDC consists of five interacting sub-models: thermal–hydraulic, aerobic decomposition, denitrification, fermentation, and plant growth (which
Model validation
Validation against experimental data is an essential part of model development. If experimental measurements agree well with model predictions, there is increased confidence that the model is correctly simulating the underlying processes. On the other hand, in cases where the model fails to predict the measurements this can help identify processes that the model simulates poorly.
DNDC has now been used to simulate various cropping, grazing and forest systems in many countries. Table 1 lists some
Sensitivity analyses
Sensitivity analysis involves testing the model performance as various inputs are changed. This helps determine which inputs are having the greatest effect on the predicted emissions and whether the model has captured observed differences in emissions under different management strategies. Identifying input parameters that have a large effect on predicted emissions can be used to quantify and/or reduce the uncertainty in the model predictions arising from uncertainty in the input parameters.
Model uncertainty
Sensitivity analyses can be used to estimate the degree of uncertainty in the model predictions resulting from imperfect knowledge of the input parameters. This is particularly relevant for regional scale simulations where inputs are derived from GIS databases. These uncertainties can be estimated using Monte Carlo simulations, in which a large number of possible scenarios are generated using random values (within a specified range) for each input parameter. The set of predicted values can then
Regional inventories
DNDC can be used to estimate greenhouse gas emissions at regional or national scales. At the regional scale, the region is first divided into smaller units (“cells”) that can be considered to have uniform soil and climate properties. Second, climate and the range of each soil property within the unit are determined, usually from GIS databases. Typical farm management practices for the major farm types within the region are then defined, and the area under each farming system within each cell is
Scenario analyses
Scenario analysis involves using the model to explore the potential impacts of changes to production systems. In some ways this is similar to sensitivity analysis, except that usually combinations of changes are compared rather than just the effects of individual parameters. There are two major areas of interest for scenario analysis using DNDC. One is the effects of climate variability and potential climate change on greenhouse gas emissions, while the other is the potential for different
Discussion
The biogeochemical processes that produce greenhouse gas emissions from soil are complex and involve many feedback mechanisms. It is therefore difficult to develop simple empirical models that can reliably predict greenhouse gas emissions over a range of different soil conditions and management practices. By seeking to simulate the underlying processes, models such as DNDC are better able to predict emissions from a wide range of systems. Already DNDC has been adapted to simulate cropping,
Conclusion
DNDC is a process-based model that simulates the soil biogeochemical processes leading to greenhouse gas emissions from soil. Originally developed to model N2O emissions and SOC levels in US cropping systems, it has subsequently been adapted to model crop, pasture, rice paddy, and forest systems in a number of countries across the world.
As a process-based model DNDC is a useful tool both for modelling the environmental impacts of agricultural management systems (including feedbacks) and for
Acknowledgements
This research was funded by the Foundation for Research, Science and Technology (FRST), New Zealand. Thanks to Kevin Tate for his comments on an early draft of this paper and to Anne Austin for internal editing.
References (57)
- et al.
Validation of DNDC for 22 long-term N2O field emissions measurements
Atmospheric Environment
(2007) - et al.
Development and application of a mechanistic model to estimate emissions of nitrous oxide from UK agriculture
Atmospheric Environment
(2002) - et al.
Modelling N2O emission from a forest upland soil: A procedure for an automatic calibration of the biogeochemical model Forest-DNDC
Ecological Modelling
(2007) - et al.
Simulation of fluxes of greenhouse gases from European grasslands using the DNDC model
Agriculture Ecosystems and Environment
(2007) - et al.
Modeling nitrous oxide emissions from agriculture: a Florida case study
Chemosphere
(1994) - et al.
Simulating trends in soil organic carbon in long-term experiments using the DNDC model
Geoderma
(1997) - et al.
Modeling nitrate leaching with a biogeochemical model modified based on observations in a row-crop field in Iowa
Ecological Modelling
(2006) - et al.
Modeling effects of temperature and precipitation on carbon characteristics and GHGs emissions in Abies fabric forest of subalpine
Journal of Environmental Sciences–China
(2008) - et al.
Assessing productivity and carbon sequestration capacity of Eucalyptus globulus plantations using the process model Forest-DNDC: calibration and validation
Ecological Modelling
(2006) - et al.
Disaggregated greenhouse gas emission inventories from agriculture via a coupled economic-ecosystem model
Agriculture Ecosystems and Environment;
(2006)