Simulating Changes in Soil Organic Carbon in Bangladesh with the Denitrification-Decomposition (DNDC) Model

Developed countries' growing awareness of greenhouse gas (CO2, CH4, N2O) emissions from agricultural soils has led to an increased interest in the management of soil organic matter (SOM), which now extends to developing countries, including Bangladesh. Bangladeshi agriculture follows a largely rice-based cropping rotation, for which insufficient site-specific information regarding gas emissions exists to identify temporal variability of SOM content.;The objective of this study was to evaluate the applicability of the 'Denitrification-Decomposition' model (DNDC, version 9.3) as a tool to better understand SOC trends in tropical agriculture. DNDC was used to simulate gas emissions from 1948 to 1969 and 1981 to 2007, under farm management practices prevalent in the Dinajpur district of Bangladesh. Forty-nine years of historical daily precipitation and temperature data were used for simulation with DNDC, such that both aerobic and anaerobic conditions were experienced in any given year. A "summer rice - monsoon rice - wheat" cropping pattern was used.;As the input parameters of annual precipitation and flooding duration would likely affect DNDC-simulated results, model outputs were categorized on the basis of the magnitude of these parameters. In each categorization scheme the output data were sorted either based on (i) mean, (ii) probability of exceedance, or (iii) standard deviation of annual precipitation or flooding duration. An analysis was then conducted of correlations among input and output variables.;Relationships between simulated variables like CO2 emissions, CH4 emissions, and change in SOC content, and input variables such as annual precipitation and flooding duration were generally similar under both of categorization schemes. In high precipitation years changes in SOC content showed a negative correlation (r = 0.90, P ≤ 0.05) with CO 2 emissions, and a positive correlation with CH4 emissions (r = 0.85, P ≤ 0.05), highlighting the importance of studying gas emissions as part of the net C balance embedded in DNDC.;When categorized according to annual precipitation, CO2 and CH4 emissions were negatively correlated; however, no significant relationship existed when emissions data were categorized on the basis of flooding duration. This discrepancy might arise from the way in which DNDC computes the soil's net C balance. In physical systems, CH4 emissions from paddy fields have an important effect on SOC; however, DNDC calculates CH4 emissions based on available organic C generated by the decomposition sub-model, but the net change in SOC is only balanced according to the CO 2 gas emissions calculated by decomposition sub-model. Thus, the CH 4 emission calculated by the fermentation sub-model is not included as a loss of SOC in the C balance. The consequence of this in the output data was a steadily increasing SOC associated with the increase in CH4 emissions from the simulated soil system.;In order to more accurately model the soil carbon balance in tropical agricultural systems with flooded soils, DNDC should be modified to take into consideration C lost through CH4 emissions in addition to those lost as CO2. DNDC might then be used in sensitivity analysis for different farm management practices under paddy-based cropping systems. Physical experimental analysis is also important for validation of the modelling work. This study showed that DNDC can serve as a rough tool to represent change in the SOC content under Bangladeshi agricultural practices. Some modifications of DNDC, however, would be desirable to make it better suited for future work of this kind.