Abstract
Fertilized rice paddy soils emit methane while flooded, emit nitrous oxide during flooding and draining transitions, and can be a source or sink of carbon dioxide. Changing water management of rice paddies can affect net emissions of all three of these greenhouse gases. We used denitrification–decomposition (DNDC), a process‐based biogeochemistry model, to evaluate the annual emissions of CH 4 , N 2 O, and CO 2 for continuously flooded, single‐, double‐, and triple‐cropped rice (three baseline scenarios), and in further simulations, the change in emissions with changing water management to midseason draining of the paddies, and to alternating crops of midseason drained rice and upland crops (two alternatives for each baseline scenario). We used a set of first‐order atmospheric models to track the atmospheric burden of each gas over 500 years. We evaluated the dynamics of the radiative forcing due to the changes in emissions of CH 4 , N 2 O, and CO 2 (alternative minus baseline), and compared these with standard calculations of CO 2 ‐equivalent emissions using global warming potentials (GWPs). All alternative scenarios had lower CH 4 emissions and higher N 2 O emissions than their corresponding baseline cases, and all but one sequestered carbon in the soil more slowly. Because of differences in emissions, in radiative forcing per molecule, and in atmospheric time constants (lifetimes), the relative radiative impacts of CH 4 , N 2 O, and CO 2 varied over the 500‐year simulations. In three of the six cases, the initial change in radiative forcing was dominated by reduced CH 4 emissions (i.e. a cooling for the first few decades); in five of the six cases, the long‐term radiative forcing was dominated by increased N 2 O emissions (i.e. a warming over several centuries). The overall complexity of the radiative forcing response to changing water management could not easily be captured with conventional GWP calculations.