Upscaling and Downscaling of Regional Methane Sources - rice agriculture as a case study

Abstract

Methane (CH4) is a potent greenhouse gas. Wetland rice fields constitute one of the major anthropogenic CH4 sources but the source strength is surrounded by a large uncertainty. The work presented in this report aims at reducing the uncertainty in emissions from wetland rice fields by making independent estimates of regional CH4 source strengths: "up-scaling" from the soil-rice ecosystem perspective and "down-scaling" from the atmosphere perspective. Case studies in Java and the Philippines described the upscaling from point to regional scale, the Chinese case study focused on the regional to national scale. A process-based field scale model for CH4 emissions from rice paddies was made, validated and coupled to a Geographic Information System to scale up regional CH4 emissions from rice paddies. Potential land use changes for Java and China were quantified with a land use change model and predicted changes were evaluated using the CH4 emission model, proxy methods and emission factors. Simultaneously, trend analysis of rice field emissions indicates that (1) rice yield increases have usually not resulted in increased methane emissions per unit of harvested area and, (2) global annual emission from rice fields may be considerably lower than generally assumed, partly because of recent decreases in the use of organic amendments. Rice agriculture is a dynamic activity, economical changes and technological advances influenced the emission levels and will continue to do so. This makes fixing the CH4 source strength of rice paddies difficult and increases uncertainties. Downscaling from the global scale to regional scale was pursued by inverse modelling of the sources and sinks of atmospheric CH4. Improvements were made in the availability of measurements and a priori information, the accuracy of the atmospheric transport model, and the applied inverse modeling technique. Constraining the global rice CH4 source using inverse modelling, was explored by comparing a standard rice emission scenario (80 +- 50 Tg yr-1) and a low rice emission scenario (30 +- 15 Tg yr-1) but an unambiguous answer could not be obtained. To reduce uncertainty in other CH4 sources, ice core data and simulations of pre-industrial methane were used to estimate the pre-industrial natural wetland source strength at 163 +-30 Tg (CH4) yr-1, current emissions being possibly 10% smaller. A combination of upscaling and downscaling is at present not able to validate rice emission estimates at the intermediate scale in the Asian region. Regional scale emission measurements are essential to validate and improve upscaling methods and regional CH4 budgets.