Predicting climate change impacts to peatland soil moisture in Southeast Asia

Soil moisture is a key hydrologic variable that determines peat flammability and predicts burned area. In recent decades, there has been a rise in deadly peat fires across Southeast Asia, indicating the presence of dry conditions. This has largely been attributed to the extensive deforestation, drainage, and conversion to agricultural use that has occurred in the region. Climate also plays a role in mediating soil moisture, and the most severe fire years have previously only occurred when there are droughts during strong El Niño years. Thus, climate change threatens drier peat soil moisture conditions which would increase peat fire risk. Here, we assess these potential impacts by modeling soil moisture responses to predicted climate change. To overcome the lack of regional-scale data for hydrologic variables and peat properties necessary to parametrize a physical model, we used for a statistical modeling approach. Specifically, we used an artificial neural network to relate remotely sensed observations of soil moisture (SMAP) to climate reanalysis forcings (ERA5) and other datasets that characterize peatland degradation such as tree cover and canal density. After training the neural network on data from 2015-2020, we then compared moisture regimes under recent and future climate from state-of-the-science regional climate model projections (CORDEX-CORE) under RCP 8.5. Our findings suggest that reduced precipitation and increased evaporative demand, as predicted by the regional climate models, may cause significantly drier soil moisture regimes in the future. Future mean dry season soil moisture is found to be similar to that during 2015 and 2019 El Niño years, suggesting higher baseline fire risk. We further explore geographic differences in soil moisture responses, as mediated by differences in climate sensitivity between land use types.