Causal pathways of teleconnections to drought and peatland fires in Indonesian Borneo

Abstract

Fires occurring over the peatlands in Indonesian Borneo have posed devastating impacts on human health, livelihoods, economy and the natural environment, and their prevention requires comprehensive understanding of climate-associated risk. Despite this, attempts to understand the climate risk factors of fires across the whole region and a climatological timescale of 30 years or longer have been very limited, and there are gaps in the understanding of the plausible risk of impactful fires under climate change and natural variability. In this study, gridded products of observational estimates and outputs of ensembles of model simulations are analysed to estimate the present day and future drought and fire risk over Indonesian Borneo and their large-scale climate drivers. Atmospheric drought and fire weather indicators are used for estimating the risk, and a causal network approach is adopted to quantify the strength of these drivers. The interannual variability of fire risk is most prominently driven by the El Niño-Southern Oscillation (ENSO), with drier years corresponding to developing El Niño conditions. High sea surface temperature anomalies over the eastern North Pacific to the east of the Hawaiian Islands (Pacific SST) and westerly wind anomalies over the equatorial Western Pacific during boreal spring precede dry conditions over the Indonesian Borneo by three months. There are large model biases in the rainfall, drought and fire risks in the study region due to the unrealistic amounts of local convections simulated by some of the models, yet they tend to represent the large-scale teleconnections to the rainfall better. Based on an ensemble of CMIP6 climate models, the maximum number of consecutive dry days and its linear association with Pacific SST show significant increase under the SSP585 high-emission scenario towards the end of this century, enabling improved predictability of droughts and fires at seasonal timescale. Using the Unprecedented Simulated Extremes using Ensembles (UNSEEN) methodology based on GloSea6 seasonal hindcast ensembles, an El Niño raises the chance of an unprecedented fire season at present day by a factor of around 2.7, yet a major event is possible under all ENSO phases. This study enables enhanced early warning against fire occurrence and strengthens the scientific ground for longer-term resilience building measures such as rewetting and reforestation.