In peatland wildfires, flaming vegetation can initiate a smouldering fire by
igniting the peat underneath, thus, creating a positive feedback to climate change by
releasing the carbon that cannot be reabsorbed by the ecosystem. Currently, there are
very few models of peatland wildfires at the field-scale, hindering the development of
effective mitigation strategies. This lack of models is mainly caused by the complexity of
the phenomena, which involves 3-D spread and km-scale domains, and the very large
computational resources required. This thesis aims to understand field-scale peatland
wildfires, considering flaming and smouldering, via cellular automata, discrete models
that use simple rules. Five multidimensional models were developed: two laboratory-scale
models for smouldering, BARA and BARAPPY, and three field-scale models for flaming
and smouldering, KAPAS, KAPAS II, and SUBALI. The models were validated against
laboratory experiments and field data. BARA accurately simulates smouldering of peat
with realistic moisture distributions and predicts the formation of unburned patches.
BARAPPY brings physics into BARA and predicts the depth of burn profile, but needs 240
times more computational resources. KAPAS showed that the smouldering burnt area
decreases exponentially with higher peat moisture content. KAPAS II integrates daily
temporal variation of moisture content, and revealed that the omission of this temporal
variation significantly underestimates the smouldering burnt area in the long term.
SUBALI, the ultimate model of the thesis, integrates KAPAS II with BARA and considers the ground water table to predict the carbon emission of peatland wildfires. Applying
SUBALI to Indonesia, it predicts that in El Niño years, 0.40 Gt-C in 2015 (literature said
0.23 to 0.51 Gt-C) and 0.16 Gt-C in 2019 were released, and 75% of the emission is from
smouldering. This thesis provides knowledge and models to understand the spread of
flaming and smouldering wildfires in peatlands, which can contribute to efforts to
minimise the negative impacts of peatland wildfires on people and the environment,
through faster-than-real-time simulations, to find the optimum firefighting strategy and
to assess the vulnerability of peatland in the event of wildfires.