Exploring Multi-Driver Influences on Indonesia's Biomass Fire Patterns from 2002 to 2019 through Geographically Weighted Regression

• Autor(a) principal: Wee, Shi Jun
• Data de Publicação: 2024
• Outros Autores: Park, Edward, Alcantara, Enner [UNESP], Lee, Janice Ser Huay
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.1007/s41651-023-00166-w; https://hdl.handle.net/11449/305519
• topic: El Niño–Southern Oscillation; GWR; Kalimantan; LCLUC; Sumatra; Tropical forests

In the last two decades, Indonesia recorded the most biomass fires in Southeast Asia. These fires release massive amounts of carbon and smoke haze, causing significant economic and health impacts in the region. Numerous studies have used statistical methods to investigate the factors contributing to fire occurrence in Indonesia. However, they often overlook heterogeneity in the relationship between each driver and fire occurrence, and do not use a fixed interval time-series approach to track year-to-year variations in each variable’s influence. To address these limitations and gain a better understanding of the complex and multifactorial nature of biomass fires in Indonesia, we constructed annual Geographically Weighted Regression models to analyze fire density from 2002–2019. Our models explain up to 57% and 46% of the variability in fire density at Kalimantan and Sumatra, respectively. Forest loss was the dominant driver of fire across Kalimantan (mean = 61% of total area analyzed) and Sumatra (mean = 59%), while peat was constrained to severely degraded peatland areas. Dry conditions were highly influential in El Niño years and its impacts were concentrated in degraded areas extremely vulnerable to fire. There was no distinct trend in each variable’s influence on fire over the investigated period as forest loss consistently emerged as the dominant driver. A notable exception occurred in peatland areas in Sumatra, where there was a gradual shift from forest loss to peat (an indicator of the extent of degradation) as the dominant driver. Overall, our analysis revealed significant spatial and temporal variation in each driver’s influence on fire occurrence. These findings have significant implications for mitigation strategies and monitoring efforts, as the primary driver of fires in fire-prone areas varies by region.