A drought-based predictor of recent haze events in western Indonesia

Indonesiaʹs fire and haze problem is reviewed, and a model quantifying the relationship between drought and haze from biomass burning in western Indonesia is presented. Visibility observations from weather stations in Sumatra and Kalimantan were used as a haze indicator. The Drought Code component of the Canadian Forest Fire Weather Index System was used as a drought indicator. Using meteorological data from 1994 to 1998, we obtained regional haze and drought signals for western Indonesia. Nonlinear regression analysis was performed between the two signals to obtain a model of haze potential based on the Drought Code. Using the curvature properties of the nonlinear model, we estimated that severe haze is likely above a threshold Drought Code of 388.2. Using this threshold value, we propose four levels of drought that can be used operationally as an early warning tool in managing Indonesiaʹs serious haze problem.