Assessment of Wildfire Hazards with a Semiparametric Spatial Approach

Rural households in agricultural economies are vulnerable to several environmental risks such as fires, floods, and droughts that may affect their productivity in whole or in part. These hazards are especially relevant in developing countries where farmers have few or no access to traditional risk-transfer techniques, such as insurance and finance, and where low governmental investments in rural infrastructure, risk assessment techniques, or early warning systems makes the aftermath of such hazards more expensive and results in slower recovery for those who are affected. In this paper, we use historical satellite data (Terra) of burned areas in South America to fit a semiparametric spatial model, based on kernel smoothing and on a nonlinear relationship between average time between events and damage, to assess the environmental hazard affecting the land’s productivity. The results were twofold: first, we were able to develop a spatial assessment of fire hazard, and second, we were able to evaluate how much a farmer may lose in terms of productivity per hectare due to the environmental hazard. The methodology may be easily adapted to other world regions; to different environmental hazards such as floods, windbreak, windthrow, or related land-use changes; or to integrate various environmental hazards simultaneously, as long as they can be monitored via remote sensing (e.g., satellite imagery, aerial photographs, etc).