Download (PDF 511 KB): https://research.fs.usda.gov/download/treesearch/15830.pdf
Abstract
A multivariate decision tree model was used to quantify the relative importance of complex hierarchical relationships between biophysical variables and the occurrence of tropical forest fires. The study site is the Huai Kha Kbaeng wildlife sanctuary, a World Heritage Site in northwestern Thailand where annual fires are common and particularly destructive. Thematic layers of several biophysical variables were combined in a GIS with field measurements of fuel loading and stand physiognomy. Canopy vegetation (NDVI), rainfall, geology, elevation, and forest type explain most of the variation in burned surface across the mountainous landscape. Pixels with normalized vegetation difference index values of 0.7284 best discriminated burned and non-burned areas. Less important decision tree model rules identified fire occurrence thresholds for annual rainfall of 1,285 mm, elevation of 700 m, and distinguished between most evergreen and dry deciduous formations. A map of the sanctuary was prepared using GIS to illustrate spatial variation in fire hazard probabilities predicted from the decision tree model.
Parent Publication
Citation
Ofren, Rey S.; Harvey, Edward. 2000. A multivariate decision tree analysis of biophysical factors in tropical forest fire occurrence. In: Hansen, Mark; Burk, Tom, eds. Integrated tools for natural resources inventories in the 21st century. Gen. Tech. Rep. NC-212. St. Paul, MN: U.S. Dept. of Agriculture, Forest Service, North Central Forest Experiment Station. 221-227.
