Estimation of forest fire hazard is essential in order to protect and manage the forest against unforseen damage. This will invaluably and effectively support forest protection and planning of good judgment and charting correct future programs. Quantification 2 method carried out the analysis using data observed in a forest fire damage site in Japan. Data include aerial photographs, existing forest record files, topographic maps and Landsat Thematic Mapper (TM) data.
Estimation of sites considered as fire hazard area exhibited 87% accuracy when compared the outcome with the fire damage data. Fire hazard area classified into three-level classification, which defines high fire hazard, moderate fire hazard and light fire hazard, revealed that high fire hazard attributed an accuracy of 75%, while estimate for moderate and light fire hazard showed 34% and 53% accuracy. Combining moderate and light fire hazard as low fire hazard, and high fire hazard into two-level classification, however, showed higher accuracy of 72% and 70%, respectively.
Moreover, the study underscored the significance of Normalized Difference Vegetation Index (NDVI) and other TM band data in the estimate. Analysis disclosed that NDVI and TM band 4 could produce an accuracy of at least 62%. Additional employment of TM data could further produce a relatively higher degree of agreeability. Reduction in volume of data is therefore possible while keeping the reliability of the estimates in order.