Forest fire risk estimation from time series analisys of NOAA NDVI data

Abstract

The values of the Normalized Difference Vegetation Index obtained from NOAA Advanced Very High Resolution Radiometer (AVHRR) have often been used for forestry application, including the assessment of fire risk. Forest fire risk estimates were based mainly on the decrease of NDVI values during the summer in areas subject to summer drought. However, the inter-annual variability of the vegetation response has never been extensively taken into account. The present work was based on the assumption that Mediterranean vegetation is adapted to summer drought and one possible estimator of the vegetation stress was the inter-annual variability of the vegetation status, as reflected by NDVI values. This article presents a novel methodology for the assessment of fire risk based on the comparison of the current NDVI values, on a given area, with the historical values along a time series of 13 years. The first part of the study is focused on the characterization of the Minimum and Maximum long term daily images. The second part is centered on the best method to compare the long term Maximum and Minimum with the current NDVI. A statistical index, Dynamic Relative Greenness, DRG, was tested on as a novel potential fire risk indicator.