Correlation analysis of different optical remote sensing indices for drought monitoring: a case study of Canton Sarajevo, Bosnia and Herzegovina.

Abstract

Droughts are second to hurricanes the world's most costly weather events. Damage caused by droughts in certain countries is measured in tens of billions of dollars per year. Timely detection of drought and prediction of its occurrence has the potential to reduce costs and save a large number of people from its consequences. Numerous methods that serve this purpose exist in scientific research and practice. One group of drought monitoring methods belongs to the field of remote sensing, where it is possible to monitor drought indicators over large areas in almost real-time through satellite images. This paper is focused on the optical indices of remote sensing calculated by raster algebra. The intention was to reach conclusions about the quality of individual indices used for the Canton Sarajevo area in Bosnia and Herzegovina for each month of August in the period 2008-2021 through correlational and qualitative analysis and the use of meteorological indicators. Among the used indices, NDVI (normalized difference vegetation index) and NMI (normalized moisture index) proved to be the most reliable, and their mutual correlation was very strong (r = 0.99).