Assessing native desert vegetation recovery in a war‐affected area using multispectral and hyperspectral imagery: a case study of the Sabah Al‐Ahmad Nature Reserve, Kuwait

Abstract

AbstractThe primary aim of this project was to assess vegetation changes in the Sabah Al‐Ahmad Nature Reserve, Kuwait, which is a war‐affected area following the Iraqi invasion in 1990. After the liberation in 1991, several portions of the reserve were under a restoration program. Remote sensing has been used as a tool to assess vegetation and land cover changes. We studied the feasibility of three common methods—the Mahalanobis distance (MD), maximum likelihood (ML), and support vector machine (SVM)—for classification of the multispectral imagery (Landsat) and hyperspectral (Hyperion). The reserve was also compared to the demilitarized zone (DMZ) located at Umm Nigga at the northern border of Kuwait, as it had recovered naturally, to distinguish between an autogenic recovery and a restored area. We discovered that the location was damaged during the military occupation, but a rapid recovery of the vegetation was then recorded in the reserve after the war from less than 1% measured in 1991 to 42% in 1998. Then, the vegetation cover significantly decreased in 2002 (26%) and slightly increased in 2013 (28%). We found that similar rapid increase in vegetation cover occurred in most parts of the reserve that was under the restoration program, and in the DMZ, which was naturally recovered. We concluded that remote sensing technologies are helpful tools in understanding the process of vegetation recovery as it provides information on location and timing of recovery, particularly where optimal condition exists.