Mapping sub-pixel occurrence of an alien invasive Hyptis suaveolens (L.) Poit. using spectral unmixing technique

Abstract

Hyptis suaveolens (L.) Poit. is an alien invasive plant species in the Indian subcontinent introduced from tropical America, and is an emerging invader of the Himalayan foothills. This study discusses the application of spectral unmixing techniques for detection and discrimination of H. suaveolens invasion at sub-pixel level in the Doon valley and Shiwalik of India using Earth Observation 1 (EO-1) Advanced Land Imager (ALI) satellite data. Multi-date satellite images corresponding to the life cycle of H. suaveolens were processed for geometric and environmental effects. Jeffries–Matusita spectral separability analysis helped in identifying the optimal period (i.e. late October and early November) for better discrimination of H. suaveolens. The applied analysis spectral analytical process (AASAP)-based spectral unmixing of ALI data during the month of November detected H. suaveolens cover as low as 40–50% of the ALI pixel with a producer's accuracy of 76.67%. The area percentage accuracy for abundance class (40–50%) was 82.56%. The information on infested localities in the study area is presented in the form of a fractional cover map of H. suaveolens occurrence. The study outcomes have significance in understanding the pattern of invasion such as dispersal routes, severity of infestation, categories of land use affected, and its monitoring and management.