Spatial and temporal distribution of pontederia crassipes mart. in Chitwan Annapurna Landscape (CHAL) area, Nepal using satellite imageries

Abstract

Invasive alien plant species (IAPS) are escalating in Nepal and harming species, communities, environment and human health. In recent years, remote sensing has become widely used method for mapping IAPS as remote sensing data provide broad aerial coverage of intermittently inaccessible area. Pontederia crassipes Mart. (formerly Eichhornia crassipes) is one of the invasive species established in wetlands of Nepal. It is a floating aquatic macrophyte, commonly known as water hyacinth. It is one of the members of family Pontederiaceae. The main objective of the research work was to study spatial and temporal distribution of P. crassipes from 1990 to 2020 A.D. at the gap of 10 years in Chitwan Annapurna Landscape (CHAL), Nepal. World View2 and Landsat series imageries were used to map P. crassipes. The spatial resolution of

Landsat image is 30 m × 30 m and of World View-2 image is 2 m × 2 m. Iso cluster

unsupervised

classification and maximum likelihood classification (supervised classification) by using training sample were done in Arc MAP 10.5. Knowledge-based classification and fusion of several terrain information including digital elevation model (DEM), water resources, human activities, and accessibilities like several ancillary data were integrated in ERDAS IMAGINE 2014. In 2018 A.D. it was found that 27.58 km vi 2

(0.086 %) of CHAL was infested by P. crassipes. Analyzing the map of 1990 A.D., it was found that P. crassipes was only in 9.6 km 2 (0.030%). In 2000 A.D., infestation area decreased to 2.93 km 2 (0.0092). In 2010 A.D. the area of infestation increased to 23.44 km 2 (0.073%). Again, that increased to 37 km 2 (0.116%) in 2020 A.D. World View-2 imageries were also used in small area of interest (AOI). The same size of AOI was clipped in Landsat imageries. The overall accuracy varied from 81.25 to 86.96 % in World View-2 imageries. While overall accuracy in Landsat imageries was 62.96 to 68 %. The Kappa coefficient in World View-2 varied from 0.61 to 0.72, which is good accuracy. But kappa coefficient in Landsat is less and varied from 0.21 to 0.29. Not only World View-2 image (high resolution sensor) but also Landsat image (moderate resolution) can be used to map invasive species. Keywords: Pontederia crassipes, World View-2, Landsat, Remote sensing, Iso cluster unsupervised classification, Maximum likelihood classification, Knowledge-based classification