A multinomial logistic model-based land use and land cover classification for the South Asian Association for Regional Cooperation nations using Moderate Resolution Imaging Spectroradiometer product

The land-use and land-cover (LULC) mapping is crucial for the management and planning of a nation. The land use and land cover are the two separate terminologies often used interchangeably. We mapped LULC for the nations of South Asian Association for Regional Cooperation represented by Afghanistan, Bangladesh, Bhutan, India, Nepal, the Maldives, Pakistan and Sri Lanka covering an area of ca. 5,100,000 km(2). Mapping LULC is a cumbersome process for a larger area; nevertheless, in recent decades the satellite-based Earth observation has provided a great impetus and various approaches have been adopted to map LULC. In this study, we used the multinomial logit model (Mnlogit) to classify Moderate Resolution Imaging Spectroradiometer (MODIS)-derived time-series data for mapping LULC. The MODIS-archived images for the years 2004-2007 and 2016 were used to map LULC for corresponding years. For the same years, maps were available from IGBP, FAO and NR Census, India, for comparison; hence, these years were selected. The classification of images for each of the years was found to be more consistent and accurate in comparison with the existing LULC maps of IGBP and FAO, and hence a robust approach. The method can be adopted to map global resources. We achieved overall accuracy ranging from 76 to 82% with kappa statistic (K-hat) ranging from 0.64 to 0.73, which is better than previous works. The study demonstrates an effective way to utilize remote sensing data for mapping LULC for a larger area with reasonable accuracy.

Automated summary

This study introduces a method for mapping land use and land cover using remote sensing data, utilizing a multinomial logit model (Mnlogit) to classify MODIS-derived time-series data and achieving high accuracy. The results demonstrate that this method can effectively map LULC for a large area.