Continuous land cover and land use maps for the East Asian Monsoon region from 1982 to 2015

Keywords: Land cover and Land use, East Asia Summer Monsoon, GIMMS NDVI3g, phenology, random forest
Session Type: Virtual Paper Abstract
Day: Friday
Session Start / End Time: 2/25/2022 08:00 AM (Eastern Time (US & Canada)) - 2/25/2022 09:20 AM (Eastern Time (US & Canada))
Room: Virtual 21

Abstract

Recent intensification of the East Asian summer monsoon (EASM) has resulted in more frequent floods and intense dry spells, influencing some 1.6 billion people. Attributing the causes of EASM variability is thus paramount. The land cover and land use change (LCLUC) could perturb surface incoming heat energy, heat flux partition, and terrestrial carbon storage, resulting in changing regional climate. However, the impact of LCLUC on the EASM remains unclear due to scarcity of temporally continuous LCLU maps of the monsoon region. To address this limitation, in this study, we produced annual LCLU maps of the EASM region for 34 years (1982-2015), using random forest classification of 19 phenological information derived from AVHRR GIMMS NDVI3g and terrain information from ALOS World 3D-30 m Digital Surface Model. The random forest classification was trained by land cover information with unchanged pixels from the 2001-2010 MODIS product and performed for each subregion based on the FAO global ecological zones. For assessing the reliability of the classified maps, we conducted visual interpretation of very-high resolution images in 2015 from Google Earth. The overall accuracy of classified maps was 73%, which is 7% higher than that of MODIS map in the same year. Based on our classified maps, this study demonstrates that croplands and forests statistically significantly increased in the EASM region, and the dominant transition was from croplands to forests over last three decades. These classified maps are valuable in the climatology study of land and atmosphere interaction in the EASM region.