Evaluation MODIS and Sentinel-2 Data for Detecting Crop Residue Burned Area

Abstract

AbstractCrop residue burning is one of the major sources of air pollution problems. Various studies have been focused on burned area detection by using 500-m MODIS (MCD64A1) because of the high temporal and ready-to-use product. MCD64A1 data was widely used and well detected in the wildfire area. But the underestimated results in agricultural fields were found in many studies. Because its capabilities of coarse spatial resolution cannot clearly detect small fire patches. This study aims to evaluate the MODIS burned area product and Sentinel-2 burned area for crop residue burning. Higher spatial of burned areas were analyzed by Normalized Burn Ratio (NBR) from Sentinel-2 to compare with MODIS burned area product. The burned areas were overlayed with land uses and sugarcane plots in the production season 2018/2019 in Kalasin province and two districts of Udon Thani province. The estimation of the sugarcane harvesting period was done by NDVI derived from 15-day composite of Sentinel-2 imageries. The result showed that the total burned from Sentinel-2 had around 11 times higher than MCD64A1. The majority of detected burned areas of both sensors were detected in the paddy field. Additional, MCD64A1 has better performace to identify the forest burn area than the Sentinel-2 where as it is not suitable to detect the field crops which has relative smaller size. The study reports that most sugarcane in the study area was harvested in the first half of January 2019 with the cutting of fresh sugarcane, while burned harvest was mostly found in the first half of April 2019. Therefore, this study could fill the gap of present knowledge about crop residue burning over the region.KeywordsBurned areaNormalized Burn Ratio (NBR)Crop residue burningMCD64A1Sentinel-2