Agricultural Crop Pattern Mapping and Change Analysis at a Sub-district Level Using Landsat satellite data: evidence from South-eastern Region, Bangladesh

Abstract

The study first time identified and analyzed winter season agricultural crop patterns (ACP) derived from Land use (LU) maps in between 2010 to 2019 of southeastern regions of Chittagong, Bangladesh. ACP identification was a challenging task in the worldwide research relevant to crop-related suidies. To overcome this, we have considered frequently used traditional unsupervised classifier, such as K-means clustering algorithm technique. This has been applied on 30m pixel Landsat satellite reflectance images to identify crop pattern of the study area using the ENVI 5.3 and ArcGIS 10.8 software's, respectively. Multiple crops with seven classes were identified with the validation of in-situ ground-truth data and Google Earth (GE) images. The overall accuracy and kappa coefficient values were found at 81.96% and 0.79. respectively. The results suggest a significant variation of crop patterns in the study area and in recent time. the area largely dependent on mixed irrigation approach. Moreover, the crop pattern change was observed in the smdied period as mixed crop 19% (9282.17 ha). Lentils (Pelon) 24.80% (11594.38 ha). Melon (Bangi) 22.37% (10461.08 ha), Chilis 17.90% (8367.48 ha), Paddy rice, unused land, and other crops, respectively. Among them, Lentils (Pelon) and Melon (Bangi) are identified as two common crops followed by mixed crops category, cultivated in the winter season as it required less irrigation compared to paddy rice area.