Multisource spatiotemporal analysis of cropping patterns on dry upland: A case study in Rubaru Sub-district, Sumenep Regency

Abstract

Dry upland agriculture is vital for securing food production in several countries. However, the research on evaluating cropping patterns using remote sensing techniques is completely neglected due to several factors such as the availability of clean imagery and the complexity of the landscape. This research primarily focused on the evaluation of data availability from three different satellite imageries: Sentinel-2, Landsat-8, and MODIS. The consistently high data availability demonstrated by Sentinel-2 established its potential as a reliable source for gap-filling analysis in remote sensing studies. Using a classification model, various land cover types were identified with an overall accuracy of 86.4%, indicating the model's efficiency in accurately classifying these areas. This research also analyzed the detailed cropping patterns, revealing seven distinct temporal cultivation patterns of various crops. This period is strategically positioned between the cultivation of maize, which spans an area of 5,943 ha in December, January, and February, suggesting a potential crop rotation system. The rotation indicated that nearly 83.7% of the cultivated land was planted between maize and shallot throughout the year. The study emphasizes the significance of continuous monitoring and adaptive management in agriculture to ensure sustainability and productivity.

Keywords

• Classification
• Cropping pattern
• Dry upland
• Multisource remote sensing
• Spatio-temporal