Assessing the spatial variation of cropping intensity using multi-temporal Sentinel-2 data by rule-based classification

Abstract

The present study was conducted to analyze cropping intensity of four blocks (Mogra-Chinsurah, Polba-Dadpur, Singur and Haripal) of the Gangetic alluvial zone of India using multi-dated Sentinel-2 data in 2018–19 cropping year. It was observed that during peak growing stage all crops ascribed higher Normalized Difference Vegetation Index NDVI values (0.4 to 0.73) and NDVI became as low as 0.06 when the fields were vacant. Sentinel-2 data acquired in the peak crop growing period during each cropping season were carefully selected, and NDVI was computed over the whole study area. Rule-based classification was applied for cropping sequence and cropping intensity classification based on the occurrence and non-occurrence of crops using NDVI threshold (0.4). Sentinel-2 images acquired on 22/10/2018, 6/12/2018, 30/1/2019 and 30/4/2019 were used for masking of trees and non-agricultural area. October 22, January 30 and April 30 imageries demonstrated peak crop growing period during kharif, rabi and pre-kharif seasons whereas December 6 image represented occurrence of no or little crop in the study area. Crop acreage was the highest in Polba-Dadpur block during all the three seasons. The crop–fallow—crop sequence occupied the highest areas (43%) followed by crop–crop–crop sequence (39%). 50% and 39% of the total cultivated land was under 200% and 300% cropping intensities. Overall, accuracies of cropping system and cropping intensity classification were 88.54% and 87.85%, respectively. Sentinel-2 data can be successfully used for cropping system analysis which helps in crop planning and management.