Mapping multi-year cropping patterns in small irrigation districts from time-series analysis of Landsat TM images

Abstract

Although it is well known that one of the main advantages of remote sensing from satellites is the synoptic and repeated collection of data, few examples of its application to map multi-year cropping patterns and/or crop rotations have been developed. Those examples reveal different conceptions of the term cropping pattern and also show the difficulty or limitations of these methods in mapping cropping patterns or crop rotations in more than two successive years. The present paper proposes a method that allows long term cropping patterns to be mapped using time-series of crop maps derived from supervised classification of remote sensing data. These maps allow implicit spatial and temporal relationships between the crops grown in an agricultural area. The cropping pattern is understood to be the spatial distribution of associations between crops or crops and uncultivated land in the same fields (although not in a particular order or sequence) during the period of the analysed time-series. The method was applied to map the multi-year cropping patterns in the Flumen irrigation district (33,000 ha), which is located in the Ebro Valley (northeast Spain). A 7-year time-series (1993, 1994, 1996, 1997, 1998, 1999 and 2000) of crop maps derived from Landsat 5 TM and Landsat 7 ETM+ images was used to obtain the yearly crop maps from which to derive the multi-year cropping patterns. To achieve this, the proposed method looks for the spatial and temporal relationships between the main crops present in the study area in their typical locations (areas where the crops are most frequently located) and the alternate crops that have been in these locations in any of the years of the analysed time-series. GIS overlay analysis operations—mainly cross-classification operations—are applied to derive the spatial and temporal relationships between crops.