Difficulties in using spectral properties to map irrigated areas in a temperate climate: A case study of potatoes in England

Abstract

Irrigation in England is supplemental to rainfall and only used on a small proportion of the cultivated land, notably on high value vegetable and potato crops. However, it is a significant water user as most of the irrigated area is located in the driest part of England. The existing data on irrigated areas are based on government and industry surveys. Recently these datasets have been used with Geographic Information Systems (GIS) to produce irrigated maps, but these can only be published at catchment level due to confidentiality constraints on the datasets. To assess the possibility of using remote sensing data for mapping the irrigated area, one Landsat image for the summer 2003 was used to compare the spectral signature between irrigated and non-irrigated potato fields in the East of England. ISODATA algorithm was used to perform unsupervised classification, and 50 spectral classes were created. A ground truth dataset was then used to identify the most representative spectral class for irrigated and nonirrigated fields. The result showed that categories both fall into the same spectral class, suggesting there are no significant differences between their spectral properties. Therefore, using satellite imagery may not yet be an appropriate method or need more research for mapping irrigated area in temperate climates such as England. The summer rainfall reduces the water stress differences between irrigated and non-irrigated potato fields such that these satellite sensors cannot yet differentiate the crops.