Land Use Classification: A Surface Energy Balance and Vegetation Index Application to Map and Monitor Irrigated Lands

Irrigated agriculture consumes the largest share of available fresh water, and awareness of the spatial distribution and application rates is paramount to a functional and sustainable communal consumptive water use. This remote sensing study leverages surface energy balance fluxes and vegetation indices to classify and map the spatial distribution of irrigated and non-irrigated croplands. The purpose is to introduce a classification scheme applicable across a wide variation in regional climate and inter-growing seasonal precipitation. The rationale for climate and inter-growing seasonal adaptability is founded in the derivation and calibration of the scheme based on the wettest growing season. Therefore, the scheme becomes a more efficient classifier during normal and dry growing seasons. Using empirical distribution functions, two indices are derived from evapotranspiration fluxes and vegetation indices to contrast and classify irrigated croplands from non-irrigated. The synergy of the two indices increases the classification proficiency by adding another classifying layer which re-characterizes misclassified croplands by the base index. The scheme was applied to a region with wide climate variation and to multiple years of growing seasons. The results presented, in cross validation with groundtruth, show an accurate and consistent approach to classify irrigation with overall accuracy of 92.1%, applicable from humid to semi-arid climate, and from dry to normal and wet growing seasons.