Multi-Decadal Remote-Sensing Analysis of Irrigated Areas in the Lower Rio Grande Valley, New Mexico

A time series of estimates of irrigated area was developed for the Lower Rio Grande valley (LRG) in New Mexico from the 1970s to present day. The objective of the project was to develop an independent, accurate, and scientifically justifiable evaluation of irrigated area in the region for the period spanning from the mid-1970s to the present. These area estimates were used in support of groundwater modeling of the LRG region, as well as for other analyses. This study used a remote-sensing-based methodology to evaluate overall irrigated area within the LRG. We applied a methodology that involved the normalization of vegetation indices derived from satellite imagery to get a more accurate estimation of irrigated area across multiple time periods and multiple Landsat platforms. The normalization allows more accurate evaluation of vegetation index data that span several decades. An accuracy assessment of the methodology and results from this study was performed using field-collected crop data from the 2008 growing season. The comparisons with field data indicate that the accuracy of the remote-sensing-based estimates of historical irrigated area is very good, with rates of false positives (areas identified as irrigated that are not truly irrigated) of only about 4%, and rates of false negatives (areas identified as not irrigated that are truly irrigated) in the range of 0.6-2.0%.