A simple method to identify potential groundwater-dependent vegetation using NDVI MODIS

Abstract

The potential groundwater-dependent vegetation (pGDV) in the Iberian Peninsula (IP) was mapped, with a simple method, hereafter referred to as SRS-pGDV, that uses only Normalized Difference Vegetation Index (NDVI) time series retrieved from the Moderate-Resolution Imaging Spectroradiometer (MODIS) Terra V6 product, covering the period February 2000 to April 2018. NDVI was standardized, to minimize the effect of the different land cover types. The extreme drought event of 2004/2005 was used to perform the classification. Considering the water scarcity that affected vegetation in the IP during this event, it was postulated that vegetation showing a high standardized NDVI should be classified as pGDV. Irrigated vegetation and areas with sparse vegetation were eliminated. A cluster analysis was performed, in order to classify the pixels as more/less likely to be pGDV. The results obtained were compared with modeled water table depth, and a propensity of pixels identified as pGDV in areas with low water table depth was clearly observed. However, based on CORINE Land Cover types, some areas identified as pGDV are likely irrigated, such as fruit-tree plantations; this inference is in line with the postulated criterion of vegetation access to sources of water other than precipitation. SRS-pGDV could also be applied to regional studies, using NDVI with a higher spatial resolution.