Patrícia Páscoa (Portugal) 1; Célia M. Gouveia (Portugal) 1,2; Ana Russo (Portugal) 1; Cathy K. Besson (Portugal) 1
1 - Instituto Dom Luiz; 2 - Instituto Português do Mar e da Atmosfera
Vegetation activity depends and may be limited by water availability, temperature and radiation, among other factors. In arid and semi-arid areas, water scarcity may exist due to precipitation seasonality or the frequent occurrence of droughts. Thus, to overcome surface water scarcity, some species access water from other sources, such as groundwater.
The goal of this work is to map potential groundwater dependent vegetation (GDV) in the dry and semi-arid regions of the Iberian Peninsula (IP), using only the Normalized Differences Vegetation Index (NDVI). The climate in the IP ranges from humid to semi-arid, and drought episodes are frequent. GDV have been identified in some areas of the IP, using in situ methods. Elsewhere in semi-arid areas the existence of aquifers can be an indicator of the possible occurrence of GDV. This has not yet been analysed.
Satellite data, namely NDVI, has been used to map potential GDV in several world regions, due to its ability to detect vegetation greenness. Vegetation stress caused by drought events may be identified with this index, since negative NDVI anomalies reflect the lack of greenness shown by vegetation experiencing water scarcity in comparison with the normal behaviour of vegetation over the analysed regions. For this reason, vegetation showing positive NDVI anomalies during drought events are less affected by rain scarcity and are likely GDV. A cluster analysis was performed on standardized NDVI, during the extreme drought event of 2005. NDVI retrieved from MODIS Terra V6 was used. Data are available since 2000 with 250m of spatial resolution. Areas classified as humid or sub-humid were excluded, as well as areas with a mean NDVI value lower than 0.3 in August, since this such a low value likely represents dead vegetation. The Corine Land Cover map was used to exclude artificial, irrigated areas, and rice fields.
The results of the cluster analysis were compared with the hydrogeology of the region, as well as the climate classification. Pixels showing the highest (lowest) likelihood of being GDV are mainly located in areas where geology allows (does not allow) for the existence of aquifers, pointing to the suitability of the method in identifying potential GDV. These areas are mainly occupied by forest and semi natural vegetation.
Acknowledgments: This work was partially supported by national funds through FCT (Fundação para a Ciência e a Tecnologia, Portugal) under Projects PIEZAGRO (PTDC/AAG-REC/7046/2014) and IMDROFLOOD (WaterJPI/0004/2014.