Detection and prediction of illicit drilling by combination of drought indices and NDVI: case of the Nadhour-Saouaf aquifer (Northeast Tunisia)

In Tunisia, illicit drilling exists since the 1980s in both southern and central Tunisia. Since 2011, with the weakening of the administration control, farmers and landowners have been drilling in a frantic manner in the various aquifers throughout the country. According to DGRE (Direction Générale des Ressources en Eau), aquifers are in advanced overexploitation, more accentuated in case of drought. It would therefore be relevant to monitor, detect and forecast this phenomenon by combining drought (SPEI and SGI) and remote sensing (NDVI) indices.

The objective of this work is the monitoring, detection and prediction of illicit drilling by the cross-checking and exploitation of hydroclimatic and vegetation indices, derived from remote sensing for Nadhour-Saouaf aquifer. SPEI makes it possible to consider precipitation and evapotranspiration on different time scales which facilitates the analysis of drought impact on water resource demands. The same is true for SGI, which makes it possible to analyse the water table level. As for NDVI calculated on time series, it allows monitoring the state of crops and their development over time.

These different indices have been used in different contexts to detect anomalies that would be interpreted as illicit drilling, since the detection of significant plant cover in drought conditions could indicate the existence of such drilling. We first computed SPEI and SGI from time series of rainfall (series of more than 30 years derived from DGRE Zaghouan database), temperatures and piezometric levels. The analysis allowed to (1) describe, summarizing, and evaluating the data and (2) Fill gaps. Then, we extracted NDVI time series from LANDSAT data. Finally, the multidimensional exploratory analysis of the three indices combined (SPEI, SGI, NDVI) revealed relationships between them.

A robust indicator is currently being constructed to detect the probable existence of illicit drilling. The variation of this indicator will make it possible to detect the over-exploitation of aquifers caused by these drillings, based on DGRE data. This analysis, moreover, considered the social typology of farmers in the study area and was supported by interviews with farmers. This index makes it possible to analyse the socio-spatial and temporal evolution of the study area.