Effects of land use on groundwater recharge of a loess terrace under long-term irrigation

Highlights: • Incorporating growth cycle of crop aids in land use detection by remote sensing data. • High groundwater level is induced by high irrigation intensity for vegetable fields. • Local GWT is strongly correlated with irrigation intensity governed by land use type. • High spring discharge can affect local GWT close to the margin of the loess terrace. Understanding the relation between land use and groundwater recharge is of direct interest in areas that are prone to geohazards. This study was performed to characterize the effects of land use on groundwater table (GWT) distribution in the Heifangtai (HFT) loess terrace under long-term flood irrigation, as irrigation intensity was governed by the crop types. In light of the regional temperature-vegetable dryness index (TVDI) and optical images obtained in different seasons over four years, the dataset incorporated the growth cycles of the local crops, which in turn improved the workload and accuracy of land use detection. Irrigation intensities for different crops were recorded during field investigations. A total of 26 electrical resistivity tomography (ERT) profiles were conducted to estimate the GWT distribution of HFT terrace, which was further utilized to reveal the relation between land use (i.e. irrigation) and GWT fluctuations. The results indicated that high local GWT was associated with vegetable fields which had the highest irrigation intensity, and by contrast, low local GWT was attributed to the perennial fruit tree fields with the lowest irrigation intensity. The discharge was monitored intermittently for over three years to analyze the effect on the local GWT close to margin of HFT. The result showed a strong correlation between spring discharge and GWT fluctuations along the margin of HFT, indicating that groundwater recharge was affected by both irrigation and spring discharges. These results suggested that incorporating growth cycles of crops can effectively facilitate the interpretation of remote sensing data for land use detection, and the results can provide useful guidance for improving irrigation programs and for alleviating geohazards in regions that are under long-term irrigation.