Coupling a water balance model with a groundwater flow model and its appllication on Ergene Basin

Abstract

Groundwater is of great importance to support the ecosystem life on planet earth. The reliable simulation of groundwater recharge and identification of its dominant influencing factors is indeed the key control to understand the groundwater system, its potential accessibility, and its sustainability. During this research, a GIS grid based water budget model (mGROWA) was applied to estimate the long-term groundwater recharge and other water balance quantities based on basin characteristics such as hydro-climatic data, land cover, soil textures, geology, and topography. The study was performed in Ergene river catchment (11000 km2) which is one of the Mediterranean basins located in Turkey. The model simulations were done for the spatial resolution of the grid cell of 100 m for the period of 20 years (1991-2010). The hydrograph separation techniques, based on the base flow indices (BFI), were applied to split up the modeled total runoff into two main components namely groundwater recharge and direct runoff depending on the catchments attributes (e.g. geology, groundwater depth). The model was validated in 10 sub-catchments possessing long-term daily stream records. Thereafter, the statistical methods including Pearson correlation method and principal component analysis (PCA) were integrated for assessing the most controlling basin factors that are influencing the groundwater recharge. Simulated groundwater recharge rates were displayed as long-term yearly mean and as long-term monthly levels with the purpose of denoting periodic variation of groundwater recharge. mGROWA has shown the ability for estimating the spatial and temporal variability of groundwater recharge with high confidence, however, it does not simulate the distribution of hydraulic heads. Against this background, a new method for integrating the mGROWA, with the regional subsurface water flow model (MODFLOW) was adopted. The recharge values resulted from mGROWA was provided to MODFLOWas input data, in a steady state, and the simulated groundwater depths by MODFLOW become the input file of mGROWA. In this coupling, the iteration ends when acceptable groundwater heads are reached. The model validation and statistical analysis have indicated that the differences between simulated and observed values are generally under 20% and no valid inconsistency was observed. PCA revealed that the recharge variation is mostly influenced, in order of significance, by vegetation land cover, soil textures, and climate variables. The generated results may conclusively contribute to the establishment of quantitative status of groundwater system exploitation and enhancement of water resource management policy of the study area. In addition, this study draws special attention to the advantage of applying coupled procedures to adjust and validate the recharge rates as input data of the upper boundary in sub-surface water flow models.