Application of a novel cascade-routing and reinfiltration concept with a Voronoi unstructured grid in MODFLOW 6, for an assessment of surface-water/groundwater interactions in a hard-rock catchment (Sardon, Spain)

Abstract

Integrated hydrological modelling (IHM) can reliably characterize surface-water/groundwater interactions in complex hydrological systems such as hard-rock systems (HRS), located in water-limited environments (WLE). Such HRS-WLE conditions are represented by Sardon catchment (~80 km2) in Spain, where the MODFLOW 6 modelling environment was tested, applying the following improvements as compared to previous works in that catchment: a new conceptual model, driving forces redefined based on remote sensing data, an unstructured Voronoi grid, and, most importantly, a novel cascade-routing and reinfiltration (CRR) concept. In the standard MODFLOW 6, rejected infiltration and groundwater exfiltration have always been considered as sinks (evaporation). However, in reality, that water can not only evaporate but also reinfiltrate back to the subsurface or move as runoff towards drainage water bodies. The CRR improves surface–unsaturated-zone interactions and also surface-water/groundwater interactions. The standard and new capacities of MODFLOW 6 are presented in the transient model of the Sardon catchment, calibrated using 7 years of daily groundwater heads and streamflows. The results showed: the large spatio-temporal variability of the groundwater fluxes, the substantial role of groundwater exfiltration, the low catchment storage, the fast reaction of the water table and streams to rainfall, and the mosaic character of the net recharge. These characteristics are typical for HRS-WLEs with a shallow water table. MODFLOW 6 has many improvements compared to previous MODFLOW versions, so with the proposed CRR concept (still can be improved), the single-environment MODFLOW 6 has modelling capacity comparable with multienvironment IHMs, while being more flexible and more efficient.

Partitioning and sourcing of evapotranspiration using coupled MARMITES-MODFLOW model, La Mata catchment (Spain)

“…The MM-MF modeling results of partitioning and sourcing of ET are consistent with corresponding experimental results of Balugani et al (2017) and with modeling results of Hassan et al (2014), Daoud et al (2022), emphasizing the relevance of Exf g in the water balance, as well as generally low, but highly spatio-temporally variable Rn, all representing typical characteristics of hard rock systems with shallow water table such as the La Mata catchment.…”

“…For that purpose, we selected the commonly simulated hydrological years . All the three models show similar results, especially the MM-MF and the one by Daoud et al (2022), despite different codes and different approaches used. The model of Hassan et al (2014) provided some differences in water fluxes, mainly because of simplifications in the assignment of driving forces (Daoud et al, 2022), implying higher ET soil than in the .…”

“…This study uses the La Mata catchment (∼4.8 km 2 , Figure 1), located west of Salamanca (Spain), to present the functionality of the MM-MF model. The La Mata catchment is a sub-catchment of the Sardón catchment (∼80 km 2 ), where extensive research on subsurface water fluxes has been carried out (Lubczynski and Gurwin, 2005;Lubczynski, 2009Lubczynski, , 2011Reyes-Acosta and Lubczynski, 2013;Francés et al, 2014;Hassan et al, 2014;Balugani et al, 2016;Daoud et al, 2022). The La Mata catchment was selected because of a semi-arid, water-limited environment, negligible groundwater use, availability of monitoring data and shallow water table enhancing E g , T g and Exf g .…”

“…Subsurface water fluxes in the Sardón catchment, in which La Mata is a sub-catchment, were modeled by Hassan et al (2014), who used a GSFLOW model (Markstrom et al, 2008) and also by Daoud et al (2022), who used MODFLOW 6 (Langevin et al, 2017), both representing transient, distributed solutions. As there is an uniformity of geology, soil, land use and climatic conditions in the two areas, the results of the three models can be compared.…”

“…However, such assessment is complex and uncertain (Kinzelbach et al, 2002;Hendricks et al, 2003;Xu and Beekman, 2003;Lubczynski and Gurwin, 2005;Lubczynski, 2011) because: (i) subsurface water fluxes, due to their inaccessibility, are by far more difficult to quantify than surface water fluxes and consequently the methods of their estimation are highly uncertain; (ii) subsurface water fluxes are controlled by spatio-temporal variability of precipitation (P) and evapotranspiration (ET) but also by surface and subsurface heterogeneity of geology, soil texture, topography, drainage, vegetation, etc. ; (iii) groundwater fluxes are generally small and Rg cannot be reliably determined by subtracting ET from P, since unavoidable small errors in the two lead to high inaccuracy of Rg (Lubczynski, 2011); (iv) ET g is generally neglected or underestimated (Lubczynski, 2000(Lubczynski, , 2009, which results in the overestimation of the net recharge (Rn = Rg − Exf g − ET g ) and therefore erroneous model calibration; and (v) standard, standalone groundwater flow models do not take into account interactions with unsaturated zone, so also do not quantify Exf g that can represent important component of groundwater balance, especially in shallow water table environments (Hassan et al, 2014;El-Zehairy et al, 2018;Daoud et al, 2022).…”