Calibration of a lumped karst system model and application to the Qachqouch karst spring (Lebanon) under climate change conditions

Abstract

Abstract. Flow in complex karst aquifers is challenging to conceptualize and model, especially in poorly investigated areas, in semiarid climates, and under changing climatic conditions; however, it is necessary in order to implement long-term sustainable water management practices. Thus, the objectives of this work were to propose a calibration approach based on time series analyses for a karst aquifer and to assess the impact of climate change on spring discharge. Based on more than 3 years of high-resolution continuous monitoring, a semi-distributed lumped model was calibrated and validated for the Qachqouch karst spring, north of Beirut (Lebanon). Time series analyses and decomposition of spring hydrographs revealed that the system has a high regulatory function, with considerable storage capacity providing stable flow (minimum flow of 0.2 m3 s−1) during the dry season and with flow rates exceeding 10 m3 s−1 during the wet season, which is similar to other karst aquifers in the region. Based on this detailed understanding of the hydrodynamics of the system, the model geometry and parameters were validated. Three linear reservoirs were implemented to reproduce the combined contribution of the different flow components of the system. A satisfactory simulation (Nash–Sutcliffe efficiency coefficient, NSE, of 0.72) of the measured spring flow rates was obtained after calibration. Climate change conditions (+1 to +3 ∘C warming, −10 % to −30 % less precipitation annually, and the intensification of rain events) were added to a baseline climatic year to produce scenarios of expected spring flow responses. Results show that the Qachqouch karst aquifer is sensitive to decreasing rainfall, which is associated with more pronounced recessions, with flow rates decreasing by 34 % and 1-month longer dry periods. Because of the limited influence of snow on the spring flow rate, a warming climate has less impact on spring flow conditions than a reduction in precipitation. Although the model shows that increasing rainfall intensity induces larger floods, recessions, and shorter low-flow periods, the real impact of high-intensity precipitation events remains uncertain, as the model does not account for complex unsaturated and epikarstic processes. This work shows that calibrating a semi-distributed lumped model using time series analyses can be an efficient approach to improve simulations of complex karst aquifers, thereby providing useful models for long-term sustainable water management.

Comparison of artificial neural networks and reservoir models for simulating karst spring discharge on five test sites in the Alpine and Mediterranean regions

“…Qachqouch spring (64 m asl) is located in the Nahr el-Kalb catchment and originates from a Jurassic karst aquifer. The recharge area is estimated to be about 56 km 2 with altitudes ranging from 60 to over 1500 m asl (Doummar and Aoun, 2018;Dubois et al, 2020). The catchment is primarily exposed to a Mediterranean climate, with snow influence at higher altitudes (Dubois et al, 2020).…”

“…The recharge area is estimated to be about 56 km 2 with altitudes ranging from 60 to over 1500 m asl (Doummar and Aoun, 2018;Dubois et al, 2020). The catchment is primarily exposed to a Mediterranean climate, with snow influence at higher altitudes (Dubois et al, 2020). The lithology mainly consists in Jurassic karstified limestone and dolomitic limestone (on the higher plateaus) changing to more massive micritic limestone in the lower part of the catchment.…”

“…Artificial neural networks (ANN) have been successfully used to simulate karst spring discharge (Kurtulus and Razack, 2007;Hu et al, 2008;Meng et al, 2015;Wunsch et al, 2022), predict and forecast water flood/inrush (Wu et al, 2008;Kong A Siou et al, 2011) and manage the exploitation of karst aquifers (Yin et al, 2011;Kong A Siou et al, 2015). Reservoir models also have been successfully used to simulate karst spring discharge (Fleury et al, 2007;Dubois et al, 2020), manage the exploitation of karst aquifers (Fleury et al, 2009;Zhou et al, 2021), as well as characterise specific functioning in karst systems (Tritz et al, 2011;Perrin et al, 2003;Bittner et al, 2020;Jukić and Denić-Jukić, 2009). Although several authors compared the performance of different ANN models (Cheng et al, 2020;Kurtulus and Razack, 2010;Kovačević et al, 2018) and studied structure and parameters' equifinality in reservoir models (Hartmann et al, 2012;Gondwe et al, 2011;Mazzilli et al, 2012;Makropoulos et al, 2008), only few studies have been conducted on the comparison of both approaches (Jeannin et al, 2021;Kong A Siou et al, 2014;Sezen et al, 2019).…”

Karst spring recession curve analysis: efficient, accurate methods for both fast and slow flow components

“…From our analysis, Qachquoch spring is classified as medium karstified system by most method combinations. However recent study by Dubois et al (2020) categorise the system as poorly karstified with a very large regulation capacity. Meanwhile, if we also consider two standard deviation distances from the calculate mean K values, a regulation capacity >0.5 will be obtained and the system will be equally classified as poorly karstified by REM-POA pairs used.…”

Effects of passive-storage conceptualization on modeling hydrological function and isotope dynamics in the flow system of a cockpit karst landscape

“…A wide range of hydrological models have been developed for karst areas, ranging from lumped models at the catchment scale to (semi-)distributed models with hydrological function parameterized for grid scales or landscape unit scales (Martínez-Santos and Andreu, 2010;Hartmann et al, 2013;Husic et al, 2019;Dubois et al, 2020;Ollivier et al, 2020;Xu et al, 2020;Jeannin et al, 2021;Wunsch et al, 2022). A key function of karst hydrological models is to capture the dual-or multi-phase flows in a complex porous medium, thereby capturing low velocities in the matrix and small fractures as well as very high velocities in large fractures and conduits (White, 2007;Worthington, 2009;Jourde et al, 2018;Ding et al, 2020).…”