Semi-distributed Model Research Articles

The dilemma of spatial representation for urban hydrology semi-distributed modelling: Trade-offs among complexity, calibration and geographical data

Summary Semi-distributed models are widely used in urban hydrology, supported by the abundance and detail of geographical data. The inclusion of these data into hydrological models requires however an increasing complexity of the model structure with spatially distributed parameters, potentially driving to over-parameterisation issues. In this paper, different configurations and model structures, including an increasing quantity of geographical information, are tested for both water quantity and water quality on the widely used SWMM5 model for a 2.3 km2 catchment. The Nash criterion is used to calibrate the model and compare alternative configurations. Results for water quantity modelling show that the inclusion of some basic geographical information, particularly on land uses, clearly improves performances, but further refinements are less effective. Uncalibrated models with sufficient land use information reach performances comparable with those of calibrated models. For water quality modelling (suspended solids concentration), the best modelling performance is obtained by a compromise solution with moderate spatial distribution of parameters: no spatial distribution drives to limited performances, while an excessive one to severe over-parameterisation. A comparison to suspended solids measurements realized on a single road of the catchment shows that parameters providing good performances at the catchment scale are a realistic, although non optimal, representation of local scale processes.

Parameter uncertainty and identifiability of a conceptual semi-distributed model to simulate hydrological processes in a small headwater catchment in Northwest China

Abstract Introduction Conceptual hydrological models are useful tools to support catchment water management. However, the identifiability of parameters and structural uncertainties in conceptual rainfall-runoff modeling prove to be a difficult task. Here, we aim to evaluate the performance of a conceptual semi-distributed rainfall-runoff model, HBV-light, with emphasis on parameter identifiability, uncertainty, and model structural validity. Results The results of a regional sensitivity analysis (RSA) show that most of the model parameters are highly sensitive when runoff signatures or combinations of different objective functions are used. Results based on the generalized likelihood uncertainty estimation (GLUE) method further show that most of the model parameters are well constrained, showing higher parameter identifiability and lower model uncertainty when runoff signatures or combined objective functions are used. Finally, the dynamic identifiability analysis (DYNIA) shows different types of parameter behavior and reveals that model parameters have a higher identifiability in periods where they play a crucial role in representing the predicted runoff. Conclusions The HBV-light model is generally able to simulate the runoff in the Pailugou catchment with an acceptable accuracy. Model parameter sensitivity is largely dependent upon the objective function used for the model evaluation in the sensitivity analysis. More frequent runoff observations would substantially increase the knowledge on the rainfall-runoff transformation in the catchment and, specifically, improve the distinction of fast surface-near runoff and interflow components in their contribution to the total catchment runoff. Our results highlight the importance of identifying the periods when intensive monitoring is critical for deriving parameter values of reduced uncertainty.

Testing the realism of a topography-driven model (FLEX-Topo) in the nested catchments of the Upper Heihe, China

Abstract. Although elevation data are globally available and used in many existing hydrological models, their information content is still underexploited. Topography is closely related to geology, soil, climate and land cover. As a result, it may reflect the dominant hydrological processes in a catchment. In this study, we evaluated this hypothesis through four progressively more complex conceptual rainfall-runoff models. The first model (FLEXL) is lumped, and it does not make use of elevation data. The second model (FLEXD) is semi-distributed with different parameter sets for different units. This model uses elevation data indirectly, taking spatially variable drivers into account. The third model (FLEXT0), also semi-distributed, makes explicit use of topography information. The structure of FLEXT0 consists of four parallel components representing the distinct hydrological function of different landscape elements. These elements were determined based on a topography-based landscape classification approach. The fourth model (FLEXT) has the same model structure and parameterization as FLEXT0 but uses realism constraints on parameters and fluxes. All models have been calibrated and validated at the catchment outlet. Additionally, the models were evaluated at two sub-catchments. It was found that FLEXT0 and FLEXT perform better than the other models in nested sub-catchment validation and they are therefore better spatially transferable. Among these two models, FLEXT performs better than FLEXT0 in transferability. This supports the following hypotheses: (1) topography can be used as an integrated indicator to distinguish between landscape elements with different hydrological functions; (2) FLEXT0 and FLEXT are much better equipped to represent the heterogeneity of hydrological functions than a lumped or semi-distributed model, and hence they have a more realistic model structure and parameterization; (3) the soft data used to constrain the model parameters and fluxes in FLEXT are useful for improving model transferability. Most of the precipitation on the forested hillslopes evaporates, thus generating relatively little runoff.

Improved semi-distributed model for phosphorus losses from Irish catchments

Phosphorus (P) is one of the important limiting elements governing the quality of natural waters in Ireland. Good models that simulate the loss of P from catchments to water bodies are needed to understand the mechanism and transport of P and to assist in formulating appropriate policies and practices to control or manage the problem. In the present study, a new combined model (NCM) is developed as a hybrid between soil and water assessment tool (SWAT) and Hydrological Simulation Program FORTRAN (HSPF) packages. The developed NCM model was tested, with the data from the Oona catchment in Ireland, and evaluated to insure it is capable of predicting flow and total P loads at the same level or better than SWAT and HSPF packages when used alone. The NCM produced better flow simulation in terms of Nash–Sutcliffe efficiency (R2) than SWAT when used alone for both calibration and validation periods when hourly time steps are used for rainfall and evapotranspiration. In addition, calibration for the total phosphorus load gave better R2 values than previously reported result for either HSPF or SWAT and showed better performance for most of the validation period. Using multi-objective optimization methods for optimizing both flow and total P loads gave better simulated results than using sequential optimization of flow followed by total phosphorus. Hence, the NCM model developed here can be used to improve the prediction of phosphorus loss to water bodies in Ireland.

Changes of rural to urban areas in hydrograph characteristics on watershed divisions

This study examined changes in hydrograph characteristics of rural statuses to urban statuses on watershed divisions in Taiwan. The main approach was to relate applicable model parameters with the corresponding imperviousness based on calibration and verification using a semidistributed model and 102 events. The model structure is conceptual linear reservoirs with parallel-type cascaded storages which is represented by overland and channel coefficients. The hourly mean rainfall of the watershed and its divisions were calculated using the Kriging method. The time-variant rainfall losses were calculated using the constant percentage method. The spatial and temporal model inputs, division effective rainfall, were obtained by subtracting mean rainfall of divisions from the rainfall losses. In calibration, the storage values of 50 events derived using appropriate parameter bounds were more reasonable than those using inappropriate bounds. Based on the optimal interval method, the overland storages displayed more marked change than did channel storage in response to imperviousness changes. By contrast, the channel storages were unaffected by the changes in urbanization. The overland storages were related with the imperviousness by using the regression equations for determining their relationships in continuous changes of urbanized divisions. The verification of the regression relationships was based on 52 events. The results indicated that power linkage was an available selection for linking division parameters with the corresponding imperviousness. Finally, the study concludes (1) appropriate bounds in parameter calibration are useful for obtaining significant storage values and (2) the study results using these suitable storages indicated large changes in imperviousness on the downstream divisions, marked urbanization resulted in reduced the time to peak at least 10 %, the peak discharge exceeded an increment of 20–30 %.

Effects of urbanization variables on model parameters for watershed divisions

A semi-distributed model with a parallel connection was applied to examine the effects of urbanization variables. Data were obtained from watershed divisions that were characterized by various degrees of urbanization. The mean rainfall was calculated using the kriging method. The model inputs were obtained by subtracting mean rainfall from Ф-index values, based on the spatially uniform loss assumption. Regression analysis was applied to determine the relationship between the parameters of 64 calibrations and urbanization variables among the divisions. The results showed that overland parameters produced more consistent change in response to imperviousness than to population. Conversely, the channel parameter was unaffected by changes in urbanization. The verification results of 46 cases showed that power linkage was a potential option for linking division parameters with the corresponding imperviousness based on four evaluation criteria. The changes in imperviousness on overland parameters show the hydrological effects of division urbanizations.Editor D. Koutsoyiannis; Associate editor T. WagenerCitation Chen, R., Chuang, W.-N., and Cheng, S., 2014. Effects of urbanization variables on model parameters for watershed divisions. Hydrological Sciences Journal, 59 (6), 1167–1183. http://dx.doi.org/10.1080/02626667.2014.910305

Analysing the effect of land-use/cover changes at sub-catchment levels on downstream flood peaks: A semi-distributed modelling approach with sparse data

This paper aims to evaluate how varying degrees of land-use/cover (LULC) changes across sub-catchments affect a flood peak at the catchment outlet. The study site was the Konar catchment, a part of the upper Damodar Basin in eastern India. A HEC-HMS model was set up to simulate rainfall–runoff processes for two LULC scenarios three decades apart. Because of sparse data availability at the study site, we used the Natural Resource Conservation Service (NRCS) curve number (CN) approach to account for the effect of LULC and soil on the hydrologic response of the catchment. Although a weak (r=0.53) but statistically significant positive linear correlation was found between sub-catchment wise LULC changes and the magnitude of the flood peak at the catchment outlet, a number of sub-catchments showed marked deviations from this relationship. The varying timing of flow convergence at different stream orders due to localised LULC changes makes it difficult to upscale the conventional land-use and runoff relationship, evident at the plot scale, to a large basin. However, a simple modelling framework is provided based on easily accessible input data and a freely available and widely used hydrological model (HEC-HMS) to check the possible effect of LULC changes at a particular sub-catchment on the hydrograph at the basin outlet.

When does higher spatial resolution rainfall information improve streamflow simulation? An evaluation using 3620 flood events

Abstract. Precipitation is the key factor controlling the high-frequency hydrological response in catchments, and streamflow simulation is thus dependent on the way rainfall is represented in a hydrological model. A characteristic that distinguishes distributed from lumped models is the ability to explicitly represent the spatial variability of precipitation. Although the literature on this topic is abundant, the results are contrasting and sometimes contradictory. This paper investigates the impact of spatial rainfall on runoff generation to better understand the conditions where higher-resolution rainfall information improves streamflow simulations. In this study, we used the rainfall reanalysis developed by Météo-France over the whole country of France at 1 km and 1 h resolution over a 10 yr period. A hydrological model was applied in the lumped mode (a single spatial unit) and in the semidistributed mode using three unit sizes of subcatchments. The model was evaluated against observed streamflow data using split-sample tests on a large set of French catchments (181) representing a variety of sizes and climate conditions. The results were analyzed by catchment classes and types of rainfall events based on the spatial variability of precipitation. The evaluation clearly showed different behaviors. The lumped model performed as well as the semidistributed model in western France, where catchments are under oceanic climate conditions with quite spatially uniform precipitation fields. By contrast, higher resolution in precipitation inputs significantly improved the simulated streamflow dynamics and accuracy in southern France (Cévennes and Mediterranean regions) for catchments in which precipitation fields were identified to be highly variable in space. In all regions, natural variability allows for contradictory examples to be found, showing that analyzing a large number of events over varied catchments is warranted.

鄱阳湖区流域水文模型比较研究

选用目前具有代表性的三水源新安江模型、半分布式TopModel和分布式VIC模型，开展降雨径流模拟计算，分析比较各种水文模型在平原湖区的适用性。选择鄱阳湖未控区间内的博阳河和西河流域，分别进行三个水文模型的参数率定和检验。应用结果表明，VIC模型的效率系数比TopModel模型和新安江模型高出10多个百分点，径流总量相对误差均控制在5%以内，低于其余2个模型。VIC分布式水文模型的模拟结果明显优于集总式和半分布式水文模型，能较好地模拟湖区流域的降雨径流过程和水资源量。The lumped Xin’anjiang model, semi-distributed TopModel model and distributed VIC model were selected for rainfall runoff simulation in order to compare the suitability in the plain lake region. The Boyang and Xi River sub-basins in the Poyang Lake uncontrolled intervening basin were used to calibrate and verify these hydrological models. Application results show that the model efficiency of VIC model is 10% higher than those of Xinanjiang and TopModel models, the relative error of total runoff is within 5% and much better than other two models. The distributed VIC model performs much better than lumped and semi-distributed models and is able to simulate rainfall runoff and water resources well in lake area.

Simultaneous estimation of soil moisture and hydraulic parameters using residual resampling particle filter

Land data assimilation (DA) has gradually developed into an important earth science research method because of its ability to combine model simulations and observations. Integrating new observations into a land surface model by the DA method can correct the predicted trajectory of the model and thus, improve the accuracy of state variables. It can also reduce uncertainties in the model by estimating some model parameters simultaneously. Among the various DA methods, the particle filter is free from the constraints of linear models and Gaussian error distributions, and can be applicable to any nonlinear and non-Gaussian state-space model; therefore, its importance in land data assimilation research has increased. In this study, a DA scheme was developed based on the residual resampling particle filter. Microwave brightness temperatures were assimilated into the macro-scale semi-distributed variance infiltration capacity model to estimate the surface soil moisture and three hydraulic parameters simultaneously. Finally, to verify the scheme, a series of comparative experiments was performed with experimental data obtained during the Soil Moisture Experiment of 2004 in Arizona. The results show that the scheme can improve the accuracy of soil moisture estimations significantly. In addition, the three hydraulic parameters were also well estimated, demonstrating the effectiveness of the DA scheme.

The performance of semi-distributed modelling in urban drainage: the trade-off between hydrological measurements and geographical data

Nowadays, high resolution geographical data are widely available for research and operational purposes. The inclusion of these data into hydrological models may suggest a direct and clear improvement of their performances. Different configurations and model structures, including an increasing quantity of geographical information, are tested using the widely used Stormwater Management Model 5 on a 2.5 km2 pilot catchment (located in Sucy en Brie, close to Paris, France). The Nash-Sutcliffe criterion is used to estimate the goodness of fit between model simulations and available measurements. If including some basic spatial information on landuse clearly improves the performance of the uncalibrated model, the increase in performance is less obvious if the user continues to refine geographical information on the catchment. Moreover, model predictions are comparable between a model calibrated with an efficient calibration procedure and a more physical approach including a fine spatial description and no calibration. Finally, the quality of data used for calibration and validation seems to be a key parameter to obtain a good fit between measurements and model predictions.

Implementation and testing of routing algorithms in the distributed Hydrologiska Byråns Vattenbalansavdelning model for mountainous catchments

The main purpose of this study was to implement and test routing algorithms in the distributed Hydrologiska Byråns Vattenbalansavdelning (HBV) model with the emphasis of obtaining a most suitable routing algorithm for large mountainous catchments. Two routing algorithms were built into the grid-based HBV model and tested on the Losna (11,213 km2) and the Norsfoss (18,932 km2) catchments in central southern Norway. In the first algorithm, runoff is first routed from cell to cell and hydrographs are re-calculated at each cell, and then runoff is routed by the Muskingum–Cunge method in river channels. The second algorithm is a source-to-sink method, which routes runoff of all cells to the catchment outlet as a function of local slope and a calibrated velocity parameter. The routing approaches were compared at different spatial resolutions (i.e. 1, 5 and 10 km) in daily streamflow simulation. Additionally, the elevation band-based semi-distributed model was also compared with the distributed models. The results show that the distributed HBV models are able to perform better than the elevation band-based model, and hillslope routing is crucial in the mountainous catchments. However, incorporating the Muskingum–Cunge channel routing does not add value to the simulation of daily runoff in the mountainous catchments.

Three perceptions of the evapotranspiration landscape: comparing spatial patterns from a distributed hydrological model, remotely sensed surface temperatures, and sub-basin water balances

Abstract. A problem encountered by many distributed hydrological modelling studies is high simulation errors at interior gauges when the model is only globally calibrated at the outlet. We simulated river runoff in the Elbe River basin in central Europe (148 268 km2) with the semi-distributed eco-hydrological model SWIM (Soil and Water Integrated Model). While global parameter optimisation led to Nash–Sutcliffe efficiencies of 0.9 at the main outlet gauge, comparisons with measured runoff series at interior points revealed large deviations. Therefore, we compared three different strategies for deriving sub-basin evapotranspiration: (1) modelled by SWIM without any spatial calibration, (2) derived from remotely sensed surface temperatures, and (3) calculated from long-term precipitation and discharge data. The results show certain consistencies between the modelled and the remote sensing based evapotranspiration rates, but there seems to be no correlation between remote sensing and water balance based estimations. Subsequent analyses for single sub-basins identify amongst others input weather data and systematic error amplification in inter-gauge discharge calculations as sources of uncertainty. The results encourage careful utilisation of different data sources for enhancements in distributed hydrological modelling.

Streamflow Responses to Climate Change: Analysis of Hydrologic Indicators in a New York City Water Supply Watershed

AbstractRecent works have indicated that climate change in the northeastern United States is already being observed in the form of shorter winters, higher annual average air temperature, and more frequent extreme heat and precipitation events. These changes could have profound effects on aquatic ecosystems, and the implications of such changes are less understood. The objective of this study was to examine how future changes in precipitation and temperature translate into changes in streamflow using a physically based semidistributed model, and subsequently how changes in streamflow could potentially impact stream ecology. Streamflow parameters were examined in a New York City water supply watershed for changes from model‐simulated baseline conditions to future climate scenarios (2081‐2100) for ecologically relevant factors of streamflow using the Indicators of Hydrologic Alterations tool. Results indicate that earlier snowmelt and reduced snowpack advance the timing and increase the magnitude of discharge in the winter and early spring (November‐March) and greatly decrease monthly streamflow later in the spring in April. Both the rise and fall rates of the hydrograph will increase resulting in increased flashiness and flow reversals primarily due to increased pulses during winter seasons. These shifts in timing of peak flows, changes in seasonal flow regimes, and changes in the magnitudes of low flow can all influence aquatic organisms and have the potential to impact stream ecology.

중규모 유역에서 Clark 방법과 ModClark 방법을 이용한 유출해석 시적정 소유역 분할 개수에 대한 연구

In this study, flood runoff characteristics is analyzed according to subbasin divisions by physically based rainfall-runoff model and appropriate number of subbasin divisions is suggested for midsize test basins. The Clark method, a lumped model in HEC-HMS, and the ModClark method, a semi-distributed model are used to simulate rainfall-runoff processes on Andong-reservoir basin, Imha-reservoir basin, and Pyeongchang river basin. The test basins were divided into nine subdivision cases by equal-area subdivision method such as single basin, 3, 5, 6, 7, 9, 10, 12, and 15 subbasins, and compared the simulated and observed values in terms of the peak flow and the peak time. The simulation results indicated that the peak flows tended to increase and the peak time shifted earlier as the number of subdivisions increased and this tendency weakened after the certain number of subdivisions. In this research, the specific number of subdivision was defined as the minimum number of subdivision considering both peak flow and peak time. Consequently, the minimum number of subdivisions is determined as 5 for Andong and Imha reservoir basins and 7 for Pyeongchang river basin.

The impact of different elevation steps on simulation of snow covered area and the resulting runoff variance

Abstract. This study analyses the impact of vertical model discretisation on modelling snow covered area and the consequential effects on runoff formation of the semi-distributed water balance model HQsim. Therefore, the parameters relevant for snow modelling are varied within the frame of a uniformly distributed Monte Carlo Simulation (MCS). Since the model is based on the hydrological response unit (HRU) approach, the effect of building the HRUs with different elevation steps (250 m and 500 m) is tested for two alpine catchments. In total 5000 parameter combinations were generated for simulation. The results of modelled snow covered area were compared with thirty MODIS (Moderate Resolution Imaging Spectroradiometer) snow cover maps for the melting periods in 2003–2011. Based on a contingency table the comparisons were evaluated by different skill measures. Finally, the pareto optimal parameter settings of each skill measure were detected. Evaluation of runoff variability within the MCS and their pareto optimal runs show reduced variances of model output resulting from an improved simulation of the snow covered area.

Hydrological Response of Watershed Systems to Land Use/Cover Change. A Case of Wami River Basin

Wami river basin experiences a lot of human disturbances due to agricultural expansion, and increasing urban demand for charcoal, fuel wood and timber; resulting in forest and land degradation. Comparatively little is known about factors that affect runoff behaviour and their relation to landuse in data poor catchments like Wami. This study was con- ducted to assess the hydrological response of land use/cover change on Wami River flows. In data poor catchments, a promising way to include landuse change is by integrating Remote Sensing and semi-distributed rainfall-runoff models. Therefore in this study SWAT model was selected because it applies semi-distributed model domain. Spatial data (lan- duse, soil and DEM-90m) and Climatic data used were obtained from Water Resources Engineering Department, govern- ment offices and from the global data set. SWAT model was used to simulate streamflow for landuse/landcover for the year 1987 and 2000 to determine the impact of land use/cover change on Wami streamflow after calibrating and validating with the observed flows. Land use maps of 1987 and 2000 were derived from satellite images using ERDAS Imagine 9.1 software and verified by using 1995 land use which was obtained from Institute of Resource Assessment (IRA). Findings show that there is decrease of Forest area by 1.4%, a 3.2% increase in Agricultural area, 2.2% increase in Urban and 0.48% decreases in Waterbody area between 1987 and 2000. The results from SWAT model simulation showed that the average river flows has decreased from 166.3 mm in 1987 to 165.3 mm in 2000. The surface runoff has increased from 59.4mm (35.7%) in 1987 to 65.9mm (39.9%) in 2000 and the base flow decreased from 106.8mm (64.3%) to 99.4mm (60.1%) in 1987 and 2000 respectively. This entails that the increase of surface runoff and decrease of base flows are as- sociated with the land use change.

Evaluation of different calibration strategies for large scale continuous hydrological modelling

Abstract. For the analysis of climate impact on flood flows and flood frequency in macroscale river basins, hydrological models can be forced by several sets of hourly long-term climate time series. Considering the large number of model units, the small time step and the required recalibrations for different model forcing an efficient calibration strategy and optimisation algorithm are essential. This study investigates the impact of different calibration strategies and different optimisation algorithms on the performance and robustness of a semi-distributed model. The different calibration strategies were (a) Lumped, (b) 1-Factor, (c) Distributed and (d) Regionalisation. The latter uses catchment characteristics and estimates parameter values via transfer functions. These methods were applied in combination with three different optimisation algorithms: PEST, DDS, and SCE. In addition to the standard temporal evaluation of the calibration strategies, a spatial evaluation was applied. This was done by transferring the parameters from calibrated catchments to uncalibrated ones and validating the model performance of these uncalibrated catchments. The study was carried out for five sub-catchments of the Aller-Leine River Basin in Northern Germany. The best result for temporal evaluation was achieved by using the combination of the DDS optimisation with the Distributed strategy. The Regionalisation method obtained the weakest performance for temporal evaluation. However, for spatial evaluation the Regionalisation indicated more robust models, closely followed by the Lumped method. The 1-Factor and the Distributed strategy showed clear disadvantages regarding spatial parameter transferability. For the parameter estimation based on catchment descriptors as required for ungauged basins, the Regionalisation strategy seems to be a promising tool particularly in climate impact analysis and for hydrological modelling in general.

A conceptual semidistributed model of the Coulazou River as a tool for assessing surface water–karst groundwater interactions during flood in Mediterranean ephemeral rivers

This study aims to assess surface water–karst groundwater interactions during floods in the Mediterranean karst watershed of the Coulazou River (southern France) using a conceptual semidistributed model at 5 min time steps. The kinematic wave routing approximation is used for the transfer of surface flow, while overflows from a linear underground reservoir account for karst flows along the riverbed. After calibration, values of parameters and simulated time series are compared to independent physical measurements. Results show that direct runoff can be neglected on the karst terrains. In addition, this study demonstrates that, in some cases, karst watersheds can be considered as relatively poor systems of regulation but strong systems of amplification or generation of floods and flash floods, depending on rainfall characteristics and also on groundwater level conditions prior to the flood event. Considering that the flood peak is the most important factor defining flash flood hazard, it is shown that the flood hazard regulation effect of the karst is relatively limited for low water table conditions prior to the flood, while the aggravating effect for high water table conditions may be higher than 80% with respect to expected values from surface runoff only. These results show that understanding groundwater–surface water interactions is crucial for describing the flash flood dynamics in karst terrains.

Assessing water availability in a semi-arid watershed of southern India using a semi-distributed model

Summary Appropriate groundwater resource management becomes a priority for the States of the semi-arid southern India. Because of the highly increasing groundwater demand, the number of drought-prone regions where the groundwater resource is classified as over-exploited by the Government is critically increasing. Thus there is a need to develop quantitative methodologies adapted to the regional context that are capable to assess water resources at watershed scale and the impact of management measures. This study demonstrates the calibration and use of an integrated water resource assessment model (SWAT) in an 84 km 2 representative semi-arid crystalline watershed of southern India with no perennial surface water source. The model can reproduce (i) the recharge rate estimates derived independently by a groundwater balance computation, (ii) runoff and surface water storage occurring in tanks that spread along the drainage system, (iii) groundwater table fluctuations monitored at a monthly time step. Results show that even if the calibration period (2006–2010) was more humid than long-term average, the watershed is sensitive to the monsoon inter-annual variability with water-stress during the dry years and an associated loss in agricultural production. The impact of these dry years is spatially variable with higher vulnerability for sub-basins having proportionally larger irrigated paddy areas, lower groundwater resource, and/or lower recharge potential (i.e., due to land use and repartition of percolation tanks). The scope for additional recharge by means of managed aquifer recharge structures is limited and demand-side management measures are needed to mitigate pumping. A wishful management objective may be to see groundwater reserves as a supplementary resource in case of monsoon failure and not as the main water resource to be used indiscriminately. SWAT proved to be an adequate modeling framework for the simulation of water resource in semi-arid hard-rock context where groundwater vertical fluxes largely prevail over regional lateral flows at km-scale. It provides interesting capabilities for water availability mapping and the simulation of different types of scenarios (e.g., land use changes, climate change).