Ungauged Watersheds Research Articles

Simulation of Hydrology in Ungauged Watersheds: SWAT Modeling in Kasillian Basin in Iran

Simulation of Hydrology in Ungauged Watersheds: SWAT Modeling in Kasillian Basin in Iran

Development of Synthetic Unit Hydrographs Using Spatial Proximity Regionalization for the Makiling Forest Reserve, Philippines

There is a dearth of streamflow data in the Philippines to generate hydrographs needed for flood forecasting and water resources assessment. A method of generating and calibrating synthetic hydrographs using model simulations and spatial proximity regionalization is presented. Synthetic unit hydrographs of four storm events in the gauged watershed of Makiling Forest Reserve were generated using Soil Conservation Service Unit Hydrograph, Clark Unit Hydrograph, and Snyder’s Unit Hydrograph methods. The generated synthetic hydrographs for each runoff modeling technique were calibrated with the actual hydrographs of the watershed. Snyder’s Unit Hydrograph model results were the most acceptable based on the Nash-Sutcliffe Efficiency, Index of Volumetric Fit, and Relative Error of Peak Flow. The weighted values of the calibrated watershed parameters computed using spatial proximity regionalization technique and the hyetographs derived from the Rainfall Intensity Duration Frequency Curve of the University of the Philippines Los Baños-National Agrometeorological Station were used to generate the synthetic unit hydrographs for three neighboring ungauged watersheds at 2-, 5-, 10-, 15-, 20-, 25-, 50, 100-year return periods. Total runoff volume, the magnitude of peak flows, and time to peak derived from the generated hydrographs can be used in watershed planning, water resources management, flood forecasting and the design of various water control structures.

Modeling surface water potential using the SWAT model combined with principal component analysis in the ungauged Gelana watershed, Ethiopia

Abstract This study aims to model and assess surface water potential in an ungauged watershed using the Soil and Water Assessment Tool (SWAT), principal component analysis (PCA), and regression-based regionalization techniques in the Gelana River, Ethiopia. The SWAT model was calibrated (and validated) for the 1989–2007 (2008–2015) period, and it showed a very good performance to model the river flow. The 18 physical catchment characteristics that affect the production of streamflow were selected for correlation, and to make an equation using 19 optimized SWAT model parameters. These characteristics were categorized as two climate descriptors, three soil descriptors, seven land use land cover descriptors, and six topographical descriptors. The regression equations for each SWAT parameter with a function of the physical characteristics and principal components were developed. Then, SWAT model-sensitive parameters were transferred and validated. Results reveal that the watershed has a greatest surface runoff volume of 162.14 MCM in May and a least runoff of 9.15 MCM in January. The contribution of the water balance during the spring season is highest with a total surface runoff of 103.72 mm. From the whole Gelana watershed area of 336,460 ha, 778.4 MCM of yearly surface runoff was produced. The maximum mean monthly river flow is 15.7 m3/s in May.

Estimating Freshwater Inflows for an Ungauged Watershed at the Big Boggy National Wildlife Refuge, USA

Bays and estuaries rely on freshwater inflows to maintain the salinity gradient necessary to sustain their fisheries. Reduced freshwater inflows, particularly during summer seasons, can be detrimental to the health of these systems. Despite an extensive network of streamflow gauges in the U.S., many coastal watersheds remain ungauged. The primary objective of this study was to develop methods to build a water budget for an ungauged watershed using limited data to determine the watershed contribution of freshwater to an at-risk bay system. This method was developed and tested for Big Boggy Creek, which flows into East Matagorda Bay (EMB), Texas. The streamflow into and out of Big Boggy Creek was quantified at key upstream and downstream sites. Over the summertime study period, we found average monthly freshwater inflows of 244 megaliters (ML). A simple inflow decision tool was developed to assist resource managers in estimating freshwater inflows during the summer months in the study area. Two recommendations are provided to increase freshwater inflows to EMB, with the most approachable option being purchasing water from a regional river authority. The framework developed herein can be modified and applied to ungauged watersheds to budget, model, and predict freshwater inflow contributions.

Integrating Support Vector Machines with Different Ensemble Learners for Improving Streamflow Simulation in an Ungauged Watershed

Integrating Support Vector Machines with Different Ensemble Learners for Improving Streamflow Simulation in an Ungauged Watershed

Hydrological simulation of ungauged basins via forcing by large‐scale hydrology models

AbstractThe established watershed's hydrologic simulation process, involving the calibration and validation of a model through spatiotemporal streamflow observations, has recently been adapted to encompass ungauged watersheds. The research introduces a methodology for simulating ungauged hydrosystems, suggesting the calibration and validation of regional rainfall‐runoff models using large‐scale hydrological models (LSMs) in the absence of in situ streamflow measurements. For doing so, LSM's 30‐years simulated discharges at various basins and subbasins of the case study River Basin District are assessed for their accuracy against the limited available observations. Subsequently, the rectified time series tunes the optimal parametrization of a regional lumped hydrologic model that is triggered by precipitation and temperature data from a partially dense meteorological network. The validated regional model is then applied in selective ungauged neighbouring subbasins employing the regionalization concept used in LSMs, that is in basins with similar physiographic characteristics to those where the regional model was set up. The performance of the research is assessed using statistical metrics. In cases where observations are available, the results show very satisfactory indexes (e.g., coefficient of determination: 0.654 < R2 < 0.949, Kling‐Gupta efficiency: 0.759 < KGE <0.790, Nash–Sutcliffe efficiency: 0.649 < NSE <0.745 for all the cases). In completely ungauged basins and in an order of magnitude index, the comparison of the simulated discharges with those from the LSM also indicates the satisfactory performance of the proposed approach. To sum up, the methodology underscores the alternate use of large and regional scale hydrological modelling coupling as an asset across various basin scales globally, offering reliable water quantity information.

Update of Empirical Models for Predicting Specific Degradation in South Korea and Future Sediment Management Considering Climate Change

In this study, the existing empirical models for sediment management for ungauged watersheds in South Korea developed in 2019 were updated. The existing models were developed based on the Universal Soil Loss Equation structure with domestic river sediment data obtained from 2005 to 2014 and watershed characteristic values. To update the models, the specific degradation (SD) value was calculated first in the same way as the existing method by adding sediment measurement and daily discharge data obtained from 2015 to 2020. The calculated SD values with recent data decreased compared with the past and differed from the results predicted using the existing models and the watershed characteristics investigated in this study. For this reason, the existing models were updated using the latest watershed characteristics and SD values. The root-mean-square error values of the existing models, which ranged from 131 to 90 tons/km2·year, decreased to 102 to 62 tons/km2·year in the updated models with the same structures. Multicollinearity was tested to evaluate the reliability of the updated models, and a new 95% prediction interval for usability was estimated. In addition, quantitative changes in the future sediment load of domestic rivers were analyzed using the future rainfall data obtained from climatic change scenarios and the watershed area. The results of this study reconfirmed that periodic updates are required for the existing empirical models. It is expected that the empirical models will be used in various ways, such as for the prediction of future sediment loads for river sediment management.

Exploiting satellite data for total direct runoff prediction using CN-based MSME model

This paper explores the potential to enhance the functionality of the modified Sahu-Mishra-Eldho model (MSME-CN) using indirect soil moisture measurements derived from satellite data. The current version of the MSME-CN model is not applicable in ungauged watersheds due to the necessity of calibrating the crucial parameter α, which reflects soil saturation, based on measured rainfall-runoff events. We hypothesize that the Normalized Difference Vegetation Index (NDVI) can serve as an indirect indicator of soil moisture to assess the soil saturation parameter α in the MSME model. This hypothesis was tested across five different watersheds, three located in the southeastern USA and two in southern Poland. The NDVI product, developed from data obtained from the Advanced Very High-Resolution Radiometer (AVHRR), was utilized in this study. Results indicate that NDVI is a robust indicator of soil moisture for representing the α parameter in the MSME model. The correlation coefficient between α and NDVI a day prior to a rainfall event was around 0.80 for the WS80 and Kamienica watersheds and nearly 0.60 for the other watersheds. The analysis corroborates the hypothesis that NDVI can serve as an indirect parameter of soil moisture to assess the soil saturation parameter α in the MSME-CN model. Based on Nash-Sutcliffe Efficiency (NSE) statistics, the total direct runoff predicted by the MSME-CN model, with the α parameter updated using NDVI, was rated ‘very good’ for the WS80 and AC11 watersheds, ‘good’ for the Kamienica watershed, ‘satisfactory’ for Stobnica, and ‘unsatisfactory’ for the high forest density WS14 watershed, potentially highlighting the model's limitation in such watersheds.

Climate-associated risks for reliability, resilience, vulnerability and sustainability of irrigation reservoirs in small ungauged watersheds

Climate change poses an existential threat to the functional integrity of the network of agricultural reservoirs responsible for sustaining rice production in South Korea. Anthropogenic interventions have triggered a cycle of droughts and floods by magnifying the already existing supply–demand gap, which is seriously hampering reservoir performance. The watersheds of most agricultural reservoirs in South Korea are ungauged, and inflow and outflow data are not measured. In this study, recorded volumetric water storage data from five reservoirs were used to derive observed inflow data based on the hydrological reasoning and reservoir operation rule curves. These data were then used to evaluate the performance of a lumped rainfall-runoff model in simulating the runoff generated in upstream ungauged watersheds. A water balance analysis was conducted for 400 major agricultural reservoirs spread throughout the country during 1973–2017. Performance indices were estimated based on inflow availability in the reservoirs and downstream irrigation demands. Empirical relationships were derived to predict performance indices by using the reservoirs’ physical characteristics as input variables. Generally, reservoirs received substantially higher inflows compared to irrigation demands, suggesting sufficient water availability during the rice growing season (Apr−Sep). An out-of-phase cycle of peak inflow generation and water demand periods induced frequent failure events during the rice nursery transplanting period (May−Jun), and reservoir sustainability was minimal in these months. Southern and southwestern major agricultural regions with a higher density of small- to medium-sized reservoirs would be at higher risk of climate-induced water scarcity. Upsizing storage capacity could alleviate the water deficit in reservoirs associated with watersheds featuring higher means and lower standard deviations of runoff. A watershed-to-irrigated area ratio > 5 indicated acceptable reservoir sustainability.

Geospatial modeling of hydrological ecosystem services in an ungauged upper Yamuna catchment using SWAT

Hydrological ecosystem services (HES) are vital for resource allocation, conservation prioritization, and climate change adaptation. However, research gaps persist due to limited understanding of complex hydrological systems and inadequate ground-station data, especially in ungauged watersheds with complex terrains. The present study addresses these gaps by estimating and mapping HES descriptors using regionalization techniques in an ungauged Aglar watershed. Additionally, it conducts a temporal analysis of hydrological fluxes using Soil and Water Assessment Tool (SWAT). Aglar is a constituent sub-watershed of a gauged watershed named Bausan having two discharge observation stations located at Naugaon and Bausan sites in Uttarakhand, India. A nested parameterization approach was adopted to represent the hydrological variabilities over the Bausan watershed with one outlet at Aglar. The stream flow derived from the calibrated Bausan watershed was used as synthetic observations in SWAT model setup for Aglar and found to yield very good statistical results. Calibration yielded coefficient of determination (R2) = 0.91, Nash Sutcliff Efficiency (NSE) = 0.91, and standardized root mean square error (RSR) = 0.29, while the validation yielded R2 = 0.68, NSE = 0.50, and RSR = 0.37. Parameters related to base flow or stream flow were the most sensitive in the model's output. Water balance analysis reveals 36% of precipitation transformed into stream flow, with direct runoff accounting for 72% and base flow for 28%. Forest cover contributed approximately 50.68% of precipitation through evapotranspiration. The study identifies a maximum sediment load of approximately 26 t/ha, indicating fragility of the landscape. Non-parametric tests such as Mann Kendall and Sen's slope indicated increasing trends in surface runoff, lateral flow, water yield, and groundwater recharge. The analysis of the empirical cumulative distribution function demonstrated that all hydrological components exhibit trends similar to precipitation. Spatially and temporally, variations in HES provisioning were observed, with forests surpassing non-forest areas. These findings emphasize the value of HES descriptors in analyzing spatiotemporal changes in HES provisioning and offer valuable insights for policymakers for future policy dialogues. This study lays the groundwork for further investigations into the hydrological processes of ungauged watersheds.

Identification of potential zones on the estimation of direct runoff and soil erosion for an ungauged watershed based on remote sensing and GIS techniques

An investigation of soil and water resources is essential to determine the future scenario of water management and water resources to attain food and water security. The improper management of watersheds results in a huge amount of sediment loss and surface runoff. Therefore, the present study was carried out to estimate the surface runoff and soil erosion using the Soil Conservation Service Curve Number (SCS-CN) method and RUSLE approach, respectively. These have been estimated using geospatial technologies for the ungauged Mandri river watershed from the Kanker district of Chhattisgarh State in India. The runoff potential zones, which are defined by the area's impermeable surfaces for a given quantity of precipitation were identified based on curve numbers at the sub-watershed levels. The land use data were collected from LISS IV images of 2009. The results showed that the average volume of runoff generated throughout the 16 years (2000-2015) was 14.37 million cubic meters (mM3). While average annual soil loss was found to be 17.23 tons/ha/year. Most of the eroded area was found to be around the major stream in a drainage system of Mandri River and on higher slopes of the terrain in the watershed. This study revealed that surface runoff and soil erosion are primary issues, which adversely affected the soil and water resources in this watershed. Therefore, suitable water harvesting sites and structures can be constructed based on the potential runoff zone and severity of soil erosion to conserve the soil and water in the watershed.

Runoff hydrographs using Snyder and SCS synthetic unit hydrograph methods: A case study of ungauged watersheds of Imam Turki Bin Abdullah Royal Nature Reserve (ITBA) - Saudi Arabia

Background: The estimation of surface runoff and discharge peak using the Snyder’s model and SCS Dimensionless UH (SCS-DUH) method are one of the best alternatives solutions appropriate for analyzing the hydrologic systems in the ungauged watersheds. These two methods were applied to estimate the discharge peak and design their hydrographs in the drainage basins of Royal Nature Reserve (Imam Turki Bin Abdullah) in Saudi Arabia. - Aims: To estimate the discharge peak by applying Snyder’s model and SCS-DUH method using the morphometric parameters deriving from the SRTM DEM30m and three rainstorms of 40, 50 and 60mm. - Methods: The estimation of discharge peak (Qp) using Snyder’s model was depended on the SRTM DEM30m outputs and the hydrology parameters of the duration (Td) of Unit Hydrograph (UH), time to peak (Tp), time base of the UH (Tb), UH width at 50% and 75% of the peak discharge. However, the estimation of the discharge peak by The SCS-DUH method was computed using the drainage area (A), the direct runoff (Qd) and the time to peak (Tp). The direct runoff was determined by the runoff ratio 0.20 of the unimproved areas, proposed by the Soil Conservation Service (SCS). But, the time to peak was computed using the average of the concentration time obtained from Johnstone-Cross and Dooge models. The hydrographs design was done by the time ratio method recommended by the SCS. - Results: The estimated discharge peak computed by Snyder’s model varies with a maximum of 93.4 to 2092.9 m3/s, an average of 61.1 to 1368.4 m3/s and a minimum of 28.7 to 644.0 m3/s in Al Shuwaiki and Khtal watersheds, respectively. The estimated discharge peak computed by SCS-DUH method was ranged from 21.5 to 482.3 m3/s for the rainstorm of 40mm in Al Suwayki and Khtal, respectively, from 28.5 to 602.9 m3/s for the rainstorm of 50mm and from 34.1 to 723.5 m3/s for the rainstorm of 60mm, in Um Nqy and Khtal, respectively. - Conclusion: The hydrology study of ungauged basins in Royal Nature Reserve (ITBA) shows the great potential of the integrated emplyement of GIS technics and the hydrology models to estimate the discharge peak and design their hydrographs.

A Critical Evaluation of Streamflow Estimation of an Ungauged Watershed using a Gauged Watershed of Same River Basin – Case of Nilwala River Basin

Transferring hydrological model parameters of a nearby gauged watershed to an ungauged watershed is being widely practiced to predict streamflow in the ungauged watershed. However, proper conclusions on transferring data from the main to sub-watershed and vice versa need to be established, then only watershed managers can plan where to locate or ensure optimum use of resources when establishing gauging stations. Therefore, the present study aimed to evaluate the performance of spatial transferability by using the Tank Model, which was initially calibrated to daily streamflow from water year 2008/09 to 2017/18 observed at two gauged watersheds in Nilwala River Basin, and then transferred from main to sub-watershed and vice versa. The efficiency of predictions under two parameter transferability scenarios was highly acceptable where the Mean Ratio of Absolute Error (MRAE) values were 0.35-0.39 and, the accuracy of annual water balance was 91.8%-93.3%. Most importantly, the best results were in the intermediate flow regime with an accuracy of 61%-70%. The monthly-scaled streamflow indicated an accuracy of 83%-92%. The present study demonstrated the liability of the Tank Model for successful estimation of daily streamflow with an average accuracy of 63% under spatial parameter transferability thereby with only a single wellmanaged streamflow gauging station, Tank Model can be applied for sustainable water resources management in similar ungauged watersheds of the basin.

Evaluating the Options for Streamflow Modelling in Ungauged Watersheds for Sustainable Engineering Designs - A Case Study at Attanagalu River Basin, SriLanka

Water management and infrastructure designs all over the world are mostly associated with ungauged watersheds. The critical task is to estimate the streamflow at a desired location and then to convert the streamflow as design streamflow by incorporating safety factors. In the current setting, practicing engineers find a deficiency in guidelines to select a model and also find out ways to incorporate safety factors ensuring sustainable design of water infrastructure. The guideline of Irrigation Department, Sri Lanka, provides guidance on the use of an empirical model while the Snyder’s synthetic unit hydrograph method and HEC-HMS model are two other popular process-based models opted by most recent modelers to estimate watershed streamflow. The authors carried out a critical literature review and a case study in the Kalu river basin to investigate the issue of model selection faced by ungauged watershed modelers. A streamflow modeler of ungauged watersheds requires to responsibly select a streamflow model, model parameters and safety factors. In this context, the major challenge in the Sri Lankan practice is the unavailability of catchment studies with comparative evaluations of model studies with observed streamflow. The objective of the present work was to first determine the rationale for a design engineer to select a watershed model and parameters when computing design streamflow from an ungauged watershed and then to verify the selection using observed streamflow. Accordingly, the Irrigation Department Empirical Model (IDEM), Snyder’s UH Model (SUHM) and HEC-HMS model were selected and developed for Dunamale watershed of the Attanagalu river basin using methods and parameters from responsible publications. Then the model estimates were critically evaluated by comparing the observed streamflow. IDEM estimates of monthly yield estimations closely matched with the observed data. In the case of overall hydrographs, the IDEM indicated a MRAE value of 0.43 for design rainfall estimation, while the same for HEC-HMS model and SUHM method were 0.63 and 0.61, respectively. The qualitative review based on the uncertainties associated with the computations using IDEM, HEC-HMS and SUHM models classified them into three, as Moderately uncertain, Uncertainty between moderate and high, and Highly uncertain, respectively. The study concluded that, amongst the three selected models, the IDEM is still the best option available for an ungauged watershed modeler to compute design streamflow but pointed to the urgent need for upgrading the ID guidelines with focused research. The study pointed to the importance of identifying appropriate models and parameters for each watershed for meaningful use of process based streamflow models for design of water infrastructure. It was identified that explicit safety factors should be established to incorporate the uncertainties associated with the models and their parameters.

Machine learning‐based regional flood frequency analysis of Indian watersheds

AbstractEstimates of the flood quantile for ungauged watersheds are crucial for water resources management but challenging due to the nonlinear complex hydrological system. For ungauged watersheds, estimating flood quantiles relies on various interdependent physiometeorological variables, many of which are not adequately considered in regional flood frequency analysis (RFFA). In this study, we utilized the random forest (RF) and support vector regression (SVR) algorithms, which can learn the nonlinear relationship between the physiometeorological variables and flood quantiles for RFFA. Thirteen physiometeorological variables that were not collectively employed before were used to estimate the 10‐year, 50‐year, and 100‐year return period flood quantiles (Q10, Q50, and Q100), respectively, for 39 watersheds spread across India. The RF and SVR models were trained on 29 (75%) watersheds to estimate individual flood quantiles and were subsequently tested on the remaining ten (25%) ungauged watersheds. The R2 achieved by RF is 0.862, 0.813, and 0.845, and SVR is 0.807, 0.793, and 0.789 for Q10, Q50, and Q100, respectively. Overall, the results indicate that RF can effectively learn the nonlinear relationships, while SVR with a linear kernel requires further improvement to estimate reliable flood quantiles. The study demonstrates that machine learning algorithms, with appropriate physiometeorological input datasets, can be used to estimate flood quantiles even in the sparse data region.

Comparative Application of Rain Gauge, Ground- and Space-Borne Radar Precipitation Products for Flood Simulations in a Dam Watershed in South Korea

This study presents a comparative analysis of flood simulations using rain gauge, ground- and space-borne radar precipitation products. The objectives are to assess the effectiveness of two radar-based data sources, namely the Radar-AWS Rainrates (RAR) and Integrated Multi-satellite Retrievals for GPM (IMERG), in a dam watershed with gauge observations, and explore the modeling feasibility of integrating the half-hourly IMERG satellite precipitation in regions with ungauged or limited observational area. Two types of HEC-HMS models were developed, considering areal-averaged and spatially distributed gridded data simulations utilizing eight selected storm events. The findings indicate that the RAR data, although slightly underestimate precipitation compared to the gauge measurements, accurately reproduce hydrographs without requiring parameter adjustments (Nash–Sutcliffe efficiency, ENS, 0.863; coefficient of determination, R2, 0.873; and percent bias, PBIAS, 7.49%). On the other hand, flood simulations using the IMERG data exhibit lower model efficiency and correlation, suggesting potential limitations in ungauged watersheds. Nevertheless, with available discharge data, the calibrated model using IMERG shows prospects for utilization (ENS 0.776, R2 0.787, and PBIAS 7.15%). Overall, this research offers insights into flood simulations using various precipitation products, emphasizing the significance of reliable discharge data for accurate hydrological modeling and the need for further evaluation of the IMERG product.

A Case Study on GIS Based Method to Develop Geographically Distributed Synthetic Unit Hydrograph using Geo-Spatial Tools in Ungauged Sub-watershed of Ambika River Basin, Gujarat

Abstract In this work, using Snyder, SCS, and the geographically distributed approaches, synthetic unit hydrographs are developed for the recurring flood-impacted ungauged sub-watershed in the Ambika river basin of Gujarat. Due to the scarcity of observed rainfall-runoff data, enforcement of effective flood control measures have not been implemented in this area. Hence the districts in the Ambika river basin regularly experience unexpected floods during the monsoon season, resulting in loss of life and property. The unit hydrograph of a watershed is essential for developing a flood hydrograph, which is useful for flood forecasting. The majority of Indian watersheds are not gauged, and there is not enough rainfall-runoff data to develop a unit hydrograph. However, there are several approaches available that may be used to develop synthetic unit hydrographs using different regionally based empirical equations, and some of them are mentioned above. Using geospatial tools is one of the GIS-based methods for obtaining synthetic unit hydrographs that solely take into account the digital elevation model (DEM) which eliminates the use of regional empirical constants. The geographically distributed synthetic unit hydrograph (GDSUH) approach is employed here on the idea of distributed time-area unit hydrograph and spatial distribution of overland and channel flow velocity. The cumulative travel time map has been divided into isochrones to develop a time-area curve and the resulting unit hydrograph. In this technique, a pure translation flow process is assumed. This approach is most appropriate for watersheds of smaller size. Hence, in this present study, a smaller, unmeasured sub-watershed with an area of 46 km2 is chosen, and the synthetic unit hydrographs are derived using each of the three approaches. The findings from this study were analyzed to determine the most appropriate method for this region. It was observed that the GDSUH is more suitable in comparison to the other two conventional synthetic unit hydrograph approaches. In smaller and un-gauged watersheds, this method gives a conceptual model for forecasting stream flow hydrographs that will capture the geographically distributed nature of the rainfall-runoff process. Additionally, it has the advantages of being simple, quick, inexpensive, and time-saving, especially for smaller regions that lack rainfall-runoff data.

Estimation of Runoff in Data Scare Watershed of Middle Gujarat, India Using HEC-HMS Model

The accurate estimation of the hydrologic response in water resources planning and development is crucial. In this study, the rainfall-runoff relationship and hydrologic response of the Devgadh Baria watershed in the Panam river catchment were investigated. Due to limited availability and accuracy of runoff data, Clark UH, and SCS UH models were explored for runoff estimation. Remote sensing and GIS techniques were integrated to assess the ungauged watershed in the Middle Gujarat region.
 Rainfall and discharge data, Sentinel-2 data, 12.5m ALOS PALSAR DEM, base maps, and soil maps were collected from various sources. The study area's five different events were randomly selected, and a Thiessen map was prepared to calculate weighted rainfall. Geomorphological parameters and Horton's ratios were estimated using remote sensing and GIS software ILWIS.
 The composite curve number (CN) for the watershed was found to be 77.69. Excess-rainfall hyetographs were separated using the SCS-loss method, and the time to peak estimated by the models matched well with observed data, except for a one-day delay in the HEC-HMS based Clark UH model.
 Four performance indices, namely NSE, PBIAS, RMSE, and d, were used to compare model performance. The SCS UH model demonstrated the highest NSE value (-0.28) for event-1, while the HEC-HMS based SCS UH model showed the lowest NSE value (-0.28) for event-1. The SCS model outperformed the other models based on performance indicators NSE, RMSE, and d. Overall, the SCS UH approach provided the best estimation of peak discharge and time to peak discharge. The direct surface runoff hydrographs estimated using.

How to enhance hydrological predictions in hydrologically distinct watersheds of the Indian subcontinent?

AbstractAccurate hydrological predictions are required to prepare for the impacts of climate change, especially in India, which experiences frequent floods and droughts. However, the complex hydrological processes of its distinct watersheds and limited data make it challenging to deliver highly‐performant hydrologic predictions using conventional models. Moreover, it remains uncertain where the limits of predictability are and whether recently‐popular deep learning approaches can offer significant improvements. Here, we tested the first instance of the hydrologic model based on long short‐term memory (LSTM) for 55 Indian watersheds, using a new dataset comprising forcing, attributes, and discharge data. Our results show that the LSTM model provides much‐improved performance compared to conventional models in India, providing a median Nash‐Sutcliffe efficiency (NSE) of 0.56. The LSTM model trained on all the watersheds is more favourable to those trained on individual or homogeneous watersheds, as it benefits from a broader range of hydrological processes and patterns in the input data. However, the LSTM model performs poorly for non‐perennial, large, and semi‐arid climate zone watersheds due to its inability to simulate the complex hydrological processes specific to these environments. Integrating lagged observations with the LSTM model (referred to as DI‐LSTM) improved the predictions in such watersheds and enhanced the median NSE to 0.76 by capturing the temporal dependencies and historical patterns that influence hydrological processes. Overall, the contrast of model performance across watersheds suggests major limitations could be associated with the quality of forcing data, and the slow flow or groundwater processes are highly important in the Indian subcontinent. Notably, both LSTM and DI‐LSTM models performed reasonably well for predictions in ungauged watersheds. The findings of this study demonstrate that data‐sparse countries, too, can benefit from big‐data deep learning and point out further avenues toward model improvements.

Developing an Empirical Relations between Nash Model Parameters and Watersheds Topographical Characteristics for Predicting Direct Runoff Hydrograph

The limited availability of the recorded rainfall-runoff data for many watersheds restricts the development and management of different activities of water resources. To overcome this limitation, the Natural Resources Conservation Service (NRCS) for estimating storm excess rainfall and momentum and optimization methods were combined in a mathematical model to estimate the optimal parameters of Nash Instantaneous unit hydrograph (IUH) and resulting direct runoff hydrograph (DRH), using a developed computer program in MATLAB. The available recorded data of 14 storms (out of 18) of four watersheds in northern Iraq have been applied in the calibration stage. An empirical relationship was developed between the average of each IUH optimal parameter (obtained by optimization as an optimal method according to the applied tests) and the effective watershed topographical characteristics. The developed empirical relations were used in the verification stage to estimate the IUH parameters and DRH for the verification storms and compare with that resulted from Haan’s empirical relations and optimization method. The statistical tests showed that the developed empirical relations efficiency was better than that of Haan’s method and close to that of the recorded storm by optimization method, where the average value of the Nash-Sutcliffe Efficiency for the four watersheds resulted from applying the optimization method, Haan’s method and the developed empirical relations were 0.925, 0.587, 0.883 respectively. The results indicated the developed model’s ability to estimate the IUH and direct runoff hydrograph for ungauged watersheds in northern Iraq.