A comparative study of data-driven models for runoff, sediment, and nitrate forecasting

Effective prediction of qualitative and quantitative indicators for runoff is quite essential in water resources planning and management. However, although several data-driven and model-driven forecasting approaches have been employed in the literature for streamflow forecasting, to our knowledge, the literature lacks a comprehensive comparison of well-known data-driven and model-driven forecasting techniques for runoff evaluation in terms of quality and quantity. This study filled this knowledge gap by comparing the accuracy of runoff, sediment, and nitrate forecasting using four robust data-driven techniques: artificial neural network (ANN), long short-term memory (LSTM), wavelet artificial neural network (WANN), and wavelet long short-term memory (WLSTM) models. These comparisons were performed in two main tiers: (1) Comparing the machine learning algorithms' results with the model-driven approach; In order to simulate the runoff, sediment, and nitrate loads, the Soil and Water Assessment Tool (SWAT) model was employed, and (2) Comparing the machine learning algorithms with each other; The wavelet function was utilized in the ANN and LSTM algorithms. These comparisons were assessed based on the substantial statistical indices of coefficient of determination (R-Squared), Nash-Sutcliff efficiency coefficient (NSE), mean absolute error (MAE), and root mean square error (RMSE). Finally, to prove the applicability and efficiency of the proposed novel framework, it was successfully applied to Eagle Creek Watershed (ECW), Indiana, U.S. Results demonstrated that the data-driven algorithms significantly outperformed the model-driven models for both the calibration/training and validation/testing phases. Furthermore, it was found that the coupled ANN and LSTM models with wavelet function led to more accurate results than those without this function.

Comparing critical source areas for the sediment and nutrients of calibrated and uncalibrated models in a plateau watershed in southwest China

Controlling non-point source pollution is often difficult and costly. Therefore, focusing on areas that contribute the most, so-called critical source areas (CSAs), can have economic and ecological benefits. CSAs are often determined using a modelling approach, yet it has proved difficult to calibrate the models in regions with limited data availability. Since identifying CSAs is based on the relative contributions of sub-basins to the total load, it has been suggested that uncalibrated models could be used to identify CSAs to overcome data scarcity issues. Here, we use the SWAT model to study the extent to which an uncalibrated model can be applied to determine CSAs. We classify and rank sub-basins to identify CSAs for sediment, total nitrogen (TN), and total phosphorus (TP) in the Fengyu River Watershed (China) with and without model calibration. The results show high similarity (81%-93%) between the identified sediment and TP CSA number and locations before and after calibration both on the yearly and seasonal scale. For TN alone, the results show moderate similarity on the yearly scale (73%). This may be because, in our study area, TN is determined more by groundwater flow after calibration than by surface water flow. We conclude that CSA identification with the uncalibrated model for TP is always good because its CSA number and locations changed least, and for sediment, it is generally satisfactory. The use of the uncalibrated model for TN is acceptable, as its CSA locations did not change after calibration; however, the TN CSA number changed by over 60% compared to the figures before calibration on both yearly and seasonal scales. Therefore, we advise using an uncalibrated model to identify CSAs for TN only if water yield composition changes are expected to be limited. This study shows that CSAs can be identified based on relative loading estimates with uncalibrated models in data-deficient regions.

Combining downscaled-GRACE data with SWAT to improve the estimation of groundwater storage and depletion variations in the Irrigated Indus Basin (IIB)

The growth of agricultural production systems is a major driver of groundwater depletion worldwide. Balancing groundwater supply and food production requires localized understanding of groundwater storage and depletion variations in response to diverse cropping systems and surface water availability for irrigation. While advances through Gravity Recovery and Climate Experiment (GRACE) have facilitated estimating the groundwater storage (GWS) changes in recent years, the coarse resolution of GRACE data hinders the characterization of GWS variation hotspots. Herein, we present a novel spatial water balance approach to improve the distributed estimation of groundwater storage and depletion changes at a spatial scale that can detect the hotspots of GWS variation. We used a mixed geographically weighted regression (MGWR) model to downscale GRACE Level-3 data from coarse resolution (1° × 1°) to fine scale (1 km × 1 km) based on high resolution environmental variables. We then combined the downscaled GRACE-based GWS variations with results from a calibrated Soil and Water Assessment Tool (SWAT) model. We demonstrate an application of the approach in the Irrigated Indus Basin (IIB). Between 2002 and 2019, total loss of groundwater reserves varied in the IIB's 55 canal command areas with the highest loss observed in Dehli Doab by >50 km³ followed by 7.8-49 km³ in the upstream, and 0.77-7.77 km³ in the downstream canal command areas. GWS declined by -325.55 mm/year at Dehli Doab, followed by -186.86 mm/year at BIST Doab, -119.20 mm/year at BARI Doab, and -100.82 mm/year at JECH Doab. The rate of groundwater depletion is increasing in the canal command areas of Delhi Doab and BIST Doab by 0.21-0.35 m/year. Larger groundwater depletion in some canal command areas (e.g., RACHNA, BIST Doab, and Delhi Doab) is associated with the rice-wheat cropping system, low rainfall, and low flows from tributaries.

Improvement of simulating sub-daily hydrological impacts of rainwater harvesting for landscape irrigation with rain barrels/cisterns in the SWAT model

Rain barrels/cisterns, a popular type of low impact development (LID) practice, can restore urban hydrological processes and decrease municipal water use by harvesting roof runoff for later use, such as landscape irrigation. However, tools to assist decision makers in creating efficient rainwater harvesting and reuse strategies are limited. This study improved the Soil and Water Assessment Tool (SWAT) in simulating the subdaily hydrological impacts of rainwater harvesting for landscape irrigation with rain barrels/cisterns, including the simulation of rainwater harvesting with rain barrels/cisterns, rainwater reuse for auto landscape irrigation, evapotranspiration, initial abstraction, impervious area, soil profile, and lawn management operation. The improved SWAT was applied in the urbanized Brentwood watershed (Austin, TX) to evaluate its applicability and investigate the impacts of rainwater harvesting and reuse strategies on the reductions and reduction efficiencies (reductions per volume of rain barrels/cisterns implemented) of field scale runoff (peak and depth) and watershed scale streamflow (peak and volume) for two storm events. Scenarios explored included different sizes of rain barrels/cisterns, percentages of rooftop areas with rain barrels/cisterns implemented, auto landscape irrigation rates, and landscape irrigation starting times. The performance of rainwater harvesting and reuse strategies, which is determined by features of fields, watersheds, and storm events, varied for different reduction goals (streamflow or runoff, and peak or depth/volume). For instance, the scenario with rain barrel/cistern sizes of 7.5 mm (design runoff depth from treated roof area) and the scenario with 10% of suitable area implemented with rain barrels/cisterns provided the highest peak streamflow reduction efficiency and total streamflow volume reduction efficiency at the watershed scale, respectively for the smaller storm event. To achieve sustainable urban stormwater management, the improved SWAT model has enhanced capability to help stakeholders create efficient rainwater harvesting and reuse strategies to reduce field scale runoff and watershed scale streamflow.