Assessing the impacts of climate change on streamflow dynamics: A machine learning perspective.

This study investigates changes in river flow patterns, in the Hunza Basin, Pakistan, attributed to climate change. Given the anticipated rise in extreme weather events, accurate streamflow predictions are increasingly vital. We assess three machine learning (ML) models - artificial neural network (ANN), recurrent neural network (RNN), and adaptive fuzzy neural inference system (ANFIS) - for streamflow prediction under the Coupled Model Intercomparison Project 6 (CMIP6) Shared Socioeconomic Pathways (SSPs), specifically SSP245 and SSP585. Four key performance indicators, mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination (R2), guide the evaluation. These models employ monthly precipitation, maximum and minimum temperatures as inputs, and discharge as the output, spanning 1985-2014. The ANN model with a 3-10-1 architecture outperforms RNN and ANFIS, displaying lower MSE, RMSE, MAE, and higher R2 values for both training (MSE = 20417, RMSE = 142, MAE = 71, R2 = 0.94) and testing (MSE = 9348, RMSE = 96, MAE = 108, R2 = 0.92) datasets. Subsequently, the superior ANN model predicts streamflow up to 2100 using SSP245 and SSP585 scenarios. These results underscore the potential of ANN models for robust futuristic streamflow estimation, offering valuable insights for water resource management and planning.

Intercomparison of SWAT and ANN techniques in simulating streamflows in the Astore Basin of the Upper Indus.

The current research work was carried out to simulate monthly streamflow historical record using Soil and Water Assessment Tool (SWAT) and Artificial Neural Network (ANN) at the Astore Basin, Gilgit-Baltistan, Pakistan. The performance of SWAT and ANN models was assessed during calibration (1985-2005) and validation (2006-2010) periods via statistical indicators such as coefficient of determination (R2), Nash-Sutcliffe efficiency (NSE), percent bias (PBIAS), and root-mean-square error (RMSE). R2, NSE, PBIAS, and RMSE values for SWAT (ANN with Architecture (2,27,1)) models during calibration are 0.80 (0.88), 0.73 (0.82), 15.7 (0.008), and 79.81 (70.34), respectively, while during validation, the corresponding values are 0.71 (0.86), 0.66 (0.95), 17.3 (0.10), and 106.26 (75.92). The results implied that the ANN model is superior to the SWAT model based on the statistical performance indicators. The SWAT results demonstrated an underestimation of the high flow and overestimation of the low flow. Comparatively, the ANN model performed very well in estimating the general and extreme flow conditions. The findings of this research highlighted its potential as a valuable tool for accurate streamflow forecasting and decision-making. The current study recommends that additional machine learning models may be compared with the SWAT model output to improve monthly streamflow predictions in the Astore Basin.