Accurate prediction of precipitation is of paramount importance for effective planning of future water resources. In this study, we focused on the improvement and evaluation of the European Centre for Medium-Range Weather Forecasts (ECMWF) fifth-generation ensemble-based seasonal precipitation prediction product, designated (SEASonal prediction of precipitation (SEAS5)). Three selected linear regression methods, namely ordinary least squares (OLS), flexible least squares (FLS) and the quantile-quantile (Q-Q) methods, were used to develop a correction procedure. The watershed of Lake Urmia was selected as a case study. The application of these augmentation methods has yielded encouraging results, demonstrating an improvement in the statistical metrics of SEAS5 precipitation forecasts for the first and second-coming months. However, all linear projection methods improve the performance of the SEAS5 products. The Q-Q method has shown the highest efficiency among the methods, playing a significant role in improving the accuracy of the hindcast precipitation. A variety of statistics (deterministic, forecast skill and uncertainty scores) were used to evaluate the effectiveness of both the raw and enhanced SEAS5 products. These analyses provide a comprehensive understanding of the performance of the SEAS5 product in its original form and after augmentation. The results highlight the potential of the linear projection method (specifically Q-Q method) to improve the accuracy of hindcast precipitation and provide valuable insights for water resource planning in the study area.

Due to the complicated nature of environmental processes, consideration of uncertainty is an important part of environmental modelling. In this paper, a new variant of the machine learning-based method for residual estimation and parametric model uncertainty is presented. This method is based on the UNEEC-P (UNcertainty Estimation based on local Errors and Clustering–Parameter) method, but instead of multilayer perceptron uses a “fuzzified” version of the general regression neural network (GRNN). Two hydrological models are chosen and the proposed method is used to evaluate their parametric uncertainty. The approach can be classified as a hybrid uncertainty estimation method, and is compared to the group method of data handling (GMDH) and ordinary kriging with linear external drift (OKLED) methods. It is shown that, in terms of inherent complexity, measured by Akaike information criterion (AIC), the proposed fuzzy GRNN method has advantages over other techniques, while its accuracy is comparable. Statistical metrics on verification datasets demonstrate the capability and appropriate efficiency of the proposed method to estimate the uncertainty of environmental models.