Self-organising map rainfall-runoff multivariate modelling for runoff reconstruction in inadequately gauged basins

Abstract

Water resources assessment activities in inadequateiy gauged basins are often significantly constrained due to the insufficiency or totai iaci< of hydro-meteorologicai data, resulting in huge uncertainties and ineffectuai performance of water management schemes, in this study, a new methodology of rainfall-runoff modelling using the powerful clustering capability of the selforganising map (SOM), unsupervised artificial neural networks, is proposed as a viable approach for harnessing the muitivariate correlation between the typically long record rainfall and short record runoff in such basins. The methodology was applied to the inadequately gauged Osun basin in southwest Nigeria for the sole purpose of extending the available runoff records and, through that, reducing water resources planning uncertainty associated with the use of short runoff data records.The extended runoff records were then analysed to determine possible abstractions from the main river source at different exceedance probabilities. This study demonstrates the successful use of emerging tools to overcome practical problems in sparsely gauged basins.

Non-stationary hydrological model parameters: a framework based on SOM-B

“…The upper index * indicates that the training of the SOM is finished. The SOM has been successfully applied in other studies, including the extension of flow records (Adeloye and Rustum, ), model evaluation (Herbst and Casper, ), and interpolation of precipitation (Kalteh and Berndtsson, ).…”

Role of hydrology in managing consequences of a changing global environment

“…Alvisi et al (2012) explore uncertainties in flood forecasts by comparing artificial neural network (ANN) and evolutionary polynomial regression (EPR) models using grey numbers, finding in their trial on the Tiber River in Italy that ANN performance exceeds that of the EPR model with increasing lead times. Adeloye & Rustum (2012) similarly find favour in an ANN approach, in a water resources study in Nigeria where exceptionally large uncertainties arising from a very sparse river flow gauging network can be compensated for, to some extent, by the use of longer rainfall records, and assist in enhancing the planning of water resources investments. A range of statistical models are reviewed by Harvey et al (2012) to develop a new strategy for selecting appropriate methods for the routine infilling of gaps in national river flow data sets, offering a means of enhancing the utility of national archives.…”

A novel workflow for streamflow prediction in the presence of missing gauge observations

“…Multiple imputation methods, such as multivariate imputation by chained equations, reduce this bias by replacing missing data points with predictions from an ensemble of regressors obtained from multiple subsets of the complete data (Van Buuren, 2018). Each of the imputation techniques above is used extensively in hydrology (Adeloye and Rustum, 2012;Mwale et al, 2012;Hamzah et al, 2021).…”