Run-off modelling of pervious and impervious areas using couple SWAT and a novel machine learning model in cross-rivers state Nigeria

Using the soil and water assessment tool (SWAT), runoff in pervious and impervious urban areas was simulated in this study. In the meantime, as a novel application of machine learning, the emotional artificial neural network (EANN) model was employed to enhance the SWAT obtained for this study. As a result of the EANN model's capabilities in rainfall–runoff phenomena, the SWAT-EANN couple model has been used to assess urban flooding. The pervious, impervious, and water body areas of the study area were classified and mapped to estimate the cover change over three epochs. Land use map, precipitation data, temperature (minimum and maximum) data, wind speed, relative humidity, soil map, solar radiation, and digital elevation model were used as inputs for modelling rainfall–runoff of the study area in the ArcGIS environment. The accuracy assessment of this study was excellent (root-mean-square error 1 mm of precipitation). It also revealed that (a) a land use map illustrating changes in impervious, pervious surface, and water body for 1998, 2008, and 2018; (b) runoff modelling using a historical pattern of rainfall–runoff changes (1998–2018); and (c) descriptive statistical analysis of the runoff results of the research. This research will aid in urban planning, administration, and development. Specifically, it will prevent flooding and environmental problems. HIGHLIGHT This study complimented others on exploring rainfall–runoff modelling of pervious and impervious areas in Cross River State, Nigeria. In this study, the run-up modelling using the soil and water assessment tool model, and geographical information system/remote sensing systems were utilized to extract details of the land cover and to analyse the rainfall–runoff using logistic regression and other descriptive statistical analysis.;