Automated modelling of urban rainfall-runoff: knowledge library, conceptual model and data set
Description
This dataset supports automated modelling of rainfall-runoff with the Process-Based Modelling Tool (ProBMoT) (http://probmot.ijs.si/). ProBMoT allows for the integration of domain knowledge, formalized as template components for the construction of the process-based models, into equation discovery from measured data. It automatically identifies both the structure and parameter values of an appropriate process-based model, given: a) a knowledge library (i.e., a mathematical formulation of the selected domain) in the form of model components, or, more specifically, template entities and processes, b) a conceptual model of the observed system, and c) measurements. Namely, ProBMoT can be used to find the structure and parameters’ values of alternative rainfall-runoff models that provide optimal fit against pipe flow measurements. Three alternative methods are provided to describe infiltration: the SCS CN method, the Variable UK runoff equation, and the UK Water Industry Research equation. This approach is transferable to any catchment and enables the discovery of viable rainfall-runoff models. To do so, the user has to specify a conceptual model of the selected case study area, compliant with the knowledge library. Furthermore, the user has to provide rainfall and flow measurement data and catchment characteristics (i.e., land use, topography, soil properties), based on which the values of the model constants can be determined. Data structure: 1) Knowledge library (.pbl file), 2) Conceptual model (.pbm file), 3) Settings (.xml file), 4) Measurements of rain events for model calibration and validation (.data files).