High spatial–temporal resolution and integrated surface and subsurface precipitation-runoff modelling for a small stormwater catchment

Abstract

Summary Reliable runoff estimation is important for design of water infrastructure and flood risk management in urban catchments. We developed a spatially distributed Precipitation-Runoff (P-R) model that explicitly represents the land cover information, performs integrated modelling of surface and subsurface components of the urban precipitation water cycle and flow routing. We conducted parameter calibration and validation for a small (21.255 ha) stormwater catchment in Trondheim City during Summer-Autumn events and season, and snow-influenced Winter-Spring seasons at high spatial and temporal resolutions of respectively 5 m × 5 m grid size and 2 min. The calibration resulted in good performance measures (Nash-Sutcliffe efficiency, NSE = 0.65–0.94) and acceptable validation NSE for the seasonal and snow-influenced periods. The infiltration excess surface runoff dominates the peak flows while the contribution of subsurface flow to the sewer pipes also augments the peak flows. Based on the total volumes of simulated flow in sewer pipes (Qsim) and precipitation (P) during the calibration periods, the Qsim/P ranges from 21.44% for an event to 56.50% for the Winter-Spring season, which are in close agreement with the observed volumes (Qobs/P). The lowest percentage of precipitation volume that is transformed to the total simulated runoff in the catchment (QT) is 79.77%. Computation of evapotranspiration (ET) indicated that the ET/P is less than 3% for the events and snow-influenced seasons while it is about 18% for the Summer-Autumn season. The subsurface flow contribution to the sewer pipes are markedly higher than the total surface runoff volume for some events and the Summer-Autumn season. The peakiest flow rates correspond to the Winter-Spring season. Therefore, urban runoff simulation for design and management purposes should include two-way interactions between the subsurface runoff and flow in sewer pipes, and snow-influenced seasons. The developed urban P-R model is useful for better computation of runoff generated from different land cover, for assessments of stormwater management techniques (e.g. the Low Impact Development or LID) and the impacts of land cover and climate change. There are some simplifications or limitations such as the runoff routing does not involve detailed sewer hydraulics, effects of leakages from water supply systems and faulty/illegal connections from sanitary sewer are not considered, the model cannot identify actual locations of the interactions between the subsurface runoff and sewer pipes and lacks parsimony.