Design flood estimation and utility of high‐resolution calibration data in small, heavily urbanised catchments

Abstract

Design flood estimates are often required for small, heavily urbanised catchments, which respond quickly to storm events. However, hydrological models are most frequently calibrated using daily or hourly data on larger, more rural catchments, which respond on much longer timescales. Here, we calibrate a lumped, conceptual rainfall-runoff model (ReFH2) in three small (2-6 km 2 ), heavily urbanised catchments in Swindon, UK, assessing the benefits of using high-resolution temporal and spatial data. Modelling shows that heavy urbanisation does not by itself invalidate the applicability of a lumped, conceptual model. However, we find great dissimilarities between runoff behaviour in different heavily urbanised catchments, with some types behaving similarly to rural catchments. In other cases, response and contributing catchment area can depend more on underground topology than catchment topography. Calibrated runoff response is insensitive to the temporal resolution of the calibration events in all study catchments. Future research should aim to differentiate between different types of heavily urbanised catchment, potentially through landscape metrics to measure the connectivity and isolation of different land surface types. K E Y W O R D S design flood estimation, hydrological modelling, rainfall-runoff, ReFH, urbanisation

Contribution of urbanisation to non-stationary river flow in the UK

“…For example, impervious areas that drain runoff to pervious surfaces may have less effect on river flow than impervious areas that are directly connected to the drainage system (Jacobson, 2011). Vesuviano and Miller (2019) found that in an urban catchment in Swindon, UK, the effect of draining to pervious surfaces could even reduce the peak flows to below what would be expected for a rural catchment. Other reasons may be that it is difficult to isolate the signal of urbanisation from other drivers (e.g., climate change and other headwater land cover changes) or from background hydrologic noise (e.g., the imports and exports of sewage water across catchment boundaries, upstream river regulation, and ground and surface water abstraction).…”

Improving urban flood estimation and flood risk assessment using landscape metrics

“…Secondly, a number of studies have pointed towards the value of being able to differentiate between spatial distributions of land use types for flood estimation in ungauged urban catchments (Wan Jaafar & Han, 2012; Kjeldsen et al, 2013;Vesuviano & Miller, 2019). This research has clearly demonstrated that landscape metrics could better characterize spatial properties of land classes for lumped catchment approaches such as flood estimation in small urban catchments.…”

“…The reliance of UK flood estimation methods on lumped catchment approaches and spatially generalized catchment data has led to concerns that this limits the performance of methods, particularly in small urban catchments (Faulkner et al, 2012;Miller & Hess, 2017;Vesuviano & Miller, 2019). These methods use the index of urban exent -URBEXT -as a means for estimating impervious surfaces effects within the catchment (Bayliss, 1999).…”

Estimating design flood runoff volume

“…In both cases shown in Figure 6, the recorded volume of runoff is greater than the recorded volume of rainfall. In Figure 6a, the general shape of the hydrograph is captured accurately, but the vertical scaling is different, suggesting that the gauged percentage runoff of 117% could result from overestimation at the flow gauge, underestimation at the rain gauges, inappropriate weighting of rain gauges (due to the irregular quadrilateral approximation of the catchment boundary), uncaptured spatial variations in rainfall, or underestimation of contributing area, through neglecting subsurface flows or anthropogenic modification to the catchment (Miller et al, 2014; Vesuviano & Miller, 2019). Figure 6b shows a gauged hydrograph with two equal‐magnitude peaks, while ReFH2 estimates a much smaller second peak than first peak.…”