On the sensitivity of urban hydrodynamic modelling to rainfall spatial and temporal resolution

Abstract

Abstract. Cities are increasingly vulnerable to floods generated by intense rainfall, because of their high degree of imperviousness, implementation of infrastructures, and changes in precipitation patterns due to climate change. Accurate information of convective storm characteristics at high spatial and temporal resolution is a crucial input for urban hydrological models to be able to simulate fast runoff processes and enhance flood prediction. In this paper, a detailed study of the sensitivity of urban hydrological response to high resolution radar rainfall was conducted. Rainfall rates derived from X-band dual polarimetric weather radar for four rainstorms were used as input into a detailed hydrodynamic sewer model for an urban catchment in Rotterdam, the Netherlands. Dimensionless parameters were derived to compare results between different storm conditions and to describe the effect of rainfall spatial resolution in relation to storm and hydrodynamic model properties: rainfall sampling number (rainfall resolution vs. storm size), catchment sampling number (rainfall resolution vs. catchment size), runoff and sewer sampling number (rainfall resolution vs. runoff and sewer model resolution respectively). Results show catchment smearing effect for rainfall resolution approaching half the catchment size, i.e. for catchments sampling numbers greater than 0.5 averaged rainfall volumes decrease about 20%. Moreover, deviations in maximum water depths, form 10 to 30% depending on the storm, occur for rainfall resolution close to storm size, describing storm smearing effect due to rainfall coarsening. Model results also show the sensitivity of modelled runoff peaks and maximum water depths to the resolution of the runoff areas and sewer density respectively. Sensitivity to temporal resolution of rainfall input seems low compared to spatial resolution, for the storms analysed in this study. Findings are in agreement with previous studies on natural catchments, thus the sampling numbers seem to be promising as an approach to describe sensitivity of hydrological response to rainfall variability for intra-urban catchments and local convective storms. More storms and different urban catchments of varying characteristics need to be analysed in order to validate these findings.

Critical scales to explain urban hydrological response

“…Finding a proper match between rainfall resolution and hydrological model structure and complexity is important for reliable flow prediction (Berne et al, 2004;Ochoa-Rodriguez et al, 2015;Pina et al, 2016;Rafieeinasab et al, 2015;Yang et al, 2016). High resolution rainfall data are required to reduce errors in estimation of hydrological responses in small urban catchments (Niemczynowicz, 1988;Schilling, 1991;Berne et al, 2004;Bruni et al, 2015;Yang et al, 2016). New technologies and instruments have been developed in 20 order to improve rainfall measurements and capture its spatial and temporal variability (Einfalt et al, 2004;Thorndahl et al, 2017).…”

“…, where L RA indicates the spatial resolution of the runoff model, defined as the square root of the averaged sub-catchment size (Bruni et al, 2015). Lower values of this ratio indicate that the model is unable to capture rainfall variability, while higher values indicate possible incorrect transformation of rainfall into runoff.…”

“…In this study, for 3000 m spatial rainfall resolution values are bigger than 1, so poor model performance at this resolution is expected. The runoff sampling number suggests that SD1 will not be able to capture rainfall variability, because it presents Bruni et al (2015), where this parameter varied between 2.6 for high resolution and 93 for lower resolution. The sewer sampling number applied to SD2…”

“…Dimensionless sampling numbers, presented at first by Ogden and Julien (1994), and then re-proposed by Bruni et al (2015), are presented in Table 6 for the three models (for underlying equations see Section 3.2.1). SD2 and FD model have the same contributing area and network length, hence they show that values for the catchment sampling number and sewer sampling number are the same.…”

“…Dimensionless parameters as proposed by Bruni et al (2015) and Ogden and Julien (1994) were determined to investigate the interaction and relation between rainfall resolution and different model properties and characteristic. The catchment sampling number ∆s L C was introduced as the ratio between the rainfall spatial resolution ∆s and the characteristic length of the catchment L C (square root of the total area).…”