Using Remote Sensing Based Metrics to Quantify the Hydrological Response in a City

Charlotte Wirion,Willy Bauwens,Boud Verbeiren

doi:10.3390/w11091763

Abstract

We propose a remote-sensing based metric approach to evaluate the hydrological response of highly urbanized areas and apply it to the city of Brussels. The model is set-up using 2 m resolution hyperspectral data. Next, it is upscaled to the city level, using multi-spectral Sentinel-2 data with 20 m resolution. We identify the total impervious area, the vegetation cover and the leaf area index as important metrics to derive a timeseries of spatially distributed net rainfall, runoff and infiltration from rainfall data. For the estimation of the actual evapotranspiration we use the potential evapotranspiration and the available water storage based on the interception, the depression storage and the infiltration. Additionally, we route the runoff to the outlet of selected sub-catchments. An important metric for the routing is the timing to the outlet which is approximated using the total impervious area and the hydrological distance to the outlet. We compare our approach to WetSpa model simulations and reach R 2 values of 98% for net rainfall, 95% for surface runoff, 99% for infiltration and 97% for cumulative evapotranspiration. The routing in the Watermaelbeek catchment is evaluated with discharge observations and reaches NSE values of 0.89 at a 2 m resolution and 0.88 at a 20 m resolution using an hourly timestep. At the timestep of 10 min and a 20 m resolution the NSE is reduced to 0.76. For the Roodebeek catchment we reach an NSE of 0.73 at a spatial resolution of 20 m and an hourly timestep. The results presented in this paper are optimistic for using spatial and temporal metrics retrieved from remote sensing data to quantify the water balance of urban catchments.

Highlights

Due to the highly impermeable land surface in cities, the hydrological response in an urban environment is altered: (1) reducing initial rainfall losses [1], (2) generating more surface runoff [2] and (3) altering infiltration and low flow regimes [3]
We identify the total impervious area, the vegetation cover and the leaf area index as important metrics to derive a timeseries of spatially distributed net rainfall, runoff and infiltration from rainfall data
The results presented in this paper are optimistic for using spatial and temporal metrics retrieved from remote sensing data to quantify the water balance of urban catchments

Summary

Introduction

Due to the highly impermeable land surface in cities, the hydrological response in an urban environment is altered: (1) reducing initial rainfall losses [1], (2) generating more surface runoff [2] and (3) altering infiltration and low flow regimes [3]. To manage the increase in frequency and volume of surface runoff, hydraulically efficient (storm)water drainage networks were created. These engineered systems alter the flow regime by leading high magnitudes of polluted stormwater to receiving water bodies causing high flow events and limiting the infiltration and recharge to groundwater for low-flow events [3]. To improve urban water management practices, new developments focus on local measures to reduce or disconnect the impermeable surfaces and increase the urban vegetation network [4]. Many studies exist that evaluate the local effects of new water management

Methods

Results

Conclusion