EGU2020-10916
https://doi.org/10.5194/egusphere-egu2020-10916
EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Urban Drainage Systems modelling for Early Warning Service Using Data-Driven Modelling

Solomon Seyoum1,3, Boud Verbeiren1,2, and Patrick Willems1
Solomon Seyoum et al.
  • 1Vrije Universiteit Brussel, Department of Hydrology and Hydraulic Engineering, Brussels, Belgium
  • 2Brussels Company for Water Management (SBGE/BMWB), Direction Exploitation, Brussels, Belgium
  • 3IHE Delft Institute for Water Education, Delft, Netherlands

Urban catchments are characterized by a high degree of imperviousness, as well as a highly modified landscape and interconnectedness. The hydrological response of such catchments is usually complex and fast and sensitive to precipitation variability at small scales. To properly model and understand urban hydrological responses, high-resolution precipitation measurements to capture spatiotemporal variability is crucial input.

In urban areas floods are among the most recurrent and costly disasters, as these areas are often densely populated and contain vital infrastructure. Runoff from impervious surfaces as a result of extreme rainfall leads to pluvial flooding if the system’s drainage capacity is exceeded. Due to the fast onset and localised nature of pluvial flooding, high-resolution models are needed to produce fast simulations of flood forecasts for early warning system development. Data-driven models for predictive modelling have been gaining popularity, due to the fact they require minimal inputs and have shorter processing time compared to other types of models.

Data-driven models to forecast peak flows in drainage channels of Brussels, Belgium are being developed at sub-catchment scale, as a proxy for pluvial flooding within the FloodCitiSense project. FloodCitiSense aims to develop an urban pluvial flood early warning service. The effectiveness of these models relies on the input data resolution among others. High-temporal resolution rainfall and runoff data from 13 rainfall and 13 flow gauging stations in Brussels for several years is collected (Open data from Flowbru.be) and the data-driven models for forecasting peak flows in drainage channels are build using the Random Forest classification model.

Optimal model inputs are determined to increase model performance, including rainfall and runoff information from the current time step, as well as additional information derived from previous time steps.

The additional inputs are determined by progressively including rainfall data from neighboring stations and runoff from previous time steps equivalent to the lag time equal to the forecasting horizon, in our case two hours. The data-driven model we develop has the form as shown in the following equation.

Qt = f(Qt-lag, ∑RFi,jfor i is the number of rainfall stations considered and j is the time  from t-lag to t

Where Qt  is the flow at a flow station at time t, Qt-lag is the lagged flow at the station and RFi,j is the rainfall values for station i and time j.

For Brussels nine relevant sub-catchments were identified based on historical flood frequency for which we are building data-driven flood forecasting models. For each sub-catchment, RF models are being trained and tested. More than 200,000 data point were available for training and testing the models. For most of the flow stations the data-driven models perform well with R-squared values up to 0.84 for training and 0.6 for testing for a 2-hour forecast horizon. 

To improve the reliability of the data-driven models, as next step, we are including radar rainfall data input, which has the ability to capture temporal and spatial variability of rainfall from localized convective storms to large scale moving storms.

KEYWORDS

Data driven models, FloodCitiSense, Flood Early Warning System, Urban pluvial flooding

How to cite: Seyoum, S., Verbeiren, B., and Willems, P.: Urban Drainage Systems modelling for Early Warning Service Using Data-Driven Modelling, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-10916, https://doi.org/10.5194/egusphere-egu2020-10916, 2020

Display materials

Display file

Comments on the display material

AC: Author Comment | CC: Community Comment | Report abuse

Display material version 1 – uploaded on 04 May 2020
  • CC1: Comment on EGU2020-10916, Jiada Li, 05 May 2020

    Hi: I'm jiada from the University of Utah. This is a nice project. I'm a beginner of Random Forest. Here are my questions regarding your work:

    1. Why did you use a random forest classifier instead of a random forest regression? Since you going to predict the time-series drainage flow data. Also, the DDM performance metrics like R-square and RMSE only work for regression purposes.

    2. Is the '2-hour' predicting horizon determined by the catchment lag time? What are the logics?

    3. How did your results reflect that higher frequency-inputs can produce better DDM performance?

    4. Would you mind sharing your future plan about using DDM to predict the spatial flooding extent?

  • AC1: Comment on EGU2020-10916, Solomon Seyoum, 05 May 2020

    Hi Jiada, thank you very much for your questions. I will try to answer your questions one by one as follows.

    1. We used random forest regressor not classifier. If it was mentioned as classifier in the ppt or in the abstract, we will correct it.
    2. The 2 hours prediction horizon is not determined by lag time. It was decided that 2 hours would be sufficient to issue flood warning. We are currently determining the lag time for each catchment (by correlation analysis of flow versus lagged rainfall) and we will see if that improves the result of the models.

    3. We have not yet tested the models with input frequency more than 5 minutes. We have some data with 1 minute’s resolution. We could compare the model performance to see if high temporal resolution of the input data could increase the model performance.
    4. We are using the DDM as a proxy to predict flooding per sub catchment level. We would not be able to predict spatial extent within the subcatchment by using the DDM alone
    • CC2: Reply to AC1, Jiada Li, 05 May 2020

      Hi Solomon:

      Thank you for your explanations. This is an awesome project. Looking forward to seeing more progress from your team.

       

      Jiada