High-resolution hydrological modelling

Background

Multi-basin flood forecasting models and systems are available in several regions, countries and even continents. Typically, these models are run with a daily time step and a spatial resolution of ~100 km² or (much) more, although the spatial resolution is increasing. Therefore these systems are primarily aimed at simulation and forecasting of fluvial flooding, i.e. flooding generally caused by long periods of rain or snow melt over extended areas and a gradual increase in discharge until river banks become overtopped and flooding is a fact.

Many urban areas are vulnerable to fluvial flooding but in addition the urban environment is associated with an increased risk of pluvial flooding, which is generally caused by smaller-scale, short-duration, high-intensity rainfall events producing runoff that exceeds the drainage capacity of the basin. A proper description of these events in time and space requires substantially higher resolutions than in existing multi-basin systems.

Until recently, data availability has been a limiting factor with respect to developing multi-basin hydrological model systems towards higher resolutions and thereby an increased relevance in an urban context. Both meteorological forcing data and geographical characteristics have been too coarse in time and/or space. Today, however, the situation has changed. High-resolution meteorological forcing has become available, owing to e.g. improved weather radar systems and “convection-permitting” weather forecasting models. Further, many new sources of detailed geographical data on e.g. elevation and land-use have been released.

In light hereof, it is possible to start exploring higher-resolution multi-basin hydrological modelling systems, and their applicability to support urban flood forecasting, which is one key target of MUFFIN.

The HYPE model

HYPE (Hydrological Processes for the Environment) is an open source dynamic integrated rainfall-runoff and nutrient transfer model developed and maintained by the Swedish Meteorological and Hydrological Institute (SMHI). HYPE simulates water flow and substances on their way from precipitation through soil, river and lakes to the river outlet in a catchment. The catchment is divided into sub-basins which in turn are divided into classes (calculation units) depending on land use, soil type and elevation (Figure 1). A general description of the model can be found here (link to: http://www.smhi.se/en/research/research-departments/hydrology/hype-1.7994)  and the full model documentation is available here (link to: http://www.smhi.net/hype/wiki/doku.php).

Figure 1. Schematic of the HYPE model.

The HYPE model has been set up for different parts of the world, see http://hypeweb.smhi.se/, and in MUFFIN we will use the set-ups for Sweden (S-HYPE) and Europe (E-HYPE) as starting point. In S-HYPE, the country is divided into ~40 000 sub-basins with a mean size of ~10 km². This model is used operationally to make 10-day discharge forecasts in all sub-basins (http://vattenwebb.smhi.se/hydronu/). In E-HYPE, the mean sub-basin size is ~250 km² and in both set-ups the time step is 1 day.

Development in MUFFIN

In MUFFIN we will develop the HYPE model for applications at higher resolutions.

  • Temporal resolution. In MUFFIN we will develop the HYPE model for a time step of 1 h. Generally, this time step is strictly not sufficient for pluvial flooding, which may involve a very rapid runoff increase that requires a time step on the order of 1 min to be accurately described. However, it is a substantial improvement compared with today’s daily time step and it is likely to be sufficient for meaningful support in urban flood events. Further, 1 h is a suitable choice with respect to available high-resolution forcing data (observations and forecasts). The development will include evaluation of the “time-step sensitivity” of process descriptions and parameters, adjustments if required, and finally calibration and validation for different urban or semi-urban sub-basins.
  • Spatial resolution. In MUFFIN we will explore different ways to provide a more detailed and realistic representation of the urban environment than what is currently available. This includes e.g. using more detailed land-use data such as EEA Urban Atlas (link to: http://www.eea.europa.eu/data-and-maps/data/urban-atlas). By using the most recently released DEM (Digital Elevation Model) data available it may be possible to decrease the sub-basin size in urban areas. It will be attempted to include descriptions of technical modifications of the natural flow paths in HYPE, e.g. sewer systems and pumping stations. Further, by using land-use specific runoff, rather than just the lumped basin outflow that is normally considered, a more spatially resolved result may be obtained.

Model development will primarily be performed using a HYPE sub-model for Stockholm City (Figure 2), but also other Swedish cities may be used.

Figure 2. Model domain of HYPE sub-model for Stockholm.

After the model development, HYPE sub-models will be developed for the study basins in Aalborg, Rotterdam and Helsinki. In joint experiments, these models will be run in parallel to the local model systems in order to assess possibilities and limitations of the approach.

Expected outcomes

  • A HYPE model with processes and parameters tailored for high-resolution applications
  • New approaches to communicating results with a high resolution in time and space
  • Calibrated and validated high-resolution HYPE set-ups for sub-basins in Aalborg, Rotterdam and Helsinki

End-users

Three Swedish end-users are committed to the project (Advisory Board members):

  • Länsförsäkringar (insurance company)
  • The Swedish Civil Contingencies Agency (MSB)
  • Lund University

Further some 10-15 end-users representing different categories e.g. municipalities, county boards, rescue services, national authorities and consultants are contributing to the development through active participation at workshops, interviews, etc.