The Helsinki experiment

Case study objectives

The overarching objective of this research is to improve nowcasting of extreme precipitation and the associated urban flooding in the capital area of Finland. In addition, the impact of urban densification on the susceptibility to flooding will be explored, as well as the possibilities to alleviate the adverse effects of urban development with low impact development (LID) tools.    

Study site

Länsi-Pakila is a residential area in northern Helsinki, characterized mostly with single-family houses (Fig 1.). The area is known to be prone to stormwater flooding, which has caused e.g. water entering basements and yards of individual houses in the past. The area is subject to urban development and densification in near future, which puts it at a risk of more severe problems caused by urban flooding unless due attention is given to sensible stormwater management.

Figure 1. Länsi-Pakila catchment in Helsinki, Finland. Map data: © National Land Survey of Finland, Background map series 03/2017, CC BY 4.0.

Data

The Helsinki metropolitan region is covered by three polarimetric C-band Doppler radars, namely the Finnish Meteorological Institute (FMI) operated Vantaa radar (VAN), the University of Helsinki operated Kumpula radar (KUM), and the Vaisala Oyj operated Kerava radar (KER) (Fig. 2). The University of Helsinki produces a quantitative precipitation estimate (QPE) for the metropolitan region as a composite product utilizing data from all three radars (Fig. 3). Use of a multi-radar setup mitigates the common problems in urban radar measurements by providing extra observations for filling gaps in individual radar measurements and by extending the range of low elevation observations. A high quality QPE with a Cartesian grid resolution of 250x250 m2 is achieved by using several quality control methods on calibrated polarimetric radar data and by quality based compositing. Furthermore, the accuracy of rainfall estimate in a given location is increased by using advection interpolation on the composited QPE field to achieve a temporal resolution of 1 min.

Figure 2. Länsi-Pakila catchment and the Helsinki metropolitan region radars in Kumpula (KUM), Vantaa (VAN), and Kerava (KER). Map data: © National Land Survey of Finland, Background map series 03/2017, CC BY 4.0.

Fig. 3. Helsinki metropolitan region composite QPE from the KUM, VAN, and KER radars produced by the University of Helsinki Department of Physics. Initial composite product based only on radar reflectivities. © University of Helsinki Department of Physics, 2017.

Two to three co-located high-resolution fully automatic tipping-bucket rain gauges (Decagon ECRN-100 High Resolution Rain Gauge) (Fig. 4) will be installed at the study site to provide on-site rainfall measurements for the snow-free periods of 2017 and 2018. In addition, the FMI open data portal provides continuous weather radar and rain gauge observations from the Helsinki metropolitan region, with the nearest FMI radar (VAN) 5 km from the study site and 4 rain gauges within a distance of 10 km from the site.

Fig. 4. A tipping bucket rain gauge.

On-site runoff observations will be collected in a measurement campaign during summer 2017 to provide calibration and validation data for runoff simulations. A high-resolution Starflow Ultrasonic Doppler Instrument will be installed at the catchment outlet point to provide water level and flow velocity measurements at 1 min interval.

Methods

Precipitation Nowcasting

An open source precipitation nowcasting model is to be developed at Aalto University to provide short-term radar-based ensemble predictions of precipitation. The model will consist of a component for advection extrapolating the latest radar observation into the future, and a component for evolution of the fields to describe the growth and decay processes of precipitation features. The UH composite QPE and the FMI operational radar data will serve as high-resolution input data for the model, whereas the on-site rainfall measurements will be used to test the model performance in hindcast experiments.

Runoff Nowcasting

A high-resolution Storm Water Management Model (SWMM) description for the study site will be built utilizing a novel GisToSWMM5 tool, the 2x2 m digital elevation model (DEM) available from the National Land Survey of Finland, the land cover description and pipe network data from the Helsinki Region Environmental Services Authority HSY, and other available data. The model will be run using the precipitation nowcasts to provide runoff nowcasts at the study site, and the model performance will be tested utilizing the on-site rainfall and runoff measurements in hindcast experiments.

Expected outcome

The Helsinki case study of the MUFFIN project is expected to produce

  1. An open source two-dimensional (2D) precipitation model for both nowcasting of precipitation and for producing 2D design rain events
  2. Assessment about the accuracy of precipitation and runoff nowcasts in the Helsinki metropolitan region utilizing the composite radar rainfall product
  3. Assessment about the possibilities of green infrastructure and other stormwater management solutions for the densifying, flood-prone Länsi-Pakila region