Extended Summary

Facing the phenomena of climate change, city planners and architects have to develop adaptation strategies to mitigate the impacts of extreme weather conditions on citizens to ensure human well-being outdoors and indoors. To assess the effectiveness of proposed adaptation measures, quantitative information is needed. Models designed to simulate the urban climate can serve as valuable tools to provide this information.

During the last decade, substantial progress in both mesoscale numerical weather prediction (NWP) and the description of urban atmospheric processes has been achieved. With increasing computer capacity, NWP models are now approaching the required horizontal and vertical resolution to provide high quality urban meteorological data.

The partners in the present project, MAQ and ESS-CC, have a large expertise with respect to the mesoscale models WRF1 and RAMS2, respectively. Recently, these models have been equipped with a so-called Urban Canopy Model (UCM) embedded within a land surface scheme, WRF has been coupled to the NOAH land surface model-single-layer Urban Canopy model (NSLUCM) and RAMS to the Town Energy Balance model (TEB) (Kusaka et al. 2001; Chen et al., 2004; Dandou et al. 2005; Masson, 2000; Masson et al., 2002; Lemonsu et al. 2004). In the present study currently available models for the Urban Canopy have been reviewed and the performance of the aforementioned specific models was evaluated.

The theoretical basis of all present-day models is the Surface Energy Balance (SEB) for the urban area:

Q* + QF = QH + QE + ∆QS

Where Q* is net all wave radiation, QF heat arising from anthropogenic activities, QH is the turbulent sensible heat flux, QE is the turbulent latent heat flux, and ∆QS is the net heat storage flux of buildings. A large number of urban surface schemes exists. These vary in complexity from simple schemes that represent the city as a concrete slab to multi-layer models that model energy exchanges at multiple levels within the urban canopy, thereby allowing for varying building heights. All models forecast the SEB fluxes representative of the local (or neighbourhood) scale.

Many of the models are also capable of calculating additional terms, typically air and surface temperatures and wind speed at street level, and providing more detailed flux information, for example by facet.

We participated with WRF in the Urban Surface Energy Balance: Land Surface Scheme Comparison project (PILPS4 URBAN) which started in March (2008) and was coordinated by the Department of Geography, King’s College London. By participating, we were able to assess the offline performance of the model as compared to other available models. Since the analysis of latter phases in not complete yet, this report covers only the first phase of the project. It appears that the models have best overall capability to model net radiation (Q*) and least capability to model latent heat flux (QE). None of the models performs best or worst for all fluxes. Generally speaking, the simple models in each of the classifications perform as well, if not better, than the more complex models for the daytime fluxes. For Q*, QH and ∆QS, NSLUCM-WRF performs better than the other models. However, the model has a relatively low capability to model QE during day time, but this holds for most of the participating models.

Online simulations with NSLUCM-WRF and TEB-RAMS have been carried out for the city of Rotterdam. Because characteristics of the built environment are not known yet, the default settings of each model were applied for these simulations. The main aim was to get a first impression of the model performance and capabilities under Dutch climatic conditions. Although both models differ in their results for the diurnal variation and maximum value for UHI intensity, they show a clear overall Urban Heat Island effect ranging from 1 to 5 K. Both modelling studies show a significant influence of urban areas on their neighbouring rural areas (the effect of the urban plume). The capability to quantify with mesoscale models the effect of changes in city configuration on the UHI has been demonstrated.

At present, no systematic meteorological data records for towns and cities in the Netherlands are available. Therefore, an inventory has been made of internationally available data sources. In addition, the possibility to utilize datasets provided by hobby meteorologists has been explored. We collected meteorological observations for 20 urban sites in the Netherlands including air temperature and humidity, wind speed, and for some stations also incoming solar radiation. With these data a first assessment of the UHI intensity for the Netherlands could be made. To assess the effect on thermal comfort, wet-bulb-globe-temperature (WBGT) was estimated. The preliminary analysis shows that the UHI is clearly present in the Netherlands as well. The UHI ranges from 3 to 10°C. Seven out of the 20 cities investigated experience extreme heat stress for ~7 days a year under the present climatic conditions.

Detailed observations are required for process understanding, and to support model development and evaluation, to ensure realistic simulations. To obtain meaningful observations, it is essential to clearly establish the objectives of the observations (measurement rationale). In connection with this, it is necessary to pay careful attention to urban scale and related issues. Recognition of scale differences in cities is a central key to the design of meaningful field observations. A distinction can be made between: (1) Observations in the Urban Canopy Layer (UCL) and (2) observations in the Urban Boundary Layer (UBL). The observations in the UCL provide insight into the dependence of the microclimate on street and city quarter characteristics. With the observations in the UBL, valuable information on the urban surface energy balance (SEB) can be obtained. It can be concluded that systematic observations of the urban meteorological and climatological conditions should be based on a set of complementary observations, covering detection of temporal as well as spatial variations of the meteorological conditions in the urban canopy. For model parameterization and validation more specialized observations are required as well. Satellite imagery and other remote sensing techniques can give valuable additional information that can also be used to determine optimal locations for observations within the urban canopy.

As an illustration experiences obtained with measurements that have been conducted in the framework of the ‘Heat stress in the city of Rotterdam’ project are discussed.

Read more: synthesis report
Contact: projectCOM29