Unexpected downpours in cities are a creation of their urban terrains. A project by an Anglo-French team of experts has shown that it is now possible to simulate the extent to which large areas of buildings and roads shape the weather above and around them.
Essentially the physical make-up and physical processes going on in large urban areas, ranging from 10 to 100 kilometres across, combine together to produce what are known as heat islands. This increases the convection and turbulence in the atmosphere locally and consequently affects
the large-scale meteorological processes around them. Jutta Thielen and Alan Gadian, meteorologists at the University of Manchester Institute of Science and Technology (UMIST) applied a complex fine-scale model to look at rainfall over Paris, although they believe their results will be equally valid for other cities around the world. The work implies that weather forecasts for city dwellers will often be less predictable - and hence less accurate - than for the rest of the population because this significant effect is not considered in contemporary official calculations. ''The City of London should have considerable impact on the rainfall pattern and the UK Meteorological Office does not really take this into account yet,'' said Dr Thielen. The UMIST team looked at weather on a meso-scale of between one kilometre and 100 km, much finer than the broad-brush approach modern forecasters generally use. The UK Meteorological Office is developing a meso-model ultimately for the whole country to improve local forecasting. Although the rainfall that comes via the vast frontal systems that sweep over regions are essentially unaffected by the presence of a city, the study found a very different result for convective rain.
During the day, the built-up terrain - asphalt, brick, concrete and tile - absorbs heat, making the overall city warmer than the surrounding countryside, sometimes by as much as five degrees Celsius. The model showed that this caused columns of warm air to rise, destabilising the atmosphere above and pushing what moisture there was high enough for clouds to condense. If there was enough water vapour this could lead to rainfall, often in the form of summer evening thunderstorms. The "roughness" of the urban terrain also obstructed the flow of air across the city, creating more turbulence and boosting the chances of rain clouds forming. The larger the city and the more high-rise its buildings, the more significant the effect.
It appears that there is little escape from this urban rain, even for those who choose to commute from outside; the model showed that much fell downwind of the city, soaking nearby towns and villages.
Very much in the same line of research is hydro-meteorological modelling. This is an exciting development which is now being researched in many countries in Europe. At present the Joint Research Centre of the EU is also expanding on this line of research. So far, rainfall and flood predictions are based on the extrapolation of observed rainfall fields with radar and therefore their prediction time lag is limited.
Advances in flood forecasting, or increase of flood warning time, can be achieved using the novel approach of combined hydro-
meteorological modelling. This work is being jointly proposed by UMIST in conjunction with Professor Keith Beven of nearby Lancaster University. Collaborations are also being forged with other EU countries especially France.
Atmospheric convective rainfall processes will be modelled with a high-resolution meso-scale model and hydrological processes will be described using a rainfall/run-off model. Interactive feedback processes can uniquely be assessed using this approach. Discharge patterns will be contrasted with existing data sets from the Pennines and the Cevennes region in France.
Freitag, 20. April 2001