Weather and wind forecasting typically relies upon low-resolution computer simulations, with no more detail than a 12-kilometer grid. Also, these model simulations virtually never incorporate urban features such as pavement and skylines. As a result, they cannot accurately predict how temperature varies over city neighborhoods during a heat wave, nor can they always tell what direction wind will be blowing an atmospheric contaminant. In the end, weather forecasters do an amazing job with mediocre weather models by using their long-term knowledge of what works for various situations and locations.
Sometimes this isn’t good enough, so a goal of a research collaboration between our research group and collaborators at the Naval Research Laboratory and City College of New York is to push the boundaries of dealing with these shortcomings in atmospheric prediction. Our research can help improve prediction capabilities for weather and atmospheric transport, as well as the scientific understanding of urban weather features such as the urban heat island, which often keeps temperatures 10-15 degrees (F) warmer than rural areas at night, as shown above. The influence of the ocean and sea breezes on weather is also captured much better by COAMPS than by most weather models. Previous results and publications from this research are summarized on Julie Pullen’s webpage.
We are working with the Navy’s Coupled Ocean-Atmosphere Mesoscale Prediction System (COAMPS) computer model, and using grids as high as 333 meters resolution. For checking the model results with actual weather observations, we are using the usual airport and city weather stations, maintained by NOAA, but also hundreds of civilian-run weather stations — all these are conveniently merged in City College’s NYC Met Net.
[Temperature figure data credits: NOAA (NOS-PORTS, NWS-ASOS, NWS-HADS, Urbanet), Rutgers NJ Weather and Climate Network, APRSWXNET, AWS Convergence Technologies, Inc. (WeatherBug), and Weatherflow, via Mark Arend (NYCMetNet).]