Utility of Thermal Remote Sensing for Evaluation of a High-Resolution Weather Model in a City

Abstract: The use of satellite data is explored to evaluate high-resolution numerical weather prediction (NWP) models, the latter of which will play a key role in next-generation operational weather forecasts. Urban areas are expected to be one focus for such applications, but this will require new modelling approaches and extensive evaluation.

In this study, we retrieve Landsat land surface temperature (LST) using a new technique and use this to assess 100 m-resolution NWP predictions for London. We demonstrate that the retrieved Landsat LST data are spatially highly correlated with two other LST retrieval methods. We also address the limitations imposed by the restricted viewing angle of the satellite on its ability to view the “complete” surface temperature, and discuss potential ways to improve LST retrievals in urban areas.

The Landsat LST data helps us to enhance the NWP modelling and identify where future model improvements can be made. The extensive LST spatial coverage allows major features to be explored that would not be evident if using analyses only for small areas: notably, spatial patterns visible in the 100 m NWP modelling domain that are not apparent in the Landsat imagery. The resulting investigation identified downscaling soil moisture using soil properties to be the cause of the artefacts. New 100 m model runs have more realistic spatial correlations but a larger mean difference. Correlation of the differences in LST with surface cover suggest that model performance is better for vegetated areas.

Biography: I completed my PhD in meteorology at the University of Reading investigating better ways of representing cloud structure in weather and climate models. After that, I worked at the University of Reading for 12 years in various research areas, including cloud remote sensing, tropical dynamics in climate models, and climate effects on the Indian summer monsoon. During this time, I wrote a general interest book “Introducing Meteorology: a Guide to Weather”. In 2020 I began working for the Met Office, where I am developing an improved representation of the urban surface and its interaction with the boundary layer.