Can we use remotely sensed land surface temperatures to evaluate and improve model simulations of the urban heat island?
Abstract
Extreme heat events are the leading cause of weather-related human mortality in the United States and in many countries world-wide, and the development of highly accurate urban climate models to predict heat waves and extreme heat events is critical. However, the heterogeneous urban surface with myriad energy and moisture fluxes increases model complexity and uncertainty. Remotely sensed land surface temperature (LST) offers advantages such as comparable spatial scale, global coverage, steady periodicity, and long-term observations, which can be applied to assess model simulations. This research proposes a sampling technique to select and compare MODIS LST and model-simulated radiative temperature for eight configurations of the High Resolution Land Data Assimilation System (HRLDAS) during 2003-2012 summers (JJA) for Houston, TX. The objective is to decrease comparison biases between MODIS and HRLDAS caused by clouds, view angles, and the LST retrieval algorithm, and to understand which urban surface properties are critical for accurate UHI simulations. The results show that the accurate description of urban fraction can effectively decrease more than 25% of RMSE for HRLDAS LST for both daytime and nighttime comparisons. Assuming irrigated vegetation in the urban area largely improved the RMSE by about 2K during the daytime, while there was no significant difference for the nighttime periods. In the most realistic scenario HRLDAS performed quite well at night, both temporally and spatially. HRLDAS daytime LST simulations are warmer than MODIS observations by approximately 5K but with relatively strong correlations. In summary, remotely sensed LST can be a good observational source for the assessment of UHI simulations, but requires careful pre-processing beforehand to avoid unrepresentative comparisons. The proposed sampling method is practical and effective for validation of long-term urban-scale model simulations.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2013
- Bibcode:
- 2013AGUFM.A51E0062H
- Keywords:
-
- 3333 ATMOSPHERIC PROCESSES Model calibration;
- 3360 ATMOSPHERIC PROCESSES Remote sensing;
- 3355 ATMOSPHERIC PROCESSES Regional modeling;
- 0350 ATMOSPHERIC COMPOSITION AND STRUCTURE Pressure;
- density;
- and temperature