Publications

He, XY; Wang, DD; Gao, S; Li, X; Chang, GJ; Jia, XD; Chen, Q (2024). The anisotropy of MODIS LST in urban areas: A perspective from different time scales using model simulations. ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING, 209, 448-460.

Abstract
Remote sensing is one of the effective means to obtain urban land surface temperature (LST), but the observed temperature varies with sensor viewing angle due to urban thermal anisotropy (UTA) and biased sensor viewing angle. The anisotropy of satellite -based LST products (e.g., MODIS LST) varies at different time scales. Previous researches focus on the anisotropy of MODIS LST at the seasonal scale or at other time scales (e.g., hourly) but only for individual cities. The characteristics of anisotropy at different time scales and for cities at different latitudes are still unknown. In this study, we focus on analyzing the characteristics of anisotropy at hourly, daily, seasonal, and annual time scales separately for 80 global cities using the simulation data. The simulation data were generated by coupling a microscale urban surface temperature model and a thermal remote sensing model. In addition, the impact factors (i.e., sensor factor, meteorological factor, and surface factor) at different time scales are discussed and validated using satellite observations. The results show that the daytime instantaneous anisotropy varies from -4.2 K to 1.1 K at the hourly scale. For each city, the smaller anisotropy is captured at the local time 12:00 more easily for both Terra and Aqua overpass during daytime. This is because the magnitude of urban thermal anisotropy is more influenced by the sensor factor (i.e., sensor zenith angle), which varies with the overpass time and the relative position of the city in the track. However, if the seasonally or yearly aggregated MODIS LST is utilized to study mean surface temperature, the intensity of angular effect is much smaller with the value ranging from 0.1 K to 3.3 K at the seasonal scale (from 0.2 K to 2.1 K at the yearly scale) because the variation of sensor zenith angle is smoothed. At seasonal and yearly scales, the strength of the angular effect is greater in the morning during daytime, in summer among all seasons, and in the cities with lower latitudes, which is due to larger solar radiation. Among all meteorological factors, wind speed also has a larger impact on the magnitude of surface thermal anisotropy besides the downward surface shortwave radiation. Our study provides a reference for the correction of the angular effect in the MODIS LST products.

DOI:
10.1016/j.isprsjprs.2024.02.012

ISSN:
1872-8235