Satellite observation of land surface temperature (LST) is an important tool for monitoring urban thermal environments but is prone to significant angular effects over non-isothermal pixels characterized by the substantial three-dimensional (3D) structure of urban areas. However, accurately characterizing the urban thermal anisotropy at the satellite pixel scale is still challenging. In this paper, comparing the simultaneous airborne observations and the temporal-aggregated signal from MODIS LST, we further investigate the seasonal and diurnal patterns of thermal anisotropy in urban areas with various simulation scenarios. The anisotropy was defined as the difference between off-nadir and nadir temperatures. First, a realistic morphological representation of urban surfaces as well as urban component temperatures measured by infrared thermometers (IRTs) were input into the discrete anisotropic radiative transfer (DART) model to simulate MODIS-scale anisotropy for Toulouse, France in summer. Extending this ‘IRT-DART’ method to other seasons was restricted because the IRTs' radiative source areas are affected by shadows when the sun elevation is low, meaning that measured sunlit temperatures are not reliably available. An energy balance model, TUF3D, and a sensor view model, SUM, for simplified urban geometry were coupled to replace ‘IRT-DART’ in winter. The cross-comparison shows that ‘TUF-SUM’ can simulate MODIS-scale anisotropy if urban structure and materials are relatively homogeneous, which provides a basis for assessing anisotropy when IRT sunlit temperatures are unavailable. The MODIS anisotropy is subject to a strong diurnal yet a discernable seasonal variation. For example, for a mid-latitude city, the anisotropy over all MODIS view angles reaches up to 6.6 K for Terra-Day (around 11:00 local time) and 4.9 K for Aqua-Day (around 13:00) in summer, and 6.1 K for Terra-Day and 4.0 K for Aqua-Day in winter. Nocturnal anisotropy is weak with values less than 0.2 K. The modeling methods can be used to quantify anisotropy of satellite LST in a broad range of urban environments and facilitate the use of multi-angular MODIS LST for a better assessment of urban thermal environments in the future.

卫星观测地表温度（LST）是监视城市热环境的重要工具，但在非等温像素（倾向于以市区的三维（3D）结构为特征）上容易产生明显的角度影响。然而，在卫星像素尺度上准确表征城市热各向异性仍然具有挑战性。在本文中，比较了同时进行的机载观测和来自MODIS LST的时间聚集信号，我们进一步研究了在各种模拟情况下城市地区的热各向异性的季节性和昼夜模式。各向异性定义为离最低点和最低点温度之间的差。第一的，将离散的各向异性辐射传输（DART）模型输入城市表面的真实形态学表示法和通过红外温度计（IRT）测量的城市组件温度，以模拟法国图卢兹在夏季的M​​ODIS尺度各向异性。将这种“ IRT-DART”方法扩展到其他季节受到限制，因为当太阳高度较低时，IRT的辐射源区域会受到阴影的影响，这意味着无法可靠地获得所测得的日照温度。结合了用于简化城市几何结构的能量平衡模型TUF3D和传感器视图模型SUM，以在冬天替换“ IRT-DART”。交叉比较表明，如果城市结构和材料相对均质，“ TUF-SUM”可以模拟MODIS尺度的各向异性，当没有IRT阳光照射的温度时，这为评估各向异性提供了基础。MODIS的各向异性具有很强的昼夜性，但可以识别出季节性变化。例如，对于中纬度城市，夏季，Terra-Day（当地时间11:00左右）在所有MODIS视角上的各向异性均高达6.6 K，Aqua-Day则为4.9K（大约13:00左右），冬季则为6.1 K（Terra-Day）和Aqua K（日）为4.0K。夜间各向异性很弱，其值小于0.2K。该建模方法可用于量化广泛城市环境中卫星LST的各向异性，并有助于使用多角度MODIS LST更好地评估城市中的热环境。未来。对于中纬度城市，Terra-Day（当地时间11:00左右）的所有MODIS视角的各向异性分别达到6.6 K和夏季Aqua-Day的各向异性达到4.9 K（夏天约为13,000）冬季则为Terra-Day，而Aqua则为4.0K。夜间各向异性很弱，其值小于0.2K。该建模方法可用于量化广泛城市环境中卫星LST的各向异性，并有助于使用多角度MODIS LST更好地评估城市中的热环境。未来。对于中纬度城市，Terra-Day（当地时间11:00左右）的所有MODIS视角的各向异性分别达到6.6 K和夏季Aqua-Day的各向异性达到4.9 K（夏天约为13,000）冬季则为Terra-Day，而Aqua则为4.0K。夜间各向异性很弱，其值小于0.2K。该建模方法可用于量化广泛城市环境中卫星LST的各向异性，并有助于使用多角度MODIS LST更好地评估城市中的热环境。未来。