Seasonality and view angle effects in middle infrared radiance of urbanized areas

Spanning 3-5 microns, the middle infrared (MIR) region is the mixing zone between reflected sunlight and emitted earthlight, in roughly equal proportions. While the MIR has been utilized in atmospheric remote sensing, its potential in terrestrial remote sensing—-particularly urban remote sensing—-has yet to be realized. A major advantage of the MIR is the ability to penetrate most anthropogenic haze and smog. Here we investigated MIR radiance dynamics in relation to (1) land cover type, (2) time of year, and (3) sensor view zenith angle (VZA). We used Aqua MODIS daily swaths of calibrated radiance for band 23 (~ 4.05 microns) at 1 km spatial resolution for 2009-2010. Within season and VZA class these data were maximum-value composited in preparation for analysis. We also used the USGS National Land Cover Data (NLCD) percent impervious surface area (%ISA) product from 2001 and 2006. We focused on the relationship between MIR radiance and %ISA in and nearby eight major urbanized areas in the Great Plains of the USA. The region is characterized by four distinct seasons, relatively flat terrain, and isolated urban centers surrounded by a vegetated landscape. For each city we selected transects extending from outlying agricultural areas through the urban core and again out into agricultural areas. We used the NLCD data to characterize the fine spatial structure of %ISA and the MODIS data for the coarser spatio-temporal structure of radiance at 4 microns. Time of year is a strong influence on MIR radiance because green vegetation and snow are MIR-dark in contrast to bare soils and dried vegetation that are MIR-bright and building, paving, and roofing materials that are MIR-grey. Sensor VZA is a major factor influencing observed dynamics of MIR radiance with smaller VZAs (<15 degrees from nadir) capturing more fine spatial detail than very large VZAs (>45 degrees from nadir). In addition to characterization of urban areas, MIR seasonality may provide an alternative parameterization for the seasonality of fractional vegetation.