Hyperspectral thermal IR remote sensing is an effective tool for the detection and identification of gas plumes and solid
materials. Virtually all remotely sensed thermal IR pixels are mixtures of different materials and temperatures. As
sensors improve and hyperspectral thermal IR remote sensing becomes more quantitative, the concept of homogeneous
pixels becomes inadequate. The contributions of the constituents to the pixel spectral ground leaving radiance are
weighted by their spectral emissivities and their temperature, or more correctly, temperature distributions, because real
pixels are rarely thermally homogeneous. Planck's Law defines a relationship between temperature and radiance that is
strongly wavelength dependent, even for blackbodies. Spectral ground leaving radiance (GLR) from mixed pixels is
temperature and wavelength dependent and the relationship between observed radiance spectra from mixed pixels and
library emissivity spectra of mixtures of 'pure' materials is indirect.
A simple model of linear mixing of subpixel radiance as a function of material type, the temperature distribution of each
material and the abundance of the material within a pixel is presented. The model indicates that, qualitatively and given
normal environmental temperature variability, spectral features remain observable in mixtures as long as the material
occupies more than roughly 10% of the pixel. Field measurements of known targets made on the ground and by an
airborne sensor are presented here and serve as a reality check on the model. Target spectral GLR from mixtures as a
function of temperature distribution and abundance within the pixel at day and night are presented and compare well
qualitatively with model output.
Atmospheric eddies cause transient spatial and temporal variations of surface temperature and can limit the precision of satellite surface temperature retrievals. If a thermal IR sensor has sufficiently high spatial resolution, the effects of these transient changes of temperature will be seen as variations of the thermal spatial pattern. Nine thermal IR images of a uniform emissivity area on Mauna Loa caldera are carefully compared to document spatial differences between them. These images were obtained from the Dept. of Energy Multispectral Thermal Imager satellite at about 20m GSD. Spatial patterns with a 1C - 6C magnitude are present but not repeated in any of the images. In order to better understand the characteristics and impact of turbulence induced temperature fluctuations for quantitative remote thermal IR sensing, an effort to model the spatial variation of surface temperature as driven by turbulent energy fluxes has been initiated. Stochastic models initially examined showed a close coupling between surface temperature and turbulent fluxes but were not successful. Traditional energy balance models used in this type of simulation are insufficient to model skin temperature because of the importance of the skin layer and its small depth compared to soil depths used in the models. A new treatment based on surface renewal theory is introduced.
The statistics of ground-based retrievals of cloud liquid water path using the microwave water radiometer (MWR) are typically assumed to be independent of the cloud's absolute position in the column. Furthermore, translational invariance implies statistical parity, i.e. invariance under reflection, of cloud-base height (z<SUB>bot</SUB>) and cloud-top height distributions. This symmetry is necessarily broken, especially under conditions of high boundary-layer relative humidity for which a minimum large-scale lifting condensation level leads to the generation of a significant positive skewness in the distribution function of z<SUB>bot</SUB>. We suggest that the signature of this boundary effect is visible in ARM MWR time-series collected at the TWP site. Motivated by the MWR analysis, we incorporate a minimum lifting condensation level into the analytic model of unresolved low-cloud optical variability developed by Jeffery & Austin (J. Atmos. Sci., to appear). Preliminary results indicate that the effect of cloud-base height skewness on mean oceanic low-cloud reflectivity averaged over GCM spatial scales (order 100 km) is significant.