Search

An extension to the two-step physical retrieval algorithm was developed. Combined clear-sky multitemporal and multispectral observations were used to retrieve the atmospheric temperature--humidity profile, land-surface temperature, and surface emissivities in the midwave (3-5 [micro]m) and long-wave (8-14.5 [micro]m) regions. The extended algorithm was tested with both simulated and real data from the Moderate-Resolution Imaging Spectroradiometer (MODIS) Airborne Simulator. A sensitivity study and error analysis demonstrate that retrieval performance is improved by the extended algorithm. The extended algorithm is relatively insensitive to the uncertainties simulated for the real observations. The extended algorithm was also applied to real MODIS daytime and nighttime observations and showed that it is capable of retrieving medium-scale atmospheric temperature water vapor and retrieving surface temperature emissivity with retrieval accuracy similar to that achieved by the Geostationary Operational Environmental Satellite (GOES) but at a spatial resolution higher than that of GOES. OCIS codes: 280.0280, 300.1030, 300.2140, 350.5610, 130.3060.

We made an experimental vicarious calibration of the Moderate Resolution Imaging Spectroradiometer (MODIS) Airborne Simulator (MAS) thermal infrared (TIR) channel data acquired in the field campaign near Mono Lake, Calif. On 10 March 1998 to demonstrate the advantage of using high-elevation sites in dry atmospheric conditions for vicarious calibration. With three lake-surface sites and one snow-field site, we estimated the MAS noise-equivalent temperature difference as 0.7-1.0 degrees C for bands 30-32 in the 3.68-4.13 mu m region and 0.1-0.5 degrees C for bands 42, 45, 46, and 48 in the 8-13.5 mu m region. This study shows that the MAS calibration error is within +/- 0.4 degrees C in the split-window channels (at 11 and 12 mu m) and larger in other TIR channels based on the MAS data over Mono Lake and in situ measurement data over the snow-field site.

Land surface temperature algorithms for the moderate resolution imaging spectroradiometer satellite instrument will require the spectral bidirectional reflectance distribution function (BRDF) of natural surfaces in the thermal infrared. We designed the spectral infrared bidirectional reflectance and emissivity instrument to provide such measurements by the use of a Fourier transform infrared spectrometer. A problem we encountered is the unavoidable surface heating caused by the source irradiance. For our system, the effects of the heating can cause a 30% error in the measured BRDF. The error caused by heating is corrected by temporally curve fitting the radiance signal. This curve-fitting technique isolates the radiance caused by reflected irradiance. With this correction, other factors dominate the BRDF error. It is now [approximately]5% and can be improved further. The method is illustrated with measurements of soil BRDF. Key words: Emissivity, reflectance, thermal infrared, spectrometer.

A two-step physical algorithm that simultaneously retrieves geophysical parameters from Moderate Resolution Imaging Spectroradiometer (MODIS) measurements was developed. The retrieved geophysical parameters include atmospheric temperature-humidity profile, surface skin temperature, and two surface emissivities within the shortwave (3-5-[micro]m) and the longwave (8-14.5-[micro]m) regions. The physical retrieval is accomplished in two steps: (i) The Tikhonov regularization method is employed to generate a regularization solution along with an optimum regularization parameter; (ii) the nonlinear Newtonian iteration algorithm is carried out with the regularization solution as a first-guess profile to obtain a final maximum probability solution for geophysical parameters. The algorithm was tested with both simulated and real MODIS Airborne Simulator (MAS) data. Sensitivity studies on simulated MAS data demonstrate that simultaneous retrievals of land and atmospheric parameters improve the accuracy of the retrieved geophysical parameters. Finally, analysis and accuracy of retrievals from real MAS data are discussed. [C] 2000 Optical Society of America OCIS codes: 280.0280, 030.5620.