ausblenden:
Schlagwörter:
-
Zusammenfassung:
The microwave emission of land surfaces is spatially highly variable and dominates satellite
measurements. Thus it is a crucial factor in remote sensing of precipitation over land. In
rainfall algorithms the knowledge of the contribution of the underlying surface helps to
discriminate the precipitation signal.
A new method to determine the emission characteristics of land surfaces has been developed.
For the rst time the microwave emissivity and other land surface parameters
are derived on a pixel level from multispectral measurements of brightness temperatures.
Statistical mean values of the observations are not needed, neither in space nor time. In
addition to the surface emissivity, the surface temperature as well as the wetness of the
surface and the fractional vegetation cover are estimated simultaneously. In contrast to
previous methods, an emissivity model, depending on only few free parameters, is taken
here to simulate the emission of various natural surface types. After being corrected
for atmospheric e ects, the model is compared to observations from the TRMM satellite
within an optimisation routine.
According to a long term validation analysis, the new method re
ects even short time
variations in the surface characteristics. This ability has been proven against independent
data in both time and space. However, snow and desert surfaces are still excluded. A
ground check with meteorological measurements shows the satellite derived surface temperature
values to be in good agreement with the validation data. Both the diurnal cycle
and short time temperature variations are detected. Especially the retrieved wetness effect
due to transient precipitation events proves to be reliable. The global distribution of
vegetation zones and their seasonal variations have been determined realisticly.
The new passive microwave method is the rst to give quantitative values of the di erences
between wet and dry surface emission in the frequencies between 10 und 85 GHz, which
are sensitive to precipitation. Heavy precipitation reduces the surface emissivity by about
10 %. The new method helps to discriminate the outline of regions with light precipitation
more precisely. Up to now, the signal of the wet surface has often been misinterpreted as
being light precipitation. The method presented here, is valid for microwave instruments
on di erent satellite platforms, which allows an extension of the analysis to other sensors
and extratropical regions as well. Due to the modular nature of the method, it can be used
as a benchmark to test di erent emissivity models in simulating the surface characteristics
of natural land surfaces.
