519
USE OF SS M/I PASSIVE MICROWAVE DATA FOR THE STUDY OF
ALPINE REGIONS
Daniel Hiltbrunner w , Christian Mätzler 1
1 University of Bern
Institute of Applied Physics
Sidlerstrasse 5
CH-3012 Bern
Switzerland
Tel. +41 31 631 45 90
Fax. 441 3163137 65
2 University of Bern
Department of Geography
Hallerstrasse 12
CH-3012 Bern
Switzerland
Tel. 441 31 631 85 54
e-Mail: hiltbrunner@giub.unibe.ch
ABSTRACT:
Passive microwave data from the SSM/I sensor are used to investigate the temporal behavior
of the earth surface in the Swiss Alps and the Central Plains. The inversion of the geophysical parameters from
the satellite data and their interpretation are based on long-term in-situ measurements of the spectral
microwave signatures of several different earth surface types. In order to study the temporal behavior of the
snow-covered and snow-free land surface, the surface microwave emissivity has to be derived from the SSM/I
brightness temperatures. For that purpose, the physical temperature of land surface has to be determined. We
present three different approaches for estimating land surface temperatures. A first approach uses brightness
temperatures from the thermal infrared channels onboard the NOAA and Meteosat satellites. The second
method is based on the Kriging interpolation of temperatures measured by meteorological ground-stations. The
third algorithm linearly combines SSM/I-brightness temperatures at 19 GHz as an estimator of the physical
land surface temperature. Comparisons between estimated and measured physical temperatures suggest that the
last method showing rms values between 2.5 and 3.5K, is mostly suited for our purposes. The temporal
behavior of the microwave emission of a large forested area near Mulhouse (France) is discussed using
emissivity parameters derived from the SSM/I data for Winter 1990/91. Emissivity values at 19 and 37 GHz
(vertical polarisation) are temporally very constant (0.94 and 0.93, respectively). A singular decrease in
emissivity at 37 GHz is attributed to shallow snow layers within the observed footprint
KEY WORDS: SSM/I, passive microwaves, land surface temperature, NOAA, Meteosat
1 INTRODUCTION
earth surface in Europe, especially the alpine regions, using passive microwave data from the Special Sensor
Microwave/Imager (SSM/I) onboard the Defense Meteorological Program Satellites (DMSP, Hollinger et al.,
1987). Because of the heterogenous composition of the land surface within a SSM/I footprint a mixed signature
algorithm was developed, which compares the observed SSM/I brightness temperatures with modelled values
using a-priori knowledge of the radiative properties of forest and open water. In this paper, we will focus on the
retrieval of the physical surface temperature, which is closely related to the microwave brightness temperature.
As a test of this retrieval we will show how the estimated physical temperature can be used to monitor the
microwave emission from of a large forested area at 19 and 37 GHz by means of SSM/I data. If the physical
temperature of a certain surface type is known, we can estimate its emissivity from the remotely sensed
brightness temperature at different frequencies. This spectral information allows to infer the physical state of
the observed surface types.
continuous surface temperature field over parts of the Switzerland. The first method uses thermal infared data
from the polar-orbiting NOAA-AVHRR and the geostationary Meteosat satellite recorded at the satellite
receiving station of the Department of Geography (University of Bern, Switzerland; cf. BAUMGARTNER &
FÜHRER, 1990). The second approach is based on ground-measurements of meteorological stations operated
by the Swiss Meteorological Institute. The SSM/I data used in the third method are „Wentz-tapes“ provided by
esrin (Frascati, Italy). The microwave brightness temperatures were spatially interpolated to a 8.6 x 5.4 km grid
using the algorithm proposed by POE (1990) in order to facilitate the spatial matching of the SSM/I data and
the ground-based point-measurements.
The aim of our study is to monitor the temporal behaviour of the snow-covered and snow-free
In the following sections, we will describe three different approaches to retrieve a spatially
