(B):
angle
Medium density 0.0171 K per %
High density 0.0422 K per %
CBD 0.0332 K per %
Industrial 0.0542 K per %
Track system 0.0722 K per %
Riverside -0.0365 K per %
Road 0.0431 K per %
Water -0.0670 K per %
Forest -0.0367 K per %
Park -0.0379 K per %
Regression constant 293.7553 K
Explained variance 88.4 %
Table 3: Regression results (urban area)
The Rhine river in both images appears as the struc-
turing element with the lowest radiation tempera-
tures, which separates the upper right quadrant from
the other parts. The bright (high radiation temper-
atures) areas from the center to the right and in the
middle of the upper right quadrant are affected by
track systems. An industrial area causing high radi-
ation temperatures is located in the bottom right cor-
ner.
SIMULATION OF THE INFLUENCE OF
LANDUSE CHANGES
An area of approximately 0.26km? in the upper right
quadrant (rectangle) was selected for a simulation of
change of landuse and its impact on the radiation
temperatures. The area is currently used as a track
system. Two possible landuses were assumed: park
and CBD (Central Business District). The different
landuses were implemented into the classification and
then using the regression results from table 3, radi-
ation temperatures were recalculated. The results are
displayed as images in figure 3 (C) and (D). In the left
image (C), the change in radiation temperature for a
possible use as a park to the existing landuse is an
average —6.0K. Further information is listed in table
4. With a possible use as a high density CBD-area
the average radiation temperature differs by —1.5K
to the present use. Comparison of the two possible
future landuses in the selected area results in an av-
erage difference of —4.5K in the park area.
CONCLUSION
Thought as an analysing method for linear depen-
dencies, the regression model and its application mu-
tated as a possibility to reveal latent information hid-
den in the measured satellite data on one hand and
of simulating radiation temperatures under different
427
track system | track system park
- park - CBD | - CBD
[K] [K] [K]
mean +6.0 +1.5 -4.5
c 3.5 1.2 9.3
min +0.4 +0.0 -0.4
max +10.7 +3.6 -T.1
Table 4: Different landuses and their impact on radi-
ation temperatures (differences)
landuse conditions on the other hand. Using all the
bands of LANDSAT-TM, band 6's resolution can be
enhanced to 30m*30m. Using a landuse classification,
it is possible to recalculate/simulate band 6 very pre-
cisely (figure 3 (A) and (B)) in a 120m»120m resolu-
tion. The application of this method is to get into cli-
mate modelling on a regional or local scale. Commu-
nal planning authorities normally have detailed plans
which imply a modification of the existing landuse,
e.g. definition of new industrial areas etc.. This mod-
ified future landuse can be integrated in the existing
landuse dataset, and then a further simulation of radi-
ation temperatures with the modified landuse can be
carried out (figure 3 (C) and (D)). Alternative plan-
ing variants can be simulated and then compared to
each other. So the climatological effect of planned
landuse modification can be quantified very detailed
in a local scale using a simple linear regression model.
REFERENCES
Oke, T.R., 1987. Boundary Layer Climates. 2nd ed.,
Cambridge.
Parlow, E., Scherer D., 1991. Auswirkungen von Veg-
etationsveranderungen auf den Strahlungshaushalt.
Regio Basiliensis. Heft 32/1. Basel
Scherer D., Parlow E., 1990. Calculation of synthetic
thermal images with enhanced spatial resolution from
NOAA-AVHRR satellite images. In: Proceedings
of EARSeL-Symposium at Helsinki 1989. Briissel-
Luxemburg 1990. pp. 448-453.
Schott, J.R., Volchok, W.J., 1985. Thematic mapper
thermal infrared Calibration. Photogrammetric En-
geneering and Remote Sensing 51, 9. pp. 1352-1357.
Wüthrich, M., 1991. Modelling the thermal struc-
tures of Basel (Switzerland): a combined approach
using satellite data and GIS-techniques. In: Proceed-
ings of the 11th EARSeL symposium, Graz-Austria,
pp. 298-304.
