+wo facts that the formation of these thermal schades is a 
wind effect: a) they are all oriented, even streamlined, in 
the same direction, although the fields north and south of 
the patches of wood are exactly of the same type, and b) 
they do not occur on other recordings of the same region, i.e. 
with different weather conditions. 
Similar images were described by Mattsson (1967, 1970). How- 
ever, at that stage it was not yet possible to demonstrate 
quantitatively alterations of ST in a thermal shade. By the 
use of a calibrated scanner together with digital image pro- 
cessing, this may be now achieved. Fig. 5b shows that ST in 
thermal shade areas differ > 5°C from ST in the open parts of 
the region. 
More detail is given in Fig. 6a/b which was recorded from an 
altitude of 1000 m and consequently has a better resolution 
(pixel size 2 x 2 m2). It may be noticed that even a row of 
trees along a road causes differences in ST of 69C in the ad- 
jacent fields. Another marked feature is the strict north - 
south orientation of the thermal shade contours. 
So far the argument that the described thermal shades are wind- 
induced, rests on indirect evidence. However, Fig. 3 reveals 
that the registration of wind direction supports this view: 
There was a fairly strong southern wind in the Rhine Valley 
(Bremgarten 5.3 m/sec) which prevailed, more or less unchanged 
in direction, for over 5 hours, after which, at 5.40 am the 
thermal image was recorded. This direction corresponds to a 
continuous and strong (16.0 m/sec) southwesterly wind at the 
Feldberg station. Nevertheless easterly directions prevailed 
throughout the night in Bad Krozingen and Freiburg. 
It is clear, then, that the thermal shades on the northern mar- 
gins of forests are due to southern wind. The physical explana- 
tion would be the following: Radiative cooling produces low ST 
and consequently cold air near the ground. Where this air is 
exposed to a continuous wind, it is mixed with warmer air from 
upper layers. Any obstacles disturbing the windflow, however, 
will reduce wind speed to such an extent that mixing, on their 
leeward side, is also reduced and the cold air can accumulate. 
Everywhere on the open fields ST would, without wind-induced 
mixing, correspond to those in the thermal shade areas. In this 
case, contrary to the first example, wind action generates 
higher ST in the thermal image than there would be without wind. 
Conclusion 
It has been demonstrated that the interpretation of thermal 
images can lead to the detection of wind-induced microclimatic