to an ideal curve. From these curves we see that the lower the maxima
the more they shift towards midnight. This shifting of the maxima is
mainly caused by the temperature differences in the soil; i.e. the
position of the highest values depend less on the season but on meteoro-
logical conditions of the day and of the previous days and on the heat
reservoir of deeper soil layers (fig. 4, curve of 8./9.6. and 10.6.).
The curves of the soil/vegetation layer, plotted in fig. 4, do not
allow a correlation with values of the heat flow within the soil. A
useful curve only results if AS is multiplied by the corresponding
air temperatures T Plotted versus time (fig. 5) the values of the
L°
products decrease constantly from evening to morning. For further
evaluation the resulting curves, too, have to be idealized.
Calculation of soil moisture
Heat flows (AQ) can be calculated for various densities of solid matter
in a given volume (p) on the basis of a medium specific heat capacity
for solid matter cg = 0,46 (fig. 6). But only if the temperatures of
the faces of the soil volume are known, exact values for the water con-
tents can be derived.
To apply the diagram in fig. 6 to remote sensing data it is necessary
to find the correlation between AQ, AS and T,. Graphically this is done
L'
in fig. 7. This diagram clearly shows that there exists a correlation
only if day time, too, is taken into consideration. Thus isochrones had
to be interpolated in the diagram of fig. 7.
On the basis of fig. 6 and 7 soil moisture was calculated with thermal
data of a flight in 1100 m and a flight in 4300 m above ground (30.6.76,
Kochel Moos area). The values show maximum error of 10 Z (4300 m flight)
and 5 2 (1100 m flight). Only the values for areas A and B in fig. 10
differ very much from the ground truth value of soil moisture. In this
case temperature difference in the soil and density of the soil differ
with an unknown amount from the areas with peaty soil. This example
shows, that for even small areas a lot of stations for ground truth