490 
Table 1. Values for the coefficients a and b in eq. 
(2) for a number of crops with crop height H (after 
Thunnissen 1984a) 
Crop 
H 
(cm) 
a 
or 1 ) 
b 
(K 1 -m 1 •s) 
Grass 
<15 
0.050 
0.010 
Grass 
>15 
0.050 
0.017 
Potatoes 
Sugar beet 
60-, 
60' 
0.050 
0.023 
Cereals 
100 
0.090 
0.030 
Maize 
200 
0.100 
0.047 
le 24 /le 24 = 1 - B r (T -T*). (1) 
p C C 1 
24 24 
where LE and LE are respectively the actual and 
potential 24 hour P evapotranspiration rate (mm-d - ^), 
B r (K _1 ) is a calibration constant and the subscript 
i indicates instantaneous values. By means of eq. (1) 
differences in radiation temperature of a certain 
crop derived from thermal images can be directly 
transformed into reductions in evapotranspiration. 
Values for the calibration constant B r in eq. (1) 
have been determined with the aid of TERGRA-model cal 
culations (Thunnissen 1984a). It was found that B r 
can be described by a linear function of the wind 
velocity (u) at a height of 2.0 m above ground sur 
face : 
B r = a + b . u (K -1 ) (2) 
Values for coefficients a and b were calculated for 
different types of crops and crop heights (table 1). 
For agrohydrological purposes heat images are usual 
ly taken on clear days in the summer period. It was 
found that for such days eqs. (1) and (2) can be ap 
plied for the meteorological conditions prevailing in 
the Netherlands. 
2.2 Mapping of evapotranspiration from digitally tak 
en reflection and thermal images 
The relationship between the crop temperature and 24 
hour evapotranspiration rate is crop dependent. This 
means that crop temperatures derived from thermal 
images have to be translated into evapotranspiration 
rates for each crop type separately. Consequently a 
crop map is necessary to compose an evapotranspiration 
map. Several investigators have applied methods for 
computer-assisted crop classification procedures with 
the aid of digitally taken reflection images (e.g. 
Crist and Malila 1981, Badhwar et al. 1982, Batista 
et al. 1985). 
Especially for grassland one has to account for 
crop height. The Vegetation Index as defined by Tucker 
(1977) can be applied for this purpose. 
Afterwards with eqs. (1) and (2) differences in 
radiation temperature for each crop type as derived 
from thermal images can be transformed into reductions 
in evapotranspiration. With the method developed 
digitally taken reflection and thermal images can be 
translated into evapotranspiration maps with the aid 
of operational image processing systems. 
2.3 Simulation of the water balance of a cropped soil 
With a model such as SWATRE (Feddes et al. 1978, 
Belmans et al. 1983) the use of water by agricultural 
crops can be simulated during the entire growing 
season, hence it can be used to investigate how far 
the moisture conditions at particular times as deter 
mined with remote sensing are representative for the 
entire growing season. The SWATRE-model is a tran 
sient one-dimensional finite difference soil-water- 
root uptake model. 
There are also three-dimensional models in which 
saturated and unsaturated groundwater flow is link 
ed. With these models also information about regional 
evapotranspiration is obtained. The GELGAM-model 
(De Laat and Awater 1978) is such a model. 
Remote sensing images supply a synoptic view about 
the location of areas with shortage of water and the 
degree of the occurrence of drought damage. The 
causes of drought damage can be diverse. The avail 
able soil moisture for crop evapotranspiration is 
determined by: 
- the depth of the root zone; 
- the available moisture capacity in the root zone; 
- the hydraulic conductivity of the subsoil; 
- the groundwater level during the growing season; 
- additional water supply (rainfall, sprinkling 
irrigation). 
The first three factors are mainly dependent on soil 
properties. One has to realize that no reliable in 
formation about soil physical characteristics and 
groundwater table depth can be obtained if sprinkling 
irrigation is applied. 
By simulating the water flow in the soil-plant- 
atmosphere system the influence of the mentioned 
factors on crop water supply can be determined under 
different circumstances. This means that for the ex 
planation of the occurrence of drought damage as de 
tected with remote sensing, calculations with agro 
hydrological simulation models can be applied. 
2 RESULTS AND DISCUSSION 
3.1 Determination of soil physical characteristics 
Figure 2 shows a thermal image taken after a dry 
period in the summer of 1983 of a part of the peat 
area in the northeastern part of the Netherlands 
(study area 1 in Figure 1), where mainly potatoes 
(50%) , sugar beet (25%) and cereals (20%) are grown. 
In Figure 2 boundaries as derived from the available 
soil map (1:50,000) have been indicated as well. 
Crops grown on Haplaquod soils (Hn21) show high tem 
peratures. This means that severe drought conditions 
prevailed. Relatively low-situated and more peacy 
soils (iWp, iWz and iVz in Figure 2) show lower tem 
peratures. This means that crops on these soils are 
better supplied with water. Especially crops grown 
on soils with a thick peat layer (more than 40 cm 
peaty material between 0.0 and 0.8 m below soil sur 
face; indicated with iVz) show an optimal water sup- 
Figure 2. Thermal image of a part of study area 1 
(see Figure 1) taken on 8 August 1983 at 13.30 MET. 
Black is cold and white is warm. For the indicated 
soil codes see text