iis 
particular, left 'footprints' on the ground that made 
multitemporal work impossible. Recently, tools with 
optical sensors have been developed to provide fast, 
high-resolution, contact-free measurements. These sys- 
tems work on a photogrammetric basis (ULLaH & Dickin- 
soN 1979) or rely on laser point triangulation (HuaNc & 
Bnapronp 19904). Both methods are able to meet the 
requirements for a detailed physical interpretation of 
surfaces. 
The aim of this investigation was to quantify the causal 
factors by which the microrelief influences interrill 
erosion and to register the effect on surface runoff. 
Comprehensive tests were performed in order to do 
this. In a rain simulator soil probes with different micro- 
reliefs were rained on. Each relief was evaluated before 
and after the rain, respectively. Further tests took place 
on outdoor fields to prove the universality of the results 
determined in the laboratory. From the photogramme- 
trically measured Digital Elevation Models (DEMs) indices 
for each surface are derived as well as spatial microrelief 
Structures are exposed by geostatistical processing. Fur- 
thermore, deterministic models are developed that 
permit a quantitative evaluation of the depression reser- 
voir capacity and the effective rain energy as derived 
from the microrelief. This allows to verify the above 
mentioned theories. 
TEST SERIES 
The rainfall simulator used in the laboratory had a 
drop size distribution equal to natural rainfall. It pro- 
vides variable rainfall intensity and rain energy matches 
about 9596 of the energy of similar natural rain. For a 
detailed description of the system see Roru & HELMING 
(1992). In this simulator an Ap-horizon soil of a Haplic 
Luvisol derived from loess (596 slope) was irrigated for 
two hours with an 30 mm/h intensity (figs. 1a, 1b). 
Different kinds of microrelief were produced by passing 
the soil through a sieve. Three constellations were eva- 
luated, i.e. 
* rough (r); SO mm sieve; seedbed for winter wheat 
* medium (m); 25 mm sieve; seedbed for sugar beet 
* fine (f); 10 mm sieve; seedbed for rape 
Surface runoff was measured every two minutes. Before 
and after raining the microrelief was measured on a 
0.2 m^ area within the 1 m? area that was rained on. 
DEM spacing was 2 mm, yielding 50,000 points. 
For open air verification tests were done on two sugar 
beet fields in Lower Saxony, Germany. The soil was a 
Haplic Luvisol too. Measurements on a 0.98 m? area 
with 3 mm horizontal spacing were performed just after 
seeding (March/April) and four months later at the 
end of July, respectively. 
    
   
  
  
  
   
    
   
  
  
  
  
  
  
  
  
  
  
  
  
    
    
   
  
  
  
  
  
  
  
  
  
   
    
     
   
   
    
    
   
  
   
  
  
  
Fig. 1b: Same test site after 60 mm rain 
IMAGE ACQUISITION 
Analog Images 
Choice of a suitable image acquisition configuration 
depends on requirements relating to the accuracy finally 
wanted and on limitations due to the object environ- 
ment. In this project the objective was measurement 
of rectangular areas (1 x 1 m? in the field, 0.4 x 0.5 m? 
in the laboratory rain simulator tests) in order to connect 
quantitatively rain with erosion, i.e. to register changes 
in soil relief. For change detection normally three 
epochs were used, evaluating the soil relief before and 
after different rain periods. The periods between image 
acquisition dates in the laboratory, where intensity 
and duration of the rain could be controlled and mea- 
sured accurately, were easily predefined since the simu- 
lated rain could be interrupted at any time, e.g. after 
30 minutes, to have a break of some 10 minutes for