314 
and as a consequence organic matter might settle out 
in discrete grains. In a Ca free environment organic 
carbon is more readily absorbed by clays and thus the 
color of the soils should be lighter. 
These secondary results are probably less important 
than the main relationships between organic carbon, 
iron, particle size, and reflection. What these main 
results show is that parameters can readily interact 
and thus obscure overall relationships. Under selec 
tive conditions it is however possible to predict soil 
chemical conditions that are pertinent to soil fer 
tility and moisture status in agricultural fields. 
It is suggested that individual fields be examined 
separately so as to establish what relationships are 
most significant for predictive purposes. Once these 
relationships have been established for an individual 
field, reflection values can be used to predict the 
properties for a number of years. Another parameter 
of primary importance in soils studies is moisture 
content. Preliminary results suggest that moisture 
has the greatest effect on reflection measurements 
at 1050 nm. An in depth analysis is currently under 
way to determine the effect of soil moisture on soil 
reflection. 
Multiple correlations and regressions were also 
carried out on the three data sets in order to see if 
the overall predictive accuracy can be improved when 
considering the relationships between reflection and 
multiple physical and chemical soil properties. A 
stepwise regression showed that (total amount of 
variance explained by regression equation) could be 
improved by the inclusion of several properties, but 
as can be seen in Table 1 the improvements were very 
small. 
Table 1. Comparison between single and stepwise mul 
tiple regressions. 
Regression 
Relationship 
R 2 
Set 1 (high 
size 
organic Carbon variability, 
variability) 
low particle 
Single 
Stepwise 
% Reflection (550 nm) 
vs % Carbon 
% Reflection (550 nm) 
vs % C + Ca + N0 3 
R 2 = 85% 
R 2 = 88% 
Set 2 (high particle size variability, 
Carbon) 
low organic 
Single 
Stepwise 
% Reflection (1600 nm) 
vs % Sand 
% Reflection (1600 nm) 
vs % Sand + P + NO3 
R 2 = 64% 
R 2 = 70% 
Set 3 (high 
size 
iron, no organic Carbon, low particle 
variability) 
Single 
Stepwise 
% Reflection (630 nm) 
vs % Iron 
% Reflection (630 nm) 
vs % Iron + Si 
R 2 = 87% 
R 2 = 89% 
From the above results it is evident that selected 
physical and chemical soil properties can be differ 
entiated with multispectral reflection measurements. 
But these assessments are by no means simple and are 
greatly dependent on individual site conditions. It 
is clearly evident that organic carbon exerts an 
overwhelming influence on reflection above the 2% 
concentration level. At levels below 2% organic car 
bon, other properties such as particle size and % 
iron content become dominant in influencing reflec 
tion. The first two properties are extremely impor 
tant in soil fertility management both in terms of 
direct influence on fertility as well as indirect in 
fluence via interdependence. 
Soil surface conditions were held constant in these 
laboratory tests but these will vary considerably in 
the field. The results do indicate that if we use a 
laser system we should be able to separate soil con 
ditions spectrally since differences in surface 
roughness will be discerned with the separately de 
termined height measurement capabilities of the 
laser. Attention should be given to the site con 
ditions, particularly the organic carbon content 
present, since the relationships between reflection 
and soil properties are site dependent and the re 
lationships are clearly not universally applicable. 
4 CONCLUSIONS 
In this study the remote assessment of vegetation 
via an airborne laser system was demonstrated. The 
results showed that the combination of high resolu 
tion height profile tracing and the measurement of 
laser reflection can greatly facilitate terrain 
assessment by remote sensing. Applications of this 
technique were demonstrated in a forested terrain 
and since current lasers operate at a single wave 
length range the potential of using multispectral 
laser sensing was investigated using ground based 
spectral reflection measurements. In this latter 
study applications pertinent to soil fertility 
assessments were demonstrated. Based on the results 
the following conclusions can be made. 
4.1 Terrain height and reflection measurements with 
an airborne laser 
The laser system provided a height profile tracing 
of individual tree canopies at a vertical accuracy 
of better than 20 cm. At the same time laser reflec 
tion values were generated over the same target. 
This combined data set is unique because for the 
first time we can measure surface roughness and de 
termine its influence on spectral reflection. With 
such data, vegetation and terrain differentiation is 
greatly facilitated since objects with similar re 
flection can further be separated on the basis of 
the height profile. The results showed that dis 
tinctions between broadleaf and coniferous vegetation 
and tree vs ground cover vegetation is greatly im 
proved. 
4.2 Spectral reflection measurements to differentiate 
soil fertility conditions 
It is anticipated that lasers with multispectral 
capabilities can further improve remote sensing 
assessments and the potential of such a system was 
demonstrated in a case study which involves soil 
fertility assessments. Spectral reflection measure 
ments of soil samples at 550, 630, and 1600 nm wave 
lengths were significantly related to organic carbon, 
% iron content and % sand content respectively. 
However, these relationships are not universally 
applicable and can only be used for predictions 
within a restricted concentration range. 
Organic carbon has the greatest influence on % 
reflection (at 550 nm) when the concentrations in 
the soils are above 2%. Below 2% carbon content 
other properties such as particle size and iron 
content will become dominant and depending on the 
conditions predictions of these two properties on a 
site specific basis are possible from reflection 
measurements at 1600 nm and 630 nm wavelengths. 
These relationships are thought to be useful in 
quantifying soils for sampling and fertilizer man 
agement on a field by field basis. Since there are 
different surface roughness parameters to be con 
sidered in agricultural field assessments it is 
suggested that the use of a dual or triple laser at