f an EIA 
ts (Rivas 
and (3) 
iderable, 
'Spect to 
elatively 
y can be 
n of the 
e impact 
? impact 
'abilistic 
rategies. 
carce or 
  
  
acts 
onmental 
onents 
  
ndary 
of 
ls 
rns 
cant . 
  
ruction 
Bertens, Jurjen 
  
which may represent a difficulty for communication between the two sides of the motorway, visual impact of the new 
infrastructure, land occupation or degradation in an area where available land with a high potential for use is quite limited 
and interference with natural processes and hazards. These include terrain instability (landslides, collapse, subsidence) and 
infiltration and runoff related processes (groundwater recharge and pollution, soil erosion, channel erosion, water quality, 
siltation, waterlogging and flooding (figure 2). 
3 HYDROLOGICAL IMPACTS 
3.1 Nature of impacts 
Fundamental to the geomorphological impact of transportation systems is the way in which hydrology is affected. Even if 
impacts on the processes considered are not relevant, their simulation allows for the assessment of their relevance. Changes 
imposed upon the landscape by construction and use of infrastructure will affect mechanisms such as infiltration, runoff, 
erosion and pollution. 
Full understanding of physical processes should allow for quantitative predictions with respect to the response of dynamic 
processes. A way to achieve a better understanding of these processes and making predictions is by environmental modelling 
(in this case hydrological modelling). In general, a model is fed by knowledge of a certain discipline; the use of such a model 
(and its predictions) for impact assessment and decision support, aims at transferring such knowledge across disciplines. 
If a model simulates the physical system in a satisfactory way, it is assumed the processes are represented in a manner that 
corresponds with their behaviour in the physical system. Incorporating the presence of the project (a transportation system) 
in the model can lead to a quantification of the impacts of the transportation system on hydrological processes. When spatial 
and temporal variability are accounted for in the model, the behaviour of hydrological mechanisms (in time and space) and 
the way in which they interact can be assessed with respect to the presence of a transport system. For example, large 
amounts of runoff from the highway surface will favour downslope erosion, which will decrease soil thickness, which will 
decrease storage capacity of the soil, which in turn will limit infiltration and thus favour runoff generation. 
3.2 Problems for the assessment of hydrological impacts 
It must be recognised that, although many environmental models have been developed, we will rarely have models at our 
disposal that can provide sound, quantitative estimates. This is partly due to the fact that these models have been developed 
for research areas that are usually of orders of magnitude smaller than management areas (Grayson ef al., 1993). Research 
areas are often selected on the basis of interest in a specific phenomenon or process or on the basis of data availability. This 
is not appropriate within an impact assessment framework, where the study area is provided not by an interest in a specific 
process but rather by the focus of the project (in this case the construction of infrastructure) or a particular hazard (flooding 
of a given river). Considerable database development is needed for analysis of urban watersheds (Maidment, 1996). Lumped 
models have traditionally been developed for application to large watersheds and require less data input, but they are clearly 
not capable of providing distributed information. 
In general, environmental models usually require a large amount of data input. Very often in the case of EIA, not all data one 
would desire are available or can be collected. Therefore the study of geomorphological processes does not allow for precise 
predictions with respect to the impacts on these processes (Rivas eft al, 1997). Lack of data for future conditions also 
constrains possibilities for calibration and validation. A model which calculates predictions for future conditions, will have 
to be evaluated on a non-mathematical basis. This must be done with simple reasoning (Grayson ef al., 1993) to provide 
some sort of validation, where no data is available for proper validation. This corresponds to the term "face validity" which 
describes the models credibility (Marcot ef al., 1983). 
Reliance solely on these models should be replaced with a combination of quantitative understanding of hydrological 
response and simple reasoning to assist in the decision making process. These methods will be no more accurate than 
complex models, but are simpler and more modest. Such approaches may be undertaken within or outside a GIS 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000. 177 
si 
ÿ 
a 
y 
& 
"d 
4 
ÿ 
d