422
A.H. Schumann and G.A. Schultz
provided most efficiently by remote sensing. Also distributed hydrological models
should be used for an impact assessment of land cover changes since they allow
consideration of the spatial variability in land cover.
19.2 Hydrological Modelling and Land Cover Change
Hydrological models which are used for hydrological impact assessment of land
use changes should represent the hydrological processes at the land surface physically based. Many hydrological models consider the specific land use characteristics in their parameterization. Different approaches to consider the spatial heterogeneity of landuse characteristics in hydrological modelling are possible. Examples are:
in the widely used Soil Conservation Service Method (SCS-model) (Maidment,
1993) various types of vegetation and crops, land treatments and crop practices, paving and urbanization are considered in relationship to one of four soil
types to calculate a composite runoff curve number for a drainage basin,
- a statistical approach to describe the spatial heterogeneity of the storage capacity of the upper, rooted soil zone by a distribution function which can be derived directly from an overlay of the vegetation and soil map within a GIS was
proposed by Schumann (1993) (e.g. Fig. 16.3),
- in the Precipitation-Runoff Modeling System (PRMS) (Leavesley & Stannard,
1995) the vegetation type is one of many characteristics (such as slope, aspect,
elevation, soil type and precipitation characteristics) which are used for partitioning of the watershed in so"called Hydrological Response Units which are
modelled separately,
- in the distributed SHE- model the catchment is represented by an orthogonal
grid network, where the vegetation type characterises a grid square (Abbott et
aI., 1986).
In many hydrological models the parameters of vegetation related model components (e.g. for interception and evapotranspiration) are chosen dependent on the
land use. For urbanised areas mostly special developed model components are
used to consider the specific runoff formation and concentration processes in these
areas. If these models are used to describe the hydrological effects of land cover
changes it seems easy to alter their parameters. But two problems have to be considered in such model analysis:
The first problem concerns the model approach:
Is the used hydrological model really able to describe the complex changes of
hydrological conditions for the specific catchment of interest?
Klemes (1986) proposed the following methodology to answer this question:
Find a gauged basin where a similar land use change has taken place during the
period covered by the historic record and calibrate the model on a segment corresponding to the original land use and validate it on a segment corresponding to the
changed land use. For practical purposes it is very difficult to fulfil this demand.
Not only are gauged basins with the specific land use changes rare (with exception
of urbanized catchments) also the climatic conditions of the validation and cali-
A.H. Schumann and G.A. Schultz
provided most efficiently by remote sensing. Also distributed hydrological models
should be used for an impact assessment of land cover changes since they allow
consideration of the spatial variability in land cover.
19.2 Hydrological Modelling and Land Cover Change
Hydrological models which are used for hydrological impact assessment of land
use changes should represent the hydrological processes at the land surface physically based. Many hydrological models consider the specific land use characteristics in their parameterization. Different approaches to consider the spatial heterogeneity of landuse characteristics in hydrological modelling are possible. Examples are:
in the widely used Soil Conservation Service Method (SCS-model) (Maidment,
1993) various types of vegetation and crops, land treatments and crop practices, paving and urbanization are considered in relationship to one of four soil
types to calculate a composite runoff curve number for a drainage basin,
- a statistical approach to describe the spatial heterogeneity of the storage capacity of the upper, rooted soil zone by a distribution function which can be derived directly from an overlay of the vegetation and soil map within a GIS was
proposed by Schumann (1993) (e.g. Fig. 16.3),
- in the Precipitation-Runoff Modeling System (PRMS) (Leavesley & Stannard,
1995) the vegetation type is one of many characteristics (such as slope, aspect,
elevation, soil type and precipitation characteristics) which are used for partitioning of the watershed in so"called Hydrological Response Units which are
modelled separately,
- in the distributed SHE- model the catchment is represented by an orthogonal
grid network, where the vegetation type characterises a grid square (Abbott et
aI., 1986).
In many hydrological models the parameters of vegetation related model components (e.g. for interception and evapotranspiration) are chosen dependent on the
land use. For urbanised areas mostly special developed model components are
used to consider the specific runoff formation and concentration processes in these
areas. If these models are used to describe the hydrological effects of land cover
changes it seems easy to alter their parameters. But two problems have to be considered in such model analysis:
The first problem concerns the model approach:
Is the used hydrological model really able to describe the complex changes of
hydrological conditions for the specific catchment of interest?
Klemes (1986) proposed the following methodology to answer this question:
Find a gauged basin where a similar land use change has taken place during the
period covered by the historic record and calibrate the model on a segment corresponding to the original land use and validate it on a segment corresponding to the
changed land use. For practical purposes it is very difficult to fulfil this demand.
Not only are gauged basins with the specific land use changes rare (with exception
of urbanized catchments) also the climatic conditions of the validation and cali-