412
C O R S I , D E L E E U W, A N D S K I D M O R E
this relates only to the extent of the analysis, not to its resolution. Nevertheless,
on a cost–benefit basis, it is generally more efficient to address conservationrelated issues at a coarser scale, which enables a landscape approach, than to
concentrate on a more detailed scale (e.g., individual or population level),
which requires high-resolution data to be analyzed that are either too precise
or simply too abundant in terms of storage requirements to be analyzed profitably with a landscape approach.
What economics suggests is that conservation science needs to have a
broader view of phenomena. A broad-scale approach and the possibility of predicting the potential dynamics of spatial patterns are needed to manage fragmentation of suitable environments and the inevitable metapopulation structure of the resulting population (Noss 1992). May (1994) indicates that when
multiple levels of biological organization are concerned, as in a typical conservation action, the best management approach can be achieved on the regional
landscape scale (10
3 to 10 5 km 2 ). This scale level has suffered historically from
limitations in the tools available for consistent analysis and is the one that has
gained the most from the evolution of GIS; in fact, most of the distribution
models based on GIS address problems at regional landscape level.
DATA AVAILABILITY
Data availability and quality are two of the three limiting factors in the development of GIS-based species distribution models (the other being reliability of
the models themselves [Stoms et al. 1992], which is discussed later in this chapter). The problem of developing extensive data sets of environmental variables
is limited by economic and political rather than technical constraints. Estes
and Mooneyhan (1994) list a number of different attitudes of governments
throughout the world that limit the availability of high-resolution, “sciencequality” 5 environmental data sets. These range from military classification of
the data, thereby precluding the use of the data to the scientific community, to
the low political priority that certain governments give to environmental issues.
Moreover, even when policy is not an obstacle to the production and availability of data sets, entire nationwide data sets are sometimes lost during revolutions, wars, and civil disturbances. To this it should be added that some governments (e.g., the European Union countries) ask high prices for data sets,
which are generally acquired with tax money, actually preventing their broad
use in any type of activity and more specifically in environmental research.
In many cases, high-quality site-specific data sets are generated for a particular research project but are compiled with nonstandard techniques, rendering
C O R S I , D E L E E U W, A N D S K I D M O R E
this relates only to the extent of the analysis, not to its resolution. Nevertheless,
on a cost–benefit basis, it is generally more efficient to address conservationrelated issues at a coarser scale, which enables a landscape approach, than to
concentrate on a more detailed scale (e.g., individual or population level),
which requires high-resolution data to be analyzed that are either too precise
or simply too abundant in terms of storage requirements to be analyzed profitably with a landscape approach.
What economics suggests is that conservation science needs to have a
broader view of phenomena. A broad-scale approach and the possibility of predicting the potential dynamics of spatial patterns are needed to manage fragmentation of suitable environments and the inevitable metapopulation structure of the resulting population (Noss 1992). May (1994) indicates that when
multiple levels of biological organization are concerned, as in a typical conservation action, the best management approach can be achieved on the regional
landscape scale (10
3 to 10 5 km 2 ). This scale level has suffered historically from
limitations in the tools available for consistent analysis and is the one that has
gained the most from the evolution of GIS; in fact, most of the distribution
models based on GIS address problems at regional landscape level.
DATA AVAILABILITY
Data availability and quality are two of the three limiting factors in the development of GIS-based species distribution models (the other being reliability of
the models themselves [Stoms et al. 1992], which is discussed later in this chapter). The problem of developing extensive data sets of environmental variables
is limited by economic and political rather than technical constraints. Estes
and Mooneyhan (1994) list a number of different attitudes of governments
throughout the world that limit the availability of high-resolution, “sciencequality” 5 environmental data sets. These range from military classification of
the data, thereby precluding the use of the data to the scientific community, to
the low political priority that certain governments give to environmental issues.
Moreover, even when policy is not an obstacle to the production and availability of data sets, entire nationwide data sets are sometimes lost during revolutions, wars, and civil disturbances. To this it should be added that some governments (e.g., the European Union countries) ask high prices for data sets,
which are generally acquired with tax money, actually preventing their broad
use in any type of activity and more specifically in environmental research.
In many cases, high-quality site-specific data sets are generated for a particular research project but are compiled with nonstandard techniques, rendering