Modeling Species Distribution with GIS
413
them unsuitable for combination and the achievement of more extensive
knowledge of an area.
In the past few years there has been an increasing effort to develop metadatabases of available data sets throughout the world, and the problem is being
addressed by national and international organizations (e.g., United Nations
Environmental Programme, World Bank, U.S. Geological Survey [USGS],
European Environmental Agency). These initiatives still do not address the
problem of producing high-quality data sets, but at least they are a start in collating existing data sets. An important example is given by the joint efforts of
the USGS, the University of Nebraska–Lincoln, and the European Commission’s Directorate General Joint Research Centre, which are generating a 1km-resolution Global Land Cover Characterisation (GLCC) database suitable
for use in a wide range of environmental research and modeling applications
from regional up to continental scale. All data used or generated during the
course of the project (source, interpretations, attributes, and derived data),
unless protected by copyrights or trade secret agreements, are distributed
through the Internet. This effort goes in the direction of producing and distributing homogeneous medium-resolution high-quality data sets with known
standards of accuracy.
Further aspects of raw data sets are discussed in the next section, where the
quality of the data used to build models is discussed. We do not discuss this
issue further here because we do not believe it to be a problem that can be
addressed directly by conservation biologists or ecologists, although they can
contribute to developing awareness of the need for standardization of data sets
and for their production and dissemination.
VALIDATION AND ACCURACY ASSESSMENT
Generally, the main function of a GIS-based species distribution model is to
produce a map or its digital analogue for assessment of management and conservation actions. Possibly the most important question to be asked by a user
is ‘how accurate is the distribution map that has been produced?’
Many articles have been written on the sources of error in the data layers
that may be included in a GIS. Nevertheless, few authors of papers dealing with
animal distribution include an assessment of the accuracy of their model and
a validation of the product. Because we believe this issue to be central to the
entire process of species distribution modeling, the aim of this section is to
review sources of error in GISs, to discuss methods of assessing mapping accuracy, and to evaluate the accumulation of thematic map errors in GISs, thus pro-
413
them unsuitable for combination and the achievement of more extensive
knowledge of an area.
In the past few years there has been an increasing effort to develop metadatabases of available data sets throughout the world, and the problem is being
addressed by national and international organizations (e.g., United Nations
Environmental Programme, World Bank, U.S. Geological Survey [USGS],
European Environmental Agency). These initiatives still do not address the
problem of producing high-quality data sets, but at least they are a start in collating existing data sets. An important example is given by the joint efforts of
the USGS, the University of Nebraska–Lincoln, and the European Commission’s Directorate General Joint Research Centre, which are generating a 1km-resolution Global Land Cover Characterisation (GLCC) database suitable
for use in a wide range of environmental research and modeling applications
from regional up to continental scale. All data used or generated during the
course of the project (source, interpretations, attributes, and derived data),
unless protected by copyrights or trade secret agreements, are distributed
through the Internet. This effort goes in the direction of producing and distributing homogeneous medium-resolution high-quality data sets with known
standards of accuracy.
Further aspects of raw data sets are discussed in the next section, where the
quality of the data used to build models is discussed. We do not discuss this
issue further here because we do not believe it to be a problem that can be
addressed directly by conservation biologists or ecologists, although they can
contribute to developing awareness of the need for standardization of data sets
and for their production and dissemination.
VALIDATION AND ACCURACY ASSESSMENT
Generally, the main function of a GIS-based species distribution model is to
produce a map or its digital analogue for assessment of management and conservation actions. Possibly the most important question to be asked by a user
is ‘how accurate is the distribution map that has been produced?’
Many articles have been written on the sources of error in the data layers
that may be included in a GIS. Nevertheless, few authors of papers dealing with
animal distribution include an assessment of the accuracy of their model and
a validation of the product. Because we believe this issue to be central to the
entire process of species distribution modeling, the aim of this section is to
review sources of error in GISs, to discuss methods of assessing mapping accuracy, and to evaluate the accumulation of thematic map errors in GISs, thus pro-