🚀 go-pugleaf

To be presented back-to-back at 12:30 pm in room 204C of the Evans Library,
Texas A&M University, College Station.
---------------------------------------------
INDUCTIVE EXPLORATORY SPATIAL DATA ANALYSIS: USE OF GIS FOR GENERATING
GEOGRAPHICAL HYPOTHESES AND SPATIAL HABITAT DATA.

Richard Aspinall

Macaulay Land Use Research Institute

Aberdeen, Scotland

ABSTRACT


A variety of inference engines have been integrated with GIS and used
to search spatial databases for interesting pattern/relationships.
This paper presents a method for analysis of spatial data in GIS to
generate habitat data describing suitability for a species; analysis
uses raster format but the output can be converted to topologically
coded vector format.  Examples are used to illustrate the methods.
The basic method uses bayesian probability statistics for spatial
analysis.  It is used to derive bioclimatic maps for wildlife species
from species distribution maps and sixty maps of climate.  This
spatial analytical method i) selects the climatic datasets to use for
modelling from the set of sixty, ii) establishes appropriate class
intervals for each climatic dataset, iii) presents quantitative
estimates of ecological relations between the species being analysed
and climatic variables (as testable hypotheses), and iv) carries out
monte carlo error sensitivity analysis on the input variables and
modelled relationships.

This method is also used to derive a very (spatially) detailed map of
habitat quality directly from analysis of environmental data (thematic
maps and satellite imagery) and coarse resolution wildlife survey
data.  Developing species specific distribution maps from satellite
imagery and thematic maps by direct use of wildlife survey data avoids
a number of the limitations inherent in using satellite imagery and
thematic data for wildlife habitat mapping.  The product of analysis
is a detailed map (or `information surface') at the full spatial
resolution of the raster GIS datalayers which describes the
distribution of the species as probability of occurence scores; this
also represents an index of habitat quality or suitability and can be
analysed further with GIS.  Since the map product is directly related
to the distribution of the species under investigation, the need to
establish a link between image classes and thematic categories and the
`ecological' habitat types on which habitat suitability indices
typically depend is reduced.  The strength of this relationship is
fundamental to the overall accuracy of GIS-based analyses of habitat
maps which use rules and weights to combine datalayers to produce maps
of habitat suitability from composite indices.  With the method
described here, relationships between wildlife distribution and
habitat are inherent in the map output.

The digitally derived habitat map is also fully integrated into GIS
for further spatial analysis, and, since the product generation
process is quantified, subsequent error analysis can use uncertainty
information associated with the habitat map.  Further, the information
content of the habitat map allows it to be used for analysis of
species distribution and abundance.  GIS provides an appropriate
analytical environment for this analysis, which uses `object' and
`field' data types and has a general relevance beyond the specific
data source used for this example.

The map is treated as a series of discrete habitat patches (spatial
objects) and analysed by applying the principles of landscape ecology.
Here, GIS is used for identification and quantitative description of
spatial pattern, both in habitat structures in the map, and for
functional elements identified through analysis of topological
relationships between habitat structures.  These measures can be used
to analyse habitat characteristics and wildlife abundance.  Since
analysis of structure and function using GIS retains spatial
properties of the habitat information, this leads to distribution and
abundance models which are ecologically more acceptible than the more
commonly developed models based on statistical analysis of data
restructured to conform to the standard statistical data model of
cases and variables.

Alternative forms of analysis using the habitat map as a surface of
continuous variation (a data field) is outlined.  In this case,
foraging behaviour and other activity in relation to (continuous)
variation in habitat quality are used to model distribution and
abundance of the wildlife population.  This uses both GIS and expert
systems technology.

------------------------------------------------

ANALYSIS OF VEGETATION SUCCESSION USING GEOGRAPHIC INFORMATION SYSTEMS AND
KNOWLEDGE--BASED SYSTEMS

David Miller

Macaulay Land Use Research Institute

Aberdeen, Scotland                              

ABSTRACT

The Scottish Office Agriculture and Fisheries Department has been
concerned to know the area of bracken Pteridium aquilinum in Scotland
and how rapidly it is spreading. It offers a potential danger to
animal health and a devaluation of land for agricultural practise.
Bracken is present in patches which are often difficult to delimit
because of variable density and appearance throughout the year and was
not consistently identifiable at a single minimum spatial resolution.
The encroachment of bracken on to areas of rough grazing and improved
pasture with a resultant decrease in agricultural return from such
land is widespread in upland Scotland, Wales and areas of England and
its areal extent is increasing by up to 3% per annum. The targeting of
resources for the control of bracken is difficult because of the
uncertainty about the extent of current bracken presence and where it
is spreading most rapidly.

Bracken provided a candidate vegetation type for prototyping a
knowledge--based system (KBS) which synthesized environmental data
expressed spatially, with knowledge on vegetation dynamics and
vegetation succession in particular. Temporal models of vegetation
types, which have different degrees of reliability, with analysis of
static spatial data were incorporated. The knowledge--based system
permits the user access to knowledge about botany, ecology, mapping,
image processing and spatial data analysis for an informed and
informative answer to their query.
           
Bracken was mapped at three experimental sites using land survey,
photogrammetric techniques and satellite image processing to quantify
the extent of bracken and bracken change. An existing data model was
enhanced by building--in the hierarchy of survey observations and
associated error bands. The model was then implemented within an
expert system shell with KBS routines written in POP--11 and Fortran.
The coupling of the geographic information system (GIS) and KBS
routines was of the form of: 1. organisation of the transfer of
descriptive values as parameters between the data--models of the KBS
and the GIS routines and 2. read/write of the spatial data into the
data--models for analyses by KBS routines.

Data models were designed for handling meta--data about the system
routines, data types and processing history of files. Recording the
data_lineage provided a means of assessing propagation of certain
geometric errors through the processing stages.  Environmental data
(soils, topography) around patches of bracken are represented in a
separate data model for assessment of its susceptability to changes.
Explanations of what processes have been undertaken, how and why they
were implemented are formulated during each system run.

The KBS was applied to mapping bracken in the Trossachs area of
Scotland, where the environmental niche preferences of bracken are
well documented in a qualitative sense and provide hypotheses to test
with spatial analysis techniques. These hypotheses were tested,
supported and used to improve the mapping of bracken using GIS.
Documentation of the ecological dynamics of bracken are encoded to
predict where bracken may spread over a twenty year time period.
Potential developments in the use of coupled GIS and KBS for analyses
of vegetation change are discussed.


-- 
Hal Mueller		  hmueller@orca.tamu.edu