LANDSCAPE ECOLOGY
SREM 3011
LECTURE 2
Dr Brendan Mackey
Department of Geography
The Australian National University
‘Principles of Geographic Information Systems
for Land Resource Assessments’
by Burrough,P.A,(1986/7)
Oxford Science Publications
Monographs on Soils and Resources
Survey no,12
- raster (gridded),point,and vector data types
- grid based data manipulation
- data visualisation (2d,3d)
- scale; resolution
- real and integer data types
- ancillary modelling programs
A fundamental problem!
How do we know what is where?
How can we obtain landscape-wide (ie,
spatially distributed) data about the
ecological (biophysical) resources of a
landscape?
We can only ever take measurements
from a relatively small number of field
survey sites/plots; climate,topography,
vegetation,wildlife,substrate
The ‘Integrated Landscape Paradigm’
1940s,CSIRO Land Use Research
1990s,CSIRO Water and Land
Reconnaissance survey for land use evaluation
- Grazing
- Agriculture
- Forestry- Recreation
- Nature conservation
Surveyed large areas to determine major types of
country and to assess natural resources
Analyse patterns from visual inspection of air photos
- Land units/elements that have a uniform parent
material,topography,soil and vegetation
- Land systems,however,were mapped
>1 occurrence of a land unit
Field survey undertaken to sample land units
So,survey data were spatially extended through
assumption of covariance
ie,assumed all examples of a land
unit sufficiently similar to infer
properties of an unsurveyed unit
from a sampled site
A product of the available technology
- air photo interpretation
- cartography
Basic ‘assumption of covariance’ still used in
survey (soil,geology,vegetation!)
- cartographic products often used as source data
for GIS!
Location of Chimbu Province
Sedimentary spurs north of the Sena River east of Karimur,Landscape
components are scarp (C),spur crest (D),spur slope (A) and footslope (B).
Adjoining this unit are the Sena River terraces (ART)
Location of the Queanbeyan-Shoalhaven area
Araluen (Ar) Land System (15 sq.miles)
Hills and rolling country on deeply disaggregated granite with massive
earths and texture-contrast soils and tall woodland,mostly cleared.
Wroxham (Wr) Land System (70 sq,miles)
Rolling country on granite with texture-contrast soils; dry scleorphyll
forest (E,rossii-E,mannifera) or savannah woodland (E,melliodora),
partly cleared.
Advantages of the ISP
provides a landscape-wide perspective
of biophysical resources and general
landuse capability
provides an integrated,ecological
perspective of the landscape; the land
unit is conceptually interesting!
made good use of available technology
Advantages of ISP (cont’d)
…,air -photo interpretation
information was synthesised in a readily
usable format (really a non-digital
Geographic Information System/GIS)
provided a means of placing field survey
data in a geographic context; ie
addressed spatial extension problem
the basic concept of API as the means
of spatial extension is still used in soil,
geological,and vegetation surveys!
Limitations of Integrated Survey:
1,Subjective (not explicit),lack of repeatability
2,Poor predictive capability (assumption of
covariance often invalid); descriptive
3,Primary survey data lost; data classified
4,Land units/elements not mapped
5,Single general purpose landscape classification
6,Not dynamic; what about change/flux?
7,No climate inputs - meso scale
- topo scale
The ‘Parametric Paradigm’
1,Select key environmental determinants that
‘drive’ system response
2,Generate spatially distributed data-sets for
each parameter
3,Apply quantitative data analysis
rejects notion of a single classification for,
in effect,problem-specific classifications
parameter-specific survey and spatial extension
required
Therefore,a modelling approach is required
y = f (X1,X2,X3,…,X n)
y is system response
Xs are driving variables
determinants
predictors
Therefore,if have spatial estimates of Xs,can make
a spatial prediction of y anywhere in the land
Therefore,need Geographic Information System (GIS)
use computer to spatially extend survey data
use computer to store and analyse data
realistic alternative to API and integrated survey
GIS tools to handle landscape-
wide,spatial data
Digital Elevation Models and derived
gridded estimates of terrain attributes
Interpolated Climate Surfaces
Digitized thematic maps,eg,geology
Satellite-based (and aeroplane-based)
digital sensors,eg,Landsat,TM,SPOT,
hyperspectral scanners
GIS tools cont’d
statistical/correlation analysis coupling
field survey data to GIS data base
thereby generating spatial predictions of
potential distributions (and other kinds
of response functions)
GIS makes field data MORE useful! (ie,
it in noway replaces the need for these
data!!)
SREM 3011
LECTURE 2
Dr Brendan Mackey
Department of Geography
The Australian National University
‘Principles of Geographic Information Systems
for Land Resource Assessments’
by Burrough,P.A,(1986/7)
Oxford Science Publications
Monographs on Soils and Resources
Survey no,12
- raster (gridded),point,and vector data types
- grid based data manipulation
- data visualisation (2d,3d)
- scale; resolution
- real and integer data types
- ancillary modelling programs
A fundamental problem!
How do we know what is where?
How can we obtain landscape-wide (ie,
spatially distributed) data about the
ecological (biophysical) resources of a
landscape?
We can only ever take measurements
from a relatively small number of field
survey sites/plots; climate,topography,
vegetation,wildlife,substrate
The ‘Integrated Landscape Paradigm’
1940s,CSIRO Land Use Research
1990s,CSIRO Water and Land
Reconnaissance survey for land use evaluation
- Grazing
- Agriculture
- Forestry- Recreation
- Nature conservation
Surveyed large areas to determine major types of
country and to assess natural resources
Analyse patterns from visual inspection of air photos
- Land units/elements that have a uniform parent
material,topography,soil and vegetation
- Land systems,however,were mapped
>1 occurrence of a land unit
Field survey undertaken to sample land units
So,survey data were spatially extended through
assumption of covariance
ie,assumed all examples of a land
unit sufficiently similar to infer
properties of an unsurveyed unit
from a sampled site
A product of the available technology
- air photo interpretation
- cartography
Basic ‘assumption of covariance’ still used in
survey (soil,geology,vegetation!)
- cartographic products often used as source data
for GIS!
Location of Chimbu Province
Sedimentary spurs north of the Sena River east of Karimur,Landscape
components are scarp (C),spur crest (D),spur slope (A) and footslope (B).
Adjoining this unit are the Sena River terraces (ART)
Location of the Queanbeyan-Shoalhaven area
Araluen (Ar) Land System (15 sq.miles)
Hills and rolling country on deeply disaggregated granite with massive
earths and texture-contrast soils and tall woodland,mostly cleared.
Wroxham (Wr) Land System (70 sq,miles)
Rolling country on granite with texture-contrast soils; dry scleorphyll
forest (E,rossii-E,mannifera) or savannah woodland (E,melliodora),
partly cleared.
Advantages of the ISP
provides a landscape-wide perspective
of biophysical resources and general
landuse capability
provides an integrated,ecological
perspective of the landscape; the land
unit is conceptually interesting!
made good use of available technology
Advantages of ISP (cont’d)
…,air -photo interpretation
information was synthesised in a readily
usable format (really a non-digital
Geographic Information System/GIS)
provided a means of placing field survey
data in a geographic context; ie
addressed spatial extension problem
the basic concept of API as the means
of spatial extension is still used in soil,
geological,and vegetation surveys!
Limitations of Integrated Survey:
1,Subjective (not explicit),lack of repeatability
2,Poor predictive capability (assumption of
covariance often invalid); descriptive
3,Primary survey data lost; data classified
4,Land units/elements not mapped
5,Single general purpose landscape classification
6,Not dynamic; what about change/flux?
7,No climate inputs - meso scale
- topo scale
The ‘Parametric Paradigm’
1,Select key environmental determinants that
‘drive’ system response
2,Generate spatially distributed data-sets for
each parameter
3,Apply quantitative data analysis
rejects notion of a single classification for,
in effect,problem-specific classifications
parameter-specific survey and spatial extension
required
Therefore,a modelling approach is required
y = f (X1,X2,X3,…,X n)
y is system response
Xs are driving variables
determinants
predictors
Therefore,if have spatial estimates of Xs,can make
a spatial prediction of y anywhere in the land
Therefore,need Geographic Information System (GIS)
use computer to spatially extend survey data
use computer to store and analyse data
realistic alternative to API and integrated survey
GIS tools to handle landscape-
wide,spatial data
Digital Elevation Models and derived
gridded estimates of terrain attributes
Interpolated Climate Surfaces
Digitized thematic maps,eg,geology
Satellite-based (and aeroplane-based)
digital sensors,eg,Landsat,TM,SPOT,
hyperspectral scanners
GIS tools cont’d
statistical/correlation analysis coupling
field survey data to GIS data base
thereby generating spatial predictions of
potential distributions (and other kinds
of response functions)
GIS makes field data MORE useful! (ie,
it in noway replaces the need for these
data!!)
