129
sensing system to detect change, and that this limit is related both to the
accuracy with which land cover can be mapped at a point in time and also to
the extent and rate of change on the ground. Four dimensions of change must
be considered:
Change in x- and y- (changes in extent);
Change in t- (rate of change);
Change in z- (degree of change, which might range, for example, from
complete defoliation to qualitative changes, e.g., species composition).
3.
LAND USE VS LAND COVER
The title of the paper addresses land cover specifically, although, in
practice, most applications for land information require data on land use,
rather than land cover. However, land use is rarely distinguishable from land
cover by direct observation alone, either remotely or even in the field. What
can be observed directly is land cover; it may then be possible to infer the
underlying land usage from these observations. The implications of this
generalization are explored in greater depth by Van Gils et al. (1991). An
important consequence is that there is uncertainty in any maps or statistics
derived from remote sensing which purport to describe either land use, or
changes in land use. These uncertainties can only be resolved or eliminated
by access to complementary data sets which provide more direct evidence of
usage. Often, this will require reference to information on land tenure, field
visits or even interview with land owners to establish purpose.
4.
SPATIAL ASPECTS
Spatial factors may influence the capacity to detect change reliably in
two ways. Firstly, it is important to consider the spatial resolution of the
remotely sensed data in relation to the scale of the changes to be observed.
Because of sensor design features, the minimum detectable area in a given
image is determined by a number of factors in addition to the nominal spatial
resolution of the sensor itself (Townshend 1981). It is important to consider
not only the extent of the change that may be expected, but also the degree
of fragmentation of the landscape within which change occurs. Highly
fragmented landscapes may give rise to mixed image pixels, and the
apparent degree of mixing will differ between images, even where there is
no change on the ground.
14. REMOTE SENSING OF LAND COVER AND LAND COVER
CHANGE
sensing system to detect change, and that this limit is related both to the
accuracy with which land cover can be mapped at a point in time and also to
the extent and rate of change on the ground. Four dimensions of change must
be considered:
Change in x- and y- (changes in extent);
Change in t- (rate of change);
Change in z- (degree of change, which might range, for example, from
complete defoliation to qualitative changes, e.g., species composition).
3.
LAND USE VS LAND COVER
The title of the paper addresses land cover specifically, although, in
practice, most applications for land information require data on land use,
rather than land cover. However, land use is rarely distinguishable from land
cover by direct observation alone, either remotely or even in the field. What
can be observed directly is land cover; it may then be possible to infer the
underlying land usage from these observations. The implications of this
generalization are explored in greater depth by Van Gils et al. (1991). An
important consequence is that there is uncertainty in any maps or statistics
derived from remote sensing which purport to describe either land use, or
changes in land use. These uncertainties can only be resolved or eliminated
by access to complementary data sets which provide more direct evidence of
usage. Often, this will require reference to information on land tenure, field
visits or even interview with land owners to establish purpose.
4.
SPATIAL ASPECTS
Spatial factors may influence the capacity to detect change reliably in
two ways. Firstly, it is important to consider the spatial resolution of the
remotely sensed data in relation to the scale of the changes to be observed.
Because of sensor design features, the minimum detectable area in a given
image is determined by a number of factors in addition to the nominal spatial
resolution of the sensor itself (Townshend 1981). It is important to consider
not only the extent of the change that may be expected, but also the degree
of fragmentation of the landscape within which change occurs. Highly
fragmented landscapes may give rise to mixed image pixels, and the
apparent degree of mixing will differ between images, even where there is
no change on the ground.
14. REMOTE SENSING OF LAND COVER AND LAND COVER
CHANGE