Geography Reference
In-Depth Information
Such work can be illustrated by recent work
undertaken by Sannier et al . (1998). These workers
have used NOAA-AVHRR data for real-time
vegetation monitoring as a tool for wildlife
management and food security assessment in
southern Africa.
NDVI is an indicator of the level of
photosynthetic activity in crops and other
vegetation. Using data from the FAO ARTMIS
NDVI archive, Sannier et al . (1998) processed the
information to construct the statistical distribution
of NDVI values based on each ten-day period of
the year; data for ten years were used to construct
the distributions. Using these data, it was possible
to develop a 'template' that indicated the
thresholds between years with very low, low,
average, high and very high productivity. Such
templates describe the seasonal dynamics of the
ecosystem. Sannier et al . suggest that the templates
can be used to plot the trajectory of productivity
for an area for any one year to determine whether
it represented an extreme event that may be
significant for people or wildlife. There is now
widespread interest in the application of AVHRR
data to give early warning of poor harvests, so that
steps can be taken to secure food supplies in areas
at risk.
predictions about the types of change that occur
within environmental systems.
In the late 1980s and early 1990s, for
example, the impact of commercial forestry on
the ecology of wading bird populations of the
Flow Country of Caithness and Sutherland in
Scotland was a matter of concern.
Unfortunately, large areas of forests had been
planted before the scale of the potential impacts
were recognised, so GIS and remote-sensing
techniques were used to try to gain an insight
into what had happened. The approach
employed a habitat suitability model to evaluate
the quality of the land that had been lost to
forestry and to make some judgement about the
importance of the loss.
The Flow Country is a large area of upland
peat. The flat, open landscapes are characterised by
the presence of small pool systems, which make
these areas particularly attractive to wading birds.
In fact, the Flow Country is one of the most
important breeding areas in Europe for wading
birds such as the dunlin, golden plover and green
shanks. The fear over commercial forestry in the
late 1980s was that the ecological value of these
areas would be undermined by habitat loss and
disruption of the peat hydrology.
A simple ecological model was used to
estimate the quality of dunlin habitat that had
been lost to forestry (Avery and Haines-Young
1989; Lavers et al . 1996; Lavers and Haines-Young
1997). The model was refined in several stages
and was eventually based on the integration of
remotely sensed satellite images with ground-
based measurements of the distribution of pools
across the peat surface. The first stage in the
analysis was to calibrate the model with sites at
which breeding densities of dunlin had been
estimated by field survey (Figure 40.7). The
second stage was to test the model against
another set of sites to see if the model correctly
predicted the numbers of birds observed there.
Once the model had been corroborated, GIS
could be used to apply the model to the
landscapes of the Flow Country as a whole to
map the variations in habitat quality across the
area. The study used archive satellite data from
Modelling change
When seeking to understand the nature of
environmental change, we are fortunate if we can
measure directly the particular parameter of
interest, as in the case of land cover. In some
situations, particular environmental characteristics
cannot be assessed or measured directly. Instead,
we have to use a surrogate measure and infer what
might be happening to some underlying variable
that really interests us. In other words, we may have
to model changes using our GIS.
Indices like NDVI are examples of such models.
The index is not a direct measurement of
photosynthetic activity, but it is hypothesised that
it is highly correlated with such activity. Thus
NDVI can be used to predict productivity. There
are many other examples in the literature that
illustrate how GIS can be used to model and make
Search WWH ::




Custom Search