International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004
(around week 19), perhaps due to lower temperatures in these
higher elevation areas. An abrupt decline of NDVI sets in
between weeks 36 and 39 for all cover types except arable land
(around week 22-27) and pasture in years 1995 and 1998 (week
28/29).
It is apparent that each vegetation type shows a characteristic
NDVI curve.
4.2.2 Comparison between years in each cover type: By
comparing the summarised polygons of each vegetation cover
with each other, it can generally be said that the year 1995 in all
the cover types is the one with the lowest NDVI, generally
around 0.1 lower, which is most apparent in the summer.
Figure 7 shows exemplary the time-series of 1995 to 1998 for
forest cover types (a) and Highland vegetation (b). There is also
a consistent order to be seen in the magnitude of the NDVI
curves, especially in summer and spring which has been
described above for different cover types. 1997 and 1998 have
similar and always the highest NDVI values compared to 1995
and 1996. 1996 has, like 1995, low spring values but rises
sharply to nearly reach the magnitude of the 1997 and 1998
NDVI response. This is therefore a trend that can be seen in all
cover types.
Forest covers 1995-98
0,7 1
|
08 / X |
^ |
f^ fX e]
/ - ] JN +
05 rA { \ |
/ J/ |
^
jl
k JS \ |
s 1; |
a ir 7 |
z j 77
034 P \ |
|
02 4 /
ed |
v |
04 |
|
0
0 50 Weeks sii. Jan 1 1995 150 Su
—— semi coniferous forest mainly coniferous forest
——— broadleaf forest — coniferous forest plantation, full canopy cover
(a)
Highland vegetation 1995-98
07 —
0,5
04 4 f, \
5 74
2 P
02
0,1
0
0 so 100 150 200
Weeks since Jan 1 1995
8
—— montane vegetation
—— Nardus and Molinia ( b )
—— blanket bog / peat land
Figure 7: Time-series (1995-1998) of forest covers (a) and
Highland vegetation (b)
wel heather moorland
5. CONCLUSION
This project has shown the utility of 1.1 km AVHRR NDVI
data to monitor vegetation dynamics in a northern temperate
climate using Scotland as an example. The probability of cloud
covered imagery is high in these high latitudes which means
limited datasets. Despite these cloud cover problems it is still
possible to derive meaningful NDVI curves that allow for the
detection of change in vegetation.
It has been demonstrated that the time-series approach over
several years is useful for understanding scasonal dynamics of
different vegetation types, as well as allowing inter-annual
comparisons of cover types. By taking a continuous series of
images with a short compositing period of one week, the NDVI
curves of a cover type between years can be observed in detail.
This method of change detection is straightforward and
consumes comparatively little computational time. Vegetation
can be monitored continuously and, over a longer time period,
changes can be followed.
However, it has also become evident that pre-processing and
accurate calibration. of the AVHRR imagery is extremely
important. Because of different calibration coefficients used in
1995 compared to years 1996 to 1998, care has to be taken
when comparing that year to the others in terms of NDVI
magnitude.
REFERENCES
DeFries, R.S. et al 1995. Global discrimination of land cover
types from metrics derived from AVHRR Pathfinder data.
Remote Sensing of Environment 54, pp. 209-222.
Duchemin, B., Guyon, D., Lagouarde, J.P. 1999. Potential
limits of NOAA-AVHRR temporal composite data for
phenology and water stress monitoring of temperate forest
ecosystems. International Journal of Remote Sensing 20 (5),
pp. 895-917.
European Centre For Nature Conservation (ECNC) 1992.
Council directive 92/43/EEC of 21 May 1992 on the
conservation of natural habitats and of wild fauna and flora.
http://www.ecne.nl/doc/europe/legislaV/habidire.html (accessed
July 2001).
Lambin, E.F., Ehrlich, D. 1997. Land-cover changes in sub-
Saharan Africa (1982-1991). Remote Sensing of Environment
61, pp.181-200.
Lloyd, D. 1990. A phenological classification of terrestrial
vegetation cover using shortwave vegetation index imagery.
International Journal of Remote Sensing 11 (12), pp. 2269-
2279.
Macaulay Land Use Research Institute (MLURI) (1998):
Natural Heritage management — Vegetation Dynamics. Annual
report 1998. MLURI, Aberdeen.
http://www.mluri.sari.ac.uk/ar98cont.htm (accessed July 2001).
Malingreau, J.-P. 1986. Global vegetation dynamics: satellite
observations over Asia. International Journal of Remote
Sensing 7 (9), pp. 1121-1146.
Mather, P.M. (1999): Computer processing of remotely-sensed
| puter | 8o
images — An Introduction. 2" edition. John Wiley & Sons,
Chichester.
Reed, B.C., Brown, J.E., Vander Zee, D., Loveland, T.R.,
Merchant, J.W., Ohlen, D.O. (1994): Measuring phenological
variability from satellite imagery. Journal of Vegetation Science
5. pp. 703-714.
KEY
ABS'
Befoi
impo
togetl
to ha
devel
Deve
and v
regioi
secto!
UAE
varioi
estab]
natioi
The a
all lif
in act
LI]
Being
image
under
we de
differ
This 3
image
purpo
1.2 Lc
UAE
600 ki
east it
about
censu:
The se
steep 1
Alain
the ca
The w
summ
9% We
