299
TVI-values;
ation of plot
= signifi-
sampies
amples in
amples in-
IA3
IA4
hs
hs
hs
hs
hs
hs
s
hs
”
ns
hs
hs
hs
hs
hs
hs
ns
ns
-
ns
hs
hs
hs
hs
hs
hs
hs
hs
“
ns
From our point
ay of comparing
which one
ences in TVI
B we see that
lot, the MSS
e difference
MSS measure-
It is only
given field
better sig-
3 is 'highly
ifferent points
at the per-
int of view,
however, true
y, one can
e 23,305 m^ of
nd TM 7
p-specific
shed.
IGNIFICANT
;nificant dif-
ed the means
of a large
'ith the MSS
.%, significant
:se percentages
irge relative
ir these differ-
mt or signifi
cant after the proper significance test.
Let us now take this 4% as an approximate indica
tion of the relative change in TVI which will be
assessed as not significant. We have then to estab
lish whether we can measure TVI with this accuracy.
The nominal radiometric accuracy of TM measurements
is 0.5% on reflectance. So we can calculate the
nominal accuracy on TVI by means of error analysis
of the formula defining this vegetation index. This
results in a 0.2% nominal accuracy.
We can, therefore, conclude that we can in prin
ciple measure significant differences in TVI-values
with satisfactory accuracy.
Actual radiometric accuracy of TM is, however,
quite different from nominal, If, for example, we
take the results of Slater (1986), accuracy on re
flectances is 0.5 in TM 3 and 1.7 in TM 4 at visibil
ity greater than 10 km. Accuracy on TVI will there
fore be 0.2% which confirms the nominal value.
It must finally be stressed that the multitemporal
analysis of satellite images requires that the basic
digital count values be transformed into reflectance
values (Menenti 1984, Kuipers and Menenti 1986).
Furthermore measurements on different overpass dates
have to be corrected for atmospheric effects or at
least normalized with each other by taking one of
the images as a reference.
metrization of land-surface characteristics, use of
satellite data in climate studies and first results
of ISLSCP. ESA SP-248.
Menenti, M. 1984. Physical aspects and determination
of evaporation in deserts applying remote sensing
techniques. Report ns 10 (special issue). Institute
for Land and Water Management Research (ICW), Wa-
geningen. 202 pp.
Miller, G.P., M. Fuchs, M.J. Hall, G. Asrar, E.T.
Kanemasu & D.E. Johnson 1984. Analysis of seasonal
multispectral reflectances of small grains. Remote
sensing of environment 14:153-167.
Slater, P.N. 1986. Variations in in-flight absolute
radiometric calibration. Proc. ISLSCP-conference on
parametrization of land-surface characteristics,
use of satellite data in climate studies and first
results of ISLSCP. ESA SP-248.
Steel, R.G. & J.H. Torrie 1960. Principles and
procedures of statistics. New York, Me Graw-Hill.
481 pp.
6 CONCLUSIONS
We conclude that differences in temporal and spectral
patterns relating to crop phenology and intercropping
can be obtained from LANDSAT measurements. Plots of
0.5 ha and 2 ha must be considered when applying TM
and MSS measurements respectively. It is demonstrated
that a multi-index multi-temporal crop discrimination
scheme can be established for three rather different
agricultural areas. Operational crop monitoring by
satellites in a given region, however, is only
feasible after such pre-operational investigations,
as presented in this paper.
REFERENCES
Anderson, J.W. 1986. Remote sensing finds down-to-
earth applications. Commercial Space.
Azzali, S. 1985a. A multi-index multi-temporal ap
proach to map crops in the early growing season.
An application to two Italian irrigation districts:
East Sesia and Grande Bonifica Ferrarese. Nota
1611. Institute for Land and Water Management Re
search (ICW), Wageningen. 39 pp.
Azzali, S. 1985b. Greenness and brightness formulae
applicable to East Sesia and Grande Bonifica Fer
rarese. Nota 1673. Institute for Land and Water
Management Research (ICW), Wageningen. 32 pp.
Azzali, S. 1986. Matching the analysis of LANDSAT
data to users's requirements. Nota 1711. Institute
for Land and Water Management Research (ICW), Wa
geningen. 15 pp.
Crist, E.P. 1984. Effects of cultural and environ
mental factors on corn and soybean spectral devel
opment patterns. Remote sensing of environment 14:
3-13.
Hinzman, L.D., M.E. Bauer & C.S.T. Daughtry 1984.
Growth and reflectance characteristics of winter
wheat canopies. LARS Technical report 111484. 18 po.
Jackson, R.D., P.N. Slater & P.J. Pinter 1983. Dis
crimination of growth and water stress in wheat by
various vegetation indices through clear and turbid
atmospheres. Remote sensing of environment 13:187-
208.
Kenney, J.F. & E.S. Keeping 1959. Mathematics of
statistics. Part 2. Toronto, D. van Nostrand. 429
pp.
Kuipers, H. & M. Menenti 1986. Groundwater--fed lakes
in the Libyan desert as observed by means of
LANDSAT MSS data. Proc. ISLSCP-conference on para-