×

You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

Title
Remote sensing for resources development and environmental management
Author
Damen, M. C. J.

295
August
Italy, dur-
1 1 , ¿2 ] such
is accepting
Je can final-
,ng for each
] and [1 1 ,
re not the
f the
:round cover',
which a
Figure 2. Effect of intercropping on the Transformed
Vegetation Index (TVI) as obtained with LANDSAT Multi-
spectral Scanner (MSS) measurements in band 7 (MSS 7:
0.8 to 1.1 pm) and band 5 (MSS 5: 0.6 to 0.7 pm) for
irrigation districts in the region of Mendoza in
Argentina
given fraction, 80% say, of the observed area reach
this stage.
2.3 Intercropping
Intercropping is a characteristing feature of small
holders agricultural land. Sub-pixel combinations of
different crops will give pixel reflectances which
cannot be predicted from the reflectances of the sub
pixel components. One feasible approach is by apply
ing directly the LANDSAT measurements to establish
typical temporal profiles for each main intercropping
pattern.
Such results are given in Figure 2. It can be seen
that the TVI-values for the combination 'vineyard
with olive trees' lie between the TVI-values for
'vineyard' and 'olive trees'. Our opinion is that the
practical aspects of agriculture can be efficiently
translated in terns of LANDSAT measurements by com
bining directly field inquiries on a number of plots
with the LANDSAT measurements applying to these plots.
In the following sections we will illustrate how
this can be done and how to assess the significance
of the observed differences between agricultural
units, as defined in terms of crops and of agricul
tural practices.
3 APPROACH
Differences between crops in growth cycle are better
defined than differences in spectral reflectances. So
multitemporal crop discrimination on the basis of
growth cycle is in principle easier than monotemporal
crop discrimination on the basis of spectral reflec
tances. After having established characteristic val
ues of some vegetation index for each crop and growth
stage, crop mapping is basically done by applying
simple and fast density-slicing procedures.
The required characteristic values can be estab
lished by combining published measurements of spec
tral-temporal crop properties with crop phenology in
a particular area (Azzali 1985a, 1985b, 1986). To es
tablish identification labels for the crops being
cultivated in the Po valley different vegetation in
dices have been combined to give a multi-index multi
temporal crop identification method. The identifica
tion scheme is presented in Figure 3. This scheme has
been established by comparing different selections of
overpass-dates and of index-values. The availability
of a large number of options makes crop discrimina
tion easier. Moreover, as shown in Figure 3, many
crops can be discriminated in the early growing sea
son. This is essential if one intends to apply LAND
SAT TM measurements during the second part of the
growing season to detect crop stress, e.g. due to
water shortage.
The irrigation districts of Mendoza, Argentina are
characterized by a mediterranean - small holders -
type of agriculture. Intercropping is a widespread
occurrence, thus making it impossible to establish
crop identification labels on the basis of literature,
the more so since spectral measurements on vineyards
and olive trees are quite rare. So, the characteristic
vegetation index values have been established by lo
cating a number of ground-control plots in three
LANDSAT-images during the growing season 1984-1985
and then extracting the required reflectance measure
ments from each image. The result is presented in Ta
ble 2, which shows that there is at least one combina
tion of index values and satellite overpass dates
Table 2. Characteristic values of the ratio 1^ =
(MSS 7/MSS 5). 60 and of Io - TVI.100 applying to the
main crops cultivated in Mendoza, Argentina, as ob
tained with LANDSAT MSS measurements
Crop
Dates
28-8
-1984
27-2'
-1985
15-3
-1985
T 1
I 2
X 1
T 2
T 1
h
Onion
69
83
89
87
82
86
Vineyard
73
82
140
98
125
96
Alfalfa
121
96
143
98
158
100
Olive trees
105
93
122
96
96
90
Vineyard with grass
155
100
150
100
107
93
Olive trees with
vineyard
79
85
128
97
114
94
Fruit trees
67
80
113
88
113
88
Table 3. Cultivated area (1980
farming year) as
percentage of the total agricultural
area for the 5
main crops cultivated
in Crande
Bonifica Ferrarese:
A, as estimated by means of the
multi
-temporal multi-
index method; B, according
to ISTAT agricultural
statistical data
Crop
A
B
Winter wheat
20.
5
23
Alfalfa
21.
5
14.5
Corn
9
8
Sugar beet
16
20
Trees
16
12
Rice
3
3