386
M. Menenti
in a GIS environment. In many cases the remotely sensed data have to be
combined with measurements of flow-rates or irrigation volumes to obtain
indicators of the effectiveness of irrigation (see e.g. Menenti et aI., 1989, 1996). In
other cases several thematic maps (e.g. soils, infrastructure) have to be combined
with the remotely sensed data. This is a wide research field on itself which is
covered extensively in Chap. 4.
17.2.3 Examples of applications
Some examples of applications will be described in this section to clarify the
preceding description of methods, concise by necessity.
Multi-spectral classification. Irrigated area can be detennined in most cases with
simple methods using combinations of spectral images (Colour Plate 17.A). The
spatial resolution of current sensors systems such as TM is sufficient to provide
useful information at the level of individual fields. The spectral contrast between
irrigated and not-irrigated plots is significant. Subtle features may also be
observed, however, and may lead to appreciable errors when numeric algorithms
are applied to detennine the area irrigated.
Multi-temporal classification. Spectral contrast of targets comprising a mixture
of crop and soil is often limited and multi-spectral classification may not perform
adequately. Another solution is to take advantage of phenology to detennine area
under irrigation or specific crops. When dealing with rather large areas (Colour
Plate l7.B) crop phenology leads to a seasonal cycle of spectral reflectances and
therefore of spectral indices. A combination of few observations at well chosen
dates in relation with phenology may suffice to detennine land cover type. The use
of multi-temporal measurements of spectral indices to classify crops and land
cover has been described above (17.2.2 B). Here an example is presented based on
observations of the Soil Adjusted Vegetation Index (SA VI; Huete, 1988):
SAVI= (l+L)(NIR-R)
NIR+R+L
where L= 0.5, NIR= near infrared reflectance and R= red reflectance.
(17.1)
The combination of low SA VI values in April (red component of the image),
high SA VI values in June (green component) and again lower values in August
(blue component) gives irrigated land a characteristic hue which can be used to
map irrigated land and detennine irrigated area.
Crop water stress. Actual evaporation can be estimated as described in Chap. 8.
An example is presented in Colour Plate 17.C where the evaporative fraction
(actual divided by maximum evaporation) calculated with the algorithm SEBAL
(Bastiaanssen, 1995) is presented. Irrigated fields have an evaporative fraction
close to unity, which indicates sufficient water supply, while bare soil has a
significantly lower evaporative fraction. While this is an indication of the
M. Menenti
in a GIS environment. In many cases the remotely sensed data have to be
combined with measurements of flow-rates or irrigation volumes to obtain
indicators of the effectiveness of irrigation (see e.g. Menenti et aI., 1989, 1996). In
other cases several thematic maps (e.g. soils, infrastructure) have to be combined
with the remotely sensed data. This is a wide research field on itself which is
covered extensively in Chap. 4.
17.2.3 Examples of applications
Some examples of applications will be described in this section to clarify the
preceding description of methods, concise by necessity.
Multi-spectral classification. Irrigated area can be detennined in most cases with
simple methods using combinations of spectral images (Colour Plate 17.A). The
spatial resolution of current sensors systems such as TM is sufficient to provide
useful information at the level of individual fields. The spectral contrast between
irrigated and not-irrigated plots is significant. Subtle features may also be
observed, however, and may lead to appreciable errors when numeric algorithms
are applied to detennine the area irrigated.
Multi-temporal classification. Spectral contrast of targets comprising a mixture
of crop and soil is often limited and multi-spectral classification may not perform
adequately. Another solution is to take advantage of phenology to detennine area
under irrigation or specific crops. When dealing with rather large areas (Colour
Plate l7.B) crop phenology leads to a seasonal cycle of spectral reflectances and
therefore of spectral indices. A combination of few observations at well chosen
dates in relation with phenology may suffice to detennine land cover type. The use
of multi-temporal measurements of spectral indices to classify crops and land
cover has been described above (17.2.2 B). Here an example is presented based on
observations of the Soil Adjusted Vegetation Index (SA VI; Huete, 1988):
SAVI= (l+L)(NIR-R)
NIR+R+L
where L= 0.5, NIR= near infrared reflectance and R= red reflectance.
(17.1)
The combination of low SA VI values in April (red component of the image),
high SA VI values in June (green component) and again lower values in August
(blue component) gives irrigated land a characteristic hue which can be used to
map irrigated land and detennine irrigated area.
Crop water stress. Actual evaporation can be estimated as described in Chap. 8.
An example is presented in Colour Plate 17.C where the evaporative fraction
(actual divided by maximum evaporation) calculated with the algorithm SEBAL
(Bastiaanssen, 1995) is presented. Irrigated fields have an evaporative fraction
close to unity, which indicates sufficient water supply, while bare soil has a
significantly lower evaporative fraction. While this is an indication of the