17 Irrigation and Drainage
389
17.3.2 Higb resolution mapping of irrigated lands
Crop identification and mapping. The feasibility of a remote sensing method to
determine quantity, type and water stress of crops by means of Landsat-TM images
was demonstrated by Azzali and Menenti (1989). Individual crops in two irrigation
districts in the Po valley, Italy were mapped with Landsat-TM multitemporal data
using a combination of spectral indices. An accurate crop calendar had to be
established to perform multi-temporal analysis. Reflectance measurements in
TM-bands 3, 4, 5 and 7 were a more reliable indicator of water shortages than the
radiometric surface temperature (obtained with TM-band 6 radiances). Three
practical applications were identified: (1) estimation of crop area; (2) crop water
stress; (3) estimation of water depth in rice fields.
Irrigated area. Determination of irrigated area is a mature RS application as
proven by several studies in different contexts and countries. Verdin et al. (1988)
used SPOT data to monitor irrigated lands in a mid latitude, high desert
enviromnent in the Newlands Project in W. Central Nevada. Merged 10- and
20-metre data sets were used to prepare 1 :24 000 photographic prints, which
proved suitable for delineating irrigated fields as small as 4 ha. Brightness and
Greeness transformations for XS (multispectral) data were derived and used for a
maximum likelihood classification of 8 crop types: Medicago sativa, pasture,
fallow, maize, winter cereals, Sorghum sudanense, feed grain and vegetables. An
accuracy of 91 % was obtained after application of a per-field, majority-rule
revision. Thelin and Heimes (1987) mapped irrigated cropland with Landsat Multi
Spectral Scanner data to determine use of groundwater from the High Plains
Aquifer in parts of Colorado, Kansas, Nebraska, New Mexico, Oklahoma, South
Dakota, Texas, and Wyoming. Estimates of irrigated area were combined with
irrigation application rates (from field data) to compute estimates of irrigation
water use. Carton (1991) mapped irrigation systems in Algeria using SPOT data.
The pilot study was made to develop appropriate methods for a nationwide
inventory. Two SPOT images of two very different agroclimatic zones were used:
(1) a mountainous area with relatively small irrigation systems using water pumped
from wadis; and (2) a much more arid area where the irrigation systems are
supplied from boreholes or by flooding. Digital processing of images taken in June
combined with field observations gave the following results: detailed location of
irrigation systems above 0.25 ha; a map of the system at a scale of 1 :50 000;
estimates of irrigated area and location of systems in the subwatersheds. El Kady
and Mack (1992) concluded that inventories of irrigated lands in Egypt might be
improved through an operational use of satellite data to supplement aerial
photography. Growth in irrigated area reacts to market demand. This often occurs
against regulations as for example shown by Cuevas and Gonzalez (1992) using
Landsat MSS image data. Irrigated area inside the National Park of Dofiana
(Spain) increased by 500% from 1978 to 1988 and was lower in 1988 (a fairly wet
year) than in 1989 (dry year).
389
17.3.2 Higb resolution mapping of irrigated lands
Crop identification and mapping. The feasibility of a remote sensing method to
determine quantity, type and water stress of crops by means of Landsat-TM images
was demonstrated by Azzali and Menenti (1989). Individual crops in two irrigation
districts in the Po valley, Italy were mapped with Landsat-TM multitemporal data
using a combination of spectral indices. An accurate crop calendar had to be
established to perform multi-temporal analysis. Reflectance measurements in
TM-bands 3, 4, 5 and 7 were a more reliable indicator of water shortages than the
radiometric surface temperature (obtained with TM-band 6 radiances). Three
practical applications were identified: (1) estimation of crop area; (2) crop water
stress; (3) estimation of water depth in rice fields.
Irrigated area. Determination of irrigated area is a mature RS application as
proven by several studies in different contexts and countries. Verdin et al. (1988)
used SPOT data to monitor irrigated lands in a mid latitude, high desert
enviromnent in the Newlands Project in W. Central Nevada. Merged 10- and
20-metre data sets were used to prepare 1 :24 000 photographic prints, which
proved suitable for delineating irrigated fields as small as 4 ha. Brightness and
Greeness transformations for XS (multispectral) data were derived and used for a
maximum likelihood classification of 8 crop types: Medicago sativa, pasture,
fallow, maize, winter cereals, Sorghum sudanense, feed grain and vegetables. An
accuracy of 91 % was obtained after application of a per-field, majority-rule
revision. Thelin and Heimes (1987) mapped irrigated cropland with Landsat Multi
Spectral Scanner data to determine use of groundwater from the High Plains
Aquifer in parts of Colorado, Kansas, Nebraska, New Mexico, Oklahoma, South
Dakota, Texas, and Wyoming. Estimates of irrigated area were combined with
irrigation application rates (from field data) to compute estimates of irrigation
water use. Carton (1991) mapped irrigation systems in Algeria using SPOT data.
The pilot study was made to develop appropriate methods for a nationwide
inventory. Two SPOT images of two very different agroclimatic zones were used:
(1) a mountainous area with relatively small irrigation systems using water pumped
from wadis; and (2) a much more arid area where the irrigation systems are
supplied from boreholes or by flooding. Digital processing of images taken in June
combined with field observations gave the following results: detailed location of
irrigation systems above 0.25 ha; a map of the system at a scale of 1 :50 000;
estimates of irrigated area and location of systems in the subwatersheds. El Kady
and Mack (1992) concluded that inventories of irrigated lands in Egypt might be
improved through an operational use of satellite data to supplement aerial
photography. Growth in irrigated area reacts to market demand. This often occurs
against regulations as for example shown by Cuevas and Gonzalez (1992) using
Landsat MSS image data. Irrigated area inside the National Park of Dofiana
(Spain) increased by 500% from 1978 to 1988 and was lower in 1988 (a fairly wet
year) than in 1989 (dry year).