Geoinformation, monitoring and assessment Environment

Updated December 1998

Remote Sensing Centre Series

Use of high-resolution satellite data for irrigation management and monitoring. Pilot study in Indonesia

by Messrs D. Lepoutre, A. Royer, D. Lantieri
67 pp, 22 figures, 3 plates, 6 tables
RSC Series No. 57, FAO, Rome 1991.


The objective of the pilot study using SPOT remote sensing techniques, carried out in Indonesia from July 1988 to October 1989, was to evaluate the possibilities of SPOT high-resolution satellite data for improving the management of Indonesian irrigation schemes.

The selected test site was the West Tarum area, the western section of the 240,000 ha Jatiluhur irrigation system, which is located east of Jakarta, Java. Much information was already available on these very densely populated areas. Rather than to provide additional data, preliminary analysis of the requirements indicated that SPOT should be used to collect more precise information, and to check and correct existing information.

The usefulness of four SPOT-derived data products was demonstrated in this pilot study:

An infrastructure map (product 1) updated urban areas and network, to secondary and in several cases tertiary level. This map was produced on a scale of 1:60,000 from computer-assisted interpretation of SPOT panchromatic PA taken during the dry season and digitally enhanced (textural filter). Complementary information was also extracted in certain cases from a colour composite of two multispectral vegetation index images taken at the beginning and the end of the cropping season. This infrastructure map, although limited to main features, appears particularly useful in fast-changing irrigation schemes, where it can be associated with management information obtained from farmers or cooperatives;

An updated tertiary unit map (product 2) was evaluated and compared to the actual 1:5,000 tertiary unit plan, showing tertiary and quarternary drains and channels. Most results were obtained from visual analysis of a colour composite made of SPOT digitally enhanced panchromatic channels taken during the dry season at different years. Only the biggest features of the infrastructure at tertiary level and, in particular, the limits of tertiary units, could be identified and updated;

An estimate of the acreage of irrigable land (product 3) of tertiary unit groups was extracted by computer-assisted photointerpretation of the first two products. Acreage estimates were compared with those given by the irrigation scheme authorities based on the available traditional network map. Discrepancies between these two types of data were found to be between -9% and +33% according to the group of tertiary units. A calculation based on the result of the pilot area showed that water planning and harvest forecasting were overestimated by more than 9,250 ha in a total of 43,681 ha;

An irrigation management map (product 4) was made from visual analysis of a colour composite of three vegetation index images taken on three different dates, in December (first part of the rainy season), June (first part of the dry season) and August (second part of the dry season). Each of the eight colours resulting from the three-band combination, namely red, green, blue, yellow, cyan, magenta, white and black, was associated with a presumed rice development stage (active or inactive). The link between the colour of the image and the rice development could be made from a diagram in which different hypotheses of the rice cropping calendar were compared with the dates when SPOT images were taken.

Analysis of this colour composite allowed mapping of rice development and therefore water distribution through the year, on a scale between 1:50,000 and 1:100,000. It was found that irrigation was not managed as it should have been or was claimed to be. Great heterogeneity in the water supply was noticed within the official irrigation groups or "gologans". Thus, it seemed that the supply of water depended more on the distance from the main primary canal than on any plan. Moreover, some problematic areas where apparently no rice could be grown, could be distinguished. Reasons for this failure (flooding, lack of water) would have to be checked on the ground.

From the technical point of view, it was demonstrated that visual analysis assisted by computer, using interactive enhancement or vector raster utilities, are most efficient. Best results for infrastructure identification were obtained from analysis of SPOT panchromatic imagery taken during the dry season and for monitoring irrigation management, the SPOT multispectral vegetation index and multidate data. Although very useful, high-resolution SPOT data present some limitations for analysis of an irrigation scheme like Jatiluhur; 20-metre or even 10-metre resolution is still not sufficient to analyse very small fields (cinder 0.3 ha) individually or to identify small features of the infrastructure at the quaternary or in a number of cases, tertiary level. Moreover, the almost permanent cloud cover during the rainy season did not allow the acquisition of imagery in spite of many attempts. Such imagery could have been very useful to confirm hypotheses made for interpretation of the irrigation management map.

Evaluation of the operational use of SPOT data was made in a cost benefit study. in which were compared the cost and benefit of information derived from traditional means such as aerial photographs and intensive ground survey, and from SPOT data. Two types of products were assessed: a) infrastructure maps and irrigable land statistics (which correspond to products 1, 2 and 3 above), that rely mostly on SPOT PA analysis; and b) irrigation monitoring (product 4 above), by means of multitemporal SPOT XS colour composite.

From the cost point of view, both infrastructure and irrigation monitoring products were found to be three to four times cheaper when made from SPOT data rather than from aerial photos. Depending on the type of consultancy used (national or international), cost of SPOT-derived product ranged from US$ 8.2/km2 (national consultancy) to US$12.7/km2 (international consultancy for the irrigation monitoring map, and from US$12.25/km2 (national consultancy) to US$25.25/km2 (international consultancy) for the infrastructure maps and statistics. Cost calculations were based on unit rates applied in FAO projects.

Benefit or advantages were essentially assessed on the accuracy of information provided on SPOT and by aerial photographs. It appears that aerial photos are more precise (between 95 percent and 98 percent accurate). However, SPOT data remain very satisfactory in most cases (average accuracy between 75 percent and 90 percent according to themes) except on mapping of small features.

The study concluded that SPOT-derived data are very cost-effective in most applications related to the updating and monitoring of small-scale rice irrigation schemes. They would, however, have to be complemented by an aerial survey every 10 to 15 years to map the whole tertiary and quarternary network.

The overall cost of a fail remote-sensing and mapping programme which includes acquisition and analysis of aerial photos, SPOT data, production of maps and statistics described above, report preparation, technology transfer (training, international supervision, acquisition or upgrading of local equipment) would be about US$18/km2/Year, at the condition the project is implemented in the country.

This cost appears very reasonable, even minor when compared with the cost of implementing an irrigation scheme (US$300,OOO/km2) or operating and maintaining it (US$5,OOO/km2/year in tropical wet areas); or compared with the price of rice production (average US$45,OOO/km2/year). An increase of only 0.0005 in the rice production thanks to better irrigation management due to remote sensing could economically justify the cost of using these techniques. Results of the pilot study showed that even more could be expected.

The following recommendations are made:

The above publication is available from:
Viale delle Terme di Caracalla
00100 Rome, Italy

Back to Top FAO Homepage