Improving spatial resolution of agricultural water use estimation using ALOS AVNIR-2 imagery

Thomas K. Alexandridis(1), Yann Chemin(2), Ines Cherif(1), George Tsakoumis(1), George Galanis(1), George Zalidis(1) and Nikolaos Silleos(1)

(1) Aristotle University of Thessaloniki, University Box 259, University Campus, 54124 Thessaloniki, Greece
(2) International Rice Research Institute, Metro, Manila, Philippines


Agricultural water use is estimated to exceed 70% of freshwater consumption in the Mediterranean ( Although precise information is necessary for water managers, gross measurements are only available for large regions, and rough estimations for areas without organized irrigation systems. As a result, water savings in the Mediterranean are in preliminary phase only. At the same time the European Water Framework Directive 2000/60/EC (WFD) states that Member States shall estimate and recover the costs of water services, including environmental and resource costs, and in accordance with the polluter pays principle.

Estimations of crop water demand can be different from real agricultural water use. Advanced algorithms have been developed, which couple Earth Observation (EO) data in thermal, visible and near-infrared wavelengths with a few meteorological measurements, and provide the real water use, by solving the surface energy balance equation. The advantage of these methods is their independence from site and crop specific coefficients, and the real spatial distribution of all parameters. However, due to the detailed cropping pattern of the Greek agricultural landscape, higher resolution satellite imagery is necessary, a demand which is not met by the available thermal infrared imagery (Landsat 5 TM and Terra/ASTER).

The aim of this study is to use ALOS AVNIR-2 imagery to improve the spatial distribution of estimated agricultural water use in a Greek river basin. Specific objectives are (i) to compare evapotranspiration estimates from methods based on thermal infrared imagery (such as SEBAL) with those based on AVNIR-2 imagery (such as the modified Penman-Monteith), and (ii) to merge the two datasets for improved results.

Comparisons were based on two levels: (a) on a pixel level, with a random dataset for pixel to pixel comparison; (b) on a canal command area level, with statistical comparison of the water use estimations. Merging of the two datasets was based on the modified Brovey technique, which has the advantage of maintaining the original values, but redistributes them in a higher resolution space. Concluding, ALOS AVNIR-2 imagery was successful in providing the element of high spatial detail, proving to be a useful input in modelling water use of irrigated Greek basins.

ALOS data were provided by the European Space Agency (ESA). The work presented was supported by the extension of ESA's GSE-Land project in Greece.


Symposium presentation


  Higher level                 Last modified: 07.05.06