Assessing Biophysical Potential for Water Harvesting at National and Global Scale.
Eddy F. De Pauw, Theib Oweis, Jawaad Youssef, and Nseir Bashar. ICARDA, P.O.Box 5466, Aleppo, Syria
Water harvesting is an agricultural practice with a high potential to improve water use productivity in drylands. It covers various techniques to collect rainwater from natural terrains or modified areas and concentrating it for use on smaller sites or cultivated fields to assure economic crop yields. The feasibility of different water harvesting techniques depends on local run-off characteristics, which themselves are determined by rainfall pattern, slope and soil properties. Whereas the practice has demonstrated its potential in terms of yield improvement, sustainable land management and income generation on research sites, a major knowledge gap concerns the identification of those parts of the drylands in which the chances for impact and adoption can be optimal and to which further studies could be targeted. In order to address this issue of the recommendations domain for water harvesting, a global study was undertaken to develop a rapid GIS-based analytical technique to assess suitability for various water harvesting systems. The study was conducted at two levels, at national level using a case study in Syria to develop a methodology, followed by an assessment at the level of the global drylands. At the level of Syria the assessment was undertaken by matching in a Geographic Information System (GIS) environment simple biophysical information, systematically available at country level, to the broad requirements of the specified water harvesting systems. The systems evaluated include 13 micro-catchment systems, based on combinations of 6 techniques and 3 crop groups, and one generalized macro-catchment system. The environmental criteria for suitability were based on expert guidelines for selecting water-harvesting techniques in the drier environments. They included precipitation, slope, soil depth, texture, and salinity, as well as land use/land cover and geological substratum. The dataset included interpolated surfaces of mean annual precipitation, the SRTM digital elevation model, a soil map of Syria, a land use/land cover map of Syria, and a geological map of Syria. The evaluation had two stages: scoring of the land attributes according to the individual criteria, followed by the combination of the individual scores in a multi-criteria evaluation. Fuzzy membership functions were used to evaluate suitability for continuous variables, such as precipitation and slope. For these the boundary between ‘suitable' and ‘unsuitable' forms by nature a continuum. The functions are fully defined by their shape and inflection point positions. Other relevant factors, such as soils, land use or geological materials, could at the national level only be described as qualitative constraints. In addition, for these datasets it is quite normal that the pixels contain mixtures with different properties. When several constraints occur, it becomes very difficult to estimate the total proportion of the pixel that is affected by one constraint or another. Monte-Carlo simulation of sub-pixel constraint overlap indicated that a reasonable approximation of total proportion of a pixel affected by one constraint or another could be obtained by taking the sum of the estimated proportions of the individual constraints. The individual factors were then scored on a common scale and combined through a Maximum Limitation approach as a special case of Boolean overlay. To identify areas suitable for macro-catchment systems, two separate assessments were undertaken, the first one to evaluate suitability to serve as a catchment, and the second one to evaluate suitability as a target area, with the additional constraint that both areas should be within a certain distance of each other. The evaluation for catchment suitability included fuzzy membership function for precipitation and slope, in which the scores were adjusted by taking into consideration the soil hydrological properties. In order to assess potential for water harvesting at the level of the global drylands, global climatic, land cover, soil and terrain databases were used and the method developed in Syria could be retained in essence, although some adaptations were necessary. The most important adaptations at the global level were to reduce the detail of the study to a lower resolution (1km), to adjust the precipitation thresholds for fuzzy scoring in function of the location-specific precipitation- potential evapotranspiration relationships, and to apply the evaluation to a generalized micro-catchment water harvesting system. The results of the study are available in the form of a global map with 1 km resolution and country tables with area summaries.