The authors managed well to address most of the reviewer comments. The revised version of the manuscript has improved considerably in comparison to the version submitted previously. The approach presented by the authors to map irrigated land differs from attempts published before. There is certainly enough merit and innovation so that I can recommend publication if the authors consider the following additional aspects:
1) The discussion section and the conclusion should address another reason of differences between the irrigated land mapped by the authors and irrigated land reported by agricultural statistics, FAO or other products: use of different terms and definitions. FAO and Siebert et al., 2013 map the so called area equipped for irrigation according to FAO's definition and categorization of specific water management types. According to the definition used by FAO and Siebert et al. (2013) areas are considered as equipped for irrigation when two main criteria are fulfilled:
a) presence of a permanent infrastructure for water supply and water distribution
b) use of water from external sources for irrigation (but not in-situ water concentration).
Both criteria are difficult to detect by remote sensing so that, in general, a comparison between remote sensing based irrigated area and FAO statistics has the problem of comparing different things (apples with pears). For example, almost half of the paddy rice areas worldwide use in-situ water kept on the field by retention of rainwater by earth dams. These areas are classified as rainfed wetlands and do not belong to the category of areas equipped for irrigation. This difference in the water source is difficult to detect by remote sensing and often results in a major difference in irrigated areas estimates for humid regions. Therefore, global remote sensing based approaches often report more irrigation than FAO for paddy rice areas.
Other practices that are classified by FAO as being rainfed are:
- water harvesting,
- water concentration,
- flood water recession with non-permanent structures,
- using water from non-permanent dug wells.
These practices are in particular important in arid and semi-arid regions and the areas watered with these practices are larger for many countries than the areas watered with formal irrigation, for example for most African countries. This can as well explain differences between the product presented by the authors and FAO-based statistics or products.
According to my own experience, the difference caused by the use of different terms and definitions are in general larger than differences caused by varying precision of the mapping approach. Therefore it is very important to add this aspect to the discussion.
Sources explaining FAO's definition of irrigated land:
http://www.fao.org/3/a-i3289e.pdf
http://www.fao.org/docrep/016/i2809e/i2809e.pdf
http://www.fao.org/3/a-i0936e.pdf |