Today we have actual flow data for 42 years (1948-1989), which is sufficient to calculate yield without hindcasting [CWC, 1992]. The yield from 42 years of actual data is only 22.69 MAF, and not 27.22 MAF. It is scientifically much more appropriate to use actual data rather than hindcasted data. Project authorities persist in hindcasting even though the exercise is redundant because the lower yield shown by the actual data would render their project unviable as it stands.
The use of hindcasted data has led to overestimation of yield in several projects, such as the Gandhi Sagar Project, where the reservoir has completely filled up only once in 32 years of operation, leading to net irrigated area being 35% lower than that planned. In the case of the SSP, hindcasting has resulted in mistaken estimates of river flow because it is based on inadequate rainfall data. The Khosla Commission, which was appointed to go into the Narmada water disputes question in the early 1960s, stated that the number of rain gauges available from 1891-1930 was too few to allow accurate estimation of data. The World Bank also stated that the number of rain gauges before 1925 was insufficient to allow hindcasting [WB 1985]. Thus, only 18 years of rainfall data (1930-1947) are suitable for hindcasting. In other words, the Narmada Tribunal was technically wrong when it used hindcasted data from 1891 onwards.
The decreased water yield changes the entire complexion of the project, and throws doubt on the scope and possibility of the claimed benefits. The NWDTA distributed water from the Narmada in the ratio 65:32 to Madhya Pradesh and Gujarat, with Rajasthan and Maharashtra sharing the remaining 3%. Thus, while under the original estimate of river yield at 27.22 MAF Gujarat would recieve 9 MAF of water, under the revised actual yield of 22.69 MAF, its share would drop to 7.26 MAF. This naturally means less water to go around: less water for irrigation, less water for drinking. The reduced quantum of water is specially significant for those at the tail-end of the SSP system, i.e. Kutch and Saurashtra, since they are the most likely to suffer if there is less water.
As the entire planning for the SSP is based on an assumption of using 9 MAF of water, any decrease in this amount will result in decreased benefits. Indeed, the benefit-cost ratio of the project, which at 1.13 is already at marginal acceptability according to Government of India and World Bank norms, will decrease below acceptable limits if the smaller quantum of assured water is taken into account.