In the context of irrigation, the common perception is that increasing efficiency in agriculture can provide a solution to the water crisis and result in ‘wins’ for all players (Molle and Turral 2004; Seckler et al. 2003). In contrast to the economist’s conceptualisation of efficiency, irrigation efficiency is primarily an engineering concept concerned with the volume of water diverted and consumed (Cai et al. 2001). Engineering interventions in an attempt to ‘save’ water or to ‘reduce losses’ from an irrigation system are frequently said to improve ‘water-use efficiency’. However, substantial confusion surrounds these concepts, despite the fact that they are often used interchangeably. Perry (2007: 373) argues that this confusion has frequently resulted in not only ineffective, but also undesirable, outcomes from technical interventions to ‘improve’ irrigation efficiency.
Perry (2007) traces the development and use of various conceptualisations of efficiency back to the original contribution by Israelson (1950) that came to be known as classical irrigation efficiency. Israelson defined irrigation efficiency as the ratio of the water consumed by crops of a farm or system to the water diverted. Despite later elaboration and development, [5] Israelson’s original definition, based ostensibly on the relationship between water used by the crop and the water diverted, remained the underlying basis for water accounting.
Importantly, the classical concept of irrigation efficiency ignored the potential for return flows and recycling. Later contributions to the debate emphasised the use of ratios or fractions to describe water use and to explicitly consider the impact of return flows (See, for example, Jensen 1993; Willardson 1994; Allen et al. 1996, 1997). According to these definitions, water diverted for irrigation could be divided into the consumed fraction, comprising beneficial consumption (intended purposes including environmental) and non-beneficial consumption (for example, weeds). The remainder was classified as the non-consumed fraction and this comprised two groups — recoverable flows and non-recoverable flows (Perry 2007: 372).
This approach highlights the fact that not all water purportedly ‘lost’ from a particular irrigation district in fact constitutes a loss to the hydrological system as a whole. For instance, take the case where an irrigation district in Victoria presently generates non-consumed flows that are then ‘recovered’ in the form of an environmental use in the River Murray or via extraction by a downstream or groundwater irrigator in South Australia. If actions are taken to reduce the non-consumed fraction in the irrigation district, the net impact of these activities must take into account the redistribution of water away from existing users (say the environmental use or the downstream irrigator). If the intent is to ‘save’ water, it is vital to know whether the ‘losses’ from the irrigation system are in fact losses at all. After all, when water is ‘lost’ from an irrigation district in Victoria it does not go to Mars. Similarly, in an international context, Sakthivadivel and Chawla (2002) expose the flawed reasoning that redirecting seepage losses to cities was seen as the best way to increase supply without impacting existing uses, but the ‘losses’ were found to be already tapped by other users.
The issue of scale of analysis assumes particular importance in this context and further developments in water accounting conceptualised the idea of water balance at the basin level (Molden and Sakthivadivel 1999; Perry 1999; Seckler et al. 2003). In an Australian context, Gyles (2003: 13) demonstrates the ‘illusory’ nature of water savings and argues that this derives from ‘…errors in logic and the inability or reluctance of the promoters to view water flows in a systems context’. Notwithstanding these developments, ‘improvements’ in irrigation efficiency continue to be calculated at farm or irrigation system level without regard for the overall impact on basin balances.
At the global level in the long term, evaporation from water bodies and evapotranspiration from land and vegetation must equal precipitation. However, as soon as the frame of reference is spatially or temporally narrowed, flows across borders become of vital concern (Perry 2007). Similarly, Perry (2007) notes that only where river flows are sufficient to meet demands, can irrigation efficiency be examined in isolation (as is done in classical efficiency). Thus, given the intensified sectoral competition under conditions of severely limited supply, it becomes increasingly important to conceptualise water use at the basin level. From this perspective, distinctions must be made between consumptive uses which remove water from the current hydrological cycle and non-consumptive uses which return the water for potential reuse. Moreover, ‘changing scale draws us from a mere question of cost-effectiveness of water-saving technology into a wider and thornier question of water allocation, rights to extract water and regulation of its use’ ( Molle and Turral 2004: 10).
Adopting a ‘basin-wide’ perspective invokes the ‘water-efficiency paradox’ since when water is used, a substantial part of it is not ‘used up’ but is retained within the hydrological system (Seckler et al. 2003). It is therefore possible for each component part of a water system to exhibit low water-use efficiency but when viewed from the perspective of the system as a whole, it may be quite efficient. This paradox means that there are many instances of purported water ‘savings’ that when analysed further amount to no more than a redistribution. For example, Molden and Sakthdivadivel (1999) illustrate the importance of the scale of analysis in estimations of classical efficiency, citing the example of Egyptian irrigation, which is approximately twice as efficient when measured at a basin level compared to the field level. In simple terms, this arises because water that is ‘lost’ or ‘leaks’ from upstream users is frequently recaptured by downstream users.
Seckler, Molden and Sakthdivadivel (2003: 37) argue that the potential to ‘save’ water is overestimated as the application of a majority of the concepts of water-use efficiency ‘…systematically underestimate the extent of existing efficiency by a very large amount’. Viewed from this perspective, gains to be made have been much overestimated and purported savings merely result in some users being able to increase their usage whilst others downstream face reduced availability. Thus, these interventions result in spatial shifts or reallocation of water rather than ‘savings’ (Molle and Turral 2004). The implication is that local interventions to ‘save’ water are likely to alter the flow regime and impact on other users. In the case of closed basins (defined by Molle and Turral as those with a relatively small amount of uncommitted run-off leaving the basin) with major constraints of water scarcity, gains in local efficiency eventually amount to reallocation. Clearly, the modest flows making their way to the mouth of the Murray and the much-publicised excessive demands for water in the Basin places the Murray-Darling in this category.
The literature contains a number of examples that highlight the fallacy of water savings on a basin level (see, for example, Perry 2007; 2008). Molle and Miranzadeh’s 2004 case study in Central Iran highlights the interconnectedness of water users in a closed basin. They conclude that micro level conservation through canal lining, did not eventuate in the expected water ‘savings’ but ‘only led to having more water spread and depleted locally to the detriment of users downstream’ (Molle and Miranzadeh 2004: 3). Until policy-makers understand that all water that ‘leaks’ from a channel does not automatically constitute a loss to the system as a whole then similar policy disappointments will occur in Australia. Estimates of the quantum of water to be realised by a particular ‘water-saving initiative’ are all-too-often exaggerated because the only water that can really be saved is that portion that enters a saline sink or evaporates and that which is consumed in non-beneficial consumption. Even in the case of the latter (say irrigation water consumed by weeds instead of crops), this water is seldom made available for other users since farmers invariably use the ‘saved’ water to expand production on site.
Molle and Turral (2004) refer to the supposed 1998 ‘win-win’ agreement between Southern California Metropolitan Water Authority and the Imperial Irrigation district. This agreement included the lining of canals and the transfer of usufructuary rights to Los Angeles equivalent to the amount ‘saved’ through this measure. The actual impact of this project, viewed from a basin-wide perspective, was the deterioration in the quality of the recharge to aquifers tapped by farmers on the other side of the border in Mexico (p.4). While the impetus for this type of agreement may be understandable in the context of competing national jurisdictions, it is difficult to discern its logic in an Australian setting.
In short, the purported ‘savings’ that emanate from improved storage or conservation at one point in a basin necessarily diminish that available further downstream (Molle and Turral 2004). Moreover, any analysis of water-use efficiency must take account of the particular context (location of diversions, and so on) lest the analysis become ‘worse than meaningless [causing] wrong decisions to be made economically, hydrologically and ecologically’ (Perry 2007: 369).
[5] Definitional refinements included attention to the concepts of consumptive beneficial use which comprises the quantity of water effectively used to control soil salinity (Jensen 1967) and distribution and application efficiency (Bos and Nugteren 1974; Bos and Nugteren 1982). ‘Distribution efficiency’ is the ratio of the volume of water delivered to the fields to the volume delivered to the distribution system. ‘Application efficiency’ is the ratio of the volume of water needed (and made available) to meet the evapotranspiration needs of crops compared to the volume of water delivered to the fields.