Really real ‘real water savings’

Which of two irrigation models (ISA and REWAS) accurately calculate a decrease in aggregate (net) depletion across an irrigated basin system (known as real water savings) as compared to a decrease in water withdrawals or field applications (so called paper savings)?

Before I answer that, let me ask another question. Do the authors of these models claim authority and accuracy in this objective? The answer for both ISA and REWAS is a rather obvious yes. This claim-making is done via several means; how is the method explained in words, diagrams and a workable Excel model; who is involved (and how much irrigation knowledge do they have); how is the model published (via a report or peer-reviewed paper); what does the background text argue (both ISA and REWAS warn of the myths when thinking about irrigation efficiency); and who is backing the model financially, institutionally and via a wider epistemic community.

Hence the title of my blog. We have ‘real water savings’ as the mathematical puzzle, and we have the ‘really real’ push by authors persuading readers that their model is, well,… really real.

When comparing ISA and REWAS, I am not going to refer to their background reports or papers. I’m going to directly compare their respective Excel models. The version of REWAS is REWAS_v9_8 which I downloaded from FutureWater on 26 June 2023 prior to sending my paper to the journal. The version of ISA is from 25 June 2023 submitted to the Journal Agricultural Water Management (see Appendix B).

I acknowledge that the two models are different, for example in their understanding of agro-hydrology, the design of their Excel models, and how some parts (such as the change in storage) are treated. These approaches do not compromise the models’ objectives if fully explained, and can be seen in ‘what-if’ selection of input variables.

Instead, my comparison asks a single test question: Do the models compute changes in aggregate water depletion at the basin level? This is important because changes at the field level, the farm/system level and the basin level are three different things. If we only look at the first two, we are not divining real water savings. This is because recovered flows from an irrigation system can feed another irrigation area downstream, and because over time an irrigation system can grow its area and depletion. In other words, real water savings (a reduction in aggregate depletion) can only be accurately derived at the basin level not the farm/system level.

For my test, look at the 10 questions below. For each one, the answer is ‘no’ for REWAS and ‘yes’ for ISA.

Do the ISA and REWAS models:-

1. Go beyond two scales (field and farm/irrigation system) to present a basin-scale accounting of water flows to account for aggregate depletion?
2. Include and adjust water withdrawals into the irrigation system to witness how these adjust real and paper water savings?
3. Account for non-water withdrawals to examine how these alter the calculus of real water savings?
4. Include irrigated area as initial and intermediate variables that dynamically control water depletion?
5. Adjust irrigation efficiency to see how this governs changes in aggregate depletion?
6. Divide recoverable flows into water reused by irrigation and water available for other users downstream?
7. Include beneficial consumption as part of depletion?
8. Weave in rainfall values to alter depletion?
9. Incorporate crop yield to allow for pareto checks on production versus real savings?
10. Include other irrigation practices such as deficit irrigation and irrigation duration?

ISA uses the above 10 questions (and other insights) to calculate changes in aggregate depletion at the basin scale.

REWAS only calculates changes in depletion at the irrigation system level (and not accurately in my view). It does not calculate depletion across the whole basin system. This means its claim for assessing real water savings is, erm, not really real.

In future blog posts, I will be tracking upgrades in both models.