How useful are models anyway? An example, now open for public comment

By Sonia A. Hall

Cover of the draft 2016 Long-Term Supply and Demand Forecast Legislative Report, currently available for public comment.
Cover of the draft 2016 Long-Term Supply and Demand Forecast Legislative Report, currently available for public comment.

Water, water everywhere… but will it continue to be there in the future? Will it be available when we need it? Or do we need to invest in projects or policies now, because the water in the future will not be the same as in the past? These are the issues that the collaborative research team working on the 2016 Columbia River Long-Term Supply and Demand Forecast are using models to address, at the direction of the Office of the Columbia River (OCR, part of the Washington Department of Ecology) and the Washington State Legislature.

Preliminary model results were presented at three public workshops in Richland, Wenatchee and Spokane in late June, and the draft report is available for public comment on OCR’s website until July 20, 2016. Here’s the summary of changes in water supply projected by this research:

  • Average annual supply of water for all uses across the Columbia River Basin down to Bonneville Dam is expected to increase around 12% by 2035.
  • That water would be available earlier in the spring than it has been in the past: water supply between November and May is projected to increase by almost 30%, while water supply between June and October is projected to decrease almost 11%.

What about the water needed to irrigate crops, by far the biggest destination for water diverted from rivers and pumped from the ground in the region? Well, keep in mind that as the climate warms, crops can be planted earlier in the spring, and will mature more quickly than in the past (unless varieties or even crops change, but that’s a topic that deserves its own blog post). The models quantify those effects, and project:

  • The demand for water for eastern Washington’s irrigated acres will decrease almost 5% by 2035. Why? Because we expect it to rain more in the spring, and crops would be growing earlier in the season, when it rains more, so less irrigation would be needed. Most crops would also be able to produce their yield while transpiring less water as the concentration of carbon dioxide in the air increases (sometimes called “CO2 fertilization,” another topic for its own blog post).
  • If current trends in the proportion of irrigated acres growing different crops continue, we can expect to see more acres growing things like wine grapes, that demand relatively little water per acre, and less acres growing things like pasture, that demand more water per acre. If this happens, demand for water would actually decrease even more—close to 7%.

You’ve probably noted that all these bullet points are scattered with conditional statements: “is expected”, “would be” “is projected”, “if”, “if”, “if”. So what is the use of these results, when they refer to a particular set of conditions that are not the reality of today’s world, let alone that of the world in 2035? This question was implied in many of the comments we received at the public workshops. This question may have been sparked by a very clear articulation of the conditions under which these results are valid: between now and 2035,

  • irrigated area in Washington does not increase;
  • double cropping does not become much more common (we can’t model this quite yet, but we are close);
  • producers do not shift to slower-growing varieties to take advantage of the longer growing season;
  • irrigation technology and management (and therefore efficiency in using water) does not improve.

Here’s one answer for that “what’s the use” question. It’s not the only one, but I hope it starts a conversation on other uses. These numbers provide a benchmark that we can all understand and, hopefully, agree upon.

Some people are focused on making sure we have water available to meet the needs of agriculture in different watersheds. Others are concerned with the needs of cities, towns and communities across the region. Yet others need to produce the hydropower we’ll use in the future. And some people center on protecting and restoring endangered fish species across eastern Washington. If they can all agree that under the conditions the models were run these results provide a good sense of what we can expect, then they can use this as a starting point for discussing how that benchmark would change if we changed those conditions.

Here’s an example: if renewable energy policies change in ways that would lead utilities to need more water across dams to produce electricity, we can compare that to what the models projected, and determine what else would need to change for our future use of water to be sustainable. We can do similar comparisons if water storage projects allowed expansion of irrigation to dryland areas, or if certain cities were to grow much faster than we thought. In this way, agreeing on a benchmark can help us have a more specific discussion around how to manage the multiple demands for fresh water resources in our region, and how to strategically invest in projects that meet competing water management objectives.

So can these results predict the future? No (you may have noticed that at no point have I—or the research team doing this work—said that “the models predict”). But are these results useful? Oh, yes. Especially if they can help people with a stake in how water is managed in the future have a fact-based discussion about the implications of their decisions.


2 comments on "How useful are models anyway? An example, now open for public comment"
  1. The future will likely mean less moisture storage time as snow, meaning an early runoff event. I have always been a supporter of storage reservoirs. They allow options for managing water regardless of natures will, to satisfy public needs whether it be irrigation, drinking water, fish, or recreation. Without reservoirs, late season fish runs are likely to be impacted the most.
    The summary of your climate models don’t appear to have changed much from a presentation I heard 10+ years ago.

    1. Nice one-line summary of the results, Tracy. Yes, the change in timing of when water is available is one of the main challenges for water management in the Basin, and a key result that can inform OCR’s investments moving forward. And though at this very synthetic level the results of the climate models may not appear to be much different from the past, the details are now more specific than 10+ years ago, so can be used when making decisions at finer scales – such as for particular WRIAs – and with a better understanding of where there are uncertainties and what those mean.

Leave a Reply

Your email address will not be published. Required fields are marked *