by David E. Rosenberg

Water models serve a variety of purposes. Stakeholders and managers use models to simulate the effects of new possible management operations decades into the future. Models can quantify tradeoffs between stakeholder’s conflicting objectives. Models can also help stakeholders understand how their system works. In a recent study, I created a new collaborative modeling approach to spur discussion of a combined Lake Powell-Lake Mead water bank as one possibility to adapt to the ongoing crisis on the Colorado River.  Within the collaborative modeling environment, basin partners immersed in and personified the role of water users in the basin. They devised, examined, discussed, and improved strategies for water management. This blog post explains the process used in the study. I also share 10 lessons from the new modeling approach.

Much recent Colorado River work has described basin challenges, including the division of declining flows between 4 Upper Basin states (Utah, Wyoming, Colorado, and New Mexico), 3 Lower Basin states (California, Arizona, and Nevada), and the Republic of Mexico. Current operations for Lake Powell and Lake Mead are based solely on reservoir level (that is, storage). Less consideration has been given to reservoir inflows. But as aridity and human use draw down the major Colorado River reservoirs, more of the water available for release and consumption comes from inflow, not storage. Thus, this modeling exercise sought to spur conversation on new Lake Powell and Lake Mead operations that better adapt to reservoir inflow and storage, rather than using only elevation triggers. Such conversations are particularly important now as interim guidelines for the two largest Colorado River reservoirs expire in two years. Beyond the specifics of the Colorado River, this exercise has lessons that may be useful for modeling water management everywhere.
 
The Model and the Process
 
I created a new modeling environment – immersive, online, and collaborative – using Google Sheets during a video conference session. The model had six Colorado River Basin accounts, representing different water user groups (Figure 1). These groups covered the geographic range from natural inflows to Lake Powell down to Lake Mead releases.

Twenty-six Colorado River managers and experts used the new model to experience potential uses of a combined Lake Powell-Lake Mead water bank. Groups of up to 6 collaborators at a time each individually took on water user roles. They played the roles for the duration of a 1- to 3- hour modelling session. Within a session, collaborators first articulated a management strategy for the user group they personified. Next collaborators selected an amount of unimpaired flow above Lake Powell to use for the first year. Collaborators then consumed, conserved, or traded water in response to available water (storage and inflows in their account balance), other’s choices, and real-time discussion among collaborators. The collaborators also had to protect reservoir levels in the two lakes and observe the impacts of their choices on ecosystems. The model is iterative. Each round represented one year. Thus collaborators took the output of their first round of decisions as the starting point for the next year/round of decisions. The model also included novel operations specific to the Colorado River (Box  1).

Based on the collaborative modeling exercise, I documented 7 modeling outcomes and 10 lessons from these model sessions. I also share 4 links for those interested in exploring the modeling method and outputs in more depth. One link is to a recently published journal paper on immersive modeling. A second link shares directions to download the model and hold your own model session. Two recently published articles illustrate other ways to better adapt Lake Powell and Lake Mead operations to inflow and storage conditions.
 
Modeling Outcomes
 
First and most importantly, multiple collaborators said the immersive, online modeling was fun. Yes, fun!

Collaborators improved their understanding of adaptive operations for a combined Lake Powell-Lake Mead water bank, rather than separately developing and testing competing alternatives. The collaborative approach, with ongoing discussion between participants, offered an alternative to prior offline or high-performance computing efforts that programmed water allocation rules to satisfy forecasted water demands across hydrologic scenarios. The collaboration also differed from prior efforts that excluded stakeholders, extracted data from participants, had a lead modeler or facilitation team mediate participant interactions with a model, or built a model then presented findings at the project end.

Within the online model environments, collaborators were more willing to explore extreme assumptions for natural flow above Lake Powell than is usual in public discussions, which presently focus only on 10 to 12 million acre-feet per year scenarios (Figure 2).

Collaborators also remarked:

• “I now see that more sustainable operations must include reservoir inflow.”
• “I think others will find the same value in the exercise that I have seen…. it’s thought-provoking.”
• “Sometimes the crazy ideas lead to watershed improvements.”

Collaborators also shared that the more adaptive operations were:

• “A huge leap from management today and, when we roll up our sleeves, fraught with implementation issues.”
• “I don’t know how you would ever do it.”

Many collaborators also encouraged me to share with others and suggested specific people.

10 Lessons

From the discussion and feedback, I synthesized the following lessons:

1. Model to provoke discussion and insights about new operations.  This seems to be a desirable prelude to more formal evaluation of solutions and quantifying tradeoffs.
2. Solicit feedback early to allow collaborators to improve a single management alternative and the online environment in which the alternative was modeled. This feedback kicked off a process where I met with new collaborator(s), solicited feedback, and used time before model sessions to improve the basin accounting alternative and/or its implementation in the online model environment.
3. Identify points of conflict to focus limited time during model sessions to provoke discussion on future operations rather than attempting to mediate or resolve win-lose conflicts that are inherent in Colorado River management.
4. Provide model options that show different ways to approach points of conflict. Options turned conflicts into choices. Collaborators could then think about and discuss the choices rather than try to resolve points of conflict.
5. Prorate reservoir evaporation by water account balance to more equitably treat each party. Parties with larger account balances shared more responsibility for reservoir evaporation.  This approach internalizes the evaporation costs of stored water.
6. Many options exist to prevent reservoir drawdown below protection volumes. For example, keep the shared reserve account balance at levels specified in Federal operational documents. Or trust a third party such as Reclamation to manage the account.
7. Allow trades to increase management flexibility. There was much active trading within the collaborative model environments. Trades are interesting as they require no physical water transfers.
8. Manage for combined storage in Lake Powell and Lake Mead to offer more flexibility.
9. Find common benefits such as ability to more adaptively manage water in one’s account more independently of other users.
10. Recognize the limits of a model’s acceptability and potential adoption.
    
    The immersive, online, and collaborative modeling spurred discussion of a combined Lake Powell-Lake Mead water bank as one possibility to adapt to the ongoing crisis on the Colorado River (see cartoon). I am excited to develop and use immersive, online, and collaborative models in other river basins and at different spatial and temporal scales.
     
    

For More Information on this Work:

1. Read the full journal article free – here.
2. Try out the final version of the immersive model:
   • Visit the Hydroshare model repository – here.
   • Download the Excel file.
   • Move into Google Drive.
   • Invite collaborators, friends, and/or family.
   • Follow the directions on the worksheet ReadMe-Directions. 
3. Read two other ways explored to adapt Lake Powell and Lake Mead releases to inflow and storage rather than just elevation triggers.
   • Adapt Colorado River Basin Depletions to Available Water to Live within Our Means (2023).
   • Adapt Lake Mead Releases to Inflow to Give Managers More Flexibility to Slow Reservoir Drawdown (2022).

David Rosenberg is a professor of Civil and Environmental Engineering at Utah State University, with a joint appointment at the Utah Water Research Laboratory. David and his collaborator’s mission is to make systems research more actionable. He and his collaborator’s work across the globe at scales from seconds (toilet flushes and irrigation events) to millennia (tree rings and more arid). Want to learn more? david.rosenberg@usu.edu | 435.797.8689 | http://rosenberg.usu.edu | https://utahandwesternwater.wordpress.com/

Further Reading

Resistance is Futile – Agriculture is Key to Fixing Lower Colorado River Water Shortages. CaliforniaWaterBlog.com, Posted on February 5, 2023 by Jay Lund and Josué Medellin-Azuara.

Drought and the Colorado River: Localizing Water in Los Angeles. CaliforniaWaterBlog.com, Posted on January 8, 2023 By Erik Porse and Stephanie Pincetl.

Adapt Colorado River Basin Depletions to Available Water to Live within Our Means. Published May 10, 2023. Journal of Water Resources Planning and Management. By Jian Wang and David E Rosenberg.

Adapt Lake Mead Releases to Inflow to Give Managers More Flexibility to Slow Reservoir Drawdown. Published July 19, 2022. Journal of Water Resources Planning and Management. By David E. Rosenberg.

Collaborative Modeling for Decision Support in Water Resources: Principles and Best Practices. Published May 13, 2013. Journal of the American Water Resources Association. By Stacy Langsdale, Allyson Beall, Elizabeth Bourget, Erik Hagen, Scott Kudlas, Richard Palmer, Diane Tate, and William Werick.