Lessons from Immersive Online Collaborative Modeling to Discuss More Adaptive Reservoir Operations
Publication: Journal of Water Resources Planning and Management
Volume 150, Issue 7
Abstract
This work had the purpose to provoke discussion about more adaptive reservoir operations. This work created online collaborative modeling environments by using web spreadsheets (Google Sheets) during a video conference session. In each session, up to 6 collaborators immersed in water user roles. Then the collaborators consumed, saved, and traded water in response to their available water, other choices, and real-time discussion of choices. The collaboration was an alternative to prior offline or high-performance computing efforts that programmed water allocation rules to try 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. In model sessions, collaborators improved more adaptive operations rather than separately developed and tested competing alternatives. 26 Colorado River managers and experts demonstrated use for a combined Lake Powell-Lake Mead water bank. The author used discussion and feedback to synthesize 10 lessons. For example, model to provoke discussion and insights rather than propose a solution, solicit feedback early, allow trades to increase flexibility, and recognize limits of model acceptability and adoption. To generate more actionable insights, next steps are engage multiple groups within the same model session and explore more management alternatives.
Practical Applications
Researchers, consultants, facilitators, and project leaders can build their own online collaborative model environments for their study system(s). Leaders can invite basin managers, stakeholders, colleagues, students, and the public to collaborate during video conference or in-person sessions. Leaders can also use the collaborative model environment(s) to prompt discussion of future basin operations, solicit feedback to improve operations, and/or make the model environments more user friendly. 26 Colorado River Basin managers and experts demonstrated use for a combined Lake Powell-Lake Mead water bank.
Introduction
This work had the purpose to model and discuss more adaptive reservoir operations with managers and experts. Prior simulation and optimization efforts programed water allocation rules, tried to satisfy forecasted water demands across hydrologic scenarios, maximized an ecosystem criteria, or sustained minimum flow requirements (Alafifi and Rosenberg 2020; Harou et al. 2009; Horne et al. 2016; Kasprzyk et al. 2013; Porse et al. 2015; Smith et al. 2022; Wurbs 2005; Zagona et al. 2001). Prior participatory processes either built an expert model then presented findings to participants at the project end (Horne et al. 2016), extracted data from participants (Voinov et al. 2016), had a lead modeler or facilitation team mediate participant interactions with the model across multiple sessions spaced in time (Bourget 2011; Langsdale et al. 2013; Michaud 2013; Van den Belt 2004; Wheeler et al. 2018), or created a serious or in-person game for a hypothetical water system (Babbitt 2019; Ewen and Seibert 2016; Madani et al. 2017; Schulze et al. 2015; Seibert and Vis 2012). Many prior immersive modeling efforts focused on visualization and exploration of 3-dimensional landscapes, sometimes for educational purposes (e.g., Dickes et al. 2019; Kamarainen et al. 2015; Nasr-Azadani et al. 2023). There remains a need to give the water resources community more collaborative modeling experiences where multiple managers and experts immerse in water user roles in an actual basin, articulate strategies for users, then in each time step manage water in response to their available water, other choices, and real-time discussion of choices. There is also a need to provoke discussion to improve a reservoir management alternative rather than simply view results or test competing alternatives.
This work created online collaborative modeling environments by using web spreadsheets (Google Sheets) during video conference sessions. In each model session, collaborating basin managers and experts (hereafter, collaborators) immersed in a water user role and articulated a management strategy for their user. Then in each time step, collaborators consumed, saved, and traded water in response to their available water, other choices, and real-time discussion of choices. Collaborators also protected critical reservoir levels, observed impacts of choices on ecosystems, and updated strategies for the next time step. After one or a few time steps, collaborators were asked what they liked about the model and how to improve.
This article synthesizes 10 lessons from the online collaborative modeling sessions with 26 managers and experts in the Colorado River Basin, US and Mexico. Lessons helped improve model process, build trust, grow understanding of other’s situations, increase operational flexibility, and generate more actionable suggestions for reservoir management. The next two sections describe in more detail the composition of the collaborative modeling sessions, collaborators use of the web spreadsheets to manage basin water accounts, and application for a combined Lake Powell-Lake Mead water bank in the Colorado River Basin. The following two sections share lessons from the discussions and next steps to generate more actionable suggestions for reservoir management. The final section concludes.
Composition of Online Collaborative Model Sessions
Between April and November 2021, the author invited 32 Colorado River mangers and experts to 13 video conference and 1 in-person collaborative modeling sessions. Collaborators were employed by the US Federal Government, Upper Colorado River Commission, agencies of Colorado River basin states, water districts, consulting firms, universities, a non-governmental organization, a foundation, and a First Nation. Hereafter, the term First Nations will collectively indicate the 30 Federally recognized sovereign Tribes within the Colorado River basin (Ten Tribes Partnership 2018). The 30 First Nations were represented with one basin water account. Three people engaged in two sessions. Three people started but did not complete a session. Two people declined a request to engage and one person never responded. During the same period, the author also supervised collaborative modeling sessions with 4 graduate students, 22 university colleagues, and 63 undergraduate students none of whom had Colorado River basin expertise. In Winter 2023, three basin experts completed a model session unsupervised. This article focuses on feedback from 26 Colorado River managers and experts who completed a supervised modeling session during Summer and Fall 2021.
Sessions followed the general structure:
•
Solicited collaborators through email or by invite from another collaborator.
•
Held sessions with 1 to 6 collaborators from the same organization.
•
Sessions lasted 1 to 3 h.
•
Each collaborator managed one or more basin water accounts in a combined Lake Powell-Lake Mead water bank.
•
In sessions with a small number of collaborators, the author managed one or more water accounts.
•
Collaborators sometimes managed the water account for their stakeholder group, sometimes not.
•
After play of 1 to 5 years, the author asked collaborators what they liked and how to improve water bank operations and the online collaborative modeling environment.
The next section further explains the set up and use of six basin water accounts in a web spreadsheets model for a combined Lake Powell-Lake Mead water bank.
Water Account Setup and Online Use
Six Colorado River Basin water accounts existed within an online spreadsheets model (Google Sheets) and a help guide (Rosenberg 2023). Conceptually, the water accounts existed within a combined Lake Powell-Lake Mead water bank. The bank stretched from the natural inflows to Lake Powell down to Lake Mead releases (Fig. 1). The Upper Basin water account represented the states of Utah, Wyoming, Colorado, and New Mexico. These states divert and consume water upstream of Lake Powell. Thus, the Upper Basin water account exchanged natural flow and stored water. This exchange is further explained in Step 3 of this section. The Lower Basin account represented the states of California, Arizona, and Nevada. The total of all account balances equaled the combined active storage in Lake Powell and Lake Mead.
In the initial rows of a master web spreadsheet, collaborators chose a water account to manage, entered a strategy, and registered initial assumptions such as reservoir starting levels and protection elevations. This early engagement affirmed each collaborator’s ability to interact with the web spreadsheets in real time. Subsequent rows comprised the remaining components of a water balance for a combined Lake Powell-Lake Mead water bank. The components included whole basin inflow, reservoir evaporation, available water, consumptive use, conservation, trade, and the Lake Powell release that split the combined end-of-year storage between Lake Powell and Lake Mead. Columns represented years. Collaborators entered individual choices—strategy, consumption, and conservation—into spreadsheet cells. Collaborators discussed their choices. After discussion, collaborators entered joint choices—trades and split of combined storage—into other spreadsheet cells.
The Colorado River Basin water accounts allowed more flexibility than existing operations because collaborators adapted their water consumption and conservation decisions to inflow and storage independent of other parties. The basin accounts contrasted with existing Colorado River operations that specify annual allocations to users, require increasing mandatory conservation for some users tied to declining reservoir levels, equalize reservoir storage, manage water for sovereign First Nations under state water laws, expire in 2026, and are a product of treaties, compacts, court cases, and agreements negotiated over 100 years (Carson et al. 1948; Castle and Fleck 2019; Colorado River Compact 1922; IBWC 2021; Kuhn and Fleck 2019; MacDonnell et al. 1995; Ten Tribes Partnership 2018; US Bureau of Reclamation and National Park Service 2016; USBR 2007, 2008, 2019). The online collaboration also allowed individual stakeholder groups to constructively improve basin water accounts and the spreadsheets interface rather than separately develop and test competing alternatives (Runge et al. 2015; USBR 2007). The online collaboration contrasted to distributed instances of the licensed, offline desktop RiverWare software and Colorado River Simulation System (CRSS) model (Zagona et al. 2001).
Collaborators completed 7 steps to set up and use the water accounts (Table 1).
1.
Assigned accounts and defined strategies for the next few years. Collaborators selected an account and entered their strategy. For example, an Upper Basin strategy might be to increase water use or deliver 92.3 billion cubic meters (BCM) [75 million acre-feet (maf)] to the Lower Basin every period of 10 consecutive years as specified in Article III(d) of the 1922 Compact (Colorado River Compact 1922). Collaborators who wanted advice to formulate a strategy or see current operations consulted the linked online model guide (Rosenberg 2023). Hereafter, to follow practice with the Colorado River Basin, this article will report elevations, depths, and volumes in customary units of feet and million acre-feet. The following conversion rates apply: and .
The Upper Basin, Lower Basin, Mexico, Colorado River Delta, and First Nations water accounts represented entities defined in the 1922 Colorado River Compact, 1948 Upper Colorado River Basin Compact, 1944 US-Mexico Treaty, Minutes 319 and 323, and pledges to include First Nations (Carson et al. 1948; IBWC 2021; Ten Tribes Partnership 2018; USBR 2020). The First Nations account allowed a collaborator to manage water independently from the Basin States in which the First Nations were located. This set up differed from current operations where Basin States administer water rights for the First Nations within state boundaries.
A shared reserve started with 11.6 maf of water that represented the protect volumes of 5.9 maf and 5.7 maf in Lake Powell and Lake Mead that corresponded to elevations 3,525 feet and 1,020 feet defined in the Upper and Lower Basin Drought Contingency Plans (USBR 2019). The reservoir storage-elevation relationships were extracted from the CRSS model—prior to the 2022 update to Lake Powell bathymetry (Bradley and Collins 2022).
The shared reserve prevented collaborators who drew down their water account balance to zero from further drawing down reservoir storage. At the same time, the 11.6 maf in the reserve comprised of the active storage in Lake Powell and Lake Mead at the time sessions were held. If all collaborators agreed, the reserve could transfer water to another account. When contemplating such transfers, consideration was given to the potential for reduced hydropower generation at one or both reservoirs and warmer Glen Canyon Dam release temperatures that further threaten populations of native, endangered fish in the Grand Canyon (Wheeler et al. 2021).
2.
Assigned all existing reservoir storage to water accounts. The collaborators jointly agreed on how to assign all active reservoir storage to the water accounts. The start volume varied from 21 to 16.2 maf as the actual Lake Powell and Lake Mead volumes drew down over the time period of the modeling sessions. Default assignments drew on existing agreements and operations. For example, collaborators assigned to Mexico the 0.17 maf that was the October 2020 balance in its Lake Mead conservation account (USBR 2007, 2021). Collaborators assigned the Lower Basin the 2.8 maf balance in the Lake Mead conservation accounts for California, Arizona, and Nevada (USBR 2007, 2021). Similarly, collaborators assigned the Upper Basin most of the Lake Powell storage that was not the protection volume. Collaborators assigned the shared reserve 11.6 maf as described in Step 1 (USBR 2019). The assignments gave starting water account balances to the Upper Basin, Colorado River Delta, and First Nations plus allowed the Lower Basin and Mexico to move their Lake Mead conservation account balances into a more flexible basin water account. There were many other ways to assign reservoir storage to the accounts.
3.
Selected year’s whole basin inflow and assigned to water accounts. Collaborators chose each year’s natural inflow to Lake Powell to explore a broader range of hydrologic scenarios than historical flows. Collaborators used historical data in the model guide to inform choices. For example, collaborators often chose Lake Powell natural inflows below the 2000 to 2020 average of 12.4 maf per year (Salehabadi et al. 2020) and below the Lake Powell release target of 8.23 maf per year developed in the 1970s (Fig. 2). Collaborators also changed the Lake Powell natural inflow from one year to the next. For example follow a year with recent average inflow (12 maf) by a dry year (7 maf). The Lake Powell natural inflow represented the flow if users above Lake Powell did not store, divert, or consume water (Prairie 2020; Wheeler et al. 2019). Crediting this natural inflow to the basin water accounts allowed the Upper Basin and First Nations to divert and consume Colorado River water upstream of Lake Powell, deduct consumptive use from their account, then carry over the balance to the next year. This setup allowed the Upper Basin and First Nations located upstream of Lake Powell to store and administratively recover water in Lake Powell even though they did not physically withdraw water from Lake Powell. Below Lake Powell, the model added default inflows of 0.8 maf per year for intervening Grand Canyon inflow (Rosenberg 2021; Wang and Schmidt 2020) and 0.2 maf per year for Hoover to Imperial Dam intervening inflow (Prairie 2020). The intervening Grand Canyon inflow included the Paria, Little Colorado, and Virgin rivers plus Grand Canyon seeps from Glen Canyon Dam to Lake Mead after diversions from tributary users. 0.6 maf per year of intervening Grand Canyon inflow represented a 5-year sequence average for a dry period while 1.0 maf per year was the 30-year average.
Collaborators also assigned the whole basin inflow—Lake Powell natural inflow plus downstream inflows—to the water accounts. Default assignments followed the existing priority of operations with changes for the shared reserve, Lake Havasu/Lake Parker evaporation and evapotranspiration, and First Nations that are not in current operations (Fig. 3).
The assignments were:
•
Assigned the shared reserve inflow that equaled the water account’s share of reservoir evaporation because reservoir evaporation depletes inflow before other activities. This assignment kept the shared reserve balance steady and helped protect a combined storage volume of 11.6 maf that is the sum of Lake Powell and Lake Mead protect volumes in the Upper and Lower Basin DCPs (USBR 2019).
•
Assigned inflow to equal Lake Havasu/Parker evaporation and evapotranspiration.
•
Assigned First Nations 1.9 maf per year of decreed water rights because this account managed water independently of the Basin States. The volume included 1.06 and 0.95 maf per year above and below Glen Canyon Dam (Ten Tribes Partnership 2018) and deducted First Nations in the Lower Basin’s portion of Havasu/Parker losses. The volume excluded claimed amounts.
•
Assigned Colorado River Delta 0.016 maf per year as 67% of the 9-year, 0.21 maf volume the US and Mexico pledged in Minute 323 (IBWC 2021).
•
Assigned Mexico 1.5 maf per year (1944 US-Mexico Treaty), minus the mandatory conservation volume specified in Minutes 319 and 323, minus Mexico contributions to the Colorado River Delta, and minus Mexico’s portion of Havasu/Parker losses because the US must deliver Mexico water first (IBWC 2021). The mandatory conservation volume increased as Lake Mead level declined.
•
Split the next 2.4 maf per year of whole basin inflow between the Upper and Lower Basins because the Upper and Lower Basins have 1.2 and 2.45 maf per year of pre-1922 water rights (Leeflang 2021) after deducting First Nations use.
•
Assigned the Lower Basin the next 5.3 maf per year. 5.3 maf plus 1.2 maf pre-1922 use plus 0.95 maf of water for First Nations below Hoover dam (Ten Tribes Partnership 2018) plus half the Mexico assignment resulted in 8.2 maf per year that is the Lake Powell objective release.
•
Assigned the Upper Basin all remaining Lake Powell natural inflow.
Like assigning storage, the preceding inflow assignments are one of many ways to assign whole basin inflow to the water accounts.
4.
Calculated each water account’s available water as the account balance (Step 2), plus share of inflow (Step 3), and minus share of reservoir evaporation [Eq. (1); all units maf]. An account’s share of reservoir evaporation was the combined annual Lake Powell and Lake Mead evaporation prorated by the water account’s share of combined storage. Optional purchases from other accounts increased available water while sales decreased an account’s available water. The optional trades built on a feature of the Lower Basin drought contingency plan that let Lower Basin parties transfer their Lake Mead conservation account balance to another party (USBR 2019)
(1)
5.
Parties conserved and consumed within their available water independent of other parties. Consumptive use withdrew from a basin water account. Conservation made water in the account available next year. Each party’s end-of-year account balance was their available water (Step 4) minus consumption. Account withdraws from Lower Basin, Mexico, Delta, and First Nations accounts implied a withdraw from Hoover dam or Lake Mead.
6.
Assigned remaining combined storage to Lake Powell and Lake Mead. This assignment was another joint—political—decision and gave collaborators flexibility to preferentially store water in one reservoir. The existing operations seek to equalize or split storage 50%/50% (USBR 2007). Collaborators withdrew from their water accounts whether water was physically stored in Lake Powell or Lake Mead. Two considerations to assign combined storage between Lake Powell and Lake Mead were:
a.
Maintain populations of endangered, native fish of the Grand Canyon by either (i) keeping Lake Powell elevation above 5.9 maf (3,525 feet) to maintain summertime turbine release temperatures below 18°C (Wheeler et al. 2021), or (ii) forego hydropower generation, and/or release more water through the river outlets.
b.
Keep Lake Powell and Lake Mead levels above the minimum power pool storages of 4.0 maf (3,490 feet) and 2.2 maf (955 feet) or require hydropower customers to purchase additional energy from more expensive sources.
7.
Continued to next year. All end of year water account balances carried over to the beginning of the next year (Steps 3 to 6).
| Step | Decision type |
|---|---|
| 1. Assigned water accounts and defined strategies. | Individual |
| 2. Assigned existing reservoir storage to accounts. | Joint |
| 3A. Selected year’s natural inflow to Lake Powell. | Joint |
| 3B. Assigned inflow to water accounts. | Joint |
| 4. Calculated available water for each account. | Calculated |
| 5. Collaborators conserved, consumed, and traded within their available water. | Individual |
| 6. Assigned combined storage to Lake Powell and Lake Mead. | Joint |
| 7. Continued to next year. | Calculated |
The table in Appendix summarizes similarities and differences across 16 features of existing Colorado River operations and basin water accounts.
Lessons
Discussions of basin water accounts during online collaborative modeling sessions led to 5 lessons to improve model process (Lessons #1–4 and 10) and 5 lessons to improve the substance of future operations (Lessons #5–9).
1.
Model to provoke discussion about new operations rather than propose a solution. For example, at the end of the first model session, a collaborator said to “continue to provoke thought and discussion.” During subsequent sessions, the author saw his role was to provoke thought and discussion about new operations rather than propose a solution. Further sessions elicited more discussion about features to like about basin water accounts and their implementation in an online collaborative model environment (Table 2).
One collaborator suggested:
Start asking people from different stakeholder groups to participate in the same session.
At the conclusion of a model session, another collaborator said:
I now see that more sustainable operations must include reservoir inflow.
And another collaborator later wrote:
“I think others will find the same value in the exercise that I have seen…. its thought provoking” and “Sometimes the crazy ideas lead to watershed improvements.”
Many collaborators also encouraged to share with others and suggested specific people.
2.
Solicit feedback early to allow collaborators to improve a single management alternative and the online environment in which the alternative was modeled. In the early weeks, the author shared a first version for Lake Powell with students and a colleague. They suggested to reduce the 10-year run period to 5 years. The next week, a Colorado River manager liked the exercise and asked for a more complete picture for Lake Mead and down to the Mexico border. This comment kicked off a process where the author 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.
The process resulted in 36 changes recorded in the Versions worksheet (Rosenberg 2023). The process allowed collaborators from different stakeholder groups to share ideas to improve basin water accounts as a single alternative for future operations rather than separately develop and test competing alternatives.
3.
Identify points of conflict to focus limited time during model sessions to provoke discussion on future operations (Lesson #1) and solicit suggestions to improve (Lesson #2) rather than mediate or try to resolve conflicts. Multiple collaborators raised the conflict to split Lake Powell natural inflow among the Upper and Lower Basins. Was the 75 maf every period of 10 consecutive years in Article III(d) of the 1922 Compact a delivery or non-deplete requirement (Beckstead and Hoerner 2012)? Can the Upper Basin deliver less water in the 1st model year and store more to recoup an over-delivery in the prior 9 years? How to deliver the 2.3 and 3.5 maf per year of pre-1922 water uses in the Upper and Lower Basins that have not yet been tested in 100 years since the Compact was signed (Leeflang 2021)? Can the Upper Basin store water in a basin water account for future use if Article III(e) of the Compact does not allow the Upper Basin to withhold water? One collaborator described 4 or 5 or 6 maf per year of Lake Powell natural inflow as unprecedented, never been done, and unclear what will happen.
Because collaborators kept identifying the split of Lake Powell natural inflow below 8.23 maf per year as a central conflict—a win-lose or zero-sum game—the author came to see that basin water accounts could not resolve the issue. Yet somehow in the modeling, the Lake Powell natural flow had to be split among the accounts. The split would be better resolved through separate stakeholder negotiations rather than model sessions. During the model sessions, time was better spent discussing future operations (Lesson #1) and identifying features to improve in basin water accounts and/or the online model environment (Lesson #2).
4.
Provide model options to show different ways to approach points of conflict. This lesson builds on Lesson #3. After identifying points of conflict, the author added model options to show different ways to address several win-lose conflicts inherent in Colorado River management. For example:
•
Include a water account to allow First Nations to manage their water as sovereign nations rather than under existing state water rights laws. This setup reduced allocations to the Upper and Lower Basin accounts.
•
Collaborators split existing reservoir storage among their water accounts. Assigning more storage to one account meant less to other accounts.
•
Collaborators split whole basin inflow among the water accounts. Similarly, assigning more inflow to one account meant less to other accounts.
•
Collaborators split combined storage between Lake Powell and Lake Mead.
•
Collaborators subtracted reservoir evaporation from account balances.
•
Collaborators could draw down the shared reserve and assign that water to one or more accounts.
These model options allowed collaborators to explore some of the many possibilities to resolve win-lose tradeoffs that are inherent in Colorado River management. The 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. This option was one way to address a win-lose conflict (Lesson #4) because some or all of Lake Mead and Lake Powell evaporation is not counted in current operations (Fleck and Castle 2022; Schmidt et al. 2016). Collaborators shared accolades and support for 7 spreadsheet rows that prorated the split by water account balance. Prorating evaporation by account balance may be favorable because:
•
Each party was treated equitably. Parties with larger account balances shared more responsibility for reservoir evaporation.
•
The Upper and Lower Basins could shift some of their responsibility for evaporation onto other parties and the shared reserve.
•
In model year 1, the shared reserve had the largest account balance and was charged of the combined reservoir evaporation.
Treating water accounts equitably may help parties overcome a win-lose conflict.
6.
Many options exist to govern draw down below the combined protection volume of 11.6 maf. One collaborator recommended to keep the shared reserve at 11.6 maf. Another collaborator noted that 11.6 maf is a lot of water and there may be reasons to draw down the shared reserve below the combined protection volume. A third collaborator suggested to trust a third party such as Reclamation to manage the shared reserve. There was also a suggestion to allow water account managers to sell water to the shared reserve if no other party wanted to buy. These comments suggest that multiple options exist to drawdown Lake Powell and Lake Mead below 11.6 maf.
7.
Allow trades to increase management flexibility. Within the collaborative modeling environment, most collaborators voluntarily and temporarily sold and purchased water even though few trades occurred under existing operations. During model sessions, many trades were for larger water volumes, more money, and involved more entities than the Lower Basin states and Federal government’s USD 200 million Fall 2021 plan to conserve 500,000 acre-feet in Lake Mead each year for 2 years (Allhands 2021). For example, some collaborators who role played Mexico sold water to build non-water infrastructure projects. Some collaborators who played the Upper Basin sold some water to get paid to conserve to prepare for mandatory cutbacks to meet the 10-year delivery requirement. Trades were possible because the basin water account balances defined the water each collaborator had available to trade each year. Also, trades administratively transferred water from one account to another within the combined Lake Powell-Lake Mead water bank without physical movement. There was broad support among collaborators for trades because trades gave more flexibility to acquire, consume, store, sell, or buy water.
8.
Manage the combined storage in Lake Powell and Lake Mead to offer more flexibility. Managing the combined storage offered more flexibility to store and access water in either Lake Powell or Lake Mead while sustain cold water releases from Lake Powell to benefit native, endangered fish of the Grand Canyon. Managing the combined storage also let collaborators conserve and consume independent of other collaborators and independent of where water was physically stored. Managing the combined storage helped shift discussion about Lake Powell and Lake Mead as Upper and Lower Basin reservoirs toward joint system operations.
9.
Find common benefits such as more adaptability. Lessons #5–8 combined to find common benefits as a way to escape win-lose conflicts. Each basin water account enjoyed common benefits each year of more flexibility to consume and conserve water independent of other accounts (lesson #8) and trade water with other collaborators. These common benefits treated accounts more equitably (lesson #5).
10.
Recognize the limits of a model’s acceptability and potential adoption. Collaborators identified many useful features of basin water accounts and its implementation in an online collaborative modeling environment (Lessons #1–9). Collaborators also said basin water accounts were:
•
Very different than current operations.
•
“A huge leap from management today and, when we roll up our sleeves, fraught with implementation issues.”
•
A heavy lift from existing management to whole basin management.
•
“Easy to suggest. Harder to get adopted.”
Collaborators also said:
•
“Initially I freaked out to break the existing operations.”
•
“I don’t know how you would ever do it. Hard to get traction on things that are less difficult than this.”
•
I support use so long as not a substitute for negotiations.
| Comment | Frequency |
|---|---|
| More holistic approach to basin management. | 5 |
| Drive a conversation around conservation with bad hydrology. | 5 |
| I like it/It is neat/It is fun. | 3 |
| Facilitates thought and conservation. | 2 |
| Make me think about the equity issue. How to factor in equity. | 2 |
| Interactive. I see the effect of choices. | 1 |
| See yourself in the model. | 1 |
| See effects on native fish. | 1 |
| What it means to have and use my own water account. | 1 |
These comments discounted the model’s legitimacy and actionability (Van den Belt 2004; Wheeler et al. 2018). These repeated comments suggested to stop sessions and write up lessons from the model sessions.
Discussion
This work used online collaborative modeling environments to engage 26 Colorado River managers and experts to manage and discuss Colorado River basin water accounts as an alternative to current reservoir equalization operations that expire in 2026. Using web spreadsheets during a video conference let collaborators immerse in water user roles, then consume, conserve, and trade water in response to their available water, others choices, and real-time discussion of choices. The real-time discussion identified many positive features for future Colorado River management such as prorate reservoir evaporation by account balance, allow trades, and manage storage in Lake Powell and Lake Mead as a combined system. These features gave collaborators a common set of benefits and more flexibility to manage. The discussion allowed collaborators to constructively improve basin water accounts rather than separately develop and test competing alternatives. The real-time discussion turned conflicts over reservoir management into more collaborative efforts.
This collaboration is not possible with CRSS (Zagona et al. 2001), Water Evaluation and Planning (Yates et al. 2005), R, R Shiny, Python, or cloud notebooks (Abdallah et al. 2022). The online collaboration and discussion contrasted with no/little stakeholder interaction for 42 studies of environmental water decisions (Horne et al. 2016), efforts that extracted data from participants (Voinov et al. 2016), efforts that required a lead modeler or facilitation team to mediate participant interactions with the model across many sessions (Bourget et al. 2013; Langsdale et al. 2013; Michaud 2013; Van den Belt 2004; Wheeler et al. 2018), or the build-translate approach most researchers use to build a model on their own then present findings at the project end. The online collaborative modeling also contrasted with current Colorado River basin practices where states undertake separate modeling efforts and Reclamation pushes out its model improvements in one-way communications.
Like other spreadsheet programs, the web spreadsheets made difficult version control, organize an intuitive interface, validate collaborative input, and automate tasks to support collaborative efforts. The model sessions highlighted conflicts over how to divide Lake Powell natural inflow below 8 maf per year. Discussions may not lead to agreement or consensus. The annual model assumed collaborators knew whole basin inflow before making annual consumption and conservation decisions when managers start the water year with a flow forecast that may over- or underestimate actual flow. Collaborators said the online modeling was fun and engaging. Collaborators also said basin water accounts “strayed too far from current operations.” The later comment raised issues of legitimacy and actionability.
To increase model legitimacy and actionability, researchers and facilitators can additionally ask their collaborators to:
•
Construct their own alternatives rather than chose options within a predefined alternative (Voinov et al. 2016).
•
Within a model session, collaborate with people from multiple stakeholder groups rather than a single group.
•
Screen and improve multiple alternatives rather than experiment with one alternative.
To implement these added features, an important next step is to organize efforts where creative, productive, and connected people from different stakeholder groups together design, build, and collaborate in the same online model environment. In such sessions, collaborators can learn together, build trust, generate and validate more innovative and actionable insights, and share findings with their communities (Van den Belt 2004; Voinov et al. 2016). People intending to lead or join such efforts are challenged to assemble a team with basin, modeling, discipline, integration, facilitation, guiding, communication, interpersonal, and political skills. Leaders are challenged to find money and time to support the team. The team has to convince potential collaborators to invest their time on the belief that collaboration will generate more innovative, beneficial, and actionable insights and products than if groups work solo.
Conclusions
This work had the purpose to model and discuss more adaptive reservoir operations with basin manager and experts. An online collaborative modeling environment was constructed by using web spreadsheets (Google Sheets) during video conference sessions. In model sessions, up to 6 collaborators immersed in water user roles and articulated a strategy for the user. Then in each time step, collaborators consumed, conserved, and traded water in response to their available water, other choices, and real-time discussion of choices. Collaborators also protected reservoirs and tried to sustain populations of endangered, native fish of the Grand Canyon. 26 Colorado River managers and experts demonstrated use in a combined Lake Powell-Lake Mead water bank. Collaborators gave feedback to improve (i) basin water accounts as an alternative to existing operations, and (ii) the online environment in which the water accounts were modeled.
Ten lessons were synthesized from the model sessions to improve model process and substance. Lessons included model to provoke discussion about new operations rather than propose a solution, solicit feedback early to allow collaborators to improve a management alternative and the model environment, identify points of conflict to focus limited time in sessions to provoke discussion and solicit feedback, and provide options to show different ways to approach conflicts. Further lessons were allow trades to increase flexibility, manage the combined storage in a Lake Powell and Lake Mead water bank, find common benefits such as more flexibility for all collaborators, and recognize the limits of a model’s acceptability and potential adoption.
The online collaborative model environments differed from prior studies that excluded stakeholders, extracted data from participants, required a lead modeler or facilitation team to mediate participant interactions with the model, or built a model then presented findings at the project end. The collaborative modeling also contrasted with current Colorado River basin practices where states undertake separate modeling efforts and Reclamation pushes out its model improvements in one-way communications. The collaborative model environments also allowed different stakeholder groups to constructively improve an alternative rather than separately develop and test competing alternatives. To generate more actionable suggestions, next steps are engage multiple groups within the same model session and explore more management alternatives.
Appendix. Comparison of Existing Colorado River Operations to Basin Water Accounts
| Feature | Existing operations | Basin water accounts |
|---|---|---|
| Purposes | Encourage conservation; Closer coordinate Lake Powell and Lake Mead operations; Plan for shortages; Address future controversies through consultation and negotiation not litigation (USBR 2020). | Same as existing operations; Additionally give each party more flexibility to manage their water account independent of other parties. |
| Accounts | Lake Mead conservation accounts only for 3 Lower Basin states and Mexico. | 6 water accounts in the combined Lake Powell-Lake Mead water bank. |
| Colorado River Delta | Nongovernmental organizations secure water from the US and Mexico for each pulse flow (IBWC 2021). | Separate water account in combined Lake Powell-Lake Mead water bank. |
| First Nations | Managed in trust under state water rights systems. | Separate water account in combined Lake Powell-Lake Mead water bank. |
| Account sales and trades | Only between Lake Mead conservation accounts for Lower Basin states. | Between all water accounts in a combined Lake Powell-Lake Mead water bank. |
| Reservoir protection volumes | 5.9 and 5.7 maf that correspond to elevations 3,525 and 1,020 feet in Lake Powell and Lake Mead (USBR 2019). | Shared reserve account with initial 11.6 maf volume. Collaborators can vary storage over time. |
| Adaptation triggers | Reservoir storage. | Inflow and storage (Rosenberg 2022). |
| Voluntary water conservation | Lake Mead conservation account for each Lower Basin state and Mexico (IBWC 2021; USBR 2007). | Six Basin water accounts (all parties). |
| Mandatory water conservation | Increases for Lower Basin states and Mexico as Lake Mead draws down (IBWC 2021; USBR 2007, 2019). | Each party independently consumes and conserves water within their account balance. |
| Reservoir evaporation | Ignore of Lake Mead evaporation and 0.16 to 0.23 maf of Colorado River evapotranspiration prior to build Glen Canyon Dam (Fleck and Castle 2022; Schmidt et al. 2016). | Subtracted all Lake Mead and Lake Powell evaporation in proportion to water account balances. |
| Lake Powell releases | Target 8.23 maf per year with allowances to equalize Lake Powell and Lake Mead storage (USBR 2007). | Calculated each year after collaborators choose how to split combined storage between Lake Powell and Lake Mead. |
| Manage for endangered fish of Grand Canyon | Endangered Species Act. | Collaborators choose split of reservoir storage to main cold water releases from Lake Powell to benefit native fish. |
| Expiration date | 2026. | None; Manage year-to-year. |
| Model environment | Offline, licensed, distributed RiverWare/CRSS model instances on desktop machines. | Online, open-source, collaborative spreadsheets (Google Sheets). |
| Model components | 12 reservoirs; 29 flow gages; 520 water user objects; 145 rules (Wheeler et al. 2019). | 142 rows on 1 master worksheet; 4 data support worksheets; ReadMe worksheet; Versions worksheet. |
| Model support | Reclamation staff | Linked online user’s guide to each spreadsheet row (Rosenberg 2023). |
Data Availability Statement
Acknowledgments
I thank the 26 Colorado River managers and experts for their time, engagement, collaboration, and discussion. This work benefited from a USD 50 donation from a private individual. The donation was used to purchase software to generate the online model guide. Five collaborators who participated in a model session and two anonymous reviewers gave feedback that improved this article. This work represents the views of the author, not Utah State University.
References
Abdallah, A. M., D. E. Rheinheimer, D. E. Rosenberg, S. Knox, and J. J. Harou. 2022. “An interoperable software ecosystem to store, visualize, and publish water resources systems modelling data.” Environ. Modell. Software 151 (May): 105371. https://doi.org/10.1016/j.envsoft.2022.105371.
Alafifi, A. H., and D. E. Rosenberg. 2020. “Systems modeling to improve river, riparian, and wetland habitat quality and area.” Environ. Modell. Software 126 (Apr): 104643. https://doi.org/10.1016/j.envsoft.2020.104643.
Allhands, J. 2021. “It could take at least 500,000 acre-feet of water a year to keep Lake Mead from tanking.” Accessed November 8, 2021. https://www.azcentral.com/story/opinion/op-ed/joannaallhands/2021/11/08/lake-mead-could-get-extra-water-from-lower-basin-annually/6306601001/.
Babbitt, C. 2019. “The Groundwater Game: A new hands-on way to learn about groundwater management.” Accessed June 7, 2023. https://blogs.edf.org/growingreturns/2019/07/18/new-groundwater-game-teaches-sustainable-management-trading/.
Beckstead, B., and C. Hoerner. 2012. “Does the upper basin have a delivery obligation or an obligation not to deplete the flow of the Colorado River at Lee Ferry?” Accessed March 25, 2024. http://www.waterpolicy.info/wp-content/uploads/2015/09/Delivery-Obligation-memo.pdf.
Bourget, E. C., S. M. Langsdale, and M. van den Belt. 2013. “Featured collection introduction: Collaborative modeling for decision support as a tool to implement IWRM.” J. Am. Water Resour. Assoc. 49 (3): 605–608. https://doi.org/10.1111/jawr.12071.
Bourget, L. 2011. Converging Waters: Integrating collaborative modeling with participatory processes to make water resources decisions. Washington, DC: USACE.
Bradley, D., and K. Collins. 2022. “Lake Powell 2017 area and capacity tables.” Accessed March 25, 2024. https://www.usbr.gov/uc/water/Lake_Powell_Area_Capacity_Table_Report_FINAL.pdf.
Carson, C. A., C. H. Stone, F. E. Wilson, E. H. Watson, and L. C. Bishop. 1948. “Upper Colorado river basin compact.” Accessed September 7, 2021. https://www.usbr.gov/lc/region/g1000/pdfiles/ucbsnact.pdf.
Castle, A., and J. Fleck. 2019. “The risk of curtailment under the Colorado River compact.” SSRN eLibrary. Accessed March 25, 2024. https://doi.org/10.2139/ssrn.3483654.
Colorado River Compact. 1922. “Colorado River compact.” Accessed March 25, 2024. https://www.usbr.gov/lc/region/pao/pdfiles/crcompct.pdf.
Dickes, A. C., A. Kamarainen, S. J. Metcalf, S. Gün-Yildiz, K. Brennan, T. Grotzer, and C. Dede. 2019. “Scaffolding ecosystems science practice by blending immersive environments and computational modeling.” Br. J. Educ. Technol. 50 (5): 2181–2202. https://doi.org/10.1111/bjet.12806.
Ewen, T., and J. Seibert. 2016. “Learning about water resource sharing through game play.” Hydrol. Earth Syst. Sci. 20 (10): 4079–4091. https://doi.org/10.5194/hess-20-4079-2016.
Fleck, J., and A. Castle. 2022. “Green light for adaptive policies on the Colorado river.” Water 14 (1): 2. https://doi.org/10.3390/w14010002.
Harou, J., M. Pulido-Velazquez, D. E. Rosenberg, Josue Medellin-Azuara, J. R. Lund, and R. E. Howitt. 2009. “Hydro-economic models: Concepts, design, applications, and future prospects.” J. Hydrol. 375 (3–4): 627–643. https://doi.org/10.1016/j.jhydrol.2009.06.037.
Horne, A., J. M. Szemis, S. Kaur, J. A. Webb, M. J. Stewardson, A. Costa, and N. Boland. 2016. “Optimization tools for environmental water decisions: A review of strengths, weaknesses, and opportunities to improve adoption.” Environ. Modell. Software 84 (Jun): 326–338. https://doi.org/10.1016/j.envsoft.2016.06.028.
IBWC (International Boundary Water Commission). 2021. “Minutes between the United States and Mexican sections of the IBWC.” Accessed July 22, 2021. https://www.ibwc.gov/Treaties_Minutes/Minutes.html.
Kamarainen, A. M., S. Metcalf, T. Grotzer, and C. Dede. 2015. “Exploring ecosystems from the inside: How immersive multi-user virtual environments can support development of epistemologically grounded modeling practices in ecosystem science instruction.” J. Sci. Educ. Technol. 24 (2): 148–167. https://doi.org/10.1007/s10956-014-9531-7.
Kasprzyk, J. R., S. Nataraj, P. M. Reed, and R. J. Lempert. 2013. “Many objective robust decision making for complex environmental systems undergoing change.” Environ. Modell. Software 42 (12): 55–71. https://doi.org/10.1016/j.envsoft.2012.12.007.
Kuhn, E., and J. Fleck. 2019. Science be dammed: How ignoring inconvenient science drained the Colorado River. Tucson, AZ: University of Arizona Press.
Langsdale, S., A. Beall, E. Bourget, E. Hagen, S. Kudlas, R. Palmer, D. Tate, and W. Werick. 2013. “Collaborative modeling for decision support in water resources: Principles and best practices.” J. Am. Water Resour. Assoc. 49 (3): 629–638. https://doi.org/10.1111/jawr.12065.
Leeflang, B. 2021. “Colorado River coding: Pre 1922 compact water use.” Accessed April 11, 2022. https://doi.org/10.5281/zenodo.5501466.
MacDonnell, L. J., D. H. Getches, and W. C. Hugenberg. 1995. “The law of the Colorado River: Coping with severe sustained drought.” J. Am. Water Resour. Assoc. 31 (5): 825–836. https://doi.org/10.1111/j.1752-1688.1995.tb03404.x.
Madani, K., T. W. Pierce, and A. Mirchi. 2017. “Serious games on environmental management.” Sustainable Cities Soc. 29 (Feb): 1–11. https://doi.org/10.1016/j.scs.2016.11.007.
Michaud, W. R. 2013. “Evaluating the outcomes of collaborative modeling for decision support.” J. Am. Water Resour. Assoc. 49 (3): 693–699. https://doi.org/10.1111/jawr.12066.
Nasr-Azadani, E., D. H. Wardrop, and R. P. Brooks. 2023. “Pathways for the utilization of visualization techniques in designing participatory natural resource policy and management.” J. Environ. Manage. 333 (May): 117407. https://doi.org/10.1016/j.jenvman.2023.117407.
Porse, E. C., S. Sandoval-Solis, and B. A. Lane. 2015. “Integrating environmental flows into multi-objective reservoir management for a transboundary, water-scarce river basin: Rio Grande/Bravo.” Water Resour. Manage. 29 (8): 2471–2484. https://doi.org/10.1007/s11269-015-0952-8.
Prairie, J. 2020. “Colorado River basin natural flow and salt data.” Accessed March 25, 2024. https://www.usbr.gov/lc/region/g4000/NaturalFlow/current.html.
Rosenberg, D. E. 2021. “Colorado River coding: Grand Canyon intervening flow.” Accessed April 11, 2022. https://doi.org/10.5281/zenodo.5501466.
Rosenberg, D. E. 2022. “Adapt lake mead releases to inflow to give managers more flexibility to slow reservoir drawdown.” J. Water Resour. Plann. Manage. 148 (10): 02522006. https://doi.org/10.1061/(ASCE)WR.1943-5452.0001592.
Rosenberg, D. E. 2023. “Colorado River Basin water accounts: Provoke discussion about more adaptive reservoir operations.” Accessed June 8, 2023. https://www.hydroshare.org/resource/dd17d7bdf25d4da0abeaf381c9fa2df0/.
Runge, M. C., et al. 2015. Decision analysis to support development of the Glen Canyon Dam long-term experimental and management plan. Reston, VA: USGS.
Salehabadi, H., D. Tarboton, E. Kuhn, B. Udall, K. Wheeler, E. Rosenberg, S. Goeking, and J. C. Schmidt. 2020. Stream flow and losses of the Colorado River in the Southern Colorado Plateau. Logan, UT: Utah State Univ.
Schmidt, J. C., M. Kraft, D. Tuzlak, and A. Walker. 2016. Fill mead first: A technical assessment. Logan, UT: Utah State Univ.
Schulze, J., R. Martin, A. Finger, C. Henzen, M. Lindner, K. Pietzsch, A. Werntze, U. Zander, and R. Seppelt. 2015. “Design, implementation and test of a serious online game for exploring complex relationships of sustainable land management and human well-being.” Environ. Modell. Software 65 (Aug): 58–66. https://doi.org/10.1016/j.envsoft.2014.11.029.
Seibert, J., and M. J. P. Vis. 2012. “Irrigania—A web-based game about sharing water resources.” Hydrol. Earth Syst. Sci. 16 (8): 2523–2530. https://doi.org/10.5194/hess-16-2523-2012.
Smith, R., E. Zagona, J. Kasprzyk, N. Bonham, E. Alexander, A. Butler, J. Prairie, and C. Jerla. 2022. “Decision science can help address the challenges of long-term planning in the Colorado River Basin.” J. Am. Water Resour. Assoc. 58 (5): 735–745. https://doi.org/10.1111/1752-1688.12985.
Ten Tribes Partnership. 2018. “Colorado river basin ten tribes partnership tribal water study.” Accessed March 26, 2024. https://www.usbr.gov/lc/region/programs/crbstudy/tws/finalreport.html.
USBR (United States Bureau of Reclamation). 2007. “Colorado River Interim guidelines for lower basin shortages and coordinated operations for lake Powell and Lake mead—Final environmental impact statement.” Accessed March 27, 2024. http://www.usbr.gov/lc/region/programs/strategies/FEIS/index.html.
USBR (United States Bureau of Reclamation). 2008. “Law of the river.” Accessed December 8, 2021. https://www.usbr.gov/lc/region/g1000/lawofrvr.html.
USBR (United States Bureau of Reclamation). 2019. Agreement concerning Colorado river drought contingency management and operations. Washington, DC: USBR.
USBR (United States Bureau of Reclamation). 2020. Review of the Colorado river interim guidelines for lower basin shortages and coordinated operations for Lake Powell and Lake Mead. Washington, DC: USBR.
USBR (United States Bureau of Reclamation). 2021. Boulder Canyon operations office—Program and activities: Water accounting reports. Washington, DC: USBR.
US Bureau of Reclamation and National Park Service. 2016. “Glen canyon dam long-term experimental and management plan final environmental impact statement.” Accessed April 4, 2024. http://ltempeis.anl.gov/documents/final-eis/.
Van den Belt, M. 2004. Mediated modeling: A systems dynamics approach to environmental consensus building. Washington, DC: Island Press.
Voinov, A., N. Kolagani, M. K. McCall, P. D. Glynn, M. E. Kragt, F. O. Ostermann, S. A. Pierce, and P. Ramu. 2016. “Modelling with stakeholders—Next generation.” Environ. Modell. Software 77 (Jun): 196–220. https://doi.org/10.1016/j.envsoft.2015.11.016.
Wang, J., and J. C. Schmidt. 2020. Stream flow and losses of the Colorado river in the Southern Colorado plateau. Logan, UT: Utah State Univ.
Wheeler, K., et al. 2021. Alternative management paradigms for the future of the Colorado and Green rivers. Logan, UT: Utah State Univ.
Wheeler, K. G., C. J. Robinson, and R. H. Bark. 2018. “Modelling to bridge many boundaries: The Colorado and Murray-Darling River basins.” Reg. Environ. Change 18 (6): 1607–1619. https://doi.org/10.1007/s10113-018-1304-z.
Wheeler, K. G., J. C. Schmidt, and D. E. Rosenberg. 2019. “Water resource modelling of the Colorado river—Present and future.” In Strategies. Logan, UT: Utah State Univ.
Wurbs, R. A. 2005. Comparative evaluation of generalized river/reservoir system models. College Station, TX: Texas Water Resources Institute.
Yates, D., J. Sieber, D. Purkey, and A. Huber-Lee. 2005. “WEAP21—A demand-, priority-, and preference-driven water planning model Part 1: Model characteristics.” Water Int. 30 (4): 487–500. https://doi.org/10.1080/02508060508691893.
Zagona, E. A., T. J. Fulp, R. Shane, T. Magee, and H. M. Goranflo. 2001. “Riverware: A Generalized tool for complex reservoir system modeling.” J. Am. Water Resour. Assoc. 37 (4): 913–929. https://doi.org/10.1111/j.1752-1688.2001.tb05522.x.
Information & Authors
Information
Published In
Journal of Water Resources Planning and Management
Volume 150 • Issue 7 • July 2024
Copyright
This work is made available under the terms of the Creative Commons Attribution 4.0 International license, https://creativecommons.org/licenses/by/4.0/.
History
Received: Jun 20, 2022
Accepted: Dec 20, 2023
Published online: Apr 26, 2024
Published in print: Jul 1, 2024
Discussion open until: Sep 26, 2024
Authors
Metrics & Citations
Metrics
Citations
Download citation
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.