Water Utility Climate Alliance
Purpose-Driven Climate Data Selection and Application Case Studies
As water utilities expand climate considerations across business functions and climate hazards— amid a rapidly growing landscape of climate model datasets—selecting data that are truly “fit for purpose” has become increasingly complex. To address this, the Water Utility Climate Alliance (WUCA) Climate Modeling Work Group sought to develop several case studies that would illuminate the factors behind the selection, processing, and application of climate model datasets in planning analyses. Individual case studies can be found in the links below, and the full report can be accessed here.
The purpose of the case studies was not to identify general “best practices” or create formal guidance, which has been done elsewhere (here, here, and here). Rather, it was to capture the specific circumstances and priorities that drove each utility’s decisions—what climate data to use, in what ways, and for what analyses—providing practical, real-world examples for other utilities to learn from. The case studies were informed by interviews with key utility staff and consultants as well as supporting project documents.
Each of the four case studies follows a WUCA member utility through selecting and processing climate model data, establishing a data workflow, conducting project analyses, and applying the results to planning and decision-making. Three of the projects centered on future water supply and/or demand, and the fourth focused on infrastructure flood risk. Two projects were complete at the time of writing, and the other two were in their final phases. Each case study begins with a brief overview of the utility, followed by sections addressing:
• Project context
• Project methods, including data selection and processing
• Results of the analyses
• Use of results in decision support (intended and realized)
• Lessons learned
Each case study also includes links to additional resources that describe the project, climate data, and workflow—such as utility reports and peer-reviewed studies—and a utility contact for further questions.
In all four projects, the workflows began with an ensemble of runs from 15 to 35 CMIP5 or CMIP6 climate models. From there, they followed quite different paths in processing those model runs to construct discrete climate and hydrology scenarios for the subsequent impact modeling (Table below), illustrating that there is no one “right” approach to using climate models to effectively inform planning. Unsurprisingly, the results from all four projects showed the potential for greater climate-related stresses and risks to the utility in the decades ahead—more severe droughts, larger flood events, reduced water supply, and/or increased water demand.
Key attributes of the four case studies
UTILITY | AUSTIN WATER | PHILADELPHIA WATER DEPARTMENT | PORTLAND WATER BUREAU | SAN DIEGO COUNTY WATER AUTHORITY |
|---|---|---|---|---|
Planning/ decisions to be informed | Integrated Water Resources Plan; implementation of
conservation, reuse,
and supply strategies | Planning and design of vulnerable infrastructure assets | Prioritizing water supply system projects in the Capital
Improvement Plan | System optimization and capital improvement projects
to meet future needs |
What was modeled
under climate change in the project analyses | Water supply | River flood levels | Water supply, water
demand | Water demand |
Time horizon(s) | 2030, 2040... 2080, 2120 | 2020, 2030...2090 | 2050 | 2050 |
Climate model dataset used | CMIP6, screened, custom bias-corrected and
downscaled | CMIP5-LOCA (32 models) | CMIP6-LOCA2 (23 models) | CMIP6-LOCA2 Hybrid (15 models) |
Emissions scenarios used | SSP1-2.6, SSP2-4.5, | RCP8.5 | SSP3-7.0 | SSP2-4.5, SSP3-7.0, SSP5-8.5 |
No. model runs/scenarios used in system impact modeling | 15, plus 48 add'l drought scenarios based on those runs | 1 - ensemble mean | 23 (supply); 3 (demand) | 5 |
System impact
models used | Statistical streamflow
models; basin water
availability model | Model to interpolate
peak flows, elevations,
and return intervals from FEMA data | System hydrologic
model; water demand
model; reservoir drawdown
model | Water demand model |
ADDITIONAL INSIGHTS
Practical considerations carry more weight than climate dataset attributes
The case studies show that utilities’ choices in the selection, processing, and application of climate data are conditioned by many factors—including previous experiences with climate datasets and analyses, technical capacity (in-house and consultants), risk tolerance, input requirements for system impact modeling, and whether a dataset is already “at hand.” While the inherent attributes of the climate model datasets (e.g., downscaling method, resolution, number of models) remain relevant, they generally played a secondary role compared to the practical considerations mentioned above.
System-specific impact models are key to making climate data usable
While it is natural to focus on the CMIP climate model datasets at the head of each “chain of models” deployed by the four utilities, the most important link in that chain may be the last one: a system-specific impact model that translates the climate and hydrology scenarios into corresponding system outcomes, couched in the terms familiar to internal and external audiences (e.g., reservoir drawdown dates, peak flood elevations, water supply yield). These models widely varied in their complexity, from a single-spreadsheet regression model to a water availability model involving several programs and hundreds of model nodes. Regardless of complexity, impact models need to balance simulation accuracy with ease of use and interpretability of output. (See WUCA Leading Practice: Develop tools that allow information customization [archived here] for related information and examples.)
Repeated engagements with climate data can build utilities’ internal capacity
All four utilities had previously used similar climate model data to inform their planning, and in three cases, the project workflow was adapted or refined from a previous effort, rather than developed from scratch. (In the fourth case, a similar workflow had been applied in other river basins.) Utility staff reported gaining technical capacity and confidence over time as they became more familiar with climate data and the other links in the chain of models, generally taking on more of the workflow in subsequent analyses. WUCA meetings, training, and resources were cited as important in supporting this progress. Still, every case study project involved at least some level of external expertise—consultants and/or university researchers— to aid in the selection and application of climate model data, though the degree of reliance on these resources varied. (See WUCA Leading Practice: Build and maintain in-house capacity [archived here] for related information and examples.)