Equity in Water Resources Planning: A Path Forward for Decision Support Modelers

Authors: Sarah Fletcher, Antonia Hadjimichael, Julianne Quinn, Khalid Osman, Matteo Giuliani, David Gold, Anjuli Jain Figueroa, Bethany Gordon Year: 2022 Tags: water-resources-equity, decision-support-modeling, participatory-modeling, environmental-justice, sociotechnical-systems, performance-metrics

TL;DR

A forum/position paper arguing that water resources decision support modelers embed inequitable values into their work through framing choices, metric selection, and stakeholder processes—whether or not they intend to. It offers actionable recommendations on procedural equity, equity-aware metric design, and modeling methods, illustrated by two case studies (Ethiopia's Omo River, California drinking water access).

First pass — the five C's

Category. Position/opinion paper (explicitly labeled a Forum paper by the journal; speculative and normative in parts).

Context. Water resources systems modeling and environmental justice; builds on Cash et al. 2003 (salience–credibility–legitimacy framework for knowledge systems), Jafino et al. 2021 (constructive decision aiding with equity framing), Rawls 1999 (difference principle as basis for min-max metrics), and Osman and Faust 2021 (equity vs. equality definitions in water systems).

Correctness. Load-bearing assumptions: (1) skilled facilitation can meaningfully offset power imbalances in participatory processes; (2) social science equity concepts can be operationalized quantitatively without critical loss of meaning; (3) better metrics translate into better policy outcomes. The paper acknowledges tensions in (1) and (3) but does not resolve them. Assumption (2) is taken largely on faith.

Contributions. - Synthesizes procedural and distributional equity concepts and maps them explicitly to modeling choices (framing, metric aggregation, method selection). - Demonstrates with a hypothetical but logically sound example (Fig. 2) that relative variability metrics (Gini index, COV) can penalize policies that genuinely improve outcomes for the worst-off. - Presents two illustrative case studies applying the framework in contrasting global contexts (Ethiopia, California). - Consolidates three concrete, practitioner-facing recommendations: center marginalized communities in model design, advance specific methodological gaps, and collaborate with social scientists.

Clarity. Well-organized and readable for a multidisciplinary audience; the transition from process equity to metric design to modeling methods is logical, though the case studies are loosely integrated with the recommendations that follow.

Second pass — content

Main thrust: Equity failures in water resources systems are partly an artifact of modeling choices (who is in the room, which metrics are used, how outcomes are aggregated); the paper maps these failure modes and proposes corrective practices across process, metric, and method dimensions.

Supporting evidence: - ~1 million Californians lack access to clean, affordable drinking water, concentrated in economically disadvantaged areas (Water Education Foundation 2020, cited). - Post-Katrina: lower-income and minority households received disproportionately less federal disaster assistance relative to damage sustained (Drakes et al. 2021, cited). - Ethiopia's Gibe III: environmental and social impact assessment for transboundary impacts approved only 3 years after construction began (EEPCO 2009, cited). - DAFNE project Negotiation Simulation Lab: participating stakeholders selected 6 negotiated solutions that largely traded irrigation supply and hydropower performance to safeguard recession agriculture and Lake Turkana fish yields (Castelletti et al. 2020, cited). - California Office of Environmental Health Hazards: developed 3 separate affordability indicators benchmarked against median income, county poverty level, and deep poverty level, rather than a single composite (Balazs et al. 2021b, cited).

Figures & tables: - Fig. 1 (conceptual): Contrasts idealized co-production (equal voices) with reality (power imbalances distort outcomes). No quantitative axes; purely schematic. No error bars—appropriate for this purpose. - Fig. 2 (illustrative): Bar chart of hypothetical flood damages per resident under two policies, with Gini and COV values annotated. Data are fabricated to make the logical point; no statistical significance reported or needed. The figure is effective but the hypothetical nature should be flagged to readers who may treat it as empirical. - No tables present in the paper.

Follow-up references: - Cash et al. 2003 — foundational framework on salience, credibility, and legitimacy; underpins the paper's process-equity argument. - Jafino et al. 2021 — constructive decision aiding; most technically developed equity-modeling reference cited. - Butler and Adamowski 2015 — concrete strategies for empowering marginalized communities within participatory modeling processes. - Yoon et al. 2021 (Jordan Water Project) — multi-agent coupled human–natural–engineered system with explicit equity metrics (Gini for disparity, satisficing threshold at 40 L/person/day, consumer surplus); best worked example of the paper's prescriptions.

Third pass — critique

Implicit assumptions: - Skilled facilitation is scalable and broadly accessible to water agencies — not demonstrated; highly resource-intensive and context-dependent. - Marginalized community preferences are internally coherent and articulable in ways that map to quantitative model objectives — elides deep heterogeneity within communities. - Modelers retain institutional freedom to reframe models around community values rather than those of the utility or agency paying for the study — directly contradicted by the authors' own observation that primary model clients are utilities with "neither a mandate nor the capacity to address the fundamental causes of inequity." - Min-max/satisficing metrics are unambiguously preferable to relative variability metrics — true in the constructed example, but the choice is context-dependent and not always clear-cut.

Missing context or citations: - Does not engage with data justice or critical data studies literature (e.g., D'Ignazio and Klein's Data Feminism), which scrutinizes the act of quantifying marginalization itself. - No engagement with water rate design or affordability literature (e.g., tiered pricing research) where the equity–efficiency tension is extensively analyzed. - Transportation equity toolkit examples are invoked as precedent but not evaluated for transferability; the analogy is asserted, not argued. - Legal and regulatory constraints on what utilities can optimize for are not discussed, though they directly limit implementability of the recommended approaches. - No review of how often existing peer-reviewed water resources DSMs include any equity metric, which would establish the empirical baseline gap the paper implicitly assumes.

Possible experimental / analytical issues: - Both case studies are descriptive and qualitative; neither presents outcome data demonstrating that the equity-oriented interventions actually improved distributional outcomes for marginalized groups. - The Gini/COV critique (Fig. 2) is logically sound but rests on hypothetical numbers; a scan of published studies where these metrics produced perverse recommendations would substantially strengthen the point. - The paper simultaneously endorses participatory processes and acknowledges they can "further entrench the framings and policies of the establishment" without providing a decision rule for when co-design is net beneficial versus net harmful. - "Skilled facilitation" is recommended as the remedy for power imbalances without specifying what qualifications, training, or institutional structures constitute skill in this context. - The paper's own framing—written by academics at well-resourced US and European institutions—exemplifies the power dynamics it critiques but does not reflect on this positionality.

Ideas for future work: - Systematic review or meta-analysis of water resources DSM publications to quantify the frequency, type, and aggregation choices of equity metrics actually used, establishing a defensible empirical baseline. - Controlled comparison of participatory modeling processes with and without explicit power-mapping facilitation to test whether acknowledging imbalances measurably shifts outcomes toward marginalized stakeholders. - Development of a decision-tree toolkit for metric selection—operationalizing the Gini/COV pitfall identified in Fig. 2—with guidance on when satisficing, min-max, or distributional metrics are appropriate given system structure. - Longitudinal tracking of the California drinking water monitoring tool (OEHHA GIS) to assess whether transparent equity indicators actually redirect funding and infrastructure investment toward the most disadvantaged communities over a 5–10 year horizon.