In bygone days, communities in the Greater Boston area responded to flooding the old-fashioned way: by enlarging culverts and upsizing storm drains, town by town.

As extreme storms become more frequent, that approach is no longer enough. Engineering for flood resilience increasingly requires understanding how water moves across entire regions — not just within municipal boundaries. Together, the Charles and Neponset River watersheds cover nearly 1,040km² (400 mi²), span 46 municipalities, and serve about one million residents. The two basins are
hydraulically connected: During major storms, water from the Charles can divert into the Neponset
through Mother Brook.

This interconnected system makes clear that no single community can address these challenges alone. To meet that need, watershed associations in both basins launched efforts in 2020 to build hydrologic and hydraulic models that simulate how major storms move across each watershed’s landscape and infrastructure. Updated regularly, these models give communities a broader regional view of flooding than a single community could develop on its own. However, it is important to note that the watershed-scale models provide planning-level information, but do not capture the level of detail that individual cities and towns need to design specific infrastructure improvements and/
or new stormwater infrastructure projects.

Rather, the models are designed to help clarify flood risk and evaluate the benefits of large-scale flood mitigation strategies, from nature-based solutions to policy changes, as well as to identify targeted flood-mitigation benefits from smaller scale infrastructure strategies. They also offer a cost-effective way to understand how upstream actions influence downstream communities — and vice versa — demonstrating why watershed scale planning provides a more comprehensive understanding of flood risks across jurisdictional boundaries.

Flood Risks Growing

Engineers used the models to assess how flood risk may change for different design storms under present-day and future climate conditions.

Their projections show that both watersheds are likely to see significant increases in rainfall intensity under future climate scenarios, making extreme storms increasingly frequent (Figure 1, top). For example, a 25-year storm today (a storm that has a 4% annual chance of occurrence today) could resemble a 10-year storm (a storm that has a 10% annual chance of occurrence) by 2050. Also, an event considered a 500-year storm (0.2% annual chance) now may function like a 100-year storm (1% annual chance) by 2070 — meaning rare events will become much more common.

Using this information, the flood models estimate how 2-, 10-, and 100-year, 24-hour design storms may expand under future climate conditions. Without additional mitigation, both watersheds are expected to see flooding spread over larger areas, with modest increases for smaller storms and more notable growth during extreme events. By 2070, flooded areas during a 100-year event could increase by roughly 14% in the Charles River watershed and 11% in the Neponset River watershed.

The models also show how rising flood levels could affect critical infrastructure. In the Charles River watershed (Figure 2, below), the number of critical facilities exposed to flooding during a 2-year storm is projected to rise from 33 today to 42 by 2070, potentially affecting emergency response and community resilience. In the Neponset River watershed, flooding increases are expected to fall disproportionately on buildings in environmental justice census blocks.

Flood Mitigation Strategies

The models also helped evaluate how well different flood mitigation measures — from watershed-wide strategies to site-specific projects — could reduce future flooding.

Regional-Scale Strategies.

At the watershed scale, engineers examined a range of nature-based and policy-based strategies
to see how much these strategies could reduce the additional runoff expected under future climate conditions. The goal was to avoid increasing the 2070 runoff volumes for the more frequent storm events, such as the 2-year, 24-hour storm, compared to today’s levels, and to understand which combinations of actions can help achieve this goal of “full mitigation.”

In the Charles River watershed, results showed that even ambitious one size-fits-all regional strategies would not fully offset the increases in flooding projected by 2070. The Neponset River watershed showed a similar pattern: Combining multiple approaches significantly reduced runoff under future climate scenarios but still fell short of “full mitigation” (Figure 3, below).

Overall, the modeling made one point clear: Modest or standalone measures will not be enough. Meaningful reductions in future flooding will require more aggressive action, with multiple strategies working together and supported by policy and regulatory changes that consider future climate impacts.

The results also showed that regional strategies perform differently across the watersheds. Their effectiveness varies by catchment depending on land use, soils, topography, and existing infrastructure, highlighting the need for tailored catchment-specific solutions rather than a one-size-fits-all approach.

Catchment-Scale Strategies.

Sensitivity analyses were used to identify which of three strategies — impervious cover reduction, impervious runoff storage, or pervious runoff storage — would be most effective in each catchment (Figure 4, right), and how much of each would be needed to offset future increases in runoff. In many areas, “full mitigation” was achievable, but each catchment required a different mix of approaches.

These results guided the selection of priority catchments for more detailed evaluation. Each priority catchment was then screened for opportunities to implement nature-based solutions, as well as gray infrastructure options. Site suitability was assessed using factors, such as land ownership, environmental constraints, physical characteristics, and engineering judgment.

Overall, this work was designed to help communities understand which types of strategies are feasible and how much flood reduction they might deliver for specific storm events. Results differed by catchment — some areas could achieve the “full mitigation” goal, while others could not. These findings show the importance of tailored, location-specific flood mitigation strategies.

Local-Scale Strategies.

In the Charles River watershed, local priority sites were selected and ranked using catchment-level screenings and municipal staff input, using criteria such as flood reduction potential, community priorities, and equity considerations. Nature-based solutions on these priority sites were then refined by local staff and through site visits before being presented publicly. A number of these projects are now advancing toward implementation.

In some areas, local-scale sub-models were created to provide a more detailed picture of flood behavior. These refined models captured smaller stormwater infrastructure and a more accurate representation of overland flow, offering a better understanding of local flood risks (Figure 5, below). They were also used to test site-specific interventions, allowing the project team to see how different solutions could change flooding at the neighborhood scale.

Spreading the Word

In both watersheds, municipal staff and community partners worked together to share model findings and engage residents. In the Charles River basin, a voluntary partnership of 28 communities helped review flood projections, identify mitigation actions, and propose regional project ideas, while another outreach group met residents at community events and distributed multilingual information.

In the Neponset River basin, modeling results were shared through municipal liaisons and community-based partners who learned about projected impacts and brought that knowledge back to their neighborhoods. School programs and local events added another layer of outreach, helping residents understand how stormwater, flooding, and water quality intersect.

Through a mix of virtual meetings, in-person workshops, and community events, these efforts helped residents understand that future flooding will be more severe, that risks and responses will vary by community and catchment, and that regional collaboration will be essential.

Straddling Two Watersheds

This regional work comes into sharp focus in the Town of Dedham, which sits substantially in both
the Charles and Neponset River watersheds. Its 25,000 residents have a front-row seat to how flood
risks cut across basin boundaries (Figure 6, below). Having access to two sets of model results has given the town a rare advantage — one that is helping refine and reframe its flood-mitigation approach.

While Dedham’s stormwater utility is currently structured around funding compliance with its Municipal Separate Storm Sewer System (MS4) permit, the modeling has pushed staff to consider how it could evolve to support broader, watershed-scale flood mitigation. The empirical data has helped the town pinpoint and prioritize projects in both watersheds, giving staff a more grounded basis for long-term planning.

Access to regional mapping and data has also changed the conversation with residents, elected officials, and other local stakeholders. Discussions now focus less on isolated problem spots and more on shared flooding concerns, regulatory responsibilities, and the need for both private and multi community solutions. Dedham is using its experience to encourage neighboring communities to adopt a watershed-based lens — while raising important questions about prioritizing projects, funding, long-term strategies, and how regional implementation should work.

Lessons Learned

Working across two watersheds that span much of the greater Boston area offered important insights into regional flood risk and mitigation. Key takeaways include:

• Models facilitate economies of scale. Watershed-scale models create cost and technical efficiencies while improving understanding of current and future flood risks across jurisdictions.
• Keep models practical. Adding detail on terrain, surface water, and stormwater infrastructure can improve flood risk accuracy but it can also increase complexity. Modeling requires a balance between detail and practicality.
• Flood risks don’t stop at town borders. Municipal-scale assessments alone limit mitigation options; regional models provide a more comprehensive cross-jurisdictional picture.
• Local refinement is possible. Watershed-scale models can be adapted with additional detail to evaluate site-specific or neighborhood-scale projects.
• Insights are transferable. Lessons from these models can streamline resiliency planning in other watersheds.
• Partnerships matter. Watershed organizations play a crucial role in communication and engagement, broadening participation among residents, community groups, municipal staff, and decision-makers.

Turning Data Into Action

This Boston-area modeling work is already shaping resilience planning and infrastructure investment across the region. Municipalities are using the results to prioritize projects, pursue grant funding, and coordinate with neighboring communities on strategies that extend benefits beyond local boundaries.

The models also create space for clearer conversations about equity and environmental justice. Approximately 25% of the residents in the Charles River watershed and nearly 16% in the Neponset River watershed live in environmental justice communities — areas that often face higher flood risks and have fewer resources to recover. By making these vulnerabilities visible, the work helps direct attention and investments where they are needed most.

Although the Charles and Neponset watersheds differ in size and character, their modeling, engagement, and outreach efforts converge on a shared outcome: a practical framework for understanding risk and advancing resilience that can be scaled and translated in other parts of the
country. These efforts represent a model not just for flood management, but for how collaboration itself can shape a more resilient future.

Published in WET January 2026.