Integrating social data and engineering solutions for developing resilient water infrastructure against coastal climate change

Abstract

This study combines responses from a social survey with compound flood modeling in a marginalized coastal community to assess implementation of green infrastructure, such as rain barrels and rain gardens, in a city scale. This research focuses on the City of Imperial Beach, CA, which is an underserved coastal community located near the border with Mexico. A principal objective of the present research was analyzing social survey responses and the public’s perceptions to estimate the extent to which decentralized water infrastructure might be accepted by and feasible for the target underserved coastal community. The feasibility of the proposed solution is strengthened through a collaboration with the City of Imperial Beach, which has led to the results of this study being presented in various public sources and forums. The social survey revealed that more than 4/5 of respondents are interested in receiving a rain barrel for free, and the needed financial incentives for rain barrel and rain garden installation can be the whole cost of the practice. Results of the social survey provide promising prospects for the community’s adoption of decentralized water infrastructure, but public awareness and engagement still need to be improved through appropriate outreach activities, particularly in areas at risk of future flooding and sewer overflows. The effectiveness of our proposed solution is assessed through hydrologic-hydraulic model outputs, deploying a fine resolution 2-dimensional overland flow model, present that for a stormdrain system with typical defects (e.g. 0.25% porosity), working under current sea levels (i.e., sea level rise = 0 m), and a typical storm (e.g., 1-year rainfall), the flood volume may decrease 56–99% after implementing a rain barrel system and adding a rain garden system.

Introduction

Compound flooding is experienced in low-lying coastal areas when heavy precipitation coincides with extreme sea-level events, which can increase the flooding extent when compared to these events occurring in isolation^1,2,3. Additionally, recent studies [such as Sangsefidi et al.⁴, Su et al.⁵, Teimoori et al.⁶, McKenzie et al.⁷, and Befus et al.⁸] show that coastal groundwater is dynamically connected to sea levels and with rising sea levels there is an increased risk of groundwater shoaling or emergence with shallow aquifers. Low-lying coastal areas are susceptible to various sources of marine, subsurface, and surface inundation, which can be exacerbated when considering climate change effects on water resources^8,9. Management of coastal drainage systems requires immediate attention to:

• 1. current and future interactions between marine-subsurface-surface water resources and their compound impacts on coastal infrastructure systems; and

• 2. sustainable and community-based solutions for vulnerable coastal areas and assets to adapt to climate change (i.e., sea-level rise (SLR), shallow groundwater, and precipitation intensification).

Decentralized water infrastructure is being increasingly deployed for stormwater and wastewater management because the development potential of centralized water infrastructure may not keep up with the urbanization growth^10,11,12. RainWater Harvesting (RWH) and Green Infrastructure (GI) are two popular techniques for decentralization, which can mitigate flood impacts and combined sewer overflow potentials by reducing the magnitude of precipitation-based runoff^13,14,15,16. RWH mostly refers to rain barrels connected to building gutter systems, which collect and store rainwater for future use¹⁷. Thus, in addition to runoff reduction, RWH can adjust the water budget on a local scale (i.e., providing new water sources and reducing water stresses)¹⁸. As an environmentally-friendly mitigation strategy, GI includes a variety of techniques (e.g., rain garden, infiltration basin, green roof) to mimic natural or pre-development hydrologic processes including stormwater retention, infiltration, and evapotranspiration¹⁹. Besides capturing stormwater from a variety of impervious urban surfaces, GI can remove various stormwater pollutants (e.g., oil & grease, nutrients, heavy metals, and microplastic particles) before discharging into major water bodies^20,21,22.

This Case Study will provide a methodology for how to assess the effectiveness and feasibility of decentralized infrastructure, such as RWH, in a coastal community. The effectiveness and feasibility of this mitigation strategy will be analyzed through the lens of the Intergovernmental Panel on Climate Change Working Group II (IPCC WGII). To critique and advance the framework of the IPCC WGII we reference the sixth assessment report, especially the Chapter “Cities and Settlements by the Sea”. The sixth assessment of the IPCC WGII addresses the challenges of climate action in the context of sustainable development with a specific focus on climate change impacts, adaptation, and vulnerability²³. As defined by the IPCC WGII, effectiveness is the extent to which an action is expected or observed to reduce climate-related risk. Feasibility is the extent to which a measure is considered possible and favorable in a particular setting. There are six dimensions of feasibility: geophysical, environmental-ecological, technological, economic, sociocultural, and institutional. Feasibility can also refer to a set of actions, such as governance capacity, financial capacity, public opinion, interest group pressure, and the distribution of political and economic power^24,25.

The sixth assessment report of the IPCC WGII identifies that GI has a high feasibility to support climate adaptation and mitigation efforts in cities. The report notes that institutional barriers constrain the feasibility of urban green infrastructure, such as policy resistance, to shift priorities from grey to green infrastructure. Furthermore, social and political acceptability of GI is constrained by lack of confidence in efficacy²³. This study aligns with the framework established by the IPCC WGII demonstrating a methodology for the inclusion of citizens in decision making. Through collaborations with City Officials and City Workers, we also demonstrate the ability to overcome some of the institutional barriers.

Coastal cities have adapted to meters of SLR in the past indicating that adaptation is feasible. However, future adaptation options are influenced by variations in project socioeconomic conditions and rates of SLR, which can create uncertainty, subsequently affecting judgements about the comparative effectiveness and feasibility of alternative options^24,26.Various studies have been recently conducted on social barriers to the implementation of RWH and GI, which are summarized in Table 1. According to this table, although social factors are usually difficult to quantify, they can be generally classified into three following categories:

• 1. demographic factors (e.g., age, race, income, and education);

• 2. public engagement (i.e. people’s awareness and attitudes toward environmental issues as well as their participation in decision-making and problem-solving processes); and

• 3. governance discord (i.e. inconsistent strategies within or among governance entities).

Table 1 Selected studies on social barriers in adoption of decentralized infrastructure.

The literature suggests that homeownership status is an important parameter such that renters are less likely to implement decentralized infrastructure²⁹. In addition, people with higher education levels, more positive attitudes toward the environment, higher knowledge about stormwater management, and prior experience with flooding are more likely to adopt RWH and GI practices^28,30,34. Considering the possible correlation between public knowledge and demographic factors, outreach activities may be needed for underserved communities²⁹. The discontinuation of about one-third of rain barrels in two watersheds in Indiana within five years of their installation appears that their maintenance can be inconvenient and undesirable over time. Additionally, people who garden, with the intention of water use reduction in their house yards, could be the most likely potential adopters²⁸.

Effective communication and clear distribution of responsibilities among different stakeholders can facilitate RWH and GI practices, which necessitate the involvement of the general public and the partnership of decision makers, while strengthening the feasibility of city-wide GI installation^35,36,37,38. In this regard, Shuster et al.³⁹ emphasized that besides cost–benefit evaluations of decentralized infrastructure, public engagement strengthens the networks of government agencies and non-governmental organizations (NGOs), which can lead to faster adoption of decentralized infrastructure systems on larger scales. It is worth noting that communities do not necessarily have the same priorities as infrastructure managers²⁷. For example, while stormwater management benefits remain the primary focus of decision-makers in New York City (as one of the leading cities in GI implementation by investing ~ 1 billion dollars), its residents valued other ecosystem services more (i.e., reduction of air pollution and urban heat)^31,32.

To address the compound flooding afflicting the City of Imperial Beach (Fig. 1) this study promotes coastal community resilience against climate change through mitigating flood impacts (while improving the water budget) in a watershed scale. There are low rates of decentralized infrastructure adoption by urban residents regardless of their high potential in addressing stormwater quantity and quality issues. Thus, the success of decentralized infrastructure heavily relies on understanding the barriers and motivations that communities face in implementing and operating such systems²⁴.

Fig. 1

(a) Visualization of IB’s stormdrain system along with the eight watersheds used for residency identification in the social survey, and the monitoring site location used for hydrologic validation, (b) Location of Imperial Beach (red circle) in California (grey outline) along the border of the United States and Mexico (black outline). Created using ArcGIS Pro version 3.2.2, https://www.esri.com/en-us/arcgis/products/arcgis-pro/resources.

Study area

The current study focuses on Imperial Beach (IB), San Diego County, California, USA. This low-lying coastal community is surrounded geographically by water bodies from three sides (The Pacific Ocean, San Diego Bay, and Tijuana Estuary), and it has historically experienced marine flooding due to its unique setting^4,40. The aged and defective water drainage infrastructure of IB is especially vulnerable to compound impacts of seawater-groundwater-stormwater stressors when heavy rainfall coincides with high sea-level events^4,40. In these challenging circumstances, IB’s stormdrain system (Fig. 1) is prone to compound flooding since significant portions of the stormdrain system capacity have been already occupied by SeaWater Intrusion (SWI) and GroundWater Infiltration (GWI) through immersed outfalls and defects, respectively.

IB is considered an underserved and underrepresented community based on its resident demographics, which makes it highly vulnerable to compound flooding and in strong need of community-specific solutions to build resilience against climate change impacts. California’s Department of Water Resources defines a disadvantaged community as a community with an annual median household income of less than 60% of the statewide average value⁴¹. The corresponding numbers for IB and the state of California are respectively reported as $68,917 and $119,149 by the most current (2021) income statistics obtained from the United States Census Bureau⁴².

IB has an estimated 26,059 population, of which 51.9%, 7.8%, and 5.4% are respectively from minority groups of Hispanic or Latino, Asian, and Black or African American⁴³. The owner-occupied housing unit rate in IB is 31.8%, which is substantially lower than that for San Diego County (54.1%), California (55.5%), and the US (64.6%). In addition, as low as twenty-three percent of IB residents have a bachelor’s degree or higher, compared to 40.3% of San Diego County residents. Moreover, the poverty and uninsured rates are 13.4% and 13.6% in IB while the corresponding numbers for San Diego County are 10.7% and 8.2%, respectively⁴³.

This study aims to develop resilient adaptation measures by integrating hydrologic-hydraulic criteria with responses to social surveys. As one of the first known attempts to gather the perspectives of the residents of an underserved coastal community on barriers and motivations to adopt decentralized infrastructure, the social survey is critical for this research to advance public knowledge and mitigation practices in vulnerable communities. A one-dimensional (1-D) stormwater model presented in Sangsefidi et al.⁴ is extended to simulate rainwater capture through decentralized infrastructure as well as flow propagation in two dimensions (2-D) over the land surface under coastal climate change (i.e., SLR, groundwater rise, and more intense rainfall. For decentralization, this study focuses on two popular practices for RainWater Harvesting (RWH) and Green Infrastructure (GI), specifically Rain Barrels (RB) and Rain Gardens (RG). Information and examples of these systems can be found on the Environmental Protection Agency’s website. The compound flood model is run with scenarios listed in Table 2, namely with sea-level rise scenarios, varying rainfall, varying porosity (P), and rainfall intensification. The effectiveness of these scenarios is assessed to demonstrate the use of decentralized infrastructure as a viable solution for flood mitigation. The goal of the present research is to assess flood mitigation with 24-h design storms. This approach allows for a robust analysis of peak flows for frequent flooding events. Continuous rainfall is utilized in the scenario with RBs in order to determine the long-term volume of water harvested. Thus providing information on the supplement water source of harvested rainwater over a longer period of time. Furthermore, a baseline sea level is compared with an increased sea level of 2 m to correspond with projected sea levels along the continental United States coastline⁴⁴. More detailed presentation of sea level rise scenario selection is presented in Sangsefidi et al.⁴.

Table 2 Studied scenarios in the present research.

The principal objectives of the present research are to advance the framework of the IPCC WGII by analyzing social survey responses and the public’s perceptions to estimate the extent to which decentralized water infrastructure might be accepted as a feasible solution by the target underserved coastal community; evaluating the capability and effectiveness of decentralized water infrastructure in reducing precipitation-based runoff (and subsequently mitigating the extent of overland flooding) under different climate change scenarios and adaptation measures.

Another key advancement in this paper is extending the previous base model from a 1-D model to a 2-D model. When modeling flooding in urban areas that have various urban structures and obstacles impacting flow direction and routing, 2-D modeling has been shown to have greater accuracy and model performance^46,47. Despite their computational efficiency, 1-D urban flooding models are unable to simulate the flow propagation in overland areas. For simulating overland flow propagation, PCSWMM, a numerical stormwater simulation software, solves the conservation of mass and momentum (Saint–Venant equations) in two dimensions of the land surface and one dimension along the underlying drainage infrastructure. Figure 2 illustrates how the overland system is discretized into the 2-D mesh and connected to the 1-D underlying drainage system visualized in the PCSWMM model. The model is an integrated 1-D/2-D model, where the 2-D portion computes overland surface flooding.

Fig. 2

Connections between 2D overland mesh and 1D drainage system.

The PCSWMM model, explained in the previous section, is extended to simulate the watershed-scale performance of RB and RG systems in reducing hydraulic loadings on vulnerable centralized infrastructure. A RB system also can adjust the water budget by providing new sources for both indoor and outdoor end uses. The SWMM engine is a widely used framework for developing and sizing decentralized infrastructure⁴⁸.

A schematic diagram illustrating RB flow routing in PCSWMM is presented in Fig. 3. As one of the baseline scenarios, the County of San Diego Rain Barrel Tutorial⁴⁹ recommends a 600-gallon (or 2.271 m³) volume of RB for each 1000-ft² (92.9 m²) rooftop area receiving 1-inch rainfall. The feasibility of this recommendation is confirmed in later sections, and the adoption rates of the decentralized infrastructure is estimated based on social study outputs. Health-related risks (e.g., the attraction of mosquitoes) may increase with RWH storage time exceeding 48 h (County of San Diego Standard Urban Stormwater Mitigation Plan, 2012).

Fig. 3

Schematic view of flow routing in a hydrologic unit with an RB system.

Results

Analysis of survey responses

Considering the high capabilities of decentralized water infrastructure in reducing stormwater runoff and pollution as well as their low rates of adoption by local communities, this section analyzes the responses of IB’s residents to the social survey to explore factors influencing the adoption of RWH and GI (especially RB and RG). The statistics of responses to numeric and open-ended questions of the survey are described at the end of this section. The survey served as a way for citizens to engage in the decision-making process, a vital part of ensuring a solution is feasible³⁸.

Table 3 presents demographic data between Imperial Beach, San Diego County, the State of California, and the entire United States. Although the homeownership rate in IB is less than one-third (31.8% from Table 3, more than two-thirds of the survey respondents are homeowners (68.9% of responses to Q1 of the survey), which represents their higher public participation compared to home renters. It is worth noting that during in-person data collection, some potential respondents, regardless of housing status, systematically excluded themselves from our study because they did not identify as having the ability to participate in RWH. This was due either to space limitations of where they live or other rules/restrictions prohibiting RWH such as at some condominiums or townhomes. The limited involvement from property owners affects the feasibility of this solution. During the City Council Meeting on November 1st, 2023, the Mayor and City Council members proposed that a facilitated public forum could aid in the involvement of property owners that rent out the homes in Imperial Beach, indicating that this barrier could be overcome. City Planner Reyna Ayala has also stated Imperial Beach could hold community forums and partner with non-profit organizations, such as the Water Conservation Garden, to help tenants understand how they can get involved⁵⁰. Thus, this barrier has the ability to be overcome with proper communication.

Table 3 Comparison of demographic factors [obtained from United States Census Bureau⁴³.

As expected from Q2 of the survey, the majority of respondents (80.2%) were aware of IB’s ongoing stormwater issues such as stormwater pollution. Over the past 3 decades, the city has experienced sewage pollution within the Tijuana River Watershed, which is a large binational watershed lying across Mexico–California border. The watershed encompasses the densely urbanized City of Tijuana (Mexico), and it ultimately drains untreated sewage into the Tijuana River Estuary in the City of IB⁵¹.

The interactive analysis of responses to Q1 and 2 (shown in Fig. 4) indicates that 89% of homeowners and 62% of renters were familiar with IB’s water issues (this number is 100% for business owners regardless of their homeownership status). This result implies that ownership can enhance people’s attention to their community issues in order to protect their properties. Another reason for the low participation rate of IB’s renters can be their specific limitations in implementing RB and RG (leading to their self-exclusion from the survey). Dividing the city of IB into eight aggregated watersheds, responses to Q3 are analyzed for identifying the residency of survey participants, and their spatial distribution is presented in Fig. 5. The highest survey participation across the city was respectively in watershed B (33 responses) and watershed F (25 responses). The watershed lettering refers to labels presented in Fig. 1. Since the counts are collected over unequal areas (or populations), the number of survey responses is normalized by the area of watersheds using Eq. (1), which factors out the size of the domain from survey participation. The highest rates of relative participation respectively belong to watershed W (34.3%), watershed B (19.9%), and watershed D (16.0%), which have historically experienced marine-based flooding as shown in^4,40.

$$\begin{gathered} r_{i} = \frac{{I_{i} }}{{\sum\nolimits_{m = 1}^{8} {I_{m} } }} \hfill \\ \left\{ {\begin{array}{*{20}l} {r = normalized\;participation\;rate\;(in\;percent);} \hfill \\ {I_{i} = 100\;n_{i} {/}A_{i} ;\;i = watershed\;representative\;index;} \hfill \\ {A = watershed\;area;n = number\;of\;survey\;responses\;in\;a\;watershed.} \hfill \\ \end{array} } \right. \hfill \\ \end{gathered}$$

(1)

Fig. 4

Variations of familiarity with the city’s stormwater issues and likelihood of adoption with ownership status.

Fig. 5

Spatial distribution of survey responses across the city: (a) absolute number of responses in each watershed; (b) normalized number of responses using the area of watersheds. Created using ArcGIS Pro version 3.2.2

The studies of Gallien⁵² and Merrifield et al.⁵³ revealed that even in current conditions, the thin peninsular strip in watershed W is notably vulnerable to wave-driven impacts (leading to the highest participation rate in this watershed). By the same reasoning, the low participation rate can be justified for watershed S where flood impacts are minimal (discussed in the next section). It is worth noting that public awareness of flooding issues needs to be improved particularly in watershed F, which will be at a high risk of compound flooding in the future (discussed in the next section).

Corresponding to Q1, 4, and 5 of the survey, Fig. 6a illustrates the familiarity of IB’s residents with RWH and GI concepts as well as their specific examples of RB and RG systems. As shown, for all studied decentralization practices, homeowners have a higher level of familiarity and usage compared to renters. In addition, both IB’s homeowners and renters are more aware of RB (and RG) compared to RWH (and GI). For example, about 10% (and 20%) of participants familiar with RB responded that they are not (and slightly) familiar with RWH. This point can imply that the public may have a deeper perception of an engineering application rather than the more general concept itself. Figure 6a also indicates that there is a higher level of public familiarity with RB (or RWH in general) compared to RG (or GI in general), which leads to higher rates of RB implementation and usage in both rented and owned houses. Thus, it can be concluded that an effective knowledge transfer among stakeholders (which can be achieved through public engagement programs) and sufficient considerations of ownership constraints are crucial for decentralized infrastructure to be successfully accepted and implemented by the community.

Fig. 6

(a) Public familiarity with decentralization practices and their usage in IB, (b) Residents’ interest in receiving an incentive for RB and RG installation and its correlation with the likelihood of RB and RG adoption.

Survey respondents showed higher interest in receiving an incentive for participating in RB practices (Fig. 6b) such that all participants interested in an RG incentive (Q9) already showed their interest in an RB incentive in Q7. It can be associated with the more familiarity of the residents with RB (Fig. 6a). The interactive analysis of responses to Q6, 7, 9, 11, and 12 of the survey (shown in Fig. 7a indicates that people interested in receiving financial incentives for RB and RG installation have smaller water bills compared to others. It is probably because interested people are more likely to adopt practices (b, and adopters generally have more positive toward water resources than non-adopters²⁹.

Fig. 7

Variations of (a) monthly water bill of respondents with and without interest in RB and RG incentives; (b) needed incentives for RB and RG installation.

The County of San Diego Rain Barrel Tutorial⁴⁹ recommends a 60-gallon volume of RB for each 100-ft² rooftop area receiving 1-inch rainfall (whose capability has been proven in the current study). While the estimated cost of a 60-gallon barrel system is $100–$120^54,55, Fig. 7b indicates that IB’s residents need a financial incentive of $115.10 on average (which is about the whole cost of RB installation). By assuming that (i) the catchment area of IB’s nonpublic parcels is ~ 2.7 times of that of rooftop footprints (determined by GIS analysis on urban parcels and buildings), (ii) RG surface area is 5% of the catchment area⁵⁶, and (iii) RG installation typically costs about $10 a square foot⁵⁷, the estimated cost for RG installation in a property with 100-ft² rooftop (equivalent to the catchment area for a 60-gallon RB) is $135 [= 100 ft² × 2.7 × 0.05 × $10/ft²]. With an average value of $137.49, the needed financial incentive for RG installation also can be the whole cost of the practice. It is worth noting that 6% to 8% of interested participants in receiving an incentive (answering “Yes” to Q7 and 9) selected “0” as a needed financial incentive in Q11 and 12, for whom a different rebate program can be considered to incentivize RB and RG practices.

Figure 8 (corresponding to Q13 and 14) illustrates how different potential benefits of RB and RG practices may interest IB’s residents. From Fig. 8a, garden watering by an RB system will be the most favorite benefit of decentralization, and gardeners with the intention of water usage reduction in their yards should be targeted as the most prevalent adopters of stormwater conservation practices. This finding is consistent with previous studies [i.e., Newburn et al.⁵⁸ and Gao et al.²⁸] reporting that households with vegetable or flower gardeners have much higher awareness and adoption levels than those with no gardener.

Fig. 8

Residents’ interests in potential benefits of RB and RG practices: (a) percentage of interested respondents in the city; (b) absolute number of interested respondents in each watershed. Refer to Fig. 1 for Watershed Lettering.

The second, third, and fourth reported motivators for decentralization are helping the community with expensive pollution clean-up projects, water pollution reduction, and flood prevention using a RG system. Therefore, in addition to compound flooding control, water quality also should be addressed as a serious environmental issue and a strong motivation for RG adoption in IB. Figure 8a also shows that a “beautiful low-maintenance landscape” using RG interested about ninety percent of respondents. On average, people have shown more interest in RG benefits although they are generally less familiar with it compared to RB. Thus, it is crucial for future outreach activities to include clear descriptions of decentralization benefits. Figure 8 demonstrates how to explore the values and perceptions of the target community in order to accurately develop the solution space.

The start-up and maintenance difficulty of an RG system is listed as its top barrier in the perception of people (mentioned in 31.3% of open-ended responses to Q10), which also should be addressed in future educational programs. This concern, however, is listed as a minimal barrier to RB and RWH adoption (mentioned in 3.2% and 6.5% of responses to Q8 to Q19, respectively). The first concern of respondents about RWH practices (mentioned in 35.5% of open-ended responses to Q19) is the infrastructure feasibility including installation and plumbing efforts (besides site-specific restrictions). Thus, compared to “watering plants and/or lawn”, residents will be less likely to use an RB system for “toilet flushing” [Fig. 8a]. The additional concern on the quality of harvested rainwater (including rooftop contaminants and attracting insects like mosquitos as mentioned in 12.9% of responses to Q19) makes people least interested in doing “laundry” by an RB. As a result, while IB’s residents do laundry 38.8% more frequently than garden watering (respectively with the weighted averages of 2.29 and 1.65 times per week as shown in Fig. 9), the former interests them ~ 3 times less [38.0% as mentioned in Fig. 8a]. Figure 8 indicated that watering plants/lawn and doing laundry using harvested rainwater from a RB system are respectively the most and least favorite benefits of decentralization in all eight watersheds. With the highest number of respondents, the general order of motivators in two watersheds B and F is similar to that of the whole city. Addressing the major concerns of those with reservations about RB and RG could improve the feasibility of this approach to mitigate flooding. The survey responses have been shared with the City of Imperial Beach’s City Planner’s office. The responses have influenced the guidelines set forth by the City for RB installation (Community⁵⁹).

Fig. 9

Frequency of laundry and garden watering (corresponding to Q15 and 16 of the survey).

From Tables 4 and 5, while cost is the concern of 11.3% of respondents to Q19 for integrating RWH interventions, 81.3% of the respondents to Q17 said they are interested in “receiving a rain barrel for free”. In addition, a high percentage of respondents to Q17 (77.9%) are interested in having educational resources on RWH practices, which is reflected in responses to the three open-ended questions of Q8, 10, and 19. Moreover, in the case of building their own RB or RG, people are more interested in receiving information on the latter (Q17), which requires more efforts for start-up and maintenance in their perception (Q8 and 10).

Table 4 Statistics of responses to conducted survey by the Social-Science part of the research team.

Table 5 Coded responses to the open-ended questions of the survey conducted by the Social-Science part of the research team under the supervision of Dr. Megan Welsh.

Survey participants are eventually asked in Q18 to scale the likelihood of RB or RG installation into their house or building from 1 (extremely unlikely) to 5 (extremely likely). Overall, the average value of 3.53, with a potential for improvement through public engagement and education, provides promising prospects for the adoption of decentralized water infrastructure by the community. However, this parameter is smaller for renters (3.23) with two-thirds of the city population, which can be attributed to their lower levels of familiarity with stormwater issues and decentralization practices. Figure 5a illustrates the distribution of this parameter in different watersheds.

As expected, the likelihood parameter is the most and least in watersheds of W and S (Fig. 10a), which are currently at the highest and lowest risk of flooding, respectively (discussed in the following section). This parameter is 3.04 (less the average) in watershed F, which will be substantially impacted by compound flooding in the future (discussed in the following section). Thus, appropriate outreach activities are needed to increase awareness of compound flooding impacts and mitigation solutions in the region. Likelihood parameters provide a quantifiable measure to assess the feasibility or acceptance among the citizens for a proposed solution.

Fig. 10

(a) Distribution of the likelihood of decentralized infrastructure adoption across IB; Adoption rates for (b) RB system corresponding to urban buildings; (c) RG system corresponding to urban parcels. Created using ArcGIS Pro version 3.2.2

Compound flood mitigation through use of rain barrels and rain gardens

This section presents the effectiveness of decentralization on the impacts of compound flooding extent for both current and future conditions. The outputs of the conducted social studies for modeling and implementing decentralized infrastructure are mainly:

• confirming the general acceptability of decentralized infrastructure by the community (average value of the likelihood parameter = 3.53 out of 5.00);

• routing RB outflow to the pervious area (considering the high interest of residents in garden watering) with a 24-h drain delay; and

• more importantly, estimating spatial distribution of adoption rates of decentralized infrastructure by combining the likelihood parameter for different stakeholders with urban land-use units (presented later in this section,

As pointed out previously, the County of San Diego Rain Barrel Tutorial⁴⁹ recommends 60 gallons (or 271 L) volume of RB for each 100-ft² (9.29 m²) rooftop area receiving 1-inch (or 25.4 mm) rainfall, which covers almost all hourly rainfalls and the majority of daily rainfalls in IB (marked in Fig. 1). Firstly, the capability of the above-mentioned recommendation has been verified for the City of Imperial Beach using sensitivity analysis (performed by varying one independent variable at a time). Figure 11 indicates a sharp drop in the total peak runoff (i.e., summation of peak values of stormwater runoff over all subcatchments) from a rainfall with 1-year return period by implementing an RB system and increasing its volume to the recommended value (V_RB → V_rec). However, larger RBs (V_RB > V_rec) lead to higher costs without any significant gains for capturing frequent rainfalls (i.e., a 1-year return period). From Fig. 12, the stormdrain inflow and depth (at the data measurement location specified in Fig. 1) also show the same trend with RB volume. In addition, RB impacts on peak values of flow characteristics are more remarkable compared to their average values. For more intense (i.e., larger return period) storm events there, would need to be an increase in the number of RBs over the recommended value to see any significant decrease in peak runoff.

Fig. 11

Variations of total peak runoff by changing RB volume from the recommended value.

Fig. 12

Stormdrain flow characteristics for different RB volumes (rainfall with 1-year return period; location specified with green filled star in Fig. 1a). Inflow maximum and average values appear on the left side of the figure, depth maximum and average values appear on the right side of the figure.

Figure 13 shows that by implementing the RB system, the peak stormdrain depth at the mentioned location reduces by 31.2% and 4.6% during 1- and 5-year rainfall events, respectively. Thus, the effectiveness of RBs in reducing the depth of flooding is limited to more frequent (or less intense) rainfall events. This is because the contribution of an RB system in rainwater capturing is restricted to building rooftops, which consist of ~ 25% of IB’s urbanized area. However, the mentioned percentages increase to 71.3% and 25.2% after adding the RG system. Therefore, for capturing rainfalls with higher return periods, the effectiveness of this solution could be improved through the implementation of a RG and RB systems jointly.

Fig. 13

Stormdrain flow characteristics in the presence and absence of the RB system for rainfall events with (a) 1-year and (b) 5-year return periods (location specified with green filled star in Fig. 1).

The presented results in Fig. 14 reveal that with a 2-m rise in the sea level, a slight system porosity of 0.25%, and the projected 25 percent uplift the intensity of a 1-year rainfall, the flood area (A) and volume (V) will grow up to 2.6 and 5.8 times, respectively. However, these parameters may reduce 31.4% and 51.8% by avoiding GWI into the stormdrain system (P = 0). While this figure confirms the capability of the RB system in flood mitigation for current conditions (47% and 56% reductions in A and V), the RB system implementation can reduce A and V only by 19% and 24% for the defective system at future conditions. By adding an RG system, the mitigation of future flood area and volume can be improved up to 70% and 77% while eliminating the current flooding extent. These strategies are appropriate for more frequent events such as the 1-year and 5 year events. The negligible return from RBs for the 5-year event and the limited return from the additions of RGs reveals the need to explore different strategies for reducing peak flows from more intense rain events (e.g., a 10-year return period) These results demonstrate the variability in effectiveness corresponding to varying levels of GWI. The following sections present the variable effectiveness with the consideration of additional compound coastal stressors (i.e., sea water intrusion, GWI, and stormwater runoff).

Fig. 14

Effects of RB implementation on flood extent during a 1-year return period rainfall: (a) total flooding area; (b) total flooding depth; (c) relative flooding volume in the presence and absence of RBs.

Figures 15, 16, 17 and 18 present flood maps for some selective scenarios mentioned in Table 2. As shown in the first map of Fig. 15, some areas are frequently flooding even in current conditions (mostly driven by rainfall). When comparing the first two maps in Fig. 15, it can be found that the compound coastal stressors can substantially increase the flood extent in the future conditions, in 6% of which the flood depth will be more than 60 cm (~ 2 feet) and a serious threat to urban buildings⁶⁰. Comparison of the 1st and 2nd maps of Fig. 15 with the maps of Fig. 16 shows that the RB system can be only effective for mitigating the flood extent in current conditions (not future conditions, in which 4% of the flood depth will be still more than 60 cm in 4% of the impacted areas). From the maps in Fig. 17, the combination of RB and RG system can be effective for both current and future flood mitigation. The gains will be even more by avoiding GWI through an effective maintenance of the system (P = 0 in the maps of Fig. 18).

Fig. 15

Compound flooding maps along with the frequency of future flood depth. Created using ArcGIS Pro version 3.2.2.

Fig. 16

Flooding scenarios with rain barrel implementation. Created using ArcGIS Pro version 3.2.2.

Fig. 17

Flooding scenarios with rain barrel and rain garden installation. Created using ArcGIS Pro version 3.2.2.

Fig. 18

RB and RG as well as reduced porosity. Created using ArcGIS Pro version 3.2.2.

From the first two maps in Fig. 15, the stormdrain line near/parallel to the coastline (located on Seacoast Dr. and Palm Ave.) is currently at a high risk of compound flooding. However, in the future conditions, some areas far from the surrounding water bodies will be substantially impacted by compound flooding due to emerging climate change effects on water resources (SLR, shallow groundwater, and more intense precipitation). The area shown by dashed-line rectangle (located near Imperial Beach Blvd.) will be mostly impacted, for which the flood extent is replotted in Fig. 19 for the different studied scenarios. Therefore, watershed F, which has a low likelihood for adoption (shown in Fig. 10a), needs special attention from the city to motivate the public for implementing decentralized infrastructure systems. However, SLR planning projects for the areas near the coast should be focused on marine-based stressors (e.g., living shorelines and revetments).

Fig. 19

Flood extent in the selected area on the second map of Fig. 15 (red, blue, and green polygons refer to the scenarios without RB/RG, with RG, and with RB and RG). Created using ArcGIS Pro version 3.2.2.

The benefit of implementing an RB system is twofold. Besides its contribution to flood mitigation, the captured rainwater by an RB system can adjust the water budget on a local scale (providing new sources of water supply for both outdoor and indoor end uses). RB cost and volume typically have a linear relation for nominal sizes such that the conservative prices of $100 and $1000 can be considered for 60-gallon (or 271-L) and 2000-gallon (or 7571-L) barrels, respectively⁴⁸. By combining these cost estimations with the recommendation of the County of San Diego Rain Barrel Tutorial⁴⁹ on barrel sizing, one can obtain Eq. (2) for estimating the cost of RB installation for a building with a given rooftop footprint. It should be noted that homeowners are assumed to undertake operation and maintenance efforts, which are negligible according to⁴⁸.

$$\begin{gathered} Y = \left[ {0.464 \times \left( {V_{{RB}} - 60} \right)} \right] + 100 \hfill \\ V_{{RB}} {\text{ = 0}}{\text{.6}} \times A_{B} ; \hfill \\ \end{gathered}$$

(2)

where: A_B = building rooftop footprint (ft²); V_RB = rain barrel volume (gallon); Y = approximate cost of rain barrel installation ($).

Using Eq. (2) and RB adoption rates (12), RB implementation cost in IB is estimated to be ~ 3.3 million dollars (Table 6). Additionally, the expected life of RB units may exceed 50 years with proper maintenance⁶¹. Based on the continuous simulation of the stormdrain system (performed for the 4-year historical rainfall data presented in Sangsefidi et al.⁴), the city can harvest about 1 M cubic meters over a 4-year period. According to⁶², one cubic meter of tap water costs an average of $6 in California. Thus, the value of harvested rainwater over the lifetime of the city-wide RB system (i.e., 50 years) can be ~ 63.4 million dollars, which is about 19 times larger than the implementation cost. From Table 6, it is also estimated that the RG system can occupy 2.4% of city area and cost about 15 million dollars. Although RBs have benefits for flood mitigation and financial gain, the effectiveness of RB installation is subject to a temporal variation where future projections of sea level rise and more intense rainfall events could lead to a reduced effectiveness. The combination of RB and RG can help to overcome the temporal variation and ensure an effective management strategy.

Table 6 Estimated cost of a city-wide RB system (r1–r5), value of harvested rainwater, cost and area of a city-wide RG system (r6–r11).

Conclusions

This study integrates engineering solutions with social survey responses in the City of Imperial Beach (IB; California, USA). The present research also aims to develop resilient measures for climate change adaptation aligned with the needs of the community. Although implemented in a single location, this work provides a general framework on how to determine the feasibility and effectiveness of a given solution strategy. Through our survey questionnaire we are able to assess the values and concerns of the target community and determine the limitations in the feasibility of GI (specifically RBs and RGs). Through our collaborations with Imperial Beach’s City Council Members, the City’s Public Works Department, and the City Planner’s Office, we demonstrate how to incorporate institutional and political connections to implement solutions. This Case Study has global implications, since it provides a framework on how to facilitate community driven decision making with technical compound flood modeling.

The analysis of the social survey responses reveal the values and motivations of the community which strengthen the solutions feasibility. The responses indicate that property ownership, familiarity with stormwater management practices, and residence in high-risk flood areas correlate with the likelihood of decentralized infrastructure adoption by the community. In addition, compared to RG, there is a higher level of public familiarity with RB leading to higher rates of its implementation and usage in both rented and owned houses. Moreover, “watering plants/lawn” and “doing laundry” using harvested rainwater from an RB system are respectively the most and least favorite benefits of decentralization in the study area. While the start-up and maintenance difficulties are listed as the top barrier in RG implementation, people’s concerns about the quality of RB’s harvested rainwater for indoor use need to be addressed.

Based on the conducted social studies, more than 4/5 of respondents are interested in receiving a rain barrel for free, and the needed financial incentives for RB and RG installation can be the whole cost of the practice. While the results provide promising prospects for the community’s adoption of decentralized water infrastructure, public awareness and engagement still need to be improved through appropriate outreach activities, particularly in areas that are at risk of future flooding and sewer overflow (i.e., watershed F in Fig. 12a corresponding to high-risk areas in Fig. 15).

The spatial distribution of adoption rates of the decentralized infrastructure systems is obtained from the social survey analyses. Through interactions with the local government, we anticipate that there will be an improvement in adoption rates, especially once concerns about RB functionality and installation are addressed. In addition, for rain barrel sizing, the recommendation of the County of San Diego Rain Barrel Tutorial⁴⁹ was used and verified for the City of Imperial Beach. The outputs of RWH modeling reveal that a city-wide RB system can effectively manage flooding and provide a decentralized water supply. First, for current and future conditions (SLR = 0 and 2 m and rainfall with a 1-year return period), RB implementation may decrease the flood volume by 56–57% and 24–29% for defective (with 0.25% porosity) and non-defective (with 0.00% porosity) systems, respectively. Additionally, the value of harvested rainwater over the lifetime of the RB system (i.e., 50 years) can be about $63.4 M while its implementation will only cost $3.3 M.

The model outputs also reveal that in the case of implementing both RB and RG systems, the flood volume may decrease 99% and 77% for a defective stormdrain system (with 0.25% porosity) respectively working under current conditions (SLR = 0 m rainfall with 1-year return period) and future conditions (SLR = 2 m and 25% increase in rainfall intensity). The latter can be improved up to 93% by avoiding GWI through an effective maintenance of the system. It is also estimated that the RG system can cost ~ $15 M for the city and occupy 2.4% of its area.

The economic cost of RGs can lead to indirect financial benefits from improving the aesthetic of outdoor environments, which can lead to increases in housing costs an property values^63,64. Improvements Besides the financial benefit these strategies also provide advantages when considering water quality improvements and groundwater recharge⁶⁵ . Various studies also demonstrate the advantages from a RG system in flood improvements which could result in the avoidance of property damage, thus a financial benefit⁶⁶. Widespread deployments of these systems also presents the opportunity to reduce social inequity and related distribution of environmental social benefits, by supplying green spaces in communities^64,67.

While the focus of the present study was on the vulnerability of IB’s drainage infrastructure systems to compound impacts of climate change, this community may have different environmental priorities and concerns. For example, some of the survey participants expressed that they have higher priorities for the three-decade problem of Tijuana sewage pollution, and some people refused to participate or talk to the survey team about anything else. Therefore, a bigger and richer picture needs to be painted regarding other ways that IB has been historically underserved.

This Case Study supports the framework of the IPCC WGII for the Climate Resilient Development for Cities and Settlements by the Sea by providing the people with a clear approach as to how researchers can bridge science with community engagement. This work successfully ensures the feasibility of a proposed climate adaptation strategy by providing a platform where inclusive decision-making arrangements are made to enable participation, mange conflicts and trade-offs, promote local ownership, and promote equity and justice in coastal adaptation plans and policies.

Materials and methods

Methods

Figure 20 presents the workflow carried out in the present study, which mainly aims to develop resilient adaptation solutions by integrating hydrologic and hydraulic (H&H) criteria (top part of the figure) with the community’s perception of stormwater management measures (bottom part of the figure). The H&H modeling outputs and social survey responses inform the recommended mitigation strategies. These factors also allow for an assessment of whether the solution can be considered effective or feasible for the target community.

Fig. 20

Framework of coupled engineering and social studies.

Based on this framework, decentralized infrastructure systems are implemented in the watershed-scale stormdrain model (originally presented in Sangsefidi et al.⁴) to evaluate their capabilities in flood mitigation. The findings are communicated with the Social-Science part of the research team to be reflected in the social survey. Estimating the extent to which the engineered designs might be accepted by the target community, the stormdrain model configurations are then refined to properly address stakeholders’ preferences in adaptation strategies and predict their capabilities in mitigation of overland flooding. The utilized methodologies for engineering and social studies are described in the flowing sections.

The Geometric and Geological parameters of the RG system are set by referring to local sources of The City of San Diego Whitebook⁶⁸ and The City of San Diego Stormwater Standards⁶⁹. These parameters are described in Table 7. The RG system implementation was limited to areas with a distance more than 0.601 m (or 2 ft) from groundwater table. Table 2 presents the studied scenarios for evaluating the performance of the stormdrain system with and without decentralization.

Table 7 Selected values for RG system parameters.

As the first step for the 2-D model development, a bounding layer is defined to specify the extent of the overland system. As a starting point, the bounding layer was considered equal to the IB’s urbanized area and discretized with a rough mesh. Considering the high computational cost of 2-D simulation, it was then limited to the areas prone to flooding in the most pessimistic scenario (SLR = 2 m, P = 0.25, and 25 percent more intense rainfall with no decentralized system) by increasing the model resolution. For high-resolution modeling of distributed overland flow, the ground surface (i.e., the finalized bounding layer) is discretized generally by hexagonal grid cells with the size of 4 m, which is significantly finger than the 10-m resolution of previous studies. This size is optimized based on the required resolution and computational resources (i.e., a personal computer with 128 GB RAM and 36 logical processors of an Intel Core i9-10980XE CPU @ 3.00 GHz). Each grid cell will contain basic hydrological and hydraulic inputs such as elevation (taken from the DEM layer) and Manning’s roughness coefficient (ranging from 0.025 to 0.033) (Computational Hydraulics International, 2023).

In order to perform the hydrologic-hydraulic modeling, the developed and calibrated PCSWMM model presented in Sangsefidi et al.⁴ is used as the base model. In this chapter, the base model is extended to simulate (i) rainwater capture through decentralized infrastructure and (ii) flow propagation in two dimensions of the land surface under a changing climate.

The model includes 122 fine-resolution subcatchments, 263 conduits/junctions, and 25 outfalls. The rainfall-runoff transmission over urban subcatchments (hydrological modeling) is performed using a nonlinear reservoir model coupled with Manning’s equation. Flow routing into the system (hydraulic modeling) is simulated by solving the complete form of Saint–Venant equations (referring to an unsteady and non-uniform flow) in one dimension along conduits.

As shown, a 2-D node layer is generated based on DEM layer data for representing the floodplain topography. The final mesh layer consists of more than 20 thousand 2-D cells (and junctions). Each junction from the 1-D system is connected to its closest 2-D junction through a side orifice using the “Connect 1-D to 2-D” tool in PCSWMM. This connection allows a free transfer of flow from the underlying 1-D system to the 2-D overland areas. By defining an obstruction layer based on IB’s building footprints, the size and shape of 2-D cells are modified to account for structures & barriers that impede the water flow motion (Fig. 2).

Boundary conditions are applied using the “Create Boundary Outfalls” tool in PCSWMM, which uses a downstream polyline layer to create outfalls on the 2-D domain boundary and connect them to existing 2-D nodes through new 2-D conduits. In the intersection of the bounding layer and marine inundation area, the boundary condition is set to a fixed value corresponding to the sea-level elevation at MHHW for a given scenario. For other areas, the bounding layer was large enough for water not to cross (or touch) the boundary. Finally, GIS analysis is performed on the stormdrain model outputs (e.g., flood depth and area) to provide floodplain maps and identify in-danger areas for different climate change scenarios and adaptation strategies.

In PCSWMM, the underdrain flow of an RB system is modeled as a submerged orifice and governed by Eq. (1), which is matched to meet the drain time requirement. In this equation, Ud (in/h) represents underdrain flow, and its product with RB area yields the flow rate of demands. The drain coefficient is considered C = 0.5, which allows typical 60-gallon (or 227-L) and 2000-gallon (7571-L) barrels to be drained over 24-h and 48-h periods, respectively⁴⁸. Since the survey results show that garden watering will be the most favorite benefit of decentralization in IB, a 24-h drain delay was established for models to prevent irrigation during (or right after) precipitation events.

$$\begin{gathered} U_{d} = C(H_{d} - h_{d} )^{n} \hfill \\ \left\{ {\begin{array}{*{20}l} {n = {\text{drain}}\;{\text{exponent}} = 0.5\;{\text{representing}}\;{\text{the}}\;{\text{flow}}\;{\text{through}}\;{\text{an}}\;{\text{orifice}});} \hfill \\ {h_{d} = {\text{drain}}\;{\text{offset}} = 0\;{\text{established}}\;{\text{at}}\;{\text{the}}\;{\text{bottom}}\;{\text{of}}\;{\text{barrels}}\;({\text{in}});} \hfill \\ {H_{d} = {\text{water}}\;{\text{height}}\;{\text{from}}\;{\text{the}}\;{\text{barrel}}\;{\text{drain}}\;{\text{offset}}\;({\text{in}});} \hfill \\ \end{array} } \right. \hfill \\ \end{gathered}$$

(3)

Social survey design and administration

The purpose of social surveys (led by the second and fourth authors of this manuscript, who are trained social scientists) is to gather IB residents’ and business owners’ perceptions of flooding and decentralized water infrastructure. (The material of this section (describing development, distribution, and collection of social surveys) is provided by Dr. Welsh (and/or other students working under her supervision) and edited/extended by the dissertation author.) All human subjects research activities were reviewed by the authors’ university Institutional Review Board (protocol number HS 2022-0064). A total of 110 valid responses to the social survey were collected from IB’s stakeholders (i.e., residents and business owners). The target number of survey responses (n) is determined based on the research feasibility and previous social studies by other researchers, where our sample size fell within the range of the studies presented in Table 1, and had a higher ratio of sample size to total population than similar studies. For example, in Miller and Montalto³², the purpose of the study was to examine whether New York City’s GI program can be better aligned with local values,however the sample size only covers 0.005% of the cities population. The people surveyed in our study also spanned the various major watersheds in IB allowing for a clear representation of the entire affected area. In comparison, in Finewood et al.²⁷ 30 stormwater “actors” were interviewed, including practitioners, government employees and activists. This limits the involved from the community in this study. Our study utilizes human subject activities to understand (i) perceptions of flooding and related stormwater pollution issues in IB, (ii) awareness of decentralized water infrastructure, (iii) potential incentives for improving RB and RG adoption by the community, and (iv) barriers to the implementation and maintenance of decentralized water infrastructure.

The designed survey is composed of twenty questions (provided as a supporting document) including 3 yes-or-no questions, 4 multiple-choice questions, 3 multiple-response questions, 6 numeric scale questions, and 4 open-ended questions on the community’s concerns, ideas, and preferences on decentralization and flood prevention. In addition, it includes brief and illustrative information on rainwater harvesting and green infrastructure practices, i.e., to help participants have a better understanding of survey questions and familiarize themselves with the terminology of RB and RG. A 2-min tutorial video on stormwater management practices has been included in the online version of the survey [provided by the United States Environmental Protection Agency⁷⁰]. Using screening questions at the beginning of the survey, the participants are limited to IB residents or business owners aged 18 and more.

The survey questions are developed based on past literature on assessing social barriers to the adoption of decentralized infrastructure mentioned in earlier sections and presented in Table 1 public perceptions of stormwater management practices^71,72,73,74, and water conservation attitudes and incentives^75,76,77. Mindfully adjusted to the socio-demographic context of IB, the questions progress from asking about demographic factors, awareness of decentralized water infrastructure, acceptability of financial incentive programs, and motivating benefits of RB and RG for community-based management of stormwater.

The survey was offered in English and Spanish versions, and respondents were recruited using a mix of online and in-person outreach activities. All data was collected anonymously. On average, the survey took 15 min for participants to complete. Online survey data are collected using the Qualtrics platform⁷⁸. Three in-person engagement events were conducted by setting up a table with information about the project and opportunities to take the survey either by paper or electronically via a tablet. In partnership with the City of IB, these outreach events have been held in the IB’s public library and farmers market in July–November 2022. As noted previously, the City of IB is a key partner in this work, and it has publicly expressed its commitment to utilizing our research to inform its climate change resiliency efforts.

Basic statistics of the survey responses are processed in SPSS 29⁷⁹ and presented in the next section. More descriptive and inferential statistics are calculated in Microsoft Excel. In addition to analyzing the main effects of each parameter on decentralized infrastructure adoption, this study also identifies and quantifies interactions between different parameters (implying that a parameter’s effect depends on the level of other parameters)⁸⁰. Moreover, to gain insight into residents’ perspectives on the benefits and limitations of decentralization, responses to the three open-ended questions of Q8, Q10, and Q19 are coded using an inductive approach (deriving the codes from the dataset). The responses of these surveys provide contextual evidence to determine if the solution of decentralized infrastructure has a high feasibility among the target community.

Effectiveness and feasibility assessment for adoption of rainwater harvesting methods

Feasibility of RWH for the target community is assessed using the responses of the social survey as well as feedback from stakeholders in the local government. The effectiveness of RWH to mitigate the effects of compound flooding is assessed through the model outputs. Factors that impact the feasibility of RWH include likelihood of adoption, financial incentives, qualitative responses, and institutional support from local governance. To perform this assessment, we take into account barriers, enablers, synergies, tradeoffs, while balancing diverse perspectives and values²⁵. Effectiveness is tested under various environmental scenarios, considering anticipated projections of SLR, changes in porosity, and rainfall intensification. Therefore, some uncertainty lies in determining the true effectiveness of RWH in future scenarios.

To further ensure the feasibility of this solution we have also worked directly with the Public Works Department and City Planners Office of the City of Imperial Beach. This partnership helped to determine how to best communicate our findings with the citizens and City Officials. To encourage community participation, we presented our results from the social survey and computer simulations to the City of Imperial Beach at the City Council Meeting on November 1st, 2023. The City Planner’s office has included our results in their Rain Barrel and Cistern Guidelines for Imperial Beach (Community⁵⁹). The development of this document and the support of the City Council members helps to ensure that the concerns expressed by the citizens in the social survey RBs can be addressed, thus improving the feasibility of the proposed solution.

Data

The datasets utilized for stormdrain modeling (ground elevation, sea level, stormdrain system characteristics, and rainfall specifications) are presented in Sangesefidi and Bagheri et al. (2023). Additionally, groundwater intrusion (GWI) estimations have been updated in the PCSWMM model by applying the methodology mentioned in Sangsefidi and Barnes et al. (2023) and using high-resolution estimations of groundwater table^4,4,40,40.

This paper updates the above mentioned models with a new input for stormdrain modeling, a GIS layer of urban buildings is obtained from the data warehouse of the City of Imperial Beach, which was validated through visual comparison with available aerial imagery Fig. 21. This layer has been used for (i) the determination of rooftop footprints needed for the estimation of the number of RBs for each building and (ii) the prevention of water flow in overland areas by defining an obstruction layer.

Fig. 21

Visualization of building footprints in IB. Created using ArcGIS Pro version 3.2.2.

Local sources [e.g., the County of San Diego Hydrology Manual⁸¹ and The City of San Diego Stormwater Standards⁶⁹ are used to set primary parameters for the model (e.g., subcatchment roughness, imperviousness, infiltration, and evaporation as well as conduit roughness and energy loss)^4,40. Other parameters for simulation of overland flooding (e.g., surface roughness) and decentralized water infrastructure (e.g., RB sizing and drainage) are described in the following sections.