To develop systematic recommendations for evaluating hydrologic dynamics in NBS, we first develop a conceptual model classifying NBS based on three major factors: the type or form of NBS, the scale, and the degree of naturalness. These three characteristics influence monitoring needs, including sensor type(s), the number of sensors, and their placement within the NBS to accurately determine both site water balances and effectiveness as stormwater infrastructure (Figure 2). The type of NBS strongly influences the type(s) of sensors required. For example, a constructed surface-water wetland will require surface water level sensors but generally not soil moisture sensors, whereas soil moisture sensors will be critical in measuring water storage in a green roof (Versini et al., 2016; Ouellet et al., 2021). The scale of the NBS impacts the number of sensors required. A larger natural green space will require more sensors to assess the benefits than a small park or community garden. The degree of naturalness impacts both the type and configuration of sensors. We define the degree of naturalness as the extent of natural soil and ecosystems within the site relative to the extent of constructed hard infrastructure. For example, permeable pavement is an entirely engineered system constructed from human-made (artificial) materials, whereas a native parkland or forest is considered entirely natural. NBS with intermediate degrees of naturalness include a mixture of natural and engineered elements, for example a green roof, bioretention basin, or restored wetland with engineering control of water levels. NBS that are engineered may have an outflow pipe that will be useful to instrument and may require additional in situ sensors.

Based on the key characteristics of form, scale, and degree of naturalness identified in the conceptual model (Figure 2), we selected six common types of NBS for detailed consideration: bioretention basins, rain gardens, green streetscapes, green roofs, nature preserves, and community green spaces. These selected NBS types span commonly used configurations ranging from small features located within individual residential properties to neighborhood-scale land restoration efforts, large municipal projects, and natural areas. Bioretention basins provide an example of semi-natural NBS with some engineered components that provide both surface and subsurface storage and can range in scale from approximately 5–100 m² in area, often even larger depending on the contributing catchment area. Rain gardens are similar in design to bioretention basins but tend to include deep-rooted plants and are often at a smaller scale, as they are frequently used in individual-household residential settings. Green streetscapes include several different types of small scale semi-natural NBS that can be implemented individually or, more typically, as part of a larger infrastructure project, e.g., box tree filters along a long stretch of road. Streetscapes include natural vegetation and engineered elements, such as outflow pipes connected to the municipal stormwater system, or may be combined with other infrastructure, such as permeable pavement. Green roofs represent an intermediate scale engineered NBS that provide predominantly subsurface storage and plant transpiration in an engineered soil layer. Nature preserves, including designated natural areas within urban parks, generally have native vegetation and hydrology, and we therefore consider them entirely natural even though the land may be intentionally modified or restored. These areas range in size from 1,500 m² to over 450,000 m². Lastly, community greenspaces include a variety of different forms of NBS that are designed to provide amenities for local communities, including community gardens, local parks, and other recreational areas. Many community greenspaces are embedded within highly urban areas and are designed for multiple purposes, including stormwater storage.

Recommended sensor configurations for each class of NBS are shown in Figures 3–8, and recommended types, numbers, and costs of sensors are provided in Table 1. The number of sensors needed within an NBS system is determined mainly based on the size and scale of the NBS being monitored, as well as site heterogeneity and the degree of data resolution needed for analysis of site hydrology, which is determined based on the monitoring objectives.

Bioretention basins (Figure 3), and rain gardens (Figure 4) require piezometers with water level sensors, surface water level sensors, and soil moisture sensors to estimate the water storage volume (Khan et al., 2012; Schlea et al., 2014; Paus et al., 2015; Winston et al., 2016; Stewart et al., 2017; Batalini de Macedo et al., 2019; Feldman et al., 2019; Cook et al., 2021; Meixner et al., 2021). The spacing between soil moisture sensors depends on the depth of the system, but in general these should be vertically distributed through the soil column (Stewart et al., 2017; Meixner et al., 2021). In some systems, connectivity between surface water and groundwater has been engineered so that surface and groundwater storage can be measured with one piezometer extending above the ground surface. However, in many cases the presence of an unsaturated zone or buried construction debris limits infiltration and causes a disconnect between groundwater and surface water (Shuster et al., 2014), which necessitates both a piezometer and surface water level sensor to accurately measure total storage volume (Batalini de Macedo et al., 2019). In bioretention basins or rain gardens with a defined inlet, inflow is frequently measured with a weir and water level sensor (Batalini de Macedo et al., 2019; Feldman et al., 2019). For those without a clear inlet point, inflow can be estimated using an overland flow collector (Moruza et al., 2021). In bioretention basins with an outflow or overflow pipe, a flow meter or weir and water level sensor can also be installed to measure outflow when there is interest in determining discharge to the drainage system. In low-lying landscapes, there may be concern that the effectiveness of NBS is limited by a shallow regional groundwater table (Zhang and Chui, 2019). Particularly in flat alluvial landscapes, infiltration from bioretention systems can cause the groundwater table outside the system to rise, decreasing available storage (Thomas and Vogel, 2012; Nemirovsky et al., 2015). Where this is a concern, we recommend installing one or more additional piezometers outside of bioretention basins and rain gardens to monitor the groundwater table. Both pre- and post-construction groundwater monitoring is recommended to understand the limitations imposed by the regional water table and the consequences of NBS on groundwater levels.

In a green streetscape (Figure 5), we recommend installing 1–3 soil moisture sensors distributed vertically through the soil column to measure subsurface stormwater storage within soil pore space (Grey et al., 2018). Where possible, we also recommend installing additional groundwater level sensors to measure the total subsurface water storage (Grey et al., 2018) and/or surface water flow sensors to measure the effects of tree plantings on stormwater discharge. In Table 1, sensor recommendations and cost estimates are provided for both streetscapes at the individual scale (e.g., one box tree filter) and at the city block scale (e.g., ten box tree filters). At the block scale not every box tree filter needs to be instrumented, and our cost estimates are based on instrumenting 50% of box tree filters.

Green roofs (Figure 6) are designed with a shallow substrate and provide storage in soil pore space, in addition to water uptake through plants. Soil moisture sensors are the main instrumentation required (Versini et al., 2016; Ouellet et al., 2021). In addition, many green roofs have outflow pipes that discharge excess water. Non-contact flow meters (i.e., ones that do not block stormwater drainage) or weirs with water level sensors to measure outflow provide useful information for conducting water balances and calculating water storage (Carson et al., 2013; Fassman-Beck et al., 2013; Ouellet et al., 2021).

In nature preserves (Figure 7), a combination of soil moisture sensors, piezometers, and water level sensors are recommended for determining site water balances and estimating hydrologic storage. Since surface-groundwater interactions are important in these systems, surface water level sensors should be used in concert with piezometers to capture both surface and groundwater storage (Gonzalez et al., 2023). Surface water flow measurements will often also be useful to measure discharge to urban waterways (e.g., streams, rivers, engineered drainage ditches) and downstream areas of the watershed. Placement of sensors within a large natural area is much more complex than in small-scale NBS, like a rain garden. Pre-installation modelling of site topography and flow paths is useful to determine hydrological connectivity to the surrounding urban hydrologic connections, such as natural streams and engineered stormwater conveyance ditches, and low points within the site, where pooling of stormwater is anticipated. Local LiDAR data, which is frequently available for public use, and digital elevation models (DEMs) are useful to understand site drainage and inform the optimal placement of sensors to accurately monitor site water balance.

For community greenspaces embedded within highly urbanized areas (Figure 8), on-site piezometers are recommended for monitoring site water balances and stormwater storage (Xie et al., 2021). In addition, flow meters or weirs and water level sensors are useful to document the reduction of stormwater inflow into the urban drainage system (Jarden et al., 2016). In addition, the scale of community greenspaces and NBS often requires coordination between community-based monitoring and governmental infrastructure monitoring. This is an opportunity for community science measurements in collaboration between local residents and government agencies. This is particularly useful for long-term adaptive performance monitoring to inform maintenance. In addition, flow meters or weirs and water level sensors in surrounding drainage infrastructure, like stormwater catch basins, provide data on interactions between the site and drainage infrastructure and can be used to determine the reduction of inflow into the drainage system due to the storage provided by the community greenspace.

In addition to recommendations on the type and number of sensors needed to obtain site water balances, in Table 1 we provide cost estimates for instrumentation of community-centered NBS projects based on the desired data quality. Rain gauges should be located near NBS installations and can be shared by multiple NBS installations in close proximity to one another. Tipping bucket rain gauges costs range from $700 to $900. The lowest cost estimates in Table 1 are for a minimal monitoring strategy with the fewest recommended sensors at the lowest price per sensor, and the highest cost estimates are based on achieving the greatest spatial resolution (largest number of sensors) and data quality (highest price per sensor). The cost estimates are not comprehensive, but rather indicative for typical community-centered NBS projects conducted by non-profit organizations or local governmental agencies in the United States. Larger infrastructure projects and/or more challenging settings for instrument deployment are likely to require substantially larger investments.¹

Data provided by the recommended instrumentation designs and monitoring methods is sufficient to conduct a water balance on NBS systems, which provides critical information on both ecohydrologic dynamics in greenspaces and the volume and timing of stormwater storage beyond retention and detention. This comprehensive monitoring approach allows for upscaling from individual site measurements to assess the impact of community-centered NBS on flooding and stormwater management at the catchment and regional scales. This approach is shown in the conceptual framework in Figure 9. Using water budget analysis, researchers can assess how NBS respond to different conditions such as climate, rainfall intensity or shallow groundwater tables (Zhang and Chui, 2019; Hung et al., 2020; Johnson et al., 2022). This can inform adaptive planning and selecting the most appropriate type of NBS for different settings (Kato and Ahern, 2008; Tang et al., 2020). The complexity of water budgets will vary based on the type of green infrastructure. The simplest example is a green roof, where the water input is limited to direct rainfall onto the roof, the hydrologic response time of the system is very short, and outflow can be directly measured at a discharge pipe. For this well-defined case, water mass balance can easily be obtained for a storm as ΔS=P-Q, where S is volume of stormwater stored, P is the total precipitation volume, and Q is the total outflow obtained by integrating the measured discharge hydrograph over time (Versini et al., 2016). The complexity increases when calculating the water budget for a rain garden or bioswale, in which case there is stormwater input from both on-site rainfall and runoff from an off-site contributing area. In this case, we can directly measure the surface and groundwater storage and estimate short-term water balance based on the change in water level measured by surface and groundwater level sensors. In addition, flow meters can provide an estimate of outflow from the system (Fletcher et al., 2021). A longer-term water balance incorporating evapotranspiration and slow groundwater recharge can be estimated by incorporating modeling. However, because these systems respond quickly to precipitation a short-term water balance can provide an estimate of water storage based on direct measurements through the recommended suite of sensors and instrumentation. The complexity further increases in an embedded greenspace. If the system is relatively small and directly connected to a stormwater system, the same methods for rain gardens and bioswales can be used to measure storage and outflow. However, in systems with no clearly defined or engineered inlet or outlet, the total contributing area may vary between storms, making the inflow and outflow very challenging to measure. In these cases, some estimates of short-term storage can be obtained from direct measurements of changes in groundwater and surface water levels from sensors. However, on a larger time scale, modeling is often needed to better understand the longer-term dynamics. Many of the methods for this type of analysis are still under development, but the data provided by groundwater and surface water level measurements can provide a basis for estimating an overall water budget in these systems.

To determine the stormwater storage and other benefits provided by community-centered NBS, baseline data must also be collected for a significant period of time prior to NBS construction to provide points of comparison for a variety of storms of differing intensity, duration, and frequency. Pre-implementation monitoring is useful for both constructed NBS and related activities such as restoration of natural areas. Pre-implementation data collection can range from 3 months before construction begins (Jarden et al., 2016) to more than 2 years (Boening-Ulman et al., 2022). The recommended duration of pre-implementation monitoring will vary based on local climate. Longer pre-implementation periods increase the diversity of storm conditions for comparison with post-installation performance. Our analysis of precipitation data in the Chicago region found that 1 year of pre-implementation data is needed to capture a variety of storm events and obtain a reasonable estimate of the precipitation distribution and site hydrological response (Griffin et al., 2020; Gonzalez et al., 2023). At a minimum, we recommend at least 3 months of pre-implementation data collected during the primary season of concern for high precipitation and flooding. Since sensors used for collecting baseline data can then be deployed within the NBS, no additional sensor costs are associated with collecting baseline data.

Monitoring NBS performance provides the information necessary to quantify the value of community-centered NBS for stormwater management and flood reduction, improve design, advocate for increased adoption of NBS, and inform the timing and type of maintenance needed (Ahern, 2007; Geberemariam, 2017; Gordon et al., 2018). Comparisons between pre-implementation and post-implementation monitoring data are important to quantify the additional stormwater storage provided by NBS and the resulting reduction in burden placed on the surface water drainage system (U.S. Geological Survey, 2023). These results can then be directly related to the benefits of NBS for reducing flooding in surrounding communities, and the corresponding economic value of these services. Quantitative data on the hydrological performance of NBS can also be used to advocate for additional investment in community-centered NBS from local water management authorities and municipalities. Tangible data are useful to local leaders in communicating the functionality and benefits of NBS to community members, who may not initially support or recognize the benefits of this type of infrastructure (Venkataramanan et al., 2020). In addition, when community members have access to and understand the data and can see the benefits of community-centered NBS, they can become advocates for these types of projects and provide incentives for their local representatives to provide additional support (Ando and Freitas, 2011; Lamond and Everett, 2019).

Long-term monitoring is critical to provide data on the long-term performance of NBS years after implementation and to inform maintenance needs; however, this often presents many challenges in terms of budget and staff availability (Wadzuk et al., 2021). Determining the minimum length of monitoring necessary to achieve project objectives, such as informing monitoring or quantifying the benefits of a community-centered NBS installation, should be discussed in the planning phase to allocate necessary staff time and financial resources. The most challenging aspects of long-term monitoring are deciding who should be responsible for long-term data collection and analysis, allocating staff time for this purpose, and securing budget for replacement of sensor hardware and parts. Municipalities are often understaffed, particularly in disadvantaged communities that are often more susceptible to flooding (Weller, 2023). Nonprofits and community organizations often experience the same challenges in terms of staffing. Therefore, allocating staff time and resources to long-term monitoring is often unfeasible unless resources are allocated as part of the primary NBS construction effort.

These challenges can be addressed through partnerships with local community members and by combining long-term monitoring efforts and maintenance projects. Monitoring data inform the type of maintenance required, which will in turn impact the long-term performance of the NBS (Wadzuk et al., 2021). Citizen science programs provide opportunities for community engagement and education while supporting long-term data collection for both research and operational purposes. Citizen science partnerships involve data collection, performance monitoring, and reporting by local volunteers, who have an incentive to see these sites perform well in the long term. NBS monitoring also presents the opportunity for workforce development (National Academies of Sciences, Engineering, and Medicine, 2018) and creating jobs for monitoring and maintenance of NBS in flood-prone communities. Beyond their value for both hydrologic research and operational performance monitoring, community science programs also provide value for public education and capacity building for maintenance and monitoring in environmental justice communities. These types of partnerships are particularly important for communities to improve climate resilience.

Data provided by the instrumentation of community-centered NBS will be used differently based on the goals and objectives of the institution or organization leading the monitoring. Municipalities and local non-profits or community organizations may be more interested in using data to inform the maintenance of their systems, whereas researchers may want to better understand how these systems function and inform the design of more efficient NBS.

Monitoring data are useful to inform ongoing operational maintenance, specifically when maintenance is needed and what type of maintenance is necessary. For example, Kazemi, 2014 used water level data to calculate infiltration rates and identify when clogging in a permeable pavement system had occurred. Similarly, the City of Lancaster, Pennsylvania used infiltration rates in permeable pavement to determine when vacuuming of these systems was needed and infiltration rates in bioswales to determine if additional soil testing or amendments were necessary (City of Lancaster, 2019). The early identification of maintenance needs can reduce costs over time, improve long-term performance of NBS, and help inform decision making for allocation of maintenance resources (Wadzuk et al., 2021).

Municipalities may also use monitoring data to track progress toward goals or toward meeting regulatory requirements. For example, the City of Philadelphia, Pennsylvania, uses monitoring data from NBS to assess how their program is reducing combined sewer overflows (CSOs) in the city and inform their long-term control plan for preventing CSOs (Philadelphia Water Department, 2009).

Lastly, monitoring data from multiple locations can be used to expand the assessment of NBS beyond the site scale to the catchment or regional scale. Scaling the analysis of NBS performance is critical to better understand the cumulative impacts on downstream hydrology, inform catchment management, and determine what factors impact changes in performance upon scaling (Golden and Hoghooghi, 2018). Monitoring can also inform community-centered NBS planning at the regional scale and inform decisions on the spatial configurations and network locations of NBS (Weber and Wolf, 2000; Golden and Hoghooghi, 2018; Shi and Qin, 2018; Goodspeed et al., 2022).

Beyond monitoring the stormwater storage provided by community-centered NBS, there is a need for assessment of the actual impacts on community benefits, such as flood reduction and the improved experiences and perceptions of community members. Additional measurement methods are needed for this purpose beyond the instrumentation designs detailed above. One common means of measuring community impact and benefits is through surveys (Baptiste et al., 2015; Kim and Miller, 2019; Miller and Montalto, 2019; Anderson et al., 2021). However, low-income communities and communities of color may be frequently asked to take surveys on similar topics. Therefore, surveys should be prepared with input from community members and should not be unduly long or inaccessible to reduce the burden on community members.

Practitioners and researchers measuring impact should work with stakeholders to identify and prioritize outcome measures that address local concerns. In addition, an assessment of the risks of this type of survey data collection and analysis, including breach of confidentiality or data leaks (University of Oregon, Office of the Vice President for and Research and Innovation, n.d.), should be conducted at the beginning of the project in consultation with the community and relevant Institutional Review Boards, and all risks associated should be adequately communicated to the community beforehand. In cases where community members or community organizations are involved directly in impact assessments, whether recording visual observations or participating in surveys, the effort involved should be recognized and community members should be compensated fairly for their time.

Data management and privacy are concerns among community-centered NBS monitoring projects on both private and public property. Data management should be discussed as part of community engagement early in the planning process. At a minimum, data management, usage and privacy should be discussed with the community before sensors are installed and monitoring begins. Community members may have concerns about the use and protection of data collected near their homes, particularly imagery that may show individuals and the occurrence of flooding, which could impact local property values (Gourevitch et al., 2023). Efforts should be made to balance the protection and security of data with the availability of data for community and professional use. Practitioners engaged in monitoring should make every effort to provide a space for stakeholders to determine outcomes of data collection, including making data collected in communities available to community members in an accessible format. For research purposes, access should be provided to data in raw, processed, and analysed formats.

It is important to be aware of the historical context of data collection and research conduct in the neighborhood and community where monitoring will take place, especially for imagery. Over the past several decades, many local governments have expanded surveillance of public places with the stated goals of deterring crime and enforcing the law (Brown, 2008). However, throughout United States history, surveillance has been used as a means of enforcing structural racism, including enslavement, Jim Crow laws, and segregation, to the harm of millions of people of color (Gellman and Adler-Bell, 2017; Arnett, 2020; Lee and Chin, 2022). As a result, practitioners need to approach monitoring technology with surveillance implications through authentic, responsive engagement with community stakeholders and with an overarching mandate to avoid causing harm.

Community engagement is a key component of selecting the location, type, and scale of community-centered NBS in the area in which it is implemented. However, this is outside the scope of this paper and is discussed by Ferreira et al. (2020) and Mok et al. (2021).