What Stormwater Programs Can Learn From the Covid-19 Pandemic

Date / Time:
Wednesday, Sep 16 1:30pm to 2:00pm
Track / Session:
Track: Data Management, GIS / GIS Tools, and Modeling / Session 7

Epidemiologists routinely turn to models to predict the progression of an infectious disease, just as they have during the Covid-19 pandemic. By mid-April of 2020, states were scoping plans for how to reopen the economy, and epidemiological modelers came under fire for what were criticized as overly dire initial projections of infection rates and changing projections over time . Much of this criticism was based in a poor understanding of how models work and how to interpret the outputs in a way that produces durable insights. A similar misunderstanding exists in stormwater management right now. We use models to make decisions because the systems we manage are complex, we want to know how they will respond to different management alternatives in the future, and we usually have very few applicable measurements available. Models rely on scientific understanding of systems, sets of simplifying assumptions and input data to drive and parameterize how the model structure responds to inputs. As with the much-relied upon model projections from Imperial College London during the Covid-19 pandemic , ongoing data collection to update inputs can improve the accuracy and precision of projections over time. This is a not a bug. It is a feature of good modeling practice that should guide the way that we use models to support decision making in stormwater management just as it does in epidemiology.

Given limited capacity for measuring changes associated with urban water quality improvements, stormwater managers will continue to rely on various models for planning, estimating impacts/progress, and in some cases, meeting MS4 NPDES permit requirements. Reliable estimation of the severity and patterns of water quality impacts along with benefits associated with treatment and source control allows identification of new priorities and efficient resource allocation. As when grappling with a global pandemic, in stormwater management we often begin with very little data to support models used to estimate baseline scenarios or the impact of mitigation actions. Dynamic tools that are responsive to new conditions and inputs are inherently more useful than a one-off study, with projections that cannot be refined and assumptions that are not easily scrutinized. Over time, BMP technology develops, new implementation sites are identified, BMP performance degrades, maintenance restores functioning, street sweeping practices are enhanced, etc. Given the importance of changes that we know will occur over time, adoption of approaches that preclude updating model projections with new information as it becomes available seems to be a recipe for poor decision making.

For model outputs to be meaningfully revised over time they must be updatable by those with the information readily at hand and provide outputs on timeframes and spatial scales relevant to management decision making. We provide practical examples of how this provides stormwater programs with powerful tools for optimizing implementation decisions and quantifying the aggregate catchment-scale benefits associated with green infrastructure, low impact development, street sweeping, and centralized treatment. Adoption of modeling approaches that are transparent and amenable to interim verification can open up opportunities for partnership development to accelerate investments to stormwater infrastructure. Shifting towards this more iterative approach with tools distributed into the hands of stormwater practitioners and updatable over time can provide a more reliable basis of information for realizing long-term water quality goals.

Primary Speaker:
Gary Conley, 2NDNATURE
Gary is Chief Scientist and co-founder at 2NDNATURE Software, bringing 15 years of experience working towards water quality solutions in both public and private sectors. He leads a research team developing the basis for turning data into actionable knowledge, improving watershed stewardship, and reducing the costs of clean water. With expertise in hydrology, pollution dynamics, numeric modeling, and applied math, his focus over the last decade has been on identification of environmental problems, understanding patterns of change, and building science-based geospatial tools to improve environmental decision making.