Accurate, High-Quality Data in One Place
To model stormwater flows with certainty, stormwater practitioners need accurate, actionable information and a lot of it. Yet with the sheer volume and decentralized nature of the existing data, there wasn’t a way to scale it up and out from the places where it was collected. To be useful, the Stormwater Heatmap needed to statistically link pollution measurements with landscape features, things that could be objectively measured. The data and calculations also had to be apparent and meet best scientific practices.
With access to the power of the Google Cloud Platform, we could harness tools like Google Earth Engine and BigQuery. We could integrate and efficiently compute disparate data sources, housing 30 billion lines of hydrology output data across Puget Sound alone. Hydrologic output data for hourly timesteps is now run for every hydrologic response unit that exists in Puget Sound, and over historical, current, and future climate conditions. They are run on the one square meter resolution landcover foundation layer we generated with collaborators, an undertaking that took over a year alone! From a watershed spanning thousands of square miles to a parcel covering a few square meters, the Stormwater Heatmap offers land cover, hydrology, and pollutant loading data layers that can assist with any scope of planning.
A Predictive Model for Pollution Loading
Currently, different pollution loads are assigned to different land uses like industrial, commercial, high density residential, or low density residential. These are subjectively assigned depending on county or jurisdiction. Our approach was to find statistical relationships between measurable landscape features (e.g., traffic amount, amount of impervious surface, amount of rooftops, CO2 emission estimates ), and then use those relationships to predict pollution load elsewhere.
The results spoke for themselves: our new method outperformed the current method for nearly every pollution type we modeled. Better yet, we can extrapolate pollution loading from areas where stormwater modeling is done to areas that don’t have a monitoring requirement or the funds to do it. This predictive method, using objectively measurable datasets, is a major step forward in stormwater management.
Compelling Visuals, On Demand
Fun fact: stormwater practitioners are also storytellers. Whether it’s taxpayers, regional funding bodies, or the city council, practitioners must translate huge amounts of technical data and findings into accessible, compelling stories. The heatmap does the heavy lifting for them, generating visuals and reports to bring the numbers to life.
Leave a Reply