Climate change, land development, population movement, and other dynamic factors of today’s world are leading stakeholders from all levels of government to implement geospatial solutions for stormwater management. At Ecopia AI (Ecopia), we work with federal, state, and local governments to provide a source of truth for the land cover features that impact stormwater infrastructure and planning, including highly detailed classifications of impervious, pervious, and unpaved compacted surfaces that enable decision-makers to adapt their strategies to a rapidly changing world. 

This work across the public sector has exposed us to many different stormwater mapping projects. While each government entity we work with has unique requirements, we find that most projects have a similar workflow in common. To help government agencies looking to enhance their usage of geospatial data for impervious surface and stormwater mapping, we created this checklist outlining the most common approaches we’ve seen from government clients around the world.

1. Decide on your end goal

Before beginning any project, the team involved should agree on an end goal. For stormwater mapping, this typically means defining statements like:

• optimize stormwater utility fee (SUF) or drainage fee calculations;
• build more accurate flood models;
• develop sustainable stormwater infrastructure;
• mitigate risk from natural hazards;

and other similar goals related to climate resilience. It’s important to have this simple guiding principle before getting bogged down in the details. Instead of starting out by copying the exact stormwater mapping strategy as another agency, take a step back and figure out what you ultimately want to achieve when the project is complete. For example, it is more productive to begin your project by stating “we want to optimize SUF calculations” than by deciding “we need data for every feature in our county.” Having a high-level project goal also helps justify the effort to less technical stakeholders who may need to approve it.

Once this end goal is defined, teams can better create a plan to achieve it. That’s when governments can start digging into the details of the project, like what type of data is needed and the specific requirements it should have based on the end goal. At Ecopia, we call this a fit-for-purpose approach to building the exact data our clients need.

2. Determine the scope of your project

A critical step in any project, but especially one involving geospatial information, is to determine the scope. For stormwater mapping this generally means identifying the geographic extent you will be mapping and analyzing, what features you need data for across that extent, and what timeframe you want the features to represent. As an example, the City of Detroit identified a need for 25 distinct land cover features, updated annually, in order to support their SUF calculation and general business across the entire city. 

Governments will often drill down further within each of these scope parameters depending on their project needs. For instance, the National Oceanic and Atmospheric Association (NOAA) determined that to better provide coastal communities with the geospatial information needed to boost climate resilience, they needed to increase the resolution of the data they offer through their Office for Coastal Management from 10-30 meters to 1 meter. 

3. Obtain imagery of your project AOI

Once you know what you are trying to achieve and the information needed to do so, you can start creating or sourcing data for the project. Like many geospatial initiatives, stormwater or impervious surface mapping projects often leverage imagery captured from satellites, airplanes, drones, or even street view cars. This imagery is used for a variety of purposes depending on the project goal. Some government offices rely on high-resolution imagery to serve as a basemap in their projects, providing necessary context to other information displayed in the map. Similarly, teams often georeference updated imagery onto existing maps of stormwater infrastructure they have so as to better visualize the area of interest (AOI). 

At Ecopia, we’ve seen many reasons for using geospatial imagery as the foundation of a stormwater project, as well as many different methods of obtaining it. Some of our government clients source their own aerial imagery so they can get the exact resolution they’ve defined in their project scope. We often connect clients to our global partner network of leading imagery providers if they would like assistance sourcing the right imagery for their project, but our artificial intelligence (AI)-based mapping systems are imagery agnostic, so each client ultimately chooses which imagery works best for them and their specific needs.

4. Digitize land cover features

By far the most common trend we see with government entities analyzing stormwater and climate resilience is the need to digitize geospatial imagery into interactive vector maps. While imagery provides important context for understanding impervious surfaces in an AOI, in-depth geospatial workflows can generally only be performed when raster images are digitized into vector layers. Vector maps enable teams to isolate individual land cover features and layers and conduct the analytics needed for strategic decision-making.

For example, King County in Washington identified 16 features across 2,300 square miles they needed planimetric-level detail for in order to support their flood modeling, stormwater assessment, and strategic climate action plans. Imagery alone did not provide the King County GIS Center with the depth of information or interactivity needed to be a key ingredient in their impervious surface analysis. 

Digitizing land cover features from imagery is typically achieved one of two ways. The first option is to manually digitize each feature needed across the project AOI, which involves time-consuming and tedious tracing by a GIS technician. This is a feasible option if there are only a handful of features needed across a small area, but quickly becomes a resource-intensive task when scaled up to the level most governments are digitizing stormwater data. Additionally, updating datasets is a very tedious and challenging process. Sometimes the human eye cannot fully tell when a change has occurred, so human interpretation error can lead to less accurate datasets.

The other option is to leverage AI-based mapping systems to digitize the required features efficiently and at-scale. Unlike other providers, Ecopia’s AI-powered systems do not sacrifice quality, so the output data is just as accurate and reliable for critical decision-making as the results of a GIS professional’s manual digitization. Using King County as an example again, Ecopia was able to map all 16 land cover layers across the 2,300 square mile AOI in less than 8 weeks, a task which would have taken many months of manual work by multiple government employees or contractors.

5. Classify land cover features

While digitizing imagery into vector features makes information more interactive, what makes it truly actionable is classifying each feature into different land cover type layers. Detailed classification allows teams to select all features of a specific land cover type, symbolize maps to show different types of surfaces, and understand key context about the AOI of their project. 

If manually digitizing features, it’s best to classify them as you go so that you don’t forget what they are later on. This adds another step to the manual digitization process, and can be difficult if you are unfamiliar with the area or can’t tell exactly what a feature is just from looking at the input imagery. Ecopia’s AI-powered mapping systems also classify features as they are digitized to ensure each feature is correctly added to the layers needed for a particular map. Our extensive experience working on both continental and local scale projects around the world means our AI can extract a large variety of features from imagery, even when the human eye has difficulty.

For instance, in some parts of the world the construction material of impervious surfaces like buildings is the same color as the surrounding landscape. While discerning and digitizing individual buildings just by looking at an image might be difficult in this case, Ecopia’s AI-based systems have been trained to extract and classify features across diverse and challenging landscapes to provide teams with the detailed data they need for stormwater mapping.

6. Create & analyze impervious surface maps

After the data features have been digitized and classified across an AOI, the real problem solving can begin. With comprehensive, accurate, and up-to-date land cover data representing the types of natural and impervious surfaces needed for the particular project, teams can build flood models, identify potential hazard risks, and improve the climate resilience of their community.

This step in the process is highly dependent on the end goal outlined in the beginning of a project, and at Ecopia we have seen many innovative ways to leverage land cover data for stormwater and impervious surface mapping. One government client we worked with had a particularly complex challenge when it came to stormwater management: developing an integrated flood model (IFM) to better predict and mitigate the effects of future stormwater events.

The City of Peterborough partnered with Ecopia to digitize 13 distinct layers of both natural and impervious surface land cover throughout their AOI. They then used this data as an input for a 2D surface mesh, with each grid of the mesh detailing water depth and velocity estimates for future flood events. The City achieved this by using Ecopia’s classified land cover layers to assign surface roughness and runoff coefficients to each area throughout their AOI, ultimately feeding into their IFM and enabling them to enhance their stormwater planning efforts.

This is just one example of the many government projects Ecopia has supported with land cover data. You can read the full case study here to learn more.

7. Detect change

The final (and usually ongoing) step in any stormwater project is keeping the data and maps up-to-date to reflect our dynamically changing world. Most of our clients come to Ecopia because of the need to frequently update their maps and data - and how time consuming and expensive that is if done manually.

To stay on top of change detection and make sure maps, models, and information is a true digital representation of the physical world, teams must either repeat the entire process outlined in this checklist, or enlist the help of AI-based mapping systems. That means sourcing or capturing updated imagery and then manually digitizing and classifying all features needed across the entire AOI. 

To provide an example, it took Collier County in Florida 4 years to manually digitize each one of the 132,000+ driveways and access roads in their AOI, not even accounting for any change that may have occurred during that time frame. When Ecopia digitized these features using AI-based mapping in less than a month, we identified 39,000 that had been added or changed. These changes would have taken Collier County over a year to detect manually.

Detecting change in impervious surfaces is critical for stormwater management as it helps government agencies maintain and improve community climate resilience despite how quickly our world changes. When people inevitably move, build new construction, or alter their property in some way, governments must have a system of record for those changes and be able to factor them into any models impacting their decision-making. Whether determining SUFs, building a flood model, or maintaining and improving stormwater infrastructure, impervious surface change detection is a necessary part of any climate resilience project - and if done right, should be done at least annually to ensure data freshness.

Get started with AI-based mapping

Ecopia’s AI-based mapping systems are currently helping federal, state, and local governments around the world to better understand impervious surfaces and manage their stormwater infrastructure. If you’re looking to leverage high-precision land cover maps at scale for climate-related projects, get in touch with our public sector team. We’d be happy to help you determine the exact requirements of your project and curate the data you need to predict and mitigate risks in your community.