The Center for Data innovation spoke with Jeff Allenby, director of conservation technology at the Chesapeake Conservancy, a conservation nonprofit in Annapolis, Maryland. Jeff discussed the concept of precision conservation and the value of high-resolution geospatial data for conservation efforts.
Joshua New: You help the Chesapeake Conservancy do what has been described as “precision conservation.” What does this mean?
Jeff Allenby: At its core, precision conservation means getting the right practices, in the right places, at the right scale. It’s simple in theory, but the conservation community has traditionally only been able to assess the benefits of a small number of projects that had already been proposed for implementation. While this means that we were able to choose the best projects out of what we had in front of us, we didn’t have a great understanding of what else was out there, or perspective on how these projects compared to the best possible projects in the surrounding area. The challenge has always been how to generate fine-enough information, to be able to evaluate every potential project, across a large enough geography for it to be meaningful.
To help provide this level of understanding, we have spent the last five years developing new datasets, models, and techniques that allow us to prioritize specific conservation and restoration projects across entire landscapes. We work to understand the particular needs of our partners and, in many cases, will create a customized analysis to identify and quantify the benefits of every potential project across their operational area. With this information, our partners can better target outreach efforts, evaluate what projects get funded, and identify where they are working to meet their programmatic goals. In short, we are helping our partners target their valuable resources and do more with less.
New: Other than maps, what other kinds of data technologies support precision conservation?
Allenby: While our expertise is in the mapping, remote sensing, and modeling world, we are leveraging our partners’ expertise in data science for a number of projects and see a lot of potential for integrating data streams to help verify and correlate the information from our models to information collected from real-time sensors, like water quality monitoring. Combining our models with information collected in the real-world will help improve the accuracy of our work while also helping to demonstrate how precision planning can make a difference in water quality.
New: Could you describe the Chesapeake Bay High-Resolution Land Cover Project? Why are mapping projects like this so valuable to conservation efforts?
Allenby: The Chesapeake Bay High-Resolution Land Cover Project took aerial imagery and categorized over 100,000 square miles of the Chesapeake Bay’s landscape into distinct land cover types. It has 900 times the resolution of previously available land cover data and is one of the largest freely available high resolution datasets in the world. The data was produced over a 15-month-period through a collaboration between Chesapeake Conservancy, the University of Vermont’s Spatial Analysis Lab, and Worldview Solutions to update the Chesapeake Bay Program’s models to estimate nutrient and sediment pollution.
High-resolution land cover data has been critical to our precision conservation efforts because it allows us to model the landscape down to a very fine level of detail; each pixel in our data represents a square meter on the ground, where the previous data had to summarize everything within a quarter of an acre. This breakthrough in resolution has opened up a lot of new directions in environmental modeling and we are starting to work through how it changes a lot of assumptions that had to be built into models that used the previous 30 meter land cover data. Through this project, we have eliminated a major barrier to precision planning, but more importantly, we were able to create a consistent baseline that works across the watershed and we have made it open to anyone to use it.
New: Have these maps led to any measurable impact in conservation work, or is it too early to tell?
Allenby: It’s still early to know all of the long-term impacts, but we are already seeing how it is changing the way that we are thinking about environmental planning in the Chesapeake. One of the most immediate impacts was the new data created a greater sense of trust in the models generated by the Chesapeake Bay Program, a state-federal partnership led by the Environmental Protection Agency, because landowners could look at the data going into the models and it matched up better with their knowledge of the land.
We are also seeing some great projects by county governments and local restoration partners in Pennsylvania, Maryland, and Virginia who are using our data and models to prioritize implementation funding. There is a large need to efficiently use limited resources, so having access to these tools is helping our partners ensure that projects are going to the right places as well as communicating the importance of projects to funders.
State and federal partners are also starting to think about how programs could be restructured to move away from measuring “success” based on effort, how many acres were protected or how many trees we planted, and focusing more on how differences in the landscape impact a project’s overall performance.
New: What came of the Chesapeake’s Conservancy’s 2014 partnership with Microsoft?
Allenby: Our partnership with Microsoft is ongoing as we collaborate with their AI for Earth team to better understand the role that artificial intelligence can play in generating high resolution land cover data. While the Chesapeake Bay High Resolution Land Cover Project has had an incredible benefit, it was a large, and expensive, project that would be hard to replicate on a frequent basis and the lack of high resolution data elsewhere has been a barrier to expanding our planning work to other geographies. We are working with Microsoft to try to reduce the time and cost it takes to update the dataset in the Chesapeake, so we can understand what in the landscape has changed, and expand the coverage of high resolution to other landscapes.