Charlie, ready to collect data in the waters of John Pennekamp State Park.
When we make maps for the Allen Coral Atlas
, we have a question that follows of whether the map has been validated. We ask where the data came from to inform the creation of the map. The more calibration and validation data we gather, the more confident we are in the maps.
What is difficult to reconcile, however, is how hard it is to scale this confidence where our goal is to map the entire globe. I needed to understand what was involved in doing this validation, could we find a way to scale the validation process.
One of our partners often says that one should “smell the reef.” In other words, it’s important for us to get hands-on, and really experience what work is like for those working in the coral reef space. Our Field Verification Director, Helen Fox of National Geographic Society, offered to meet me in Key Largo to teach me how to do basic field verification. With limited time, we went to the John Pennekamp State Park
where she showed me the basics of a line, point and photo transect. After renting snorkeling gear, we headed into the waters.
We did the photo transect with a GoPro, had a gps on a life jacket. We did two photo transects helping me understand how involved getting data from a reef area can be.
That wasn’t the end of the challenges, when we got to a good wifi signal we started the process of combining the photos with the gps track. A rough piece of software had been built to do this, but requires a fair amount of laptop-tweaking to make work (we’re talking installing additional libraries and tools, running ages old batch scripts, etc). After getting through that, there was the trick of syncing the gps track with the photo timestamps. With this combined data stream, you have a reference for understanding the habitat and can move to the next step of mapping.
Screenshot of one of the tools used to synchronize various sets of data from the field collection.
In other words, when our mapping team is mapping a new reef, they’ll study a series of these photo transects to understand the habitat of a reef area. Then they set rules in a piece of software that color-codes certain areas of the satellite image based on the rules. Ultimately, you could feed numerous images to the software to start creating the map of that reef. There are lots of other inputs that inform the map like depth and wave direction. But you have to start with the deep understanding of the habitat.
Over the next few months as we get more and more of the mapping production pipeline coming together, the latter part of this process will hopefully become more streamlined. As well as answering the question of how much field verification we really need to map the globe. Stay tuned for more on our progress throughout the year.