Drones for Conservation

November 14, 2014

NewKnowledge is working with Engineers for Exploration (E4E) at the University of California, San Diego (UCSD) to determine how to best use Unmanned Aerial Vehicles (UAVs) to detect how potentially endangered birds are using areas where land managers wish to treat their land. Dr. Danoff-Burg explains the project and its goals in further detail below.

Here’s a riddle for you: what’s got a bad name because of their associations with the military but could help to conserve endangered birds in San Diego? If you guessed drones, you’re right!

Hexacopter in flight over the San Dieguito River ParkThe drones that we are working with are of course drones of a different sort than what most people usually envision. In our case we do not have missiles attached, only sensors. We at NewKnowledge are working with Engineers for Exploration (E4E) at the University of California, San Diego (UCSD) to use a range of drones (better referred to as Unmanned Aerial Vehicles or UAVs) and sensors to detect whether there are birds nesting in areas where land managers wish to treat their land. The project is funded by the USDA NRCS via a California State Conservation Innovation Grant. The beauty of what we are doing is that if it is successful, it is a winning situation all around – for the land managers, the birds, science, and the bottom line.

Frequently, endangered birds nest wherever there are favorable habitats. Unfortunately this often includes plants and land that need regular tending such as fruit trees or other agricultural crops, areas that have many invasive species that need to be taken out for fire protection, or in land that is slated for restoration or remediation. All of these uses are essential for the industry or conservation activities happening on those lands. In Southern California and many other areas, they cannot happen if important species of birds are nesting in the area. Before land managers of whatever stripe can act, they need to survey to verify that no protected bird species are nesting on the lands under consideration.

Nest surveys usually require an ornithologist do an extensive search over many hours to verify that there are not nesting pairs active on the land. A daunting task, nesting surveys require exacting precision and greater concentration than most can people can easily accomplish. In addition, the very act of surveying can disrupt the very animals the ornithologist is surveying for, and in some ecosystems could lead to some habitat degradation – particularly in swamps or steep slopes prone to erosion or landslides. Similarly, the safety of the surveyor may be compromised in many of the same places prone to habitat degradation.

These are valid concerns. How to solve them? Enter the UAVs!

I’ve been working with Ryan Kastner at UCSD, one of the directors and faculty sponsors of the E4E group, and Eric Lo, the E4E project manager of the engineering components of the project, on an approach that will enable us to use UAVs to quickly and easily spot bird nests with a simple minutes-long flight. The benefits of a successful approach are significant. We will be able to cover more land, more quickly, with higher precision, without damaging habitat, without endangering surveyors, and for much less money than performing nest surveys on foot.

UAV blog 2

Researchers from NewKnowledge, UCSD E4E, and television crew from Discovery Canada

After a few field tests of mixed success, and a few challenges, we had an amazing leap forward in early November 2014. With a television crew from Discovery Canada in tow,  we came up with an approach that allows us to detect bird nests with great precision. It’s rare that you can catch your big breakthrough on camera, but that’s what happened for us.

UAV blog 3

A mock nest with rocky “birds” heated to 115°F

After trying to use the Infrared (IR) sensors with unconvincing results a few weeks ago when it was warmer out, we decided to give the IR sensor a try at daybreak. With the sun’s rays just cresting the California coastal chaparral covered hills, we heated up a few rocks to the 105-115°F temperatures characteristic of chicks, put them on a mock nest of coiled plants, and flew the hexacopter overhead. With the cameras rolling, the “nest” popped out as a bright white dot against the much cooler dark plant background. Even more amazingly, the nest was still visible even when the temperature difference between the “birds” and the environment decreased from the initial 47°F difference down to as little as only a 9o°F difference!

“Nests” as white dots against the cool dark background of the plants

“Nests” as white dots against the cool dark background of the plants

This amazing result has breathed new life into our project, and has revitalized my hope for success! Certainly, we still have several big challenges yet to surmount. However, this was our biggest, as the image analysis of the IR sensors are what we would be most likely be able to automate, and further reduce the analysis time.

The benefits of this approach for land managers in San Diego are clear and apparent, but they are only the start. Once we get this approach refined and made fail-proof (about which we are cautiously optimistic), developing future applications of it for birders, for national and state parks, for conservationists, for sustainably minded landowners, and for those who we haven’t yet thought of will be an exciting time of exploration.

Even better, all this goodness will be brought to you courtesy of the unfairly misunderstood drone!

p.s. Check back here later and we will post the link to the Discovery Channel Canada’s video piece on our research after it is posted online in the next few months!