Scorpions, shotguns, lianas, chiggers, ticks, howler monkeys, tree climbing, and 85% humidity - oh what a field experience! I am sitting on a porch overlooking Lake Gatun and the Panama Canal at the Barro Colorado Island research station of the Smithsonian Tropical ... (BCI STRI) in Panama. This is one of the oldest and most famous tropical forest research centers in the world, and includes a 50 hectare permanent plot within which every tree greater than 1 cm DBH (diameter at breast height, about 1.3m) has been tagged, mapped, and measured every 5 years since the 1980s.
I came here at the beginning of the week to work with STRI Staff Scientist Helene Muller-Landau to collect high resolution aerial photos of the 50 ha permanent plot using Ecosynth computer vision and unmanned aerial system (UAS) techniques. We flew four major missions, each about 6-8 km in total flight length, using the Arudcopter Octocopter vehicles at heights ranging from 50 m - 150 m above the forest canopy including a mission covering 1000 m x 500 m to capture the entire 50 ha plot in one go (the telemetry flight plans are in the Google Earth overlay at right.
Each day we hiked the two Octos and the main flight gear at least 2 - 3 km out to the field site on trails through the jungle. We flew out of gaps in the canopy that were identified by field teams prior to my arrival, or in aerial image datasets collected earlier in the week - planning a mission based on data collected two days prior while in the field is pretty sweet!
The weather cooperated with us every day. It was almost always overcast, I think I saw blue sky once, the winds were dead to very light, and the lighting was almost always diffuse and constant: great flying conditions all around! We did have some hiccups though. Due to a faulty wiring harness for the parallel battery assembly, for two missions the copters were not drawing off all batteries and shutdown early. On one mission the shutdown happened at the descent, about 2m above the ground at the end of the mission, so no harm done. During our third major mission however the copter descended about 200m away from us and we lost all communication, including from the dog tracker and the telemetry. We navigated to the last known coordinates from the dog tracker, and eventually the signal picked back up. We found the copter about 20m up in a tree, caught in a mat of lianas. The camera and battery pack had fallen off the copter when it crashed into the tree, but were otherwise in good shape. As we were scouting for the copter, I heard: "Jonathan, stand still! Helene, come brush this off him." WHAT!!! Walking through the jungle I had managed to get a small scorpion on my back, but fortunately we knocked it off and I am still here to write this blog post.
The next day we came back with the game wardens, climbing equipment and a long barreled shotgun. The first plan was to shoot the branches around the copter to get it to come down, but the game wardens decided they should climb instead as it would be too difficult not to shoot the copter. Using a slingshot and various diameters of line, they hoisted a climbing rope into the tree and the guard captain climbed the rope and retrieved the copter, lowering it back to the ground. Only a few broken landing gear and arms, but the on board 'brain' and telemetry were just fine!
But of course, aside from the exciting stories, the fact that my 'lab bench' was also my bed, and the exciting collaborations and contacts I have made, the most important thing is the data, and it looks good! I used Photoscan to generate a low-level orthomosaic image from one photo collection over the 50 ha plot and was able to present that in my seminar talk to the field station on Thursday. We are all very impressed with the results. From the high resolution images we can make out individual flowers on several canopy trees (see below, overlaid with the 50 ha plot boundary in red) and with images collected at multiple altitudes / resolutions over a 125 m x 125 m area, we can run some tests on the effects of altitude on the data product. I will try to push up more pictures and videos soon. It is nice to be heading home and back to Maryland, but I think I agree with the more experienced scientists here - I will leave a little piece of myself at BCI and, one day, I hope to return!