A Mountain Expedition in Panama to Test New Insect Monitoring Technology

A new, jungle-proof automated light trap for insect monitoring documented in Methods in Ecology and Evolution and Integrative and Comparative Biology. The Mothbox is like a mammal trail camera but designed for large-scale insect monitoring

Our excitement rose as the truck bounced along through the fields, the mountains looming ahead, very close now. We were following Juan Velazquez to his home at the edge of the forest, a steep ridgeline rising above us into the clouds. The sun streamed through the clouds on their leading edge, but the top of the ridge was fully hidden behind a dark curtain. Who knows what we’ll find up there? We were at the edge of Cerro Hoya National Park, a remote protected area in a corner of the Azuero Peninsula of Panama. The cloud forest was holding true to its name despite it being the start of the dry season. These mountains are a sky-island separated from similar habitat by hundreds of kilometers of deforested pastureland. Rugged terrain and remoteness mean few scientists come here. We came to test a new technology for autonomous insect monitoring.

Since 2022 I have been working with Andy Quitmeyer of Digital Naturalism Laboratories to develop a jungle-proof automated light trap for insect monitoring. The Mothbox as we call it, is like a mammal trail camera, but designed for large-scale insect monitoring. It uses light to attract insects then photographs them. Our custom data processing system ‘Mothbot’ then uses a computer vision model to identify the insects, drastically reducing time and effort required for data analysis.

Insects constitute the greatest share of terrestrial biodiversity. A study conducted by STRI’s Yves Bassett et al. estimated that there are over 18,000 insect species in a single hectare of tropical forest. They play diverse and integral roles, but they are rarely accounted for in conservation planning. We aim to change that. We are working with a reforestation project developer in Azuero. They are keen to quantify how restoring native forests boosts biodiversity and want to include Mothboxes in their biodiversity monitoring program to help secure additional funding for reforestation. 

The purpose of this expedition, however, is much simpler: show the world our system works. We wanted to conduct a study impossible with traditional monitoring methods and other technologies. We initially planned to deploy 10 Mothboxes from the base of the mountain near sea level all the way up to the clouds, programmed to simultaneously monitor insect activity across the elevation gradient. Finding 10 experienced lepidopterists to camp out at different elevations to monitor traditional traps night after night would be a tall order, and the following months required to process samples would make such a venture very resource-intensive! Why stop at 10 though? We ended up preparing all 19 Mothboxes we had at the time, planning on deploying one every 100m of elevation change, with spares at the top and bottom.

Back to the expedition:

We made it to camp at 1000m elevation exhausted, thankful to shrug out of our backpacks. They were lighter now without the Mothboxes we had deployed on our way up, but still full of 4 days’ worth of food, fuel, and camp gear. We set up our hammocks and hung our backpacks - it’s never a good idea to leave anything you care about on the forest floor in Panama. In fact, we had issues a couple days later with army ants choosing our trash bag as a place to establish a bivouac. Evicting several thousand angry, stinging tenants from their new home isn’t fun! 

After setting up camp we hiked further up to set up the final Mothboxes of the day, marveling at the transition into cloud forest. Due to the near constant mist, every surface in a cloud forest is covered in lichens, mosses, ferns, bromeliads, and orchids. We were swimming in a sea of green but occasionally caught glimpses through the vegetation of the verdant forested valley below us giving way to pasture and the blue ocean beyond. 

In addition to myself and Andy, our expedition included Bri Johns, a Fullbright Scholar studying the Mothbox as open science hardware in ecology, Sol Parra, a lepidopterist studying wasp-mimicking moths, Daniel Moreno, a keen naturalist with years of experience in the Azuero Peninsula, Joe See, a videographer shooting a documentary about the trip, and Forrest Shimazu, an entomologist working with caterpillars. We timed our trip with the new moon, so our Mothboxes wouldn’t have to compete with ambient light. We were also excited to do some conventional moth-sheeting near camp - and it didn’t disappoint. The bedsheet I strung up was covered in insects shortly after I turned on my ‘Mothbeam’ - a standalone version of the attractant lights used in the Mothboxes. Sol also set up several hanging traps for her research collections. 

The next day we set out for the summit. The trail all but disappeared, so going was significantly slowed by machete-work, but we only had 3 Mothboxes left to deploy and all day to explore the cloud forest. We took the opportunity to go on a naturalist treasure hunt, keeping our eyes peeled for an undescribed endemic hummingbird, a cousin to the Purple-throated Mountain-Gem, endemic to Cerro Hoya, and Craugastor azueroensis, a rare frog only found in the region. Wading through vegetation can be anxiety-inducing for tropical ecologists - in the lowlands just a few minutes of bushwhacking could result in collecting several ticks and chigger bites, not to mention the myriad of rash-causing or spiny plants and stinging insects. These charming qualities are missing in cloud forests, likely a result of the cooler climate and fewer grazing mammals, so we could wallow in nature to our hearts’ content. 

We followed some rhythmic cheeping sounds to a little twig adorned with the aptly named mountain-gem. Joe recorded what is likely the best video of this species, while I recorded its call and took what are likely the worst photos of it [https://www.inaturalist.org/observations/259997788]. Other highlights included a giant harlequin beetle, a velvet worm, a glass frog, and a large troop of Azuero spider monkeys. I was also excited to see my first tropical oak trees and a bromeliad endemic to Cerro Hoya with gorgeous yellow flowers. Evenings we would relax around a small campfire, Sol processing her moth samples, Daniel collecting arachnids, and everyone occasionally going to check out what’s new on the moth-sheet. 

On our way down we made a remarkable discovery - a giant caterpillar on a tree. Forrest and Sol agreed that it belonged to the family ‘Erebidae’, and we started to get excited because at this size it was possible that this could be the holy-grail of caterpillar discovery - a white witch! This giant moth species has the widest wingspan of any living insect, and despite being described over 300 years ago, no one could verify what its larvae looked like. We suspected that this caterpillar was coming down from the canopy, where it had spent its life so far, down to the soil to pupate. We brought it back down with us and set up a terrarium for it back at the Smithsonian, in which a beautiful white witch moth emerged a couple of months later. 

While the expedition was over, the fun was not - now we finally got to see what our Mothboxes had been up to for the three nights we were up on the mountain. The Mothbot data platform revealed a rich dataset of over 60,000 detections, each with a timestamp, so we could track insect activity patterns throughout each night, across the entire elevation gradient. We found that activity was greatest at the beginning of the night, but that this pattern varied between different kinds of insects at different elevations. For example, moths had early peaks of activity only at the highest elevations and fairly consistent activity throughout the night lower down, while the opposite was true for beetles. Despite the spectacular mothsheeting we had near camp, the Mothbox data showed that insect diversity was greater lower down the mountain. For a full exploration of our results, check out our paper on the expedition and the documentary shot by Joe See [https://youtu.be/1LGz3aWpeVE].

While staying up late with a Mothsheet is arguably more fun than deploying sensors, this kind of systematic data collection is essential to standardize insect monitoring and make it scalable across landscapes. As more of the Azuero Peninsula is reforested, perhaps one day Cerro Hoya will once again be connected to protected forests across the peninsula, and Mothbox will document the return of biodiversity to the regenerating forests.

References: Szczygieł, Hubert A., Johns, Brianna, Fortet, Bernat, Dent, Daisy H., Quitmeyer, Kit, and Quitmeyer, Andrew. 2026. "Mothbox and Mothbot: Automated light trap and data processing system for scalable insect monitoring." Methods in Ecology and Evolution, https://doi.org/10.1111/2041-210x.70327.

Szczygieł, H. A., Quitmeyer, A., Johns, B., Dent, D. H., Sondhi, Y., Zuleta, D., & Antonelli, A. 2026. “Simultaneous Automated Insect Monitoring Across a Remote Tropical Elevation Gradient With Mothbox.” Integrative and Comparative Biology, https://doi.org/10.1093/icb/icag072