Compact and modular autonomous surface vehicle for water research: The Naval Operating Research Drone Assessing Climate Change (NORDACC).

Research, monitoring, and management of marine and aquatic ecosystems often require surface water samples to measure biogeochemical and optical parameters. Traditional sampling with a boat and several personnel onboard can be labor-intensive and safety requirements limit sampling activities in high-risk environments. This paper describes the Naval Operating Research Drone Assessing Climate Change (NORDACC). NORDACC is an open source, light-weight, and portable autonomous surface vehicle that can acquire surface water samples while also measuring sea surface temperature and salinity for the duration of its deployment. NORDACC is ideal for operations in remote areas where resources and personnel are limited. Two sample bottles, each one liter in volume, can be filled, either at pre-programmed sampling stations or manually, using the remote control. A trimaran design provides buoyancy and stability, with hulls constructed of vacuum-formed acrylonitrile butadiene styrene (ABS) plastic. NORDACC can navigate autonomously between waypoints and features first person view capabilities for enhanced situational awareness. NORDACC's performance was validated in Aarhus Bay, Denmark, collecting multiple surface water samples in winds in excess of 8 ms-1 and steep, choppy waves.

Accurate image-based identification of macroinvertebrate specimens using deep learning-How much training data is needed?

Image-based methods for species identification offer cost-efficient solutions for biomonitoring. This is particularly relevant for invertebrate studies, where bulk samples often represent insurmountable workloads for sorting, identifying, and counting individual specimens. On the other hand, image-based classification using deep learning tools have strict requirements for the amount of training data, which is often a limiting factor. Here, we examine how classification accuracy increases with the amount of training data using the BIODISCOVER imaging system constructed for image-based classification and biomass estimation of invertebrate specimens. We use a balanced dataset of 60 specimens of each of 16 taxa of freshwater macroinvertebrates to systematically quantify how classification performance of a convolutional neural network (CNN) increases for individual taxa and the overall community as the number of specimens used for training is increased. We show a striking 99.2% classification accuracy when the CNN (EfficientNet-B6) is trained on 50 specimens of each taxon, and also how the lower classification accuracy of models trained on less data is particularly evident for morphologically similar species placed within the same taxonomic order. Even with as little as 15 specimens used for training, classification accuracy reached 97%. Our results add to a recent body of literature showing the huge potential of image-based methods and deep learning for specimen-based research, and furthermore offers a perspective to future automatized approaches for deriving ecological data from bulk arthropod samples.

A mobile observatory powered by sun and wind for near real time measurements of atmospheric, glacial, terrestrial, limnic and coastal oceanic conditions in remote off-grid areas.

Climate change is rapidly altering the Arctic environment. Although long-term environmental observations have been made at a few locations in the Arctic, the incomplete coverage from ground stations is a main limitation to observations in these remote areas. Here we present a wind and sun powered multi-purpose mobile observatory (ARC-MO) that enables near real time measurements of air, ice, land, rivers, and marine parameters in remote off-grid areas. Two test units were constructed and placed in Northeast Greenland where they have collected data from cabled and wireless instruments deployed in the environment since late summer 2021. The two units can communicate locally via WiFi (units placed 25 km apart) and transmit near-real time data globally over satellite. Data are streamed live and accessible from (https://gios.org). The cost of one mobile observatory unit is c. 304.000€. These test units demonstrate the possibility for integrative and automated environmental data collection in remote coastal areas and could serve as models for a proposed global observatory system.

Lightweight drone-deployed autonomous ocean profiler for repeated measurements in hazardous areas - Example from glacier fronts in NE Greenland.

Accelerated melting of ice in Polar Regions due to global warming increases freshwater input to coastal waters from marine terminating glaciers. Lack of measurements near the glacier terminus limits our knowledge of the mixing processes between freshwater and the underlying ocean. We present a low-cost (< € 3200) and lightweight (2.6 kg) drone-deployed, retrievable conductivity, temperature and depth (CTD) instrument for remote controlled (1 km) autonomous profiling in highly hazardous and remote areas. The instrument was deployed with a drone taking off from land and marine vessels to perform measurements near tidewater glaciers termini of the Greenland ice sheet. The free-flowing profiler is reusable due to a compact ballast based single-shot buoyancy engine and post-profiling pickup by drone. It can reach a depth of up to 250 m, and is equipped with low-cost sensors for conductivity, temperature, and depth measurements. During decent the profiler reaches a velocity of about 0.48 m/s, resulting in about 3.5 data points pr. m depth, but is designed to easily vary the velocity by changing buoyancy setup before deployment. Successful tests were conducted at marine terminating glaciers in Northeast Greenland in August 2021.

Deep learning and computer vision will transform entomology.

Most animal species on Earth are insects, and recent reports suggest that their abundance is in drastic decline. Although these reports come from a wide range of insect taxa and regions, the evidence to assess the extent of the phenomenon is sparse. Insect populations are challenging to study, and most monitoring methods are labor intensive and inefficient. Advances in computer vision and deep learning provide potential new solutions to this global challenge. Cameras and other sensors can effectively, continuously, and noninvasively perform entomological observations throughout diurnal and seasonal cycles. The physical appearance of specimens can also be captured by automated imaging in the laboratory. When trained on these data, deep learning models can provide estimates of insect abundance, biomass, and diversity. Further, deep learning models can quantify variation in phenotypic traits, behavior, and interactions. Here, we connect recent developments in deep learning and computer vision to the urgent demand for more cost-efficient monitoring of insects and other invertebrates. We present examples of sensor-based monitoring of insects. We show how deep learning tools can be applied to exceptionally large datasets to derive ecological information and discuss the challenges that lie ahead for the implementation of such solutions in entomology. We identify four focal areas, which will facilitate this transformation: 1) validation of image-based taxonomic identification; 2) generation of sufficient training data; 3) development of public, curated reference databases; and 4) solutions to integrate deep learning and molecular tools.

A cost-efficient low-weight autonomous profiler for measurements in polar coastal waters and other regions with strong density gradients.

The rapid warming of our planet has resulted in accelerated melting of ice in polar regions. Currently we have limited knowledge on how, where and when the surface meltwater layer is mixed with the underlying ocean due to lack of observations in these remote areas. We present a lightweight (17 kg) and low-cost (6000€) instrument for autonomous profiling across the strongly stratified upper layer in Arctic coastal waters, freshened by the riverine input and meltwater from glaciers, icebergs, and sea ice. The profiler uses a specially designed plunger buoyancy engine to displace up to 700 cm3 of water and allows for autonomous dives to 200 m depth. It can carry different sensor packages and convey its location by satellite communication. Two modes are available: (a) a free-floating mode and (b) a moored mode, where the instrument is anchored to the seafloor. In both modes, the profiler controls its velocity of 12 ± 0.3 cm/s resulting in 510 ± 22 data points per 100 m depth. Equipped with several sensors, e.g. conductivity, temperature, oxygen, and pressure, the autonomous profiler was successfully tested in a remote Northeast Greenlandic fjord. Data has been compared to traditional CTD instrument casts performed nearby.

An affordable and miniature ice coring drill for rapid acquisition of small iceberg samples.

Icebergs account for approximately half of the freshwater flux from the Greenland Ice Sheet and they can impact marine ecosystems by releasing nutrients and sediments into the ocean as they drift and melt. Parameterizing iceberg fluxes of nutrients and sediments to fjord and ocean waters remains a difficult task due to the complexity of ice-ocean interactions and is complicated by a lack of observations. Acquiring iceberg samples can be difficult and dangerous, as icebergs can break apart and roll without warning. Here we present open source design files for a small, lightweight ice coring drill that can be reproduced using modern computer numerical control (CNC) machining and 3D printing technology. This ice core drill can rapidly acquire small ice samples from icebergs and bergy bits using a standard commercial, off-the-shelf battery-operated hand drill. Design files and a recent field expedition to Northwest Greenland are described. Ice core collection required only 30 s, thereby minimizing risks to scientists.