Table of Contents

• Coverage
• Dataset Description
  • Methods & Sampling
  • Data Processing Description
  • BCO-DMO Processing Description
  • Problem Description
• Data Files
• Supplemental Files
• Parameters
• Instruments
• Project Information
• Funding

This dataset contains image files and a metadata table in the "Data Files" section and additional "Supplemental Files" including processing code and a file inventory.

For in situ mysid behavioral observations, we have developed and deployed a custom-built stereoimaging system capable of capturing synchronized high-quality images at a high frame rate (up to 21 fps). Our current setup features Teledyne Blackfly S GigE cameras with 5 MP resolution, IEEE1588 clock synchronization, and PoE support. Navitar lenses, with focal lengths of 12 mm or 25 mm, are attached via C-type lens mounts to ensure compatibility. Lighting is provided by a TrustFire DF50 scuba diving light with white and red LED modes. The stereo housing, designed in-house, comprises galvanized aluminum structures and acrylic end caps housing optical-grade glass lenses. It offers flexibility in inter-camera distance and angle. Custom-made underwater DragonFish Ethernet cables facilitate power and data transfer, enabling live feed monitoring, and synchronized camera frame storage, while also simplifying power management, thus extending deployment time. 

The stereo housing and lighting system are affixed to an aluminum bar frame, which is lowered into the water using a crane. Initially, the frame is positioned at a shallow depth for calibration purposes. An 8x9 checkboard pattern with a 10 mm square size, printed on an aluminum composite panel, is submerged for calibration. Synchronized images are captured by both cameras as the checkerboard is moved to cover the common field of view and focused depth. Once calibration images are recorded, the camera rig is lowered to the desired recording depth. Live feed monitoring ensures proper lighting and clear visibility before recording synchronized images commences. After sufficient images are captured, recording is stopped, and the stored data undergoes quality and synchronization analysis. Finally, details such as the data recording location, folder identification, camera synchronization status, and the number of recorded images are documented in our field notes before proceeding to the next data recording session. Calibration is typically required only once at each location.  File naming conventions: Maine2023_Pxx or Maine2023_Sxx as the naming protocol (for primary and secondary cameras, where xx represents the number of the dataset, e.g., P80, S80).

The original field data comprises 8-bit monochrome TIFF or BMP image files recorded synchronously by two cameras. The uploaded data consists of movies (MP4 video format) of the raw images, i.e., each movie is a compilation of all the raw images in a single dataset. 

A Python script was used to process a sequence of image files from a specified folder (a single dataset) and generate an output video using the 'mp4v' codec. The video's frame rate matches that of the image recording rate. You can find the Python script code we used within a python notebook for this purpose within the supplemental file 'video_from_imgs.txt

Mysid videos were bundled up into two zip packages with zip64 support using 7z. These two zip files were added to the "Data File" section (Mysid_movies_summer2023_Maine_P.zip, Mysid_movies_summer2023_Maine_S.zip).
* File inventory was created containing the imagename, filesize and checksums. Added columns to contain identifiers within the image name: image set id (primary or secondary folder), and Primary_or_Secondary indicating if it was the primary or secondary camera. (file_inventory.csv)

Sheet "Metadata" within file Maine_dataset_report_5.xlsx was imported into the BCO-DMO data system. With additional added columns with the associated primary and secondary movie name. Metadata table added as a supplemental file.
* Date converted to ISO 8601 format, ISO timestamps with timezone added (UTC) in addition to local times.
* Lat lon converted to decimal degrees

* Column names adjusted to conform to BCO-DMO naming conventions designed to support broad re-use by a variety of research tools and scripting languages. [Only numbers, letters, and underscores.  Can not start with a number]

NSF Award Abstract:
Mysid shrimp are a group of zooplankton that are distributed in diverse aquatic environments – freshwater, marine and estuarine – in both shallow and deep waters throughout the world. Due to their ubiquitous presence, mysids form a critical link in marine food webs, yet little is known about their ecology and interactions with their local surroundings. An interdisciplinary team of investigators will develop and deploy a novel in situ imaging system in coastal Maine over the span of two summers, conducting a comprehensive investigation of mysid behavior in their natural environment, including their interactions and distribution under varying local flow, food, and light conditions. The investigators will train two graduate students at the interface of fluid mechanics, instrumentation, and ecology. In addition, ten students from Florida Atlantic University and the University of Southern Maine will be invited to participate in annual workshops on ecological fluid mechanics, with first-generation college students or those from other underrepresented communities encouraged to apply. The cutting-edge instrumentation suite, including the imaging system being developed as part of this project, can be used to address different questions associated with spatial patterns and zooplankton or fish behavior in the future. Image datasets will be openly accessible to the scientific community and the public. Project results will be disseminated through public outreach lectures at the Harbor Branch Oceanographic Institute, Darling Marine Center, and the Woods Hole Oceanographic Institution.

Mysids have been historically understudied compared to other zooplankton groups (e.g., copepods, krill), despite their ubiquitous presence and importance to marine food webs. This project will investigate mysid behavior, feeding ecology, and swarm distributions in relation to prey availability, light levels, and background flow conditions, focusing on a single mysid species – Neomysis americana. The investigators will use a state-of-the-art suite of instruments in the field, including two separate imaging systems and a high resolution Acoustic Doppler Profiler, to conduct field experiments at the Damariscotta River Estuary. This instrumentation suite will collect data to analyze mysid distributions and animal-fluid interactions in their natural environment, as well as ancillary data on benthic particle/plankton community composition and physical parameters (including currents, waves, temperature and depth). Multi-spectral analysis will find correlation patterns with water turbidity, flow conditions, bottom topography, mysid abundance, and food source concentrations. Mysid samples will be collected several times a week throughout the summer months to monitor changes in the mysid population abundance, sex ratio, and stage structure. The investigators will determine how seasonal changes in environmental variables and Neomysis population structure affect distribution within the estuary, as well as small-scale swimming and aggregation behavior. They will answer questions related to how swarm organization, aggregation, and swimming behavior differ under conditions of varying zooplankton abundance and swarm composition. Results will ultimately improve our understanding of marine ecosystem dynamics.

This project is jointly funded by the Biological Oceanography Program and the Established Program to Stimulate Competitive Research (EPSCoR).

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.