Data Management Plan (Deprecated)

This data package contains georeferenced underwater photomosaics for more than 600 sites in 0 to 50 meter depths around the islands of Guam and Saipan. The photographs used to create these photomosaics were acquired from August 2021 to May 2022 using digital single-lens reflex camera (DSLR) cameras. Green lasers provided a measurement scale (10 cm). A Trimble GeoXH GPS provided the location of the vessel, and a Blueprint SeaTrac ultra short baseline (USBL) X150 beacon and X010 transponder provided the location and depth (XYZ) of the drop camera. The photographs were georeferenced (i.e., given a latitude and longitude), and mosaicked using a custom Python script (Pierce and Winians, 2023; performed using Agisoft Metashape application programming interface). The field work in Saipan was funded by NOAA National Centers for Coastal Ocean Science (NCCOS) and the NOAA Coral Reef Conservation Program (CRCP). The field work in Guam was funded by the Naval Facilities Engineering Command Marianas (NAVFAC Marianas). Both efforts were conducted in partnership with and in consultation with local partners and the territorial governments.

Lineage Statement:
This data was collected by NOAA National Centers for Coastal and Ocean Science (NCCOS). The photographs and positioning information were acquired from August 20, 2021 to May 12, 2022. The photographs were acquired using Nikon D7500 camera in 2021 and a Sony alpha7 IV mirrorless camera in 2022). Green lasers provided a measurement scale (10 cm). A Trimble GeoXH GPS recorded the location of the vessel, and a Blueprint SeaTrac ultra short baseline (USBL) X150 beacon and X010 transponder recorded the location and depth (XYZ) of the cameras. The resulting photographs were georeferenced (i.e., given a latitude and longitude), and mosaicked to produce photomosaic and 3D model of the seafloor using a custom Python script (Pierce and Winians, 2023; performed using Agisoft Metashape application programming interface).

Process Steps:

• The photographs and positioning information were acquired at approximately 668 sites from August 20, 2021 to May 12, 2022. A minimum of 4 square meters of the seafloor was photographed at each site. The process for collecting these overlapping, underwater photographs was identical at each site. A handheld Garmin GPS 76 unit was used to navigate to each site via a small boat. A drop camera system was then lowered to within 1.5 to 2 meters of the seafloor. This drop camera system included: (1) a downward-looking digital single-lens reflex (DSLR) camera (i.e., a Nikon D7500 camera in 2021 and a Sony alpha7 IV mirrorless camera in 2022), (2) two green lasers spaced 10 cm apart, and (3) a Blueprint Subsea SeaTrac ultra-short baseline (USBL) X150 beacon and X010 transponder. In 2022, an oblique-looking GoPro HERO10 Black in a Spot X Squid housing was also included, which provided real-time video topside for situational awareness. The downward-looking DSLR cameras collected still photographs of the seafloor every 0.5 to 1 s. The camera settings included: focal lengths = 18 mm, white balance = auto, and shutter speed = 1/200 s. Both jpeg and .RAW files were recorded. The photographs overlapped by 60 to 80%. The lasers provided a measurement scale (10 cm) and were visible in some but not all of the photographs. The USBL transponder provided the location and depth (XYZ) of the drop camera every 5 s, and a Trimble GeoXH GPS provided the location of the vessel every 1 s. The Trimble GPS antenna was positioned directly over the USBL transmitter pole to minimize lever arm offsets. For more information, see: Costa, B., and Sweeney, E. (2024). Characterizing submerged lands around Naval Base Guam, Mariana Islands. NOAA Technical Memorandum NOS NCCOS 329. https://doi.org/10.25923/zwwa-h562 Costa, B., Sweeney, E., and Kraus, J. (2024). Characterizing benthic habitats of western Saipan, CNMI. NOAA Technical Memorandum NOS NCCOS 328. https://doi.org/10.25923/psjm-h924
• Underwater photographs were visually reviewed for quality and color corrected (performed using Adobe Lightroom software). The GPS data were differentially corrected using a Continuously Operating Reference Station on Guam and reprojected to the WGS 1984 UTM 55 North coordinate system (performed using GPS Pathfinder Office software). The USBL data were exported from SeaTrac PinPoint software and matched with the corrected GPS locations using timestamps. Underwater photographs were georeferenced using these GPSUSBL locations (ESRI, 2023; performed using ArcGIS Pro, Geotagged Photos To Points function). Not all images were georeferenced because sampling frequencies differed among the USBL (approximately 5 s), GPS (approximately 1 s), and cameras (approximately 0.5 to 1 sec). To georeference the remaining photographs, a custom Python script was created to interpolate positions between the GPS-USBL locations. This script also mosaicked and developed 3D models from the resulting georeferenced photographs (Pierce and Winians, 2023; performed using Agisoft Metashape application programming interface). For more information, see: Costa, B., and Sweeney, E. (2024). Characterizing submerged lands around Naval Base Guam, Mariana Islands. NOAA Technical Memorandum NOS NCCOS 329. https://doi.org/10.25923/zwwa-h562 Costa, B., Sweeney, E., and Kraus, J. (2024). Characterizing benthic habitats of western Saipan, CNMI. NOAA Technical Memorandum NOS NCCOS 328. https://doi.org/10.25923/psjm-h924
• Five to seven substrate and 12 to 14 biological cover types were identified visually in the above georeferenced photographs. The amount of annotated area (four meters squared) was standardized in each photograph so that it matched the spatial resolution (i.e., two x two meter pixels) of the environmental predictors. Multiple substrate and cover types were often present at each site. Percent cover was estimated to the nearest 1% for each habitat type. Percent covers were also converted to presences (1) and absences (0) and used to train or validate the habitat predictions. In Saipan, n =1,109 annotations, where 617 annotations were used for model training and 492 annotations were used for model validation. In Guam, n = 674 annotations, where 346 annotations were used for model training and 328 annotations were used for model validation. For more information, see: Costa, B., and Sweeney, E. (2024). Characterizing submerged lands around Naval Base Guam, Mariana Islands. NOAA Technical Memorandum NOS NCCOS 329. https://doi.org/10.25923/zwwa-h562 Costa, B., Sweeney, E., and Kraus, J. (2024). Characterizing benthic habitats of western Saipan, CNMI. NOAA Technical Memorandum NOS NCCOS 328. https://doi.org/10.25923/psjm-h924