Abstract:

These 1-m vertically binned concentration data were collected with the In Situ Ichthyoplankton Imaging System (ISIIS) to describe the distribution and concentration of organisms along a transect immediately downstream of the Mobile Bay plume. Data were collected from the R/V Point Sur cruise PS16-20 as part of the CONCORDE program to examine how variable river discharge affects larval fish fine-scale distributions and predator-prey spatial relationships. Larval fishes, their planktonic prey (calanoid copepods), and gelatinous planktonic predators (ctenophores, hydromedusae, and siphonophores) were sampled across multiple freshwater pulses exiting Mobile Bay in the northern Gulf of Mexico (USA) during a large freshwater discharge event (2016-04-09 - 2016-04-11). Sampling transect 023 was carried out during daylight (April 9, 2016 10:13-13:52 CDT), while transects 024 (April 9-10, 2016 21:27-02:11 CDT) and 026 (April 10-11, 2016 21:04-00:42 CDT) were conducted at night. A sparse Convolutional Neural Network (sCNN) automated classification of different organisms within the ISIIS images. The data include 1-m vertically binned concentrations for each listed plankton taxa along with corresponding average environmental data (temperature, salinity fluorescence, oxygen, PAR) and example images of the inventoried taxa. Due to its size, the entire library of images is not available but may be requested from the dataset author.

Suggested Citation:

Axler, Kelia, Christian Briseño-Avena, Su Sponaugle, Frank Hernandez, Adam Greer, and Robert Cowen. 2020. In Situ Ichthyoplankton Imaging System (ISIIS) classified taxa concentrations and environmental metadata for 3 transects from the Mobile Bay plume process study, R/V Point Sur cruise PS16-20 (PTS02), 2016-04-09 to 2016-04-11. Distributed by: GRIIDC, Harte Research Institute, Texas A&M University–Corpus Christi. doi:10.7266/9211C8TM

Purpose:

In situ imagery data and associated environmental metadata were collected during the CONCORDE Spring 2016 cruise to examine how variable freshwater discharge can affect fine-scale larval fish distributions and predator-prey spatial relationships.

Data Parameters and Units:

The concentration data are contained in four worksheets. The first contains an explanation of parameters for each transect. Each of the remaining worksheets details the concentrations for one of the three transects. Parameters are: TIME_START (CDT, [Matlab format]), TIME_STOP (CDT, [Matlab format]), BIN_MIDDEPTH_M [m], LATITUDE_DEG [decimal degrees N], LONGITUDE_DEG [decimal degrees E], TEMPERATURE_DEG_C [degrees C], SALINITY, FLUORESCENCE_VOLTS [V], PAR_UE_PER_M2 [uE/m^2], O2_mg_PER_L [mg/L], DENSITY_Sigma (sigma, [kg/m^3]), and observed taxa concentrations [individuals/m^3]: fish, Engraulidae, Gobiidae, Microdesmidae, Ophidiidae, Unknown, Plueronectiformes, Scorpaenidae, Sciaenidae, Calanoid copepods, Ctenophores, Hydromedusae, and Siphonophores.

Methods:

Images of fish larvae and zooplankton were captured using the ISIIS, a towed shadowgraph imager that uses a line-scan camera to sample large volumes of water (150-185 L s-1). Two cameras imaged zooplankton between approximately 500 µm and 12 cm in length while simultaneously measuring salinity, temperature, and depth (Sea-Bird Electronics 49 FastCAT), dissolved oxygen (SBE 43), Chl-a fluorescence (Wet Labs FLRT), and photosynthetically active radiation (PAR; Biospherical QCP-2300). The images and oceanographic data are linked by a common timestamp, which enables a detailed description of the physical environment for each individual organism. A sparse Convolutional Neural Network (sCNN) was used to automate the identification of imaged taxa following Luo et al. (2018). Although 89 different classes of plankton taxa in the CONCORDE data and 693 million images were automatically identified from the three plume transects in our study, only key taxa that were deemed to be ecologically important predators or prey are presented in this data set. Plankton was combined into higher taxonomic groupings to enable the comparison of fish larvae with their prey (calanoid copepods) and predator groups (ctenophores, hydromedusae, and siphonophores). Images of larval fishes from each transect were extracted by the sCNN using the same automated methodology as the zooplankton imagery. However, automatically classified larval fishes were manually reviewed by a human expert to verify correct identifications and to achieve lower taxonomic classifications than the sCNN was trained to do. Taxa concentrations [ind. m^-3] were calculated by using the volume of water imaged, average tow speed, and time spent in each 1-m vertical bin. The length of each horizontal bin was roughly similar but varied slightly (<10s m) depending on the tow speed of the ISIIS vehicle and time spent in each 1-m vertical stratum.

Instruments:

Sea-bird Electronics SBE 43, Sea-bird Electronics 49 FastCAT, Biospherical QCP-2300, Wet Labs FLRT.

Individual Files: