Odontocete detections and corresponding values of environmental variables in the Hawaiian Archipelago
Data files
Oct 19, 2023 version files 106.95 MB
Abstract
This dataset contains detections of echolocation clicks at two sites in the Hawaiian Archipelago. These sites are Hawaii and Manawai (also known as Pearl and Hermes Reef). Echolocation clicks have been labeled using a neural network classifier that was trained and tested on data from the Hawaiian Islands and can successfully identify many species of regionally present odontocetes. During the labeling process, clicks were grouped into five-minute bins and each bin was given a class label. The data provided here is further binned at a daily level, where counts of a given class represent the number of five-minute bins within a given day that were labeled as that class. One file is provided per site, and files are in .csv format that can be read using any desired coding language.
In addition to acoustic counts, values for environmental variables considered in the corresponding manuscript are provided in the CSV files. The final file in this dataset contains satellite-derived chlorophyll-a concentration values from NASA MODIS for the Hawaiian region (used to create Supplementary Fig. 1). Details on all variables and how they were accessed can be found in the manuscript and in the README file accompanying this dataset.
README: Odontocete detections and corresponding values of environmental variables in the Hawaiian Archipelago
https://doi.org/10.5061/dryad.3n5tb2rq4
Description of the data and file structure
This dataset consists of daily counts of odontocete echolocation clicks recorded on HARPs at two sites in the Hawaiian Archipelago and corresponding values of environmental variables. These sites are:
Hawaii: off the western side of Hawaii Island
Manawai (Pearl and Hermes Reef): southwest of Pearl and Hermes Reef
File descriptions follow.
Hawaii_counts_finalData.csv and Manawai_counts_finalData.csv
Each .csv contains a data table with counts of odontocete detections at that site (per day), as well as values of corresponding environmental variables. These variables are:
-sea surface height (meters)
-sea surface salinity (practical salinity units)
-sea surface temperature (degrees Celsius)
-ENSO index
-PDO index
-NPGO index
Surface variable values were determined using HYCOM data (https://www.hycom.org/dataserver/gofs-3pt1/analysis). ENSO index was accessed from
National Oceanographic and Atmospheric Administration (NOAA) Physical Sciences Laboratory: https://psl.noaa.gov/enso/mei/. PDO index was accessed from the NOAA National Center for Environmental Information: https://www.ncei.noaa.gov/access/monitoring/pdo/ and NPGO index from the National Science Foundation and the National Aeronautics and Space Administration: http://www.o3d.org/npgo/.
Odontocete detections are given as daily counts for each of eight classes. These were derived from recordings of echolocation clicks which were labelled using a neural network classifier that was trained and tested on data from the Hawaiian Islands and can successfully identify many species of regionally present odontocetes. During the labelling process, clicks were grouped into 5-minute bins and each bin was given a class label. The data provided here is further binned to the daily level, where counts of a given class represent the number of 5-minute bins within a given day that were labelled as that class.
For site and deployment info, neural network class distinctions, and details on analyses, please refer to:
Ziegenhorn, Morgan A., et al. "Discriminating and classifying odontocete echolocation clicks in the Hawaiian Islands using machine learning methods." PloS one 17.4 (2022): e0266424.
Class codes are as follows:
- class1- Blainville's beaked whale, Mesoplodon densirostris
- class2- Cuvier's beaked whale, Ziphius cavirostris
- class3- false killer whale, Pseudorca crassidens
- class4- Rough-toothed dolphin, Steno bredanensis
- class5_6- short-finned pilot whale, Globicephala Macrorhyncus
- class7_8- stenellid dolphins, mix of Stenella longirostris, Stenella attenuata, and Stenella coruleoalba
- class9- common bottlenose dolphin/ melon-headed whale, Tursiops truncatus/Pepnocephala electra
- class10- Kogia spp (pygmy and dwarf sperm whale), Kogia breviceps and Kogia sima
chlamonth.mat
In addition to species detections, a .mat file of monthly satellite-derived chlorophyll-a concentration from NASA MODIS (https://modis.gsfc.nasa.gov/data/dataprod/chlor_a.php) used to create Supplementary Fig. 1 in the related manuscript is included. This file contains the following variables:
-lats: latitude for all data points
-lons: longitude for all data points
-monthvar: satellite-derived chlorophyll-a concentration at each lat-lon point; each cell is for a unique month from 2009-2019 (milligrams per meter cubed, mg m-3)
-unmonth: time (MATLAB datenum) corresponding to each monthly data point
Notes
More information about all variables and their calculations and derivations can be found in the related manuscript.
Sharing/Access information
Links to other publicly accessible locations of the data:
- https://github.com/MZiegenhorn/Odontocetes-and-Climate-
Related Code
This dataset also contains original code written for the purpose of the paper. The scripts included (and files necessary to run them) are as follows:
- allClimate.mat: a MATLAB file containing values of ENSO, PDO, and NPGO during the years 2009-2019 (monthly).
- climate_correlationAnalysis.mat: MATLAB code used to produce Supplemental Figs. 1-4 in the related manuscript.
- final_timeseries_anom.mat: MATLAB code used to produce Figs. 2 and 4 in the related manuscript.
- gmt_HAWAII.mat: MATLAB code (using Generic Mapping Tools) to produce Fig. 1 in the related manuscript.
- Hawaii_allclasses_counts_finalData.mat; Manawai_allclasses_counts_finalData.mat: MATLAB files containing a table of odontocete detections and corresponding values of oceanographic variables (same data as the .csv files in this dataset, just a different format).
- masterModellingScript.R - An R script used for creating all models in the related manuscript.