Derville et al. 2025: Humpback whale photos (batch 3)

Dataset prepared for publication in Figshare on March 30th, 2025
Dataset prepared by: Solène Derville ([e-mail])
Associated reference: Derville, S., Calambokidis, J., Palacios, D. M., Tackaberry J., Hayslip C., Flynn K., New L., Baker C. S., Steel D., Martien K. K., Moore J. E., Lauf M., Wickman L., Wall A., Newell C., Torres L. G. (2025) Rates of entanglement inferred from scarring prevalence of humpback whales photographed in US Oregon waters, Endangered Species Research
These photos are the property of the laboratory that have collected and curated them. For questions relative to these photos datasets, please contact:
WHET Lab contact: Craig Hayslip ([e-mail])
GEMM Lab contact: Leigh G. Torres ([e-mail])
CRC contact: Jenn Tackaberry ([e-mail] ) or John Calambokidis ([e-mail])

Keywords are added to the photo metadata to code for various variables used in the study. Keywords useful for the analysis are listed and described below. All other keywords found in the photo metadata can be ignored.
Keywords coding for photo source
Source-CRC (photos contributed by Cascadia Research Collective), Source-GEMMLab (photos contributed by the GEMM Lab -Oregon State University- and their field partners), Source-WHETLab (photos contributed by the WHET Lab – Oregon State University)
Keywords coding for photo quality
From best to lowest quality: QUA1, QUA2, QUA3, QUA4

Keywords coding for photo type
Dorsal tailstock, Fluke underside, Forward tailstock, Perpendicular tailstock, Ventral tailstock

Keywords coding for Wall et al. (2019) categories of entanglement likelihood
From least to most likely: ENT4-Unlikely, ENT3-Possibly, ENT2-Likely, ENT1-Most likely

Keywords coding for Robbins & Mattila (2001) scar codes per region
Perp-code-LFlank = left flank scar code
From least to most likely: LFlank-S0, LFlank-S1, LFlank-S2, LFlank-S3, LFlank-S4, LFlank-S5
Perp-code-RFlank = right flank scar scode
From least to most likely: RFlank-S0, RFlank-S1, RFlank-S2, RFlank-S3, RFlank-S4, RFlank-S5
Perp-code-LFluke = left fluke trailing edge scar code
From least to most likely: LFluke-S0, LFluke-S1, LFluke-S2, LFluke-S3, LFluke-S4, LFluke-S5
Perp-code-RFluke = right fluke trailing edge scar code
From least to most likely: RFluke-S0, RFluke-S1, RFluke-S2, RFluke-S3, RFluke-S4, RFluke-S5
Perp-code-VP = ventral tailstock scar code
From least to most likely: VP-S0, VP-S1, VP-S2, VP-S3, VP-S4, VP-S5
Perp-code-DP = dorsal tailstock scar code
From least to most likely: DP-S0, DP-S1, DP-S2, DP-S3, DP-S4, DP-S5

Keywords coding for individual ID
All keywords starting with “CRC”… (eg., CRC13626, CRC11523 etc.) are unique codes given to individual whales. For most photos, these IDs are incorporated directly as keywords. For those that do not have such code, refer to the score_df dataset provided as supplementary material (https://github.com/SoleneDerville/HW_scarring_analysis) to find the unique code of individuals found in each photo.

Other keywords of interest
4-Fresh = keyword present if photo shows fresh / raw injuries
6-Strike = keyword present if scars indicative of a potential ship strike are found on photo
7-Mirrored_fluke_notches = keyword present if opposite cuts or deep indentation are found on the trailing edge of the fluke (criteria not used in the present study)
8-Fluke-entirely visible = keyword present for fluke underside photos if the fluke is entirely above water
NB: Although this dataset is licensed under CC BY 4.0 due to Figshare limitations, the authors explicitly prohibit any commercial use. Any reuse must include proper citation of the source and authors. For commercial licensing, please contact the authors.