Contents

We provide a zip file with all the observer seabird count data, along with supporting material. The zip file includes:

  1. A description of the count data in text format (readme.txt),
  2. The count data as a comma separated values (CSV) file, suitable for loading into a spreadsheet or other analysis software (seabird-counts.csv),
  3. All the updates performed on the count data as a CSV file (grooming.csv),
  4. Definitions of the seabird taxa referred to in the counts data as a CSV file (seabird-codes.csv),
  5. A copy of a report outlining the methods used, and containing a preliminary exploration of the data (seabirds-around-fishing-vessels.pdf),
  6. A copy of the Creative Commmons Attribution 3.0 NZ licence (licence.txt).

Recommendations

The data on seabird counts were collected by government fisheries observers. The original paper records have been double entered and then lightly groomed to produce a consistent data set. Identifying seabirds at sea is a specialized skill, and the observers have varying levels of experience. For this reason, the identifications should be treated with caution.

All the data were collected from fishing vessels and the counts will depend on the distribution of the seabird taxa, how attracted they are to fishing vessels, the visibility of the birds, how readily they may be identified, and the distribution of observed fishing effort. In general, inshore species will be under-represented, as observer coverage on inshore fishing vessels has been relatively low. Interpretation of the data is further complicated by changes in the protocol. Initially, a single count was made for each species around the vessel, however a split count was then introduced, with birds being counted close to and further away from the vessels.

We recommend that the accompanying report is read for further information before analysing the data.

Licence

Data are made available for reuse by the Department of Conservation, under a creative commons 3.0 New Zealand attribution licence, following the recommendations of NZGOAL. This licence allows the data to be used for any other purpose and republished, provided only that attribution is given to the source.

Data files

seabird-counts.csv

A comma-delimited text file containing the bird count data. These data were collected by government observers on-board commercial fishing vessels while fishing within New Zealand's Exclusive Economic Zone. The observers carried out observations at certain times, generally at the first tow of the day, during which they recorded the number of birds present in the proximity of the vessel. Each row of the dataset represents the number of birds of a single species or species group (identified by a 3 or 4-letter code), recorded during an observation. Since 2007, counts have been split between the birds close and far from the fishing vessel, with the threshold being predominantly at 100 m (but sometimes 50 m). In these cases, there may be two counts for a single observation of a species or species group, but counts can also be only for birds close to the vessel, because there was no birds further away during the observation.

Seabirds were identified to the most accurate taxonomic level possible. Because of the inherent difficulties of counting seabirds around vessels, the variation in the experience of observers, and changes in the protocol with time, the counts should be regarded as indicative only. The data will inevitably contain misidentified birds, and errors in transcribing the raw counts.

Fields

  • id unique number identifying each bird count

  • trip unique identifier to distinguish fishing trips

  • event number of the fishing event during a trip, unique within each fishing trip

  • event_stage stage of the fishing event: After hauling, Before setting, Fishing, Hauling, Setting, Towing

  • event_stage_part part of the event stage: Start, Middle, End

  • observation number of the observation, unique across all observations

  • form_row unique number identifying the row on the original forms where counts were recorded. This number corresponds to the observation until distance-dependent counts were introduced, as observations were split between birds that were close and far from the vessel.

  • date date of the observation (YYYY-MM-DD, New Zealand Standard Time).

  • time time of the observation (HH:SS, New Zealand Standard Time).

  • distance distance from the vessel at which the count was made, either close or far from the vessel, or unspecified

  • threshold distance threshold (in metres) defining 'close' or 'far' in the 'distance' field (mostly 100 m but with some exceptions)

  • code unique code identifying the taxon seen by the observer. The codes are described in 'species-codes.csv'

  • code_description description of the code, i.e., the common name of the species, subspecies, species group, or family. Where possible, the common names follow the convention of the Ornithological Society of New Zealand

  • species scientific name of the species for codes identifying a species or sub-species. Where possible, the taxonomy used by the Ornithological Society of New Zealand is followed

  • genus genus name for codes identifying a genus, species, or sub-species

  • family family name for codes identifying a family, genus, species, or sub-species

  • count number of birds of the type identified by the recorded code

  • longitude, latitude coordinates of the observation, rounded to 0.2 degrees

  • sea_state sea state during the observation, measured on the Beaufort scale

  • fishing_method fishing method used during the fishing event: Trawl, Bottom longline, Surface longline, Set net, Purse seine, Troll.

  • target_fishery identifier of the fishery group. Groups were defined according to the fishing method, vessel length, and target species:

    • SMLT: Trawl vessels less than 28-m long
    • HOKT: Large trawl vessels (>28m) targeting hoki
    • HAKT: Large trawl vessels (>28m) targeting hake
    • LINT: Large trawl vessels (>28m) targeting ling
    • SBWT: Large trawl vessels (>28m) targeting southern blue whiting
    • SQUT: Large trawl vessels (>28m) targeting squid
    • DPWT: Large deepwater trawl vessels (>28m) targeting (oreo species, orange roughy, cardinal fish, and patagonian toothfish)
    • MACT: Large trawl vessels (>28m) targeting blue and Jack mackerel
    • MIDT: Large trawl vessels (>28m) targeting middle-depth species (barracouta, bluenose, alfonsino, warehou, gemfish, ghost shark, spiny dogfish, rubyfish, frostfish, and sea perch)
    • OTHT: Large trawl vessels (>28m) targeting other species, or with missing target
    • ALLB: All bottom-longline fisheries
    • SMLS: Surface-longline vessels less than 45-m long
    • LRGS: Surface-longline vessels over 45-m long
    • SHAN: Set net fisheries targeting shark species
    • OTHN: Set net fisheries targeting species other than shark
    • ALLP: All purse-seine fisheries
  • fishing_year Fishing year, starting on the 1st of October

  • observers identifier for the observers on-board of the vessel during the fishing trip (the observer carrying out the observations is unknown). The names were replaced by random numbers for confidentiality. Where more than one observer was present the identifiers are separated by a semi-colon

Confidentiality

Several modifications were made to the data to follow the regulations of the Ministry for Primary Industries regarding the confidentiality of fishing-related information:

  • Data were removed when there were less than three different fishing vessels for each year and target fishery combination

  • The unique identifiers to distinguish fishing trips and events were recoded and therefore do not correspond with the identifiers used by the Ministry for Primary Industries

  • The observers' names were recoded into unique identifiers

  • The latitude and longitude of the observations were rounded to the nearest 0.2 degrees

grooming.csv

Comma-delimited text file presenting all the changes to the original data made during the grooming process. Changes are listed in the order they were applied to each record.

Fields

  • id number of the count in seabird-counts.csv that was modified

  • action type of modification applied to the count. Either "update" when the value of a field was changed, or "delete" when the count was removed from the dataset

  • field name of the field in seabird-counts.csv that was modified

  • oldval, newval modification applied during an update of the counts, where the value in the field was changed from oldval to newval

  • reason short description of the reason for the modification

species-codes.csv

Comma-delimited text file identifying the taxa that are used in for the seabird counts. The codes are organised into a nested hierarchy, and this file also describes these relationships.

Fields

  • species_code code used to identify the seabird taxon

  • common_name common name

  • scientific_name scientific name

  • common_family common name for the family the taxon is in

  • scientific_family scientific name for the family the taxon is in

  • code_hierarchy level of the taxon within the hierarchy, ranging from one for sub-species to nine for all birds

  • dragonfly whether the code was generated by Dragonfly Science for this project ("t"), or is a code recognised by the Ministry for Primary Industries and Department of Conservation ("f")

  • parent_code the code of the parent taxon within the hierarchy