Kakila database of marine mammal observation data around the French archipelago of Guadeloupe in the AGOA sanctuary - French Antilles

Study Extent

Sampling Area And Frequency: The data collected were produced by seven different stackeholders. First, two non profit organisations (NPOs) collated marine mammal observations during daily trips regularly made over years by volunteers: OMMAG (Observatoire des Mammifères Marins de l’Archipel Guadeloupéen or Observatory of marine mammals of the Guadeloupe archipelago), and BREACH Antilles, whose observations where part of the dataset analyzed by Nadège Gandhillon. The older data are been recorded in 2000. Second, several professional wale-watchers recorded all the observations made during their daily tours. These wale-watchers are Guadeloupe Evasion Découverte (located in Deshaie), Cétacés Caraïbes (located in Bouillante) and Aventures Marines (located in Gourbeyre). They all gave us a free access to their entire datasets. We also used publicily available data coming from observation campains conducted by the IFAW (International Fund for Animal Welfare) in 1995, 1996, 200 and 2006.

Sampling

Sampling consisted first in conducting a preliminary survey of the different associations and professional whale-watchers know to collate cetacean observation data for years aroud the Guadeloupean Archipelago, and whose expertise was previously recognized (e.g. by co-authorship of scientific publications, by the defense of a PhD (Gandhillon 2012) and by the publication of books for the general public. We got in touch with several of these different associations and whale-watchers, to obtain their collaboration and agreement to share their own data in a more global aggregated database using newly designed dataset. Following this first survey, an authorization of use and analysis of the data was sent to each organisation. All agreed to share and open data once the agregated database woud be finalized. The observations were all collected during boat excursions: marine mammals were spotted and identified at a greater or lesser distance from the boat depending mainly on the visibility and the sea state. Geolocation coordinates were also generally collected thanks to a GPS. The observation data were collected in French language. Each collector had its own tabulated file for entering observations with no central data store and access interface. But all these datasets possesed many variables are in common.

Quality Control

The observation data come from multiple sources with varying levels of expertise. A specific level of expertise was assigned to each observer (ie. beginner, intermediate, confirmed levels) in order to attest to the quality of his observation. More globally, an effort to centralize and harmonize siloed data has been made with quality controls checking links between tables with pivot tables. Content quality controls were also used such as a controlled dropdown menu for many fields that avoid potential input errors. Geolocations, often transformed into decimal degrees, were verified using the Geographic Information System QGIS 3.10 Desktop (long term release) software. The structure of the Kakila database was based on the existence of variables common to all datasets and on the functional dependencies between the data. First, variables were defined with the help of a working group constituted by the scientific team who coordinated the project. This new list of variable was approved by the data providers. Then a data dictionnary was defined. The aim of this dictionnary was to produce a precise definition or description of each of the fields, based on validated scientific frameworks. The data dictionary is essential to guarantee the reusability of the database. For example, in the case of future data entry by different operators, the data dictionary ensures the perfect understanding of fields and limits input errors. In addition, data were checked for errors: 10% of the entries were randomly selected and checked by two persons. One person carried out the random draw among the “observation” table and the other operator checked the selected lines in the database against the original datasets provided by the data owners. The data entry was invalidated if it contained an error in any field. The error rate was calculated as follow : it was the rate of the number of data entries containing an error on the total number of checked data entries and was estimated at 0.073 in the Kakila database.

Method steps

1. Citizen-science approach. Data integration made into a database complying with the FAIR principles using a step-by-step community approach and a pragmatic method taking into account the constraints of the stakeholders. Observation collected over several years by local NPOs or whale watchers have been integrated in a database named "Kakila".

1. Cetacea
   rank: order
2. Balaenopteridae
   rank: family
3. Delphinidae
   rank: family
4. Kogiidae
   rank: family
5. Physeteridae
   rank: family
6. Ziphiidae
   rank: family
7. Balaenoptera acutorostrata
   rank: species
8. Balaenoptera borealis
   rank: species
9. Balaenoptera edeni
   rank: species
10. Balaenoptera physalus
    rank: species
11. Megaptera novaeangliae
    rank: species
12. Feresa attenuata
    rank: species
13. Globicephala macrorhynchus
    rank: species
14. Grampus griseus
    rank: species
15. Lagenodelphis hosei
    rank: species
16. Orcinus orca
    rank: species
17. Peponocephala electra
    rank: species
18. Pseudorca crassidens
    rank: species
19. Stenalla coeruleoalba
    rank: species
20. Stenella attenuata
    rank: species
21. Stenella clymene
    rank: species
22. Stenella frontalis
    rank: species
23. Stenella longirostris
    rank: species
24. Steno bredanensis
    rank: species
25. Tursiops truncatus
    rank: species
26. Kogia breviceps
    rank: species
27. Kogia sima
    rank: species
28. Physeter macrocephalus
    rank: species
29. Mesoplodon densirostris
    rank: species
30. Mesoplodon europeaus
    rank: species
31. Mesoplodon mirus
    rank: species
32. Ziphius cavirostris
    rank: species

AGOA Sanctuary