Where the cetaceans are in French waters: the maps of the DCSMM assessment

Every six years, France must report to Europe on the state of its marine waters. For cetaceans, that assessment rests on low-altitude aerial surveys, the MEGASCOPE surveys of the Observatoire Pelagis, whose sightings feed distribution models. The result has just been published in open access: 58 maps covering seven species, by season and by year, over the Atlantic, the English Channel and the western Mediterranean. Here they are, made navigable.

A probability, not a count

This is the first thing to understand, and it changes everything. Each grid cell carries a probability of presence between 0 and 1, that is, the chance of encountering the species there, and not a number of animals. A cell at 0.8 does not hold eight times more dolphins than a cell at 0.1: it holds them more often.

The practical consequence is that two species are not compared with one another on these maps. The common dolphin shows a mean presence of 0.42 in the Atlantic and the fin whale 0.12, but that does not say that there are three and a half times more dolphins than fin whales: it says that the model expects the common dolphin almost everywhere and the fin whale in more restricted areas. Each species is read against itself, from one season to the next.

Each cell also carries its standard error, which can be consulted by clicking the map. It is the measure of what the model does not know, and it is often of the same order as the differences one would like to comment on.

The September peak, and the species that escapes it

Recomputing the mean of each map, one pattern stands out: in the Atlantic, six species out of seven peak in September, and fall back in May as in November.

Species, Atlantic seaboard	May	September	November
Common dolphin	0.386	0.467	0.406
Common minke whale	0.259	0.249	0.254
Risso’s dolphin	0.192	0.220	0.198
Common bottlenose dolphin	0.156	0.186	0.162
Long-finned pilot whale	0.157	0.183	0.161
Striped dolphin	0.151	0.174	0.156
Fin whale	0.104	0.133	0.109

The exception is instructive. The common minke whale is the only species whose presence does not rise in September: it is even very slightly lower than in May. It is also the only one on the list that is not a great seasonal traveller in these waters, and the most coastal of the seaboard’s whales.

One reservation is called for, however: a distribution model renders the conditions in which the flying was done as much as the season. September offers a calmer sea and better detectability than November. Correcting for survey effort is part of the method, but it is never perfect, and this pattern deserves to be read as a tendency rather than as a measurement.

In the Mediterranean, summer changes everything

The seasonal contrast there is far sharper than in the Atlantic, and it concerns the fin whale above all. Its mean presence goes from 0.009 in January 2019 to 0.040 in July 2018, a ratio of one to four between two neighbouring dates. The long-finned pilot whale follows the same movement, from 0.065 to 0.111.

This is no surprise to anyone who knows the Pelagos Sanctuary: the fin whales gather in summer in the Ligurian Sea to exploit the krill swarms that the upwellings concentrate there. The map shows it at the scale of the basin, and this matches what our page on the cetaceans of the Mediterranean describes.

What a comparison between two editions is worth

The map makes it possible to subtract the first year from the last, and we should explain why we labelled this mode difference between two model outputs rather than change over time.

The differences observed between 2015 and 2020 are small: at most 0.018 point for the common dolphin in September, the largest movement in the set. Yet the standard error of a cell frequently exceeds that value. In other words, at the scale of the seaboard, these maps show no clear redistribution over five years, and to claim otherwise would be to over-interpret noise. The comparison mode indeed flags, cell by cell, when the difference is smaller than the standard error.

What changes between two editions also depends on the input data: different surveys, different effort, sometimes a revised model. Five years, for animals that live for decades and travel thousands of kilometres, is a short time step.

Where these maps come from

From the MEGASCOPE surveys, conducted by the Observatoire Pelagis aboard aircraft flying at low altitude along predefined transects. Observers record each group sighted, its species, its size and its distance from the track. These sightings, corrected for the probability of detection, feed models that relate presence to environmental conditions, then predict across the whole area, including where the aircraft did not pass.

It is the same data foundation as that of our map of sensitivity to offshore wind, and the same observatory as that of the SAMM surveys. The indicator carries the code D1C4 of the Marine Strategy Framework Directive, and feeds the OSPAR assessment of the North-East Atlantic.

Three limits, one of which we had to settle ourselves

Seven species, not all of them. The harbour porpoise, the sperm whale and the beaked whales do not appear in this indicator, for want of a sufficient number of sightings to fit a reliable model. Their absence from the map means nothing as to their presence at sea.

Two different grids. The Atlantic is divided into 4,565 regular rectangular cells, the Mediterranean into 1,541 slightly oblique cells, inherited from a different projection. We straightened the latter, the departure from the exact outline being some ten metres.

Duplicates in the Mediterranean. The Mediterranean layers deliver 2,429 features for 1,541 cells: 770 cells carry two to four superimposed estimates, of identical geometry and differing values. Neither the metadata nor the lineage explains this, and the most likely hypothesis is a mosaic of overlapping model outputs. We averaged these values, the least presumptuous treatment for repeated estimates of the same quantity. The choice is debatable; it is flagged here so that everyone can judge it.

Sources

Data: 2024 DCSMM assessment, indicator D1C4 “cetacean distribution”, MEGASCOPE surveys of the Observatoire Pelagis (La Rochelle Université and CNRS), distributed by Ifremer through the WFS service Sextant under the Licence Ouverte. Each layer carries its own DOI in the catalogue. Indicator attached to the OSPAR QSR 2023 assessment of the abundance and distribution of cetaceans. The means, maxima and differences shown on this page are recomputed by us from the original layers.