On the 14th of August 2019, my colleagues and I published a paper in PLoS ONE, wherein we talk about the way that the IUCN Red List deals with species complexes and taxonomic uncertainty, and the real meaning of ‘Data Deficient’. As I have written about this topic here before, I wanted to share a little more about this topic here (I also posted a thread on twitter on this topic on the 19th of August).
A less technical companion article to this one has been published in The Conversation.
The IUCN Red List: a vital but blunt instrument of conservation
The Red List is perhaps the single most widely-used resource for conservation. When we talk about how threatened animals or plants are, we are almost always referencing their threat statuses in the Red List. To briefly summarise how it works: species are given a rank on a scale of Least Concern (no conservation action needed), Near Threatened (almost qualifies for one of the threatened categories), Vulnerable (at risk of falling down the scale), Endangered (conservation action needed), and Critically Endangered (urgent conservation action needed). Two more ‘threat’ ranks exist: Extinct in the Wild, and Extinct, but we hope that we never have to use these. Finally, there are two more categories that do not have any bearing on the conservation status of the organisms; these are Data Deficient (we don’t know enough to assign a threat rank) and Not Evaluated (we haven’t figured it out yet, but wanted to keep its name in the list).
Where do those ranks come from? Well, they’re given in consultation between Red List Assessors and experts on the organisms. This is something the Red list does really well—getting expert opinions. Who knows the threats better than an expert on the group? The IUCN has Species Survival Commission (SSC) Specialist Groups, who work on coordinating conservation assessment of specific groups of animals and plants. I, for example, am a member of the IUCN SSC Amphibian Specialist Group, and have been actively involved in assessment of most of Madagascar’s frogs.
Assessments are done according to the IUCN Red List criteria. There are five criteria, lettered A–E. These are surprisingly easy to understand at a basic level; you can download a simple reference sheet of what the criteria refer to in English, French, and Spanish here. The full, detailed guidelines can be found here in the same three languages.
If you take a look at these reference sheets or guidelines, you will note that all of the criteria except criterion B refer to population size or some kind of prediction of extinction risk. Criterion B, on the other hand, refers to distribution area, the so-called ‘spread of risk’ to a species. This criterion rests on the logic that organisms concentrated in one tiny area have a concentrated risk, while those spread over a large area have a low risk as a whole, even if individual populations are extirpated. Thus, we can guesstimate the extinction risk of an animal based on the area over which it is found, and the threats facing it or its habitat within that area.
As area is so much easier to quantify than population size, a huge number of species are assessed under Criterion B. Of all 6022 assessed frogs, 1909 are in threatened categories (Vulnerable, Endangered, Critically Endangered). Of these 83.6% are assessed under criterion B! A small percentage are classified under Criterion D2, which is the only other criterion that is also based on distribution area.
Criterion B acts as our blunt instrument in the conservation assessment of animals where we do not (and sometimes simply cannot) have an idea of how large their population sizes are, and modelling of extinction risk is not possible, because we do not know enough to do so.
Indeed, Criterion B can be applied in cases where nearly nothing is known about a species except its whereabouts. Animals assessed under Criterion B are sometimes known from a single specimen. This is because the current IUCN guidelines strongly discourage the use of the category Data Deficient (DD), and encourage the use of even minimal data to assess the species. The result is the following: technically, as long as an animal is known from at least one geo-referenced specimen, it is found in a tiny area and known from a single location. As long as some kind of threat can at least be hypothesised for that location and/or species, the animal qualifies as Critically Endangered under Criterion B1ab(iii). Thus, based on this single data point, we can avoid using DD, even though our knowledge is, in fact, data-deficient in every possible way.
As you can imagine, this is quite a problem. There is no way to judge the quality of data behind a Red List assessment without diving into that assessment itself, so a critically endangered rhinoceros looks the same on the list as a critically endangered frog, even though the former can have every individual tagged and collared, while the latter is known from a single data point.
Species complexes are… complex
So what happens when we have a species complex? A species complex is a group of species that are difficult or impossible to discern based on current evidence. A huge number of ‘species’ are actually turning out to be species complexes as we look closer. They are a physical and conceptual manifestation of taxonomic uncertainty. We don’t necessarily know how a species complex is going to turn out until we have put the energy in to resolve it.
It helps to understand that there is a crucial difference between species names, and the species entities themselves. Each name ideally refers to a single species, but as soon as more than one species might be masquerading under the same name, we have a complex. Often, we have a collection of names that have been published in the scientific literature over the years, but it is not quite clear which name refers to which species, and there is a mismatch in the number of names and species-level units (genetic lineages). Such species complexes can be sprawling nightmares and require dozens of years of work from the bright brains of dedicated scientists to resolve (all the while being underfunded and often lacking job security). We cannot know the number of species in a complex before putting in the taxonomic work to resolve it.
Let me give you an example: the Calumma nasutum species complex, a group of beautiful but taxonomically complicated chameleons. In 2012, Gehring et al. showed that there were seven species described from this group (seven names), but it contained 27 deep genetic lineages. That could mean there would be as many as 27 species in the complex as a whole—a dramatic mismatch between names and species-level genetic lineages. Are there simply 20 species awaiting description?
Well, my colleagues and I started working on the taxonomy of this group in 2014, mostly led by Dr. David Prötzel, who did his PhD on them (you might know him as the guy who described bone-based UV-fluorescence in chameleons). David has so far published 5 new species from this group since 2014, 2 of which were discovered after Gehring et al.’s 2012 trees, and we now have a manuscript in review that describes a further 3 new species and revalidates an old name. After this, the group will be mostly resolved (one ‘sub-complex’ within it, the Calumma gallus complex, remains untouched—that group, however, is not likely to yield more than four species).
The 27 lineages of Gehring et al. have resulted not in 27 species, but instead in just 12 (at most 16)! As I mentioned before, we cannot know the number of species in a complex until we resolve it. Increasingly, algorithmic methods are being put forward to do this kind of work for us, but right now, they simply do not cut it; we have to do the leg work.
Species complexes and the IUCN Red List
So, what does the Red List have to say about such species complexes? Well, their current guidelines allow two alternatives: Option 1, leave them off the Red List altogether—after all, out of sight, out of mind, right?; or Option 2, assess them as if they were good species, i.e. all available names as though they referred to real biological units. The former is rarely chosen, as we don’t like to omit entries from lists. The latter is easier, for, as I’ve already explained, it is bound to result in an assessment under Criterion B.
What does this mean for our species complexes? Well, a species complex always has an area larger than any one lineage it contains; this is part of Set Theory or whatever. As a result, the species complex, assessed as a single species, gets a huge area, and, under Criterion B, is not threatened! Phew, what a relief, we get to list it as Least Concern!
The trouble is, any of the actual species within that complex almost certainly has a smaller distribution area, and could be threatened in its own right.
We argue in our paper that the very fact of belonging to a species complex should immediately qualify a species as DD. Before we can hope to manage the conservation of those species, we must figure out their taxonomy. By listing species complexes as DD, we can highlight the fact that taxonomy is a vital part of the conservation of species; that it is important, and needs to be funded and promoted. The current policy is sweeping taxonomic uncertainty under the rug, and we may be silently losing species as a result. We need to put more emphasis on the importance of a solid taxonomic foundation for effective conservation, and stop dismissing it as an afterthought! We also hope that this would make it clear to governments that taxonomy-focussed fieldwork is critical; all too often permitting agencies are refusing permits looking at species complexes, because they perceive the name to be clarified when it appears on the Red List.
Species complexes are Data Deficient! Except when they’re not…
There are a few exceptional groups where the taxonomy of a species complex has more or less been figured out, even though the species are not already described and named. The frogs that we showcased in our paper were such a case. In Madagascar, DNA barcoding of all reptiles and amphibians is nearly complete, so we know how many deep genetic lineages there are. These lineages are given Candidate Species numbers, which help us track data on these lineages before they are addressed taxonomically.
The mantellid frog genus Mantidactylus is divided into several subgenera, one of which is Hylobatrachus—torrent frogs found along Madagascar’s east coast and in one location in the west. Until our paper, there were two species: M. (H.) lugubris and M. (H.) cowanii. But several candidate species were also known. When the Red List assessment of M. (H.) lugubris was done, candidate species were included within its circumscription (part of assessing a species complex as a ‘good species’). This resulted in a huge range, and as a result, the species was listed as Least Concern. In our new paper, we described two of the candidate species as new species, and we also provided the map on the right. It shows hatched the area that M. (H.) lugubris occupies according to the Red List, and the area that it actually occupies, excluding other members of the complex.
This case differs from your average species complex, because this is a ‘characterised’ complex; we knew more or less how many lineages were in it, and in fact at the time of the last assessment (resulting in this huge distribution), it would easily have been possible to exclude most, if not all of the undescribed candidate species from the distribution of the species. But following the IUCN policy of treating this as one single species, we wound up with a huge distribution that very poorly captures the distribution of any of the lineages it contains. For this reason, we argue that species belonging to species complexes where data allow us to confine the assessment to a single species lineage should be assessed based on those lineages, and all other lineages omitted until taxonomically addressed and described. Again, this would put the emphasis on the requisite taxonomy, rather than broad and inaccurate circumscriptions.
But what about the plight of the Data Deficient species?
You may be surprised to read a post so strongly recommend that we assess more species as Data Deficient. After all, we have just managed to move a huge number of species from DD into some kind of ranking. When I wrote previously about ‘data deficiency and the plight of the unknown‘, I highlighted the fact that it is hard to protect what we do not understand. However, I also cautioned against the treatment of DD species as something more than they are; you cannot easily mine DD species for information about their distribution, because that data is going to be of low quality, by which virtue they are DD in the first place. If we take a look at the current red listing policies, what we see, however, is pressure to take species that should otherwise be DD, and place them in threat categories, no matter how inadequate the data behind that decision. It is my opinion that this practice simultaneously de-values the Red List, masks bad quality data, and discourages any work to improve that data.
At present, we have not yet called for changes to be made to the Red List’s policies, but this is something I am considering. If you have thoughts on this practice, please share them with me!