Start with what those percentages are shares of. The 30% in 30x30 is a share of national territory [1]. The 16% and the 39% are shares of the modeled habitat of 28 terrestrial carnivores [2][3][6]. Those are different denominators, so nobody should read 16% as "Mexico is 14 points short of its target" [22]. It is a statement about how well the existing network sits on top of the places carnivores actually live.
Inside that frame, the arithmetic is what makes the paper a policy document rather than a map. Going from 16% to 39% is 23 percentage points, about 2.4 times the coverage, and not one hectare of it is newly acquired [17][18]. The land doing the work is already managed: wildlife management units and Indigenous territories, categories that do not officially count toward Mexico's biodiversity goals [6][7]. The team looked at four categories of managed land in all, including national parks and territories demarcated more by culture than by any particular land use [8]. Harris, one of the authors, puts the choice plainly: with 2030 close, the question is what counts and what is actually working [9].
Carnivores are the instrument here, chosen for ecological, social and economic weight, with the jaguar as the headline species [10]. The map flags several carnivore hotspots with no protection at all [11], and those are the same places other researchers reached using different biodiversity criteria, which the authors read as support for carnivores as umbrella species, since a jaguar's range covers a great many other things [12]. That convergence is genuinely reassuring, and worth stating precisely: it is agreement between maps built on different criteria, not evidence that either map predicts whether populations persist.
Which is the thing the 39% does not tell you. Spatial overlap is geometry, and the study says so itself, calling for on-the-ground monitoring to judge whether designation reduces habitat loss and sustains vulnerable populations [13]. Lead author Germar González is more specific about the bill: outcomes depend on enforced capacity, sustained funding, and whether the communities managing these lands have institutional support to do it well [14]. Harris makes a parallel point about governance, noting that community-managed areas run on different structures, audiences and scales, with objectives that may not match a national park's, and still contribute [16].
One gap I would want closed before anyone quotes 39% in a negotiation. Coverage is lower for threatened and endangered species specifically, and the account we have reports no number for that subset [4]. That is exactly where a portfolio argument is weakest, because the species with the least habitat left are the ones for whom a generous accounting rule does the least. The phys.org write-up also breaks off mid-sentence while reporting that overlap with critical habitat was greatest on Indigenous land [19], so I am not leaning on it here.
The paper, in Conservation Science and Practice, describes Mexico's existing network as strong enough to model for other countries in Central and South America [5][15]. If the model includes counting lands the state does not directly administer, then it is worth separating the two costs the paper actually names: recognition is cheap, support is expensive, and coverage maps only show the first one.
