Staying alive in a shrinking cloud forest

The story of the bird, the frog and the mouse

Thanks to climate change and land conversion, Mexico’s famous cloud forests are shrinking. By 2080 it’s expected they may be a splinter of their former glory. The forests provide habitat for a disproportionate amount of the country’s biodiversity (see the box: ‘Head in the clouds’) but the impacts of forest loss will be felt differently by different species. Understanding these differences will be important when forming conservation plans. How, for example, will a local bird cope as opposed to a frog or a mouse?

A frog of the Mexican cloud forest. Climate change and land conversion may mean its days are numbered. (Photo by Omar Ordoñez)

A frog of the Mexican cloud forest. Climate change and land conversion may mean its days are numbered. (Photo by Omar Ordoñez)

One of the main objectives of conservation planning is ensuring the persistence of species. One way to measure persistence is by estimating a species extinction risk – that is, the probability of a species becoming extinct during a time frame. Some ecosystems, like cloud forests, are more vulnerable to climate and land-use change than other ecosystems but this has different implications for different animals and plants. I recently led an analysis that calculated the extinction risk of three very different types of animal found in the cloud forests of Mexico (Ponce-Reyes et al. 2013).

Different animals

Researchers have developed a range of models to calculate the extinction risk of species in different situations. I used a metapopulations model that relates the extinction risk to the ecology of species. It incorporates factors such as the area they require to live (home range), the distance they are able to travel between fragments (dispersal distance) and their fecundity rate (number of offspring that make it to adulthood and reproduce). As a general rule, populations in fragmented landscapes are generally more susceptible to human-related threats than those in continuous landscapes.

The highly fragmented nature of the cloud forest lends itself to applying a metapopulation approach to calculate the extinction risk of its species. In this analysis we chose three very different animals to explore what a shrinking forest means for different life strategies. We looked at a bird, a frog and a mouse. The bird we selected was the resplendent quetzal (Pharomachrus mocinno), a famous species in this region (and Guatemala’s national bird). Unfortunately, we couldn’t find reliable population-ecology data on amphibians or mammals living in the cloud forests so we parameterized the model with similar vertebrate species found in other forest types. We used a frog based on the cliff chirping frog, Eleutherodactylus marnockii, and a mouse based on the deer mouse Peromyscus maniculatus.

A bird like the quetzal can fly long distances, however their home range is very large and the number of ‘chucks’ that make it to adulthood is very low. In contrast, the frog has a very small home range, but their dispersal ability is very low. On top of this, the number of froglets that survive and leave descendants is also low. Finally, the mouse’s home range is small but can travel long distances and their fecundity rate is the highest of all three species.

Different regions

Cloud forests have different vulnerabilities to climate and land-use changes depending on where they are. We analysed two scenarios: 1. climate change as the only driver of habitat loss and fragmentation; and 2. climate change and land-use change outside protected areas as the drivers of habitat loss. In this second scenario we assumed that only the cloud forest-suitable areas that are currently in a protected area would remain, those outside of these areas will be converted to other uses.

We analysed both scenarios in three regions of Mexico with different landscape configurations: (a) a stable region of cloud forest (Oaxaca) with the largest total area, the least fragmented and few protected areas (around 4% of cloud forests at present); (b) the cloud forest region with the smallest total area of the three regions, although not very fragmented (Chiapas South) but heavily protected (about 72% of cloud forests in this region are currently in a reserve); and (c) a very fragmented region with about 7% of their cloud forests within a protected area (Chiapas North).

Our models predicted that the more drastic reduction of cloud forest suitable areas in Oaxaca would occur by 2080, with about 40% of the current total cloud forest area becoming climatically unsuitable for cloud forest when assuming only climate change as the only threat. However, if land-use change was assumed as well, around 2% of the currently suitable areas for cloud forest might remain in a single patch. This is because this region has the smallest proportion of their cloud forest under protection.

Chiapas South is currently not very fragmented, but we estimated that by 2080 the suitable areas for cloud forest will contract dramatically to around 3% of the forest’s current extent (and this would be as two patches that are relatively close together). When we assumed that only the cloud forest within a protected area might remain our models predicted that the total area of protected forests in this region might decrease severely (by 98%) into only one patch by 2080.

For Chiapas North, we expect that only 16% of current cloud forest might remain in climatically suitable areas by 2080. Although the number of fragments in this area is only predicted to decrease slightly, the average size of these fragments will shrink dramatically (the mean patch size in 2080 will only be 16% of what they were in 2010). From the cloud forest currently within a reserve in this region, we expect that less than 0.5% might remain by 2080.

Different strategies, different outcomes

When only climate change is considered as a threat, the species with the highest extinction risk in the Mexican cloud forest are those with the short dispersal distances. In our case, this means the frog (fig 1b). This is because fragments will become more isolated in the mountains and species with low dispersal abilities will encounter problems in colonising new fragments in case their original fragment becomes environmentally unsuitable. In fig 1b the probability of extinction of the frog seems to decrease for Chiapas South in 2080. This is because the metapopulations model does not predict the extinction risk of less than five patches accurately. For those cases, we estimated the ‘joint probability of all populations becoming extinct’ which only takes into account the home range of the species and the extent of the remaining habitat. The landscape configuration prediction for Chiapas South, in 2080 is of two relatively big fragments. Because the frog has a very small home range, the probability of all frog populations disappearing in this region is small (as they still have enough habitat remaining). However, if land-use change is added to the pool of threats, then species with big home range requirements are the most threatened (here, the quetzal, fig 1d).

The different vulnerabilities to threats of the cloud forests patches in each region affects the extinction risk of the species. When climate change was the only threat, the region where the species had the highest probability of becoming extinct was in Chiapas South (the currently most fragmented of the three, fig 1a-c) and again, the frog was the most vulnerable species.

But when we analysed the patterns within protected areas (climate and land-use change together as threats), Chiapas North always had the highest probability of extinction for all three species (fig 1d-f ); with the quetzal having the highest extinction risk in this area. Although we predict that only one patch of cloud forest will remain protected for each of the three regions, Chiapas North is predicted to lose more of its suitable habitat (compared to the other two regions).

Modelling the future

Models for extinction risk provide a quantitative and process-based link between patterns of ecosystem change and species persistence, and form a sound basis for conservation decision-making. Unfortunately, no matter which variant of our model you look at, the Mexican cloud forest’s species do not have a promising future. The loss and fragmentation of their habitat in the next 70 years under climate change is projected to be considerable, resulting in an increased extinction risk for most cloud forest species.

Figure 1. Annual probabilities of extinction of the metapopulations of the quetzal, frog and mouse in three regions of the Mexican cloud forests (a-c) and in the protected areas (d-f). (From Ponce- Reyes et al. 2013)

Figure 1. Annual probabilities of extinction of the metapopulations of the quetzal, frog and mouse in three regions of the Mexican cloud forests (a-c) and in the protected areas (d-f). (From Ponce- Reyes et al. 2013)

Our modeling approach, however, does reveal the relative risks of extinction from the combined impacts of climate change and land clearance. Approaches such as this have the potential to allow us to identify which regions and patches are more vulnerable to the threats, and potentially rank management strategies.

More info: Rocío Ponce- Reyes


Ponce-Reyes R, E Nicholson, P Baxter, R Fuller &HP Possingham (2013). Extinction risk in cloud forest fragments under climate change and habitat loss. Diversity and Distributions. Special Issue: Risks, Decisions and Biological Conservation 19: 518–529.

Leave a Reply