Intact but empty forests? Patterns of hunting-induced mammal defaunation in the tropics

Abstract

Tropical forests are increasingly degraded by industrial logging, urbanization, agriculture, and infrastructure, with only 20% of the remaining area considered intact. However, this figure does not include other, more cryptic but pervasive forms of degradation, such as overhunting. Here, we quantified and mapped the spatial patterns of mammal defaunation in the tropics using a database of 3,281 mammal abundance declines from local hunting studies. We simultaneously accounted for population abundance declines and the probability of local extirpation of a population as a function of several predictors related to human accessibility to remote areas and species' vulnerability to hunting. We estimated an average abundance decline of 13% across all tropical mammal species, with medium-sized species being reduced by >27% and large mammals by >40%. Mammal populations are predicted to be partially defaunated (i.e., declines of 10%-100%) in ca. 50% of the pantropical forest area (14 million km2), with large declines (>70%) in West Africa. According to our projections, 52% of the intact forests (IFs) and 62% of the wilderness areas (WAs) are partially devoid of large mammals, and hunting may affect mammal populations in 20% of protected areas (PAs) in the tropics, particularly in West and Central Africa and Southeast Asia. The pervasive effects of overhunting on tropical mammal populations may have profound ramifications for ecosystem functioning and the livelihoods of wild-meat-dependent communities, and underscore that forest coverage alone is not necessarily indicative of ecosystem intactness. We call for a systematic consideration of hunting effects in (large-scale) biodiversity assessments for more representative estimates of human-induced biodiversity loss.

Fig 1. Geographic variation in hunting-induced defaunation for (A) all species, (B) small-sized species (<1 kg, e.g., Sciurus spp.), (C) medium-sized species (1–20 kg, e.g., Allouatta spp.), and (D) large-sized species (>20 kg, e.g., Tapirus spp.).

The insets represent the total area (y-axis) under different levels of defaunation (x-axis, from DI = 0 [blue] to DI = 1 [red]). Note that the y-axes in the four insets have different scales. Available at https://figshare.com/projects/Intact_but_emtpy_forests_Patterns_of_hunting-induced_mammal_defaunation_in_the_tropics/31118. DI, defaunation index.

Fig 2. Percentage of mammal species per grid cell with DI >0.7 or abundance reductions of at least 70%.

Colors range from green (low relative number of species with DI >0.7) to red (high relative number of species with DI >0.7). Orange to red areas showcase hotspots of hunting-induced mammal defaunation (with at least one third of the species with DI >0.7). Available at https://figshare.com/projects/Intact_but_emtpy_forests_Patterns_of_hunting-induced_mammal_defaunation_in_the_tropics/31118. DI, defaunation index.

Fig 3. DI for different trophic groups: (A) carnivores, (B) herbivores, (C) frugivores, and (D) insectivores.

The dashed gray line indicates the mean DI across the pantropical forest zone. The y-axes have different scales. Available at https://figshare.com/projects/Intact_but_emtpy_forests_Patterns_of_hunting-induced_mammal_defaunation_in_the_tropics/31118. DI, defaunation index; sq., square.

Fig 4. Geographic variation and spatial extent of hunting-induced defaunation for all species (left-hand side; A,C), and for large-bodied species (right-hand side; B,D) in WAs (A,B) and IFs (C,D).

The insets represent the total area of forest (y-axis) predicted to be defaunated (DI > 0.1, red) and intact (DI ≤ 0.1, green) in the respective graphs. Available at https://figshare.com/projects/Intact_but_emtpy_forests_Patterns_of_hunting-induced_mammal_defaunation_in_the_tropics/31118. DI, defaunation index; IF, intact forest; WA, wilderness area.