Multispecies occupancy models unravel reduced colonization probabilities in plants from the unique Amazonian cangas

Abstract

Species extinction rates have surpassed background rates in the contemporary times, triggering a sixth global mass extinction event. Failure to prevent massive extinctions may be related to the lack of efficient tools to estimate local and regional population decline and to identify where declining species likely occur. We used a set of 22 plant species distributed in a globally unique ecosystem, the eastern Amazonian mosaics of forests, iron-rich open “cangas” and iron mine lands, as a model to test whether dynamic multispecies occupancy models can be used to assess population decline. Based on the metapopulation equilibrium between colonization and extinction probabilities compared across 90 mining plots and 60 control plots, we estimate that 45.4 % of the sampled species show no evidence of population decline. For 18.2 % of the species, we found negative equilibrium for both mining and control plots. For 36.4 % of the species, we found negative equilibrium only in 2.22–23.3 % of the mining plots, although equilibrium values were similar to those estimated for most control plots. Negative equilibrium was more associated with the reduced colonization probabilities than with elevated extinction probabilities. Our results indicate that for some wide-ranging species, negative metapopulation equilibrium was restricted to mining plots, indicating sensitivity of these species to environmental changes caused by mining. For some endemic species, apparent population declines are more likely to be occurring due to natural causes than to the negative effects of mining.