Passive recording technologies have become a valuable tool for ecologists to accumulate large data sets on vocalizing species assemblages and to study the response of biodiversity to regional and global environmental change. However, the process of interpreting these recordings and identifying species can be time-consuming. To automate this process, and to derive diversity metrics directly from the audio files, soundscape ecologists developed and published many bioacoustics indices, which are calculated using computer algorithms that mathematically summarize the acoustic energy in a recording. Of these indices, the acoustic complexity index (ACI) is perhaps the most widely proposed surrogate of bird species richness. Although several studies have shown the ACI to have positive correlations with species richness, it remains unknown how well the ACI will perform over large, habitat-diverse, regions. Our aim, here, was to examine whether site-level bird species richness (alpha diversity) correlated with the ACI across a large heterogenous mountain region. Further, we explored whether the elevational trends for alpha diversity and the ACI exhibited similar unimodal patterns with corresponding maximums and whether any shifts in the elevation of these maximums could be tracked over time in response to climate change. We deployed recording devices along two extensive transects on the Pacific Crest Trail in California, USA; these included a 689-km trail section in northern California and a 531-km section in southern California. Point counts for birds from the field were combined with the interpretation of the recordings to estimate alpha diversity using a Bayesian multispecies occupancy model. We found that alpha diversity exhibited a positive linear relationship with the ACI on both trail sections. However, these relationships were not strong (low R2 values). Similar mid-elevation maximums for alpha diversity and the ACI were observed along the northern trail section. Power analysis revealed that with repeated annual surveys using our protocols, we could detect an average shift in the elevation of either maximum species richness or ACI of approximately 100 m over 20 years (i.e., 5 m/yr), which could be improved in precision to 50 m (i.e., 2.5 m/yr) with a 5-fold increase in survey effort. Although there was considerable unexplained variation in our diversity- and ACI-elevation trends, which we suspect was due to microhabitat and microclimate effects, we concluded that the ACI was a useful, albeit coarse, surrogate of alpha diversity across large regions. The ACI warrants consideration by ecologists and land managers for application in large-scale monitoring programs.
Read full abstract