Investigating the effects of tree species diversity and relative density on bird species richness with acoustic indices

Mixed-species forest stands may provide suitable habitats for more bird species than tree monocultures by increasing the number of niches and resources available. However, assessing the effects of forest diversity on birds over time and across multiple sites using traditional bird survey methods can be expensive, time-consuming and impractical. Alternatively, we used passive acoustic monitoring (PAM) and acoustic indices to investigate how tree species diversity, relative density, and other plot characteristics affect the species richness of avian communities in temperate forests. Acoustic monitoring was conducted in Nationalpark Hainich, Germany, using plots which differed in tree species diversity and relative species density. Using a subset of our data, five acoustic indices, the normalised difference soundscape index (NDSI), acoustic complexity index (ACI), bioacoustic index (BI), the acoustic evenness index (AEI) and temporal entropy (Ht), were initially assessed using audio recordings of different soundscape elements to determine the degree to which they were influenced by non-biophony soundscape elements. Of these, NDSI, ACI and BI were considered least biased and displayed significant positive relationships with bird species richness assessed using audio recordings. However, NDSI values plateaued at around 5 bird species when measuring bird species richness and ACI values for dawn choruses were lower than expected, potentially due to the high density of birdsong saturating the soundscape. Plot tree species diversity had significant positive effects on ACI and BI during dawn, dusk and daytime periods. Increased basal area of sycamore (Acer pseudoplatanus) and, to a lesser degree, beech (Fagus sylvatica), had a negative effect on acoustic indices. These results could potentially be due to the earlier leaf out timings of these tree species in relation to the dates when surveys were performed. While the use of PAM and acoustic indices enabled surveys across multiple sites over longer time periods than would previously have been achievable, our results show that care should be taken when selecting indices and interpreting acoustic index results.