Forest age is a primary trait filter for saproxylic beetles in the southeastern United States

Many forests throughout the world consist of regenerating mature stands. Although these forests differ in many respects from old-growth (with a history of minimal human disturbance), they typically develop similar structural attributes over time. As a result, older mature forests may be of particular conservation value if they contain resources and microhabitats benefitting saproxylic (deadwood dependent) species. Species' response to forest age may be driven by traits that relate to ecological functions or habitat preferences, such that species with less compatible traits for a local forest environment are "filtered" out. Thus, forest age may influence species' distributions and the trait composition of assembled communities. The Piedmont region of the southeastern United States has experienced widespread forest regrowth over the past century due to agricultural abandonment. Today's landscapes are largely characterized by mature forests that are becoming increasingly fragmented by suburbanization. Here, we assessed the filtering effects of forest age, landscape forest cover (LFC), and deadwood volume on saproxylic beetles in northeastern Georgia. Using historic aerial imagery to distinguish forest age (young = regrown after 1938; old = mature in 1938), we sampled beetles in mature forests of both age classes occurring along an LFC gradient. We measured five traits with hypothesized functional roles (body length, body width, body roundness, antenna length, eye length) for the 472 species captured. Using a joint species distribution model (JSDM), we tested trait-niche relationships (i.e., how traits influence species' responses) and estimated community trait composition (mean and dispersion of trait values) along gradients of environmental filters. We found that forest age is a filter for several traits (six supported relationships with > 95 % posterior probability), but LFC and deadwood volume were less strongly related to fewer traits. Most notably, large species (typically having lower population sizes and requiring stable larval habitat) were filtered from young mature forests and low LFC. Thus, old mature forests with high LFC showed higher mean and dispersion of beetle body length. Sensory traits also showed responses, likely reflecting adult life under bark (eye length) or ability to detect resources or mates (antenna length). Body width and roundness showed inconsistent responses with regard to indicated functional roles. Our results show that forest age is a strong filter on saproxylic beetle communities in the southeastern United States. Old mature forests, despite their scarcity in the region, are important for species requiring habitat stability and for maintaining communities with diverse trait composition.