Sensitivity of secretive marsh birds to vegetation condition in natural and restored wetlands in Wisconsin

Abstract

Wetland loss, biological invasions, and ecological restoration are major factors altering wetland resources in the Midwestern United States. Large‐scale plant‐community change associated with these factors is an under‐investigated, potentially strong driver of habitat suitability for wetland‐dependent wildlife, such as secretive marsh birds (SMBs), which are of widespread conservation concern. We employed multi‐year, hierarchical Bayesian occupancy modeling to investigate sensitivity of 3 SMB species (American bittern, sora, and Virginia rail) to habitat and vegetation characteristics in Wisconsin, USA. We contrasted habitat characteristics and SMB occupancy in natural wetlands with those restored under the federally funded Wetlands Reserve Program (WRP). We also examined the extent to which SMB occupancy was explained by 3 levels of habitat assessment that encompassed different spatial scales and levels of sampling effort (landscape, rapid, and intensive). All species were significantly associated with variables derived from intensive assessment, and showed high sensitivity to differences in plant‐community composition and vegetation quality. Both American bittern and Virginia rail were negatively associated with abundance of the invasive wetland grass, Phalaris arundinacea (reed canarygrass), and were positively associated with an indicator of plant‐community quality (mean coefficient of conservatism, or C‐value). Sora occupancy was positively associated with Typha (cattail) abundance. For all 3 species, occupancy was greater in natural sites than in restored sites, which were characterized by greater Phalaris abundance and lower mean C‐values. Our results show broad agreement between botanical and avian indicators of wetland quality, suggesting that enhancing the condition of wetland vegetation can yield ancillary benefits for SMBs. In this region, efforts to control Phalaris and restore diverse, native‐dominated plant communities are likely to increase wetlands’ capacity to support SMBs. © 2015 The Wildlife Society.

Habitat models to predict wetland bird occupancy influenced by scale, anthropogenic disturbance, and imperfect detection

“…, Glisson et al. ). This may indicate that field‐collected data more accurately reflect wetland conditions at a given location, but not necessarily that the 100‐m scale is optimal for assessing species–habitat relationships, because a comparison with similarly collected data at larger scales is not often included in these studies.…”

“…For example, our results indicate that a 100-m buffer is not the most appropriate scale for modeling relationships between remotely sensed wetland data and wetland bird occupancy. However, a ≤ 100-m buffer is often used to assess specieshabitat relationships for wetland birds (Winstead and King 2006, Conway and Sulzman 2007, Darrah and Krementz 2009, Harms and Dinsmore 2013, and field-collected habitat variables at this spatial scale (e.g., vegetation structure and composition) are good predictors of wetland bird occupancy (Rush et al 2009, Valente et al 2011, Glisson et al 2015a). This may indicate that fieldcollected data more accurately reflect wetland conditions at a given location, but not necessarily that the 100-m scale is optimal for assessing species-habitat relationships, because a comparison with similarly collected data at larger scales is not often included in these studies.…”

Nocturnal distance sampling all‐terrain vehicle surveys for nonbreeding rails

“…The National Marshbird Monitoring Protocol was designed to optimize detection probability during the breeding season because detection is so low for many rails and other marshbirds (Conway ). Many factors can affect wetland bird detection, including ambient temperature, wind speed, cloud cover, moon phase, and observer, and often these factors go untested in new survey methods (Anderson , Bolenbaugh et al , Budd and Krementz , Conway and Gibbs , Glisson et al ). When working with rails that are difficult to detect, understanding how individuals react to the survey methodology and estimating detection probability is important.…”

Predictive multi‐scale occupancy models at range‐wide extents: Effects of habitat and human disturbance on distributions of wetland birds

“…As an example, studies have often been conducted at wetlands containing the necessary emergent vegetation, but also within watersheds that were already heavily modified by agriculture, human development, modification of hydrologic regimes or all of these factors (e.g. Alexander & Hepp, ; Baschuk, Koper, Wrubleski, & Goldsborough, ; Bolenbaugh, Cooper, Brady, Willard, & Krementz, ; Darrah & Krementz, ; Darrah & Krementz, ; Darrah & Krementz, ; Glisson, Brady, Paulios, Jacobi, & Larkin, ; Krementz, Willard, Carroll, & Dugger, ; Valente, King, & Wilson, ; Winstead & King, ). Wetland degradation over broad spatial scales has been hypothesized as a driver of population declines for marsh birds (Eddleman, Knopf, Meanley, Reid, & Zembal, ), but little work has quantified these effects.…”

“…Studies have attempted to tease apart effects of habitat over a range of spatial scales, but they typically measured different attributes at each scale (e.g. Glisson, Brady, et al, ; Harms & Dinsmore, ; Naugle et al, ; Pickens & King, ; Roach & Barrett, ; Valente et al, ), thus confounding scalar effects with effects of the habitat features themselves (McGarigal et al, ; Wheatley & Johnson, ). As a consequence, prioritizing wetlands for marsh bird conservation over broad extents via extrapolation of site‐specific and fine‐scale results assumes that results are transferable to other regions and scale up linearly to broader spatial extents; these assumptions may not be valid and, hence, are unlikely to result in reliable prediction (Miller, Turner, Smithwick, Dent, & Stanley, ; Roach, Hunter, Nibblelink, & Barrett, ; Schneider, ).…”