The power to detect regional declines in common bird populations using continental monitoring data

Abstract

Anthropogenic environmental change is driving the rapid loss of biodiversity. Large declines in the abundance of historically common species are now emerging as a major concern. Identifying declining populations through long‐term biodiversity monitoring is vital for implementing timely conservation measures. It is, therefore, critical to evaluate the likelihood that persistent long‐term population trends of a given size could be detected using existing monitoring data and methods. Here, we test the power to detect declines in Australia's common landbirds using long‐term citizen science monitoring. We use spatially explicit simulations of occupancy dynamics and virtual sampling, designed to mimic bird monitoring in better‐sampled regions of Australia, to assess likely power in these data to detect trends relevant for conservation. We predict the statistical power for 326 common species that meet minimum requirements for monitoring data across 10 regions of Australia, estimating the number of species for which we would have a high (≥80%) chance of detecting declines of different sizes. The power to detect declines of ≥30% per decade was predicted to be high for at least one‐third of the common species in 7 of 10 regions, with a total of 103 (32% of 326) unique species sufficiently monitored in at least one region. These species spanned 12 taxonomic orders, four orders of magnitude in body mass, and a broad diversity of dietary guilds, suggesting the current species pool will likely serve as robust indicators for a broad range of environmental states and pressures. Power was strongly affected by species' detectability, and power to detect even large declines was negligible when species are detected on ≤50% of visits to an occupied site. Predicted power for many species fell just short of the 80% threshold in one or more regions, which suggests an increase in effort targeting these species could greatly enhance the species and regional representation of these data. Against the backdrop of unprecedented biodiversity losses, this study shows how critical evaluation of existing monitoring schemes is valuable both for assessing the contribution of citizen science schemes to biodiversity monitoring and for designing strategic monitoring to significantly improve the knowledge these schemes provide.

Quantifying the statistical power of monitoring programs for marine protected areas

“…We found that sampling effort strongly affected the ability to detect change in a timely manner, with the lowest level of sampling unable to detect change within the time tested, whereas increasing effort could potentially detect change within 13 yr for our example species. This agrees with other research showing that insufficient within-site sampling can have a large impact on the power to detect trends (e.g., Perkins et al 2016, Barker et al 2019. However, we found that increases in sampling effort in the study area with the lowest starting densities (Montara, and the reference site in particular) achieved only marginal increases in power over time, despite sites at Montara being smaller and thus the percentage area sampled with an equivalent number of transects being almost four times that of the largest sites at Bodega Head.…”

Modeling avian full annual cycle distribution and population trends with citizen science data

“…The spatially and seasonally explicit trend estimates also provide new information about interannual changes in Wood Thrush population size. The spatial resolution of the trend estimates presented here is relatively high compared to other studies with similarly broad spatial extents (Sauer et al 2017, Baker et al 2019, Meehan et al 2019, Rushing et al 2019. Both breeding and non-breeding trend maps show significant spatial variation in the pattern of declines, with the steepest declines within the 5% FDR regions (Figs.…”

“…, Baker et al. , and Meehan et al. ) across broad spatial extents have turned to the use of less structured survey data collected by citizen scientists.…”

“…Recent efforts to estimate species' distributions (K ery et al 2010), abundance (Johnston et al 2015), and trends (Horns et al 2018, Baker et al 2019, and Meehan et al 2019) across broad spatial extents have turned to the use of less structured survey data collected by citizen scientists. Citizen science projects have been very successful at collecting species observation data with high spatial and temporal resolution across broad extents throughout the year (Dickinson et al 2010).…”

Performance of imputation-based models in predicting breeding population trend of a near-threatened bird in changing water regime: A 36-year long-term case study of Painted Stork, Mycteria leucocephala

“…These have been used on terrestrial as well as water birds, including waterfowls and waders. Most of the long-term studies have been conducted on North American and European birds, as a systematic census of birds is available from these regions (Baker et al, 2019;Freeman et al, 2007, Kamp et al, 2021Sauer et al, 2017;Shipley et al, 2020;Weimerskirch 2018). Painted Stork breeding population is well monitored at KNP and hence can be used to model the long-term temporal trend and effect of water availability.…”