Spatial models to account for variation in observer effort in bird atlases

Abstract

To assess the importance of variation in observer effort between and within bird atlas projects and demonstrate the use of relatively simple conditional autoregressive (CAR) models for analyzing grid‐based atlas data with varying effort. Pennsylvania and West Virginia, United States of America. We used varying proportions of randomly selected training data to assess whether variations in observer effort can be accounted for using CAR models and whether such models would still be useful for atlases with incomplete data. We then evaluated whether the application of these models influenced our assessment of distribution change between two atlas projects separated by twenty years (Pennsylvania), and tested our modeling methodology on a state bird atlas with incomplete coverage (West Virginia). Conditional Autoregressive models which included observer effort and landscape covariates were able to make robust predictions of species distributions in cases of sparse data coverage. Further, we found that CAR models without landscape covariates performed favorably. These models also account for variation in observer effort between atlas projects and can have a profound effect on the overall assessment of distribution change. Accounting for variation in observer effort in atlas projects is critically important. CAR models provide a useful modeling framework for accounting for variation in observer effort in bird atlas data because they are relatively simple to apply, and quick to run.

North American Breeding Bird Survey underestimates regional bird richness compared to Breeding Bird Atlases

“…BBA data are also used to analyze the potential consequences of land‐use modification (Shoffner et al, 2018; van der Hoek et al, 2015) and climate change (Jarzyna et al, 2015, 2016; Zuckerberg, Woods, & Porter, 2009). While our results support earlier studies showing limitations of BBA data, especially regarding variable survey effort (Beck et al, 2018; Kujala et al, 2013; Robertson et al, 2010; Wilson et al, 2017), modeling approaches and corrections to such limitations are possible (Isaac et al, 2014; Robertson et al, 2010; Sadoti et al, 2013; Wilson et al, 2017). For example, while we find that local richness is higher in BBS data (Figure 4), we also show that a simple correction for the effectiveness of survey effort in the BBA (Appendix S2: Figure S5) leads to an opposite pattern with a higher richness in the BBA (Appendix S2: Figures S6 and S7).…”

The traits that predict the magnitude and spatial scale of forest bird responses to urbanization intensity

“…For each forest bird species, we selected the landscapes within the species’ range (Fig 1). To determine species ranges, we used PBBA block occurrence probabilities calculated using simple conditional autoregressive models accounting for the survey hours completed by each observer and land cover class proportions [54]. Block occurrence probabilities were assigned to the center of each block and then spatially interpolated for each species using kriging in ArcGIS to account for variation within survey blocks [55].…”

Range contractions of the Broad-winged Hawk in the Northeast United States

“…The protocols ideally ensure consistency in coverage among blocks and requires that a significant amount of time is invested in each sampled block (Porter and Jarzyna 2013). Although atlas methodologies are designed to ensure consistent coverage, in reality, coverage is inconsistent, both spatially (between blocks) and temporally (between atlas projects) (Wilson et al 2017).While measures of effort can be used to account for such variation, it is important to note that effort information, such as number of visits, search effort (e.g. total hours), and years of visits are not available for all ten atlas projects used in our analysis.…”