Data files

This project was designed to estimate site occupancy of forest birds in Puerto Rico during a winter before (2015) and after (2018) passage of two significant tropical storms: Hurricane Irma and Hurricane Maria. This project includes two data files, one containing data on species occurrence gleaned from repeated point-counts and one containing data on characteristics of each point-count survey. A data dictionary is also included.<br>The species-occurrence file (occurrenceData.csv) includes occurrence data for all species of forest birds encountered during surveys in 2015 and 2018 across the island of Puerto Rico. Data collected in 2018 represent a snapshot of conditions shortly after the passage of hurricanes Irma and Maria in September 2017. Data collected in 2015 represent a snapshot of baseline, pre-hurricane conditions. Surveys consisted of standardized 10-minute counts at each point, beginning shortly after sunrise. Each point count was divided into four 2.5-min intervals conducted in immediate succession, with 1-min playbacks of Bicknell’s Thrush vocalizations broadcast before the second and fourth periods. Individuals of all species detected in each count period were recorded in four distance bands (1-10 m, 10-25 m, 25-50 m, &gt;50 m). No counting occurred during the two 1-min playback periods. <br>The point-characteristics file (surveyPointsAllYears.csv) describes the location of each point and conditions at the time of each survey. Points were established using a randomized, spatially balanced network of survey locations using a generalized random tessellation stratified (GRTS) scheme. With this approach, we selected 60 1-km2 cells from across the island as potential areas in which to conduct surveys. Once we had drawn a sample of cells to survey, we visited each cell and identified 3-5 locations suitable for point-count surveys. Suitability was based on the extent of forest cover – at least 50% of the area in a 50-m radius around each point was forested – and accessibility; all points were along public roads or trails. To maintain independence of counts conducted at different points, we placed each point at least 250 m from its nearest neighbor. Point locations were spatially referenced via GPS and marked with metal tree tags. Due to time limitations – the original survey protocol called for surveys to end in March, prior to any potential pre-migration movement by Bicknell’s Thrush – we only sampled points in 43 of the 60 cells. In total, we sampled 186 points in the 43 cells in both 2015 and 2018. Twenty-five points were surveyed in 2015 but could not be accessed due to storm damage in 2018. To make up for some of these points, we surveyed 16 new points in 2018 within the cells that contained previously surveyed points that had become inaccessible. This yielded 211 points with data from 2015 and 202 points with data from 2018, for a total of 227 points with data from at least 1 year. <br>Due to the original goal of the surveys, our sample was weighted towards areas more likely to contain habitat for Bicknell’s Thrush based on the winter-habitat model for that species and thus oversamples wet, high-elevation broadleaf forest. Although our survey locations were not a representative sample of all forest types on the island, the sample did include cells at lower elevations in drier forest types. Points surveyed in 2015 ranged in elevation from 0 - 1,297 m, with a median elevation of 705 m (interquartile range = 408 - 825 m). In 2018, surveyed points ranged in elevation from 5 - 1,297 m, with a median elevation of 760 m (interquartile range = 393 - 843 m). Thus, we believe our sampling frame can be described as predominantly forested areas on Puerto Rico that were accessible by roads or trails. The two important sources of potential bias in considering the scope of inference allowed by this sampling frame are 1) the oversampling of high-elevation forests, which will tend to yield low-precision estimates for bird assemblages characteristic of dry, lowland forest and 2) the reliance on trails and roads to access survey points, an unavoidable trade-off given the difficulty of moving through these tropical forests, especially following the hurricanes.