Waterbird count data

Counts of a waterbird species from a multi-year, multi-site monitoring program. The data are incomplete: many site-year combinations were not surveyed and are recorded as missing values. Counts also contain a large proportion of zeros, which makes the data a suitable example for the zero-inflated Poisson log-normal model implemented in the package. Sites and years are anonymized.

Usage

data(waterbird)

Format

A named list with two elements:

Y

Numeric count matrix with 419 rows (sites, site_001 to site_419) and 17 columns (years, year_01 to year_17). Unobserved site-year combinations are NA.

X

Numeric design matrix with \(419 \times 17 = 7123\) rows (site-year combinations, stacked year by year, i.e. by blocks of 419 sites) and 7 columns: intercept (column of ones), and the standardized covariates dist_coast (distance to the coast), dist_city (distance to the nearest city), elevation, year (linear time trend), spring_temp_anomaly (spring temperature anomaly) and agricultural_land (percentage of agricultural land).

Source

Anonymized data from a waterbird monitoring program.

Examples

data(waterbird, package = "colvR")
str(waterbird, max.level = 1)
#> List of 2
#>  $ Y: num [1:419, 1:17] NA 0 NA NA NA 0 6 1 6 NA ...
#>   ..- attr(*, "dimnames")=List of 2
#>  $ X: num [1:7123, 1:7] 1 1 1 1 1 1 1 1 1 1 ...
#>   ..- attr(*, "dimnames")=List of 2
mean(is.na(waterbird$Y))                # proportion of missing entries
#> [1] 0.4443353
mean(waterbird$Y == 0, na.rm = TRUE)    # proportion of zeros
#> [1] 0.6705407