Modelling bird atlas reporting rate as a function of density in the southern Karoo, South Africa

Abstract

The relationship between true population density and abundance indices derived from atlas data can be used to estimate population size, a metric that is vital for assessing the conservation status of animal species. However, this relationship remains understudied. Meanwhile, there is growing concern for the avifauna of South Africa’s arid southern Karoo region, which is threatened with an increasing frequency of heat waves as well as the expansion of infrastructure in association with uranium mining, renewable energy production and shale gas extraction. We calculated density (individuals km⁻²) for 49 bird species using data from a standardised distance-sampling survey conducted across 64 pentads (sampling areas of approximately 81 km²) in the southern Karoo region. Reporting rates from the Southern African Bird Atlas Project (SABAP) were explained by individual density for 32 of the 49 species using logistic regression models, and for 37 species using group (cluster) densities. Within this set of species, there was a positive correlation between mean reporting rate and log-transformed mean density in a linear model that included log-transformed mass as a covariate. For seven species with a high uncertainty associated with their density estimates (high relative standard error), we found that we could use our model to produce intuitively reasonable density estimates based on their SABAP2 reporting rates. This relationship suggests that where density estimate–reporting rate relationships for communities are known, they are useful for cross-checking density estimates. We illustrate the utility of this relationship by calculating population estimates for a set of species, and discuss the importance of SABAP2 for conservation purposes in the Karoo.

Keywords: Karoo, arid zone, distance sampling, citizen science, conservation