Population metrics for fynbos birds, South Africa: densities, and detection and capture rates from a Mediterranean-type ecosystem

Estimates of bird numbers through quantification of density and range sizes are necessary for decisions regarding conservation status, yet counts of birds are often confounded by uncertainty of detection. The status of the endemic birds of the Fynbos biome is of interest due to their conservation value in a global biodiversity hotspot, the ecological services they provide, and their importance for the avitourism industry. We conducted an extensive repeated point-count survey across the Fynbos biome, South Africa, to determine probability of detection and covariates of site occupancy for 27 bird species. Detection of most species was influenced by time of day, temperature or vegetation height. Important covariates influencing site occupancy were vegetation height, altitude, time since fire and habitat type. Site occupancy for four fynbos-endemic species was positively associated with increasing altitude. We further conducted point counts and mist-netting in eastern regions of the Fynbos biome to calculate local density and standardised capture rates. A linear regression analysis showed that capture rates were a function of bird densities, but that several species were captured at higher rates than expected, notably nectarivorous species. During mist-netting a relative abundance count was conducted. We expected deviation of the fit of the regression of capture rates on relative abundance to correlate with detection probability, because we expect this index to underestimate skulking and cryptic species, but there was no correlation. Estimated species richness indices were highest for the biome-wide survey, and lowest for mist-netting due to the body size limit imposed by the capture technique. Overall, we showed that point counts are an effective method for surveying birds in the fynbos and that mist-netting can be used to create an index of relative abundance for smaller species, but can be significantly affected by net placement.