BRANTA — Sally Dorothy Hofmeyr
Impacts of environmental change on large terrestrial bird species in South Africa: insights from citizen science data
Institution: University of Cape Town, South Africa
Supervisors: L.G. Underhill, P. Barnard
Details: PhD 2012 (Completed)
Address: School of Animal, Plant and Environmental Sciences, University of the Witwatersrand and Animal Demography Unit, Department of Biological Sciences, University of Cape Town (Jan 2013) Email
Subject Keywords: Large terrestrial birds, bustards, cranes, korhaans, land-use change, environmental change, conservation, population trends, habitat use
Species Keywords: Blue Crane Anthropoides paradiseus, Denham's Bustard Neotis denhami, Southern Black Korhaan Afrotis afra, Northern Black Korhaan A. afraoides, Karoo Korhaan Eupodotis vigorsii, Blue Korhaan E. caerulescens
Large terrestrial bird species, especially cranes and bustards, have adapted to low intensity agriculture to varying degrees, but large-scale industrial agriculture is in general inimical to these species. Cranes are charismatic and well studied, but bustards are retiring and in general cryptically coloured, and little is known of most species. Of South Africa's 10 bustard species, two are endemic and three subspecies are endemic or near-endemic. Six species are threatened or near-threatened. Three crane species occur in South Africa, one of which is near-endemic; all are threatened. This thesis used data from two long-term public participation bird monitoring projects to improve our understanding of six of these 13 species. The first and second Southern African Bird Atlas Projects (SABAP1, 1987-1992, and SABAP2, 2007-) provide two sets of presence/absence data which can be compared. The Coordinated Avifaunal Roadcounts (CAR; 1993-) project provides roadcount data spanning a similar period.
To assess the variability inherent in CAR data, additional, consecutive-day surveys of six CAR routes were conducted. The variability of the data was assessed by measuring the relative mean absolute difference between daily totals for route sections of a range of lengths, from 1 km up to the whole route. Species whose survey totals were most reliable were non-flocking, sedentary, and abundant, in descending order of importance. Exemplars were Northern Black Korhaan Afrotis afraoides, and Blue Crane Anthropoides paradiseus, especially in summer and where it was abundant. Flocking species, such as Helmeted Guineafowl Numida meleagris, and relatively uncommon species that range over large areas, such as Secretarybirds Sagittarius serpentarius, were the least reliably surveyed.
Methods for analysing CAR and SABAP data were developed and applied to data for the Southern Black Korhaan A. afra, Blue Crane, Denham's Bustard Neotis denhami stanleyi, Blue Korhaan Eupodotis caerulescens, Karoo Korhaan E. vigorsii vigorsii and Northern Black Korhaan.
CAR data were used to produce population trends for CAR precincts (groups of routes with similar habitat characteristics). Precincts and seasons (summer and winter) were analysed separately. Missing data were imputed using the Underhill index. Annual survey totals were converted to birds/100 km of road. These figures were plotted and a smoothed curve was fitted, using weighted linear regressions, to aid interpretation. Southern Black Korhaan totals decreased in the Western Cape and increased in the Eastern Cape. Blue Crane trends were increasing in the Western Cape and in one precinct in the Free State, but were decreasing or stable in other precincts. Denham's Bustard totals increased in one Eastern Cape and one Western Cape precinct, and decreased in one KwaZulu-Natal precinct. Trends for the remaining three korhaan species were stable or slightly increasing overall.
CAR habitat use data, separated by precinct and season, were summarised graphically and habitat selection was assessed. Habitat availability data were obtained from the National Land-Cover maps for 1994, 2000 and 2009. However, most habitat categories in these maps were different to those used in the CAR project. Habitat use and availability data were therefore summarised into two broad land-cover categories: natural and transformed. Habitat selection by each species was assessed using the Jacobs index. Southern and Northern Black Korhaans showed little preference for transformed habitats; in contrast, Blue Cranes selected transformed over natural habitats in most precincts and seasons. Other species showed intermediate levels of habitat selection. Preference for transformed habitats was strongest in the Fynbos biome and weakest in the Grassland biome.
Atlas reporting rates are related to abundance, although the exact nature of the relationship is not known. Therefore, changes in reporting rates for one species in one area for the same season or set of seasons are likely to reflect changes in abundance. Differences in reporting rates between SABAP1 and SABAP2 were analysed using maps. A statistic which takes into account the difference in reporting rates and sample sizes of checklists was calculated for each atlas grid cell. Grid cells were shaded according to the value of this statistic so as to highlight areas of potential conservation concern and those of possible population increase. SABAP comparison maps largely supported the CAR data population trends and thus provided useful information for areas not well covered by the CAR project. Reporting rates for the Southern Black Korhaan decreased in c. 80% of its range. Blue Crane reporting rates increased in the Fynbos and grassy Nama Karoo, but decreased in the rest of its range. The pattern was similar for the Denham's Bustard but the area of increase was smaller. Blue Korhaan reporting rates decreased in most of the north-easterly third of its range, but were stable elsewhere. Karoo Korhaan reporting rates were approximately stable but SABAP2 coverage of this species' range was poor. Northern Black Korhaan reporting rates increased in the Free State but decreased slightly overall in the rest of the range that had good SABAP2 coverage.
Seasonal national population indices based CAR population trend data were calculated for each species. Data for each precinct were multiplied by five different sets of weights to produce five different indices, which were compared. The extent of coverage of each species' range by the CAR project was assessed. Southern Black Korhaan national population indices were inconclusive, while Blue Crane indices showed that the species had increased by 200-300% within the area covered by CAR over the period 1993-2010. Denham's Bustard indices showed the species to be stable or increasing slightly, although the range was relatively poorly covered by CAR. Indices for the remaining three korhaans suggested that the populations had remained stable or increased slightly overall, but all indices showed a high level of variability. The Blue Korhaan's range was well covered, but small proportions of the Karoo and Northern Black Korhaans' ranges were covered.
For the Southern Black Korhaan only, occupancy modelling was used to model occupancy and detection probabilities for SABAP1 and SABAP2 data for this species. Covariates used were biome and degree of land transformation. Occupancy declined over the species range as a whole, and detection probabilities declined strongly in three of the four biomes represented in the range, lending additional support to the conclusion that this species declined substantially in abundance and somewhat in range between SABAP1 and SABAP2.
Finally, results were synthesised and an analysis of the diversity of bustards throughout South Africa was conducted using SABAP1 and SABAP2 data. This analysis highlighted areas where CAR and SABAP2 coverage should be increased, to enable more confident assessment of the status of South African bustards. It was recommended that the 2012 threat categories for five of the study species remain the same, but that the Southern Black Korhaan (still listed as Least Concern in 2012) should be reclassified as Vulnerable.