City pigeons : Kererū (Hemiphaga novaeseelandiae) in the urban Dunedin environment : abundance, habitat selection and rehabilitation outcomes

Abstract

Kererū (Hemiphaga novaeseelandiae) are a species of great biological and cultural significance within New Zealand. It is an important seed disperser of large-fruited tree species and a taonga species to Māori, and is in critical decline over much of New Zealand. Despite this overall trend of decreasing abundance, there is a prominent and apparently healthy kererū population in the urban Dunedin area.

This thesis investigates density, abundance and habitat selection of the Dunedin kererū population. Causes and outcomes of kererū undergoing rehabilitation due to injury are examined through analysis of historical rehabilitation records. In an attempt to measure success of the rehabilitation programme, radio-tracking of rehabilitated kererū was undertaken. Finally, morphometric measurements were assessed to determine their value as a tool for sexing kererū, which would aid in conservation and rehabilitation.

Kererū density in six habitat types within Dunedin was investigated using Distance Sampling. Density estimates were then used to estimate abundance of kererū within Dunedin City. Density was highest in the bush/forest habitat (0.24 birds per hectare) which consists of a mosaic of forested areas. The next highest density (0.16 birds per hectare) was found in residential 1 habitat, which is characterised by mature gardens with a mixed vegetation structure. Residential 2 habitat is characterised by open-vegetated space dominated by lawns and had low kererū density (0.06 birds per hectare). No kererū were sighted in the commercial, industrial or residential 3 habitats, defined as having less than one third open space and few mature gardens. These density estimates provided an abundance estimate of 334 kererū in the urban Dunedin environment.

To investigate habitat selection count data collected during distance sampling were used to construct resource selection functions by comparing used and available resources. Resource selection by kererū was significantly non-random over a year-long period (p = 0.001), with residential 1 habitat the most preferred habitat type.

The density and habitat selection data indicate two habitats are highly favoured by the current population of kererū in Dunedin. Bush/forest and residential 1 habitat need to be maintained and enhanced if we wish to promote kererū within the city. Other habitats which are currently not utilised can be managed to increase similarities with the favoured habitats and thereby increase the habitat available for kererū.

Logistic regression on the rehabilitation data-set tested whether cause of admission, month of admission and age of bird could predict the outcome of rehabilitation (survival or death). Testing the interaction between age of bird and cause of admission allowed 74.5% of outcomes to be predicted correctly. The primary cause of rehabilitation was impact-related injury (37% of admissions), and the majority of admissions were related to human factors (67%). Small sample size (n=6) of radio-tracked individuals resulted in little ability to draw conclusions about the fate of rehabilitated kererū, although survival rates appeared very low.

Measurements of wing length, head length, bill length, bill width, and bill height were taken from a museum sample of 59 kererū and their sex was recorded. A discriminant function analysis tested if these measurements could accurately predict sex. Predictive power was low with 66.1% of cases being classified correctly, suggesting, morphometric sexing is not a viable option for sexing kererū.