Influence of season, habitat, temperature, and invertebrate availability on nocturnal activity of the New Zealand long‐tailed bat (Chalinolobus tuberculatus)

Abstract

Nocturnal patterns of activity by the New Zealand long‐tailed bat (Chalinolobus tuberculatus) varied significantly in relation to habitat, season, time of night, temperature, and invertebrate activity. Automatic bat detection units recorded 12072 bat passes containing 593 feeding buzzes during 580 nights of sampling. Overall, 46.3% of passes were along roads within forest, 42.7% along forest edges, 8.1% in open grassland and 2.9% within forest. Pass rates averaged 30.8/night (95% confidence interval (CI) = 23.3–38.4) during spring, 31.7 (CI = 21.3–42.1) during summer, 6.7 (CI = 4.4–9.0) during autumn and 1.6 (CI = 0.2–3.0) during winter. Bats were active throughout the night in all habitats, but the patterns of their activity were different in each. During spring and summer a significantly higher proportion of bat passes in edge and open habitats were recorded in the first 2 h after sunset, and activity then declined steadily towards dawn. In road habitats activity peaked in the second hour but was more consistent during the remaining hours of the night. Activity in forest did not vary through the night. Patterns were similar, though less pronounced, in autumn but in winter there was little activity during the first three hours after sunset or in the 5 h before dawn. Analysis of deviance models indicated that habitat, season, minimum overnight temperature, and invertebrate activity contributed to explaining the activities of bats. Overnight temperature was more important than invertebrate activity in explaining total bat activity in a night. Invertebrate activity was more significant than minimum temperature in the model explaining the amount of foraging activity per night, inferring that minimum temperature determines whether bats fly at night, while invertebrate activity determines how long bats feed. Techniques for indexing bat activity that are designed to survey distribution or monitor population trends should control for temporal variation and temperature to enable valid comparisons between counts.