Using Long term harvest records of sooty shearwaters (Tītī; Puffinus griseus) to predict shifts in the Southern Oscillation

by Grant Richard Woodrow Humphries

Institution: University of Otago
Year: 0
Keywords: Sooty shearwater; El Nino; Southern Oscillation; Spatial modeling; Machine learning; La Nina
Record ID: 1316998
Full text PDF: http://hdl.handle.net/10523/4932


Rakiura Māori (New Zealand’s southern-most group of indigenous peoples) have harvested the chicks of burrow-nesting sooty shearwaters (Tītī; Puffinus griseus) for generations. As part of the harvest process some families have maintained annual harvest diaries, some of which date back to the 1950s. Throughout this thesis I investigated how these diaries could be used to determine if data obtained from the sooty shearwater harvest can predict shifts in the Southern Oscillation Index (SOI). To answer this question, I had to first calculate indices of the harvest which take into account human effort. Then I had to determine if those indices could in fact predict upcoming values of SOI. I next determined what oceanographic factors explained at-sea behaviour of sooty shearwater adults, and if those variables also explained variation in the harvest indices. I finally determined if there were any relationships between oceanographic variables and upcoming values of SOI. In Chapter 2, I used generalised boosted regression models, a machine learning algorithm, to calculate a harvest index that takes into account factors which could impact the numbers of birds taken on any given hunt. The r2 of predicted versus observed values were between 0.59 and 0.90 for the nanao (first half of the season, when chicks are harvested from burrows during the day) and 0.67 and 0.88 for the rama (second half of the season, during which chicks are harvested from the surface at night). Exploration of the controlling factors of the models reveals that ‘day of season’ plays an important role in predicting daily harvest during the second half of the season (the rama). The nightly tally in the rama peaks approximately half-way through (10 – 15 days in), which is likely related to the timing of birds emerging from burrows to fledge. The models also suggest that data from the rama (when chicks are 100 – 120 days old) may be the most suitable for long-term monitoring of populations of sooty shearwaters due to consistencies in calculated harvest indices between diaries. Nanao harvest indices, though less consistent, showed similar patterns to those of the rama. When compared to the harvest indices calculated by general linear models by Clucas et al. (2012) I found that the agreement between both indices was r =0.56 and r =0.77 for the nanao and rama, respectively. Although harvest indices have been created in the past, this represents a re-analysis of these data with a different statistical technique, plus a new diary to test the relationships under new circumstances. The use of machine learning to correct for extraneous factors (e.g., hunting effort, skill level or weather) and create standardised measures could be applied to other systems such as fisheries or terrestrial resource management. The harvest indices calculated in this study were then used to examine relationships between chick quantity, quality and Pacific Climate indices. In Chapter 3, I demonstrate that shifts in sooty shearwater (Puffinus griseus) chick size and abundance occur in…