Expanding a classic woodland food chain into a geographically variable food web

Abstract

There is ample evidence that climate change is impacting on phenology and it has been suggested that this may generate trophic mismatches. A key system for investigating phenology and trophic mismatch occurs in spring in temperate deciduous woodlands, where folivorous caterpillars and their predators, insectivorous passerines, are reliant upon ephemeral resources for reproductive success and survival. However, studies are primarily conducted within single-site, oak- (Quercus sp) dominated woodland and focus on a single caterpillar species, winter moth (Operophtera brumata), despite these passerines being habitat generalists with large geographic ranges. It remains to be seen whether insights gained from these studies can be generalised on the landscape scale across different habitats. In this thesis, I explore the extent to which geographic and habitat variation operates in this system and attempt to expand the system beyond a linear single-species food chain into a more biologically realistic multi-species food web. I also identify the most important environmental factors predicting the phenology of the passerines to allow better predictions of how their phenology could alter under future climate change scenarios. To address these questions, I established a novel 220km transect of Scotland incorporating 40 field sites that vary in elevation and the type of deciduous woodland habitat, monitoring six blue tit (Cyanistes caeruleus) nestboxes, tree and invertebrate phenology and abundance, at each site throughout the springs of 2014-16.

Firstly, I assess how blue tit occupancy and productivity are affected by the variation in fine-scale woodland habitat, latitude, elevation and prey availability that exists along the transect (Chapter 2). I find that habitat variables strongly affect fledging success but not occupancy or clutch size, whilst occupancy exhibits biogeographic trends, revealing that the relationship between breeding decisions and outcomes differs among habitats and implies that it may be difficult to generalise results from one habitat to others.

Next, I aim to identify the environmental aspects which play a role in regulating blue tit reproductive phenology by examining the ability of temperature, tree phenology, invertebrate prey abundance and photoperiod to predict nest initiation and laying dates (Chapter 3). I find that night-time temperature in early spring is the most important predictor of both nest initiation and lay date (slopes ~ -3days/°C) and I suggest that this supports the hypothesis that temperature acts as a constraint on timing rather than a cue. Invertebrate abundance is also a positive correlate of lay date, possibly allowing fine-tuning of timing.

This knowledge provides clearer foundations from which to predict future phenological change and possible trophic mismatch in this system. There is the potential that the apparent effect of temperature on blue tit reproductive phenology is indirect and mediated by diet, which is largely undescribed in the period prior to breeding. Therefore, in Chapter 4 I examine how blue tit diet varies across habitat, geography and time, and whether there is a dietary cue utilised to initiate breeding phenology, using data from metabarcoding faeces collected from nestbox-roosting adults in early spring. Geographic variation in diet is substantial, with high site-to-site dietary turnover (β-diversity), as well as high turnover along the elevational and latitudinal gradients studied. Dietary α-diversity (richness) is unaffected by geographical variables, but increases over time, with significant pre-breeding dietary increases in Lepidoptera and Hemiptera signifying a possible cue. In addition, these data provide the most comprehensive next-generation insights into the diet of a wild bird to date and identify 432 prey taxa.

Finally, I analyse how biogeographic and habitat variables affect the phenology, abundance and diversity of caterpillars (Chapter 5). Host tree species’ varied significantly in their likelihood of hosting a caterpillar, with oak and willow (Salix sp.) the most likely. Biogeography had less effect on the likelihood of caterpillar occurrence, but elevation delayed peak date by 3.7 days/100m increase. There was also support for the spring caterpillar peak being dominated by a few key species, with over half of all caterpillars identified being of just three of the 62 total species, including winter moth. These findings contribute to understanding how the temporal distribution of caterpillars varies across habitats on the landscape scale.

Taken together, the findings of this thesis reveal considerable geographic and habitat variation throughout this system, in both the composition of the food web and the impacts on blue tit productivity, demonstrating why caution must be exercised when extrapolating findings from one location or habitat to others.