What makes tits hoard food? Environmental regulation of food hoarding in titmice.

Overview

Some species of animals have evolved food-hoarding behaviour to survive periods of uncertain or scarce food access. Food-hoarding titmice (e.g. coal tits) hoard food both seasonally and on an hour-by-hour basis. The seasonal hoarding serves to enrich their winter foraging niche and therefore increase the odds of finding food in the winter, while the hour-by-hour hoarding allows the birds to take advantage of temporary food abundance by postponing consumption of the extra food items until a later time, when food is unavailable. Whereas the adaptive value of the behaviour is clear, the mechanisms that drive hoarding motivation at these two disparate time scales are less clear.
In this project, you will focus on understanding the environmental conditions that induce food-hoarding behaviour. You will, in particular, test two competing hypotheses: (1) that different mechanisms control hoarding on a seasonal scale vs. short-term hoarding; or (2) that one single set of rules, responding to a combination of temperature, day length, and food availability, can account for both the seasonal and the hour-by-hour patterns in food hoarding. Understanding how the birds respond to different environmental factors would also allow us to assess how sensitive the birds are to changes in the environment and to predict how hoarding behaviour would change under conditions of global warming.

Methodology

The project will take a two-pronged approach: data on the link between environmental conditions and hoarding motivation will be collected using a citizen-science approach, while the hypotheses about how the birds respond to the environmental conditions will be built into simulation models (“agent-based modelling”).
For the citizen-science approach, you will develop an observation paradigm that can be implemented by committed volunteers watching their garden feeders. Hoarding motivation can be observed as hoarding intensity when food is abundantly available, which it is in gardens with feeders. We have made connections with the British Trust for Ornithology’s Garden Birdwatch programme and Cornell’s Lab of Ornithology’s FeederWatch programme to collect data on food-hoarding tits (e.g., coal tits, marsh tits and willow tits in the UK and Europe; black-capped chickadees in North America). We will then make use of public weather databases to connect these observations with local temperature and precipitation conditions, using state-of-the-art large data analysis techniques (e.g. machine learning algorithms).
For the simulation modelling, you will build on past work from our lab to simulate a food-hoarding bird (in NetLogo initially) and expose these model birds to different environmental conditions. The simulation allows us to implement different decision rules about when to forage, when to eat and when to hoard, depending on the bird’s internal and external state (e.g. stomach content, fat reserves, time of day, and environmental conditions). Different models will be compared to each other, and decision rules will be optimized using advanced computing techniques such as genetic algorithms and dynamic programming to find the decision rules that best replicate observations from the citizen-science project. Finally, the selected model will be used to predict behaviour and survival of food-hoarding birds under changed climate conditions, taking into account changes in food preservation and inter-specific competition with increased winter temperatures and increased variability in weather.

Project Timeline

Year 1

1. Develop and pilot a garden bird observation protocol and data-collection website with a small number of volunteers across different countries.
2. Recruit volunteers to roll out the citizen-science data collection
3. Build simulation models that responds to different environmental and physiological state variables to decide whether to forage, eat and/or hoard. Test these model against patterns of food-hoarding found in the literature.

Year 2

1. Run the citizen-science data collection.
2. Analyse the data obtained in the first year.
3. Adapt the model (possibly into a new modelling environment) in order to apply genetic algorithms to optimize model parameters.
4. Apply the genetic algorithm to evolve model parameters to match the citizen-science collected data from the first year.
5. Build an optimization model (dynamic programming) to predict theoretically optimal behaviour

Year 3

1. Continue a second round of citizen-science data collection to build a stronger database
2. Confirm the optimal model parameters to match the empirical findings – Publication 1.
3. Compare the optimal model parameters based on empirical findings to theoretically optimal model parameters that optimize winter survival in optimization models – Publication 2.
4. Use the final model to test the hypothesis that the decision rules themselves are local adaptations to different environmental factors, by comparing whether models that match empirical data from high latitudes have the same decision parameters as models that match empirical data from lower latitudes and/or different species – Publication 3.
5. Use the final models to make predictions about how birds would respond to changing environmental factors associated with Climate Change – Publication 4.

Year 3.5

Finalize data analysis and write up thesis and publications.

Training
& Skills

The student will be trained in two crucial skills: computer modelling (and associated programming skills) and public engagement and citizen science data collection. They will also be trained in the necessary multi-level statistical modelling needed to analyse the diverse and complex citizen-science data.
In addition, they will be trained in communication skills, both to professional and to public audiences, as both the methodology of the citizen-science project and its eventual findings need to be communicated to the volunteers and the wider community of amateur ornithologists.

References & further reading

L. J. Henderson, R. C. Cockcroft, H. Kaiya, T. Boswell, T. V. Smulders, Peripherally injected ghrelin and leptin reduce food hoarding and mass gain in the coal tit (Periparus ater). Proc. R. Soc. B 285, 20180417 (2018).
V. V. Pravosudov, T. C. J. Grubb, Management of fat reserves in tufted titmice (Parus bicolor): evidence against a trade-off with food hoards. Behav Ecol Sociobiol 42, 57-62 (1998).
V. V. Pravosudov, J. R. Lucas, A dynamic model of short-term energy management in small food-caching and non-caching birds. Behav Ecol 12, 207-218 (2001).
V. V. Pravosudov, J. R. Lucas, Daily patterns of energy storage in food-caching birds under variable daily predation risk: a dynamic state variable model. Behavioral Ecology & Sociobiology 50, 239-250 (2001).
T. V. Smulders, A game theoretical model of the evolution of food hoarding: Applications to the Paridae. Am Nat 151, 356-366 (1998).

Further Information

Tom Smulders:
Tom.smulders@ncl.ac.uk
+44 191 208 5790

Apply Now