Accessibility navigation

Understanding the information content in diverse observations of forest carbon stocks and fluxes for data assimilation and ecological modelling

Pinnington, E. M. (2017) Understanding the information content in diverse observations of forest carbon stocks and fluxes for data assimilation and ecological modelling. PhD thesis, University of Reading

Text - Thesis
· Please see our End User Agreement before downloading.

[img] Text - Thesis Deposit Form
· Restricted to Repository staff only


It is advisable to refer to the publisher's version if you intend to cite from this work. See Guidance on citing.


Land surface carbon uptake and its many components (e.g. its response to disturbance from fire, felling and insect outbreak) constitute the most uncertain processes in the global carbon cycle. This uncertainty arises from significant gaps in current direct observations and poor parameterisations or missing processes in current modelled predictions. Data assimilation provides a methodology for combining observations with modelled predictions to find the best estimate of the state and parameter variables for a given system. In this thesis we implement four-dimensional variational data assimilation to combine a simple model of forest carbon balance with observations from the Alice Holt forest in Hampshire, UK. The first aim of the thesis is concerned with understanding the information content in observations for data assimilation. It is important to understand which observations add most information to data assimilation schemes in order to better constrain future model predictions. We show that the information content in carbon balance observations can vary with time and different representations of error. We next seek to improve the characterisation of uncertainties for prior model estimates and observations. We propose including correlations between errors within ecosystem carbon balance data assimilation schemes. We find including correlations allows us to retrieve a more physically realistic set of parameter and initial state values for our model, leading to a 44% reduction in error for our 14-year model forecast of forest carbon uptake. Finally, we use the data assimilation techniques developed, with additional observations of leaf area index and woody biomass, to investigate the effect on forest carbon dynamics of selective felling at Alice Holt. We show selective felling had no significant effect on forest carbon uptake. Our most confident estimate suggests this is possible due to reductions in ecosystem respiration counteracting a predicted 337 g C m−2 reduction in gross primary productivity after felling.

Item Type:Thesis (PhD)
Thesis Supervisor:Quaife, T., Dance, S., Lawless, A., Morison, J. and Nichols, N.
Thesis/Report Department:Department of Meteorology
Identification Number/DOI:
Divisions:Science > School of Mathematical, Physical and Computational Sciences > Department of Meteorology
ID Code:73339


Downloads per month over past year

University Staff: Request a correction | Centaur Editors: Update this record

Page navigation