Nicole Augustin (University of Bath)
Tue 17 Sep 2019, 11:00 - 11:55
JCMB 5323

If you have a question about this talk, please contact: Tim Cannings (tcannin2)

Note that we'll start promptly at 11:00am

Forest health is monitored in Europe by The International Co-operative Programme on Assessment and Monitoring of Air Pollution Effects (ICP Forests) in cooperation with the European Union. More recently climate change has contributed to the decline in forest health and monitoring data are increasingly being used to investigate the effects of climate change on forests in order to decide on forest management strategies for mitigation. We talk about two modelling projects utilising parts of these spatio-temporal data for official reporting. 

The first one uses a generalized additive mixed model for estimating spatio-temporal trends of defoliation, an indicator for tree health. The temporal trend of defoliation differs between areas because of site characteristics and pollution levels, making it necessary to allow for space-time interaction in the model. 

In the second project, tree survival is modelled as a function of predictor variables on tree defoliation, climate, soil characteristics and water budget. We use a smooth additive Cox model which allows for random effects taking care of dependence between neighbouring trees and non-linear functions of spatial time varying and functional predictors. At each of 2400 grid locations 24 trees were observed between 1985 and 2013, with not all locations being observed yearly, some locations ceased being observed and new locations started being observed throughout. Altogether about  58000 trees are observed making the analysis computationally challenging due to the large number of parameters and large sample size.  The large number of correlated time varying environmental predictors with non-linear effects means full models are too big for traditional backward selection. We discuss how we address these challenges on computation and model selection and present the main scientific results.