Modelling flowering of plants using time-to-event methods

Climate and environmental factors are known to influence the propensity of plants to flower. In order to capture the time dynamics in the underlying biological processes, this paper proposes to use a grouped-data survival analysis methodology. We formulate a semi-parametric model for the hazard rate of flowering, which allows all available climatic and environmental data to be included. Estimation can be carried using a generalized linear model approach. In addition, the model can be validated by means of recently developed residual-based model-checking procedures similar to what is available for ordinary regression models. The main finding is that the model proposed captures the major trends and is a first step in the direction of a model-based quantification of some climatic and environmental factors that influence the flowering of plants. (C) 2009 Elsevier B.V. All rights reserved.