Bud burst of birch in Finland and the United Kingdom - Logistic regression analysis and modeling
Abstract: The day of bud burst (DBB) of different tree species are known to be affected by factors such as growing degree days and temperature. In this paper a two state Markov chain is used to model DBB for birch. The model is fit using logistic regression and LASSO regularization is used to evaluate which of many potential factors best forecast DBB. Data of birch from both Finland and the United Kingdom is studied and differences between the models adapted to the two countries are investigated. For modeling purposes to capture the environment of forecasting, estimated interpolated gridded climate data was used and not directly measured climate data. It is found that the models give very accurate predictions on the DBB. For Finland it is little more than 2 days in mean absolute error (MAE). The model is also fairly compact having less than 10 explaining covariates. The covariate, accumulated growing degree days, was as expected part of the models as well as among others variation of precipitation.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)