Spatial modelling of phenological observations to analyse their interannual variations in Germany

We present a method to quality control and analyse a series of phenological observations and, from there, construct a spatial map for a given phenological stage or phase. In this case, data for full bloom of apple, stemming from a volunteer network managed by the German Weather Service, were used as an example. The data span the period from 1951 to 1997 and cover an area of around 14,000 km(2) in western Germany. First, the method itself is presented. It includes quality control of raw data, construction of a field for the mean date over all years of a given phenological stage and construction of the respective anomaly fields for each year. Then, the determination of a statistically correct spatial resolution of the final field is presented, which in this case is around 10-15 km. The final analysis of the time mean field and the respective anomaly fields indicates a possible way to extract all relevant climate factors determining phenological events overcoming the problems of a standard site-specific adaptation of a model. (C) 2002 Elsevier Science B.V. All rights reserved.