ANALYSIS OF PINES GROWTH BY MEANS OF REGRESSION-MODELS BASED ON TIME-SERIES REGARDING LONG-TERM CLIMATE DATA

Pines on two sites and two classes of damage were examined in order to establish whether a change occurs in the long-term course of growth in thickness and whether this change is connected with the visible damage. In order to eliminate the effects of the natural ageing process and - as far as possible - climatic influences a regression model which takes both autoregressive components and external climatic regressors into consideration was adapted to the sequence of annual ring breadths. Adaptation was undertaken only up to a fixed point in time (1950) after which the further growth was prognosticated and compared with the actual growth. The precipitations of the current vegetation period and the temperatures of the preceding spring proved to be important climatic influences. The severely damaged pines exhibited a considerably slower growth in thickness from a long-term viewpoint than those that were outwardly undamaged, but this tendency is seen long before 1950 from a low tree-age onwards. A simultaneous, equidirectional change in growth behaviour of several trees belonging to one class of damage cannot be perceived. The estimated parameters of the regression models are in some cases typical for the sites. Differences between the damage classes occur in the absolute growth in thickness and in the reactions to the amounts of precipitation received.