Abstract
Response function (RF) analyses were used to study the climatic signals in eleven kauri (Agathis australis) tree-ring chronologies from sites throughout northern New Zealand. Single station and regionally averaged monthly rainfall and mean temperature data over the period 1910–1982 were used. The RFs were selected on the basis of three significance tests. The climate regression equations (RFs) derived from different chronologies using different climatic data sets were generally similar, allowing them to be combined into a summary RF diagram. The common variance in the tree-ring chronologies (% Y) ranged from 20-35 % and was related only weakly to the variance due to climate (% r c) in the RFs. Percent variance explained by climate ranged from 17-53 % (mean = 34 % ). Three years of prior growth accounted for a further 2-47% (mean = 16 %). A negative first year lag effect over the period analysed contrasted with positive autocorrelations in the chronologies as a whole. This effect requires fo.rther study before the chronologies are used for climatic reconstruction. Such reconstructions are likely to return higher correlations with the true climatic parameters if single station (rather than regionally averaged) climatic data are used.
The analyses clearly indicate that good grbwth (a wide ring) in kauri is associated with dry conditions, especially close to the start (September-October) and the end (April) of the growing season. Narrow rings are less likely to indicate drought years than the reverse, although late summer drought may reduce the growth rate on some sites. The temperature response suggests that warm winters favour subsequent growth, while warm conditions in early summer are detrimental. The temperature effect may act through soil moisture and/or by initiating phenological events such as cone production. The RFs support the concept of kauri as a species favouring dry sunny sites growing well in periods of relatively dry climate with mild winters.