Abstract:
Quantitative stand descriptions, multivariate analyses, population structures and dendrochronological aspects of kauri were explored during the present studies.
Twenty-five stands of mature kauri, throughout its natural range, were sampled by the point-centred quarter (p.c.q.) method and described separately. Multivariate data on density, basal area and frequency were then subjected to TWINSPAN (a polythetic method of classification) and DECORANA (detrended correspondence analysis ordination). On the basis of these analyses four overlapping kauri forest types were recognised. No significant correlation occurred between stand groups (defined by TWINSPAN and DECORANA) and dendrochronological attributes. This implies that the vegetation composition of stands does not provide information about dendrochronologically sensitive sites.
Density, basal area, frequency and size class distribution provided information about the structure of the kauri population in each stand. The study suggests that kauri does not compete with associated canopy species in its later life. Regeneration gaps were not evident in mature kauri forests, indicating that they are not unstable. These characteristics increase the suitability of kauri for tree-ring investigations. Structural variability is related to variability in growth rate and age and these are discussed for various kauri forests.
Dendrochronological methods were applied to 19 kauri stands, and dated chronologies (with maximum period from 1580 to 1981 AD) from eight sites were obtained. Ring-width sequences of cross-matched sites were standardised using INDEX and the various chronology and sample statistics are discussed. These chronologies show from 20 to 35 percent common chronology variance ('Y' in ANOVA; due to climate). Noncross-matched sites were treated the same way for comparison. It is indicated that high sensitivity does not reflect dendrochronological suitability unless the cores are properly cross-matched, and this emphasizes the need for better site selection and cross-matching techniques in forests such as those in New Zealand.
Highly significant correlations (P<.001) between dated chronologies (from 1790 to 1976 AD) and similar patterns of correlation between cores among various sites suggested that all sites were influenced by similar climatic factors. A master chronology was constructed, but found to be less useful than the separate chronologies.
It is concluded that north-facing, steep slopes are most suitable for tree-ring studies. These were the only sites from which dated chronologies were developed; other sites showed various problems which are discussed.
Kauri growth-climate relationships were evaluated by running various types of response functions. Seventy four response functions are presented and their results are discussed. Variance due to climate ranged from 21 to 53 percent, and is comparable with other studies in New Zealand. Overall, in all sites growth-climate response is similar; at the beginning and end of the growing season low rainfall promotes growth while in the middle of the growing period it may reduce growth. It is concluded that, despite the limitations discussed, these results may be used in dendroclimatic reconstruction.