Relationships of Ring Widths in Arid‐Site Conifers to Variations in Monthly Temperature and Precipitation

Abstract

Two multivariate techniques for evaluation of ring—width and climatic relationships are described. One technique employs multiple regression on principal components of climate. Response functions are obtained which express in mathematical form the relative effect of monthly temperature and monthly precipitation during a 14—month period on ring—width variations. Response functions are calculated for 127 coniferous tree sites in western North America. The other technique is a cluster analysis which is used to identify similarities among the response functions and classify them on the basis of their similarities and differences. Higher—than—average precipitation most commonly results in higher—than—average growth, though on cold sites the effects of precipitation during the cooler parts of the year are sometimes lacking or inverse. Precipitation is directly related to growth throughout the entire year for 32% of the sites, representing many that have been used in dendrochronology. In the remaining 68% of the sites the effects of precipitation vary from season to season. Temperature is most commonly inversely related to ring width during autumn, spring, and summer. Inverse relationships occur in mid—winter for trees on the warmest sites and on certain high—altitude sites. However temperature effects are often direct ones, especially during winter and for many sites at high altitudes, high latitudes, or north—facing slopes. Site factors appear most responsible for variations in the growth responses. Aspect of slope appears to be the most critical, followed by altitude and latitude. There are fewer differences between species than between factors of the site, although certain species such as bristlecone pine have a more or less unique growth response. Some speculations are offered on the effects of the sites and their microenvironments on biological processes linking the climatic variables and growth. The variety of ring—width responses to variations in climatic factors suggest that more physiological study of trees on extreme sites may reveal unique growth—environment relationships that have not been observed on more optimum sites. The median percent of tree—growth variance accounted for by climate is approximately 60 to 65%. With such a high percentage of variation related to climate, ring—width variability, if adequately sampled, dated, and calibrated with climate, can be used successfully to estimate past climate, even though all linkages of cause and effect have not been demonstrated by physiological research. It is proposed for reconstruction of past climate that many ring—width chronologies be used with diverse tree—growth responses, because the differences can be handled by multivariate techniques and provide more information on various climatic factors than chronologies with the same growth response. It is also proposed that response functions have applications in the modeling of productivity in forest ecosystem. They can be used to ascertain the effects of climatic variation on tree growth and to help in extrapolating information gathered in one area to other tree sites.