Dendroclimatology is one of the most important methods for examining past global climate change. Tree-ring data have been used widely as important archives in paleoclimatic research because of their precise dating, annual resolution, and high correlations characteristics with instrumental meteorological data. In recent years, new techniques, such as X-ray densitometry to measure wood density have been used in dendroclimatologic research to show the relationship with the variation of wood density and climate change. However, few laboratories have X-ray densitometry instrumentation, and few dendrochronologists can use such systems, due to high costs and complexity. Tree-ring gray values can be obtained through a tree-ring image analysis system (WINDENDRO^TM 2005), which includes a high resolution scanner, a PC computer and a software program. Variation in tree-ring gray values may be regarded as an approximation of variation in tree-ring density. So far there are a few studies that rely on the data of tree-ring gray values. However, studies to correlate tree-ring gray values with tree-ring densities in the same sampling site have not been conducted. Thus, it is important to determine accurate correlations between these measurements, and to verify the potential of tree-ring gray values in dendroclimatology.
Healthy spruce (Picea schrenkiana) trees grow in sparse ground vegetation. P. schrenkiana with little or no apparent evidence of human or other disturbances were selected for sampling. Two thin cores (5.15 mm diameter) and one thick core (12.00 mm diameter) were extracted from each tree in different directions for cross-dating. A total of 62 thin cores and 31 thick cores from 31 spruce trees were collected from AKS sampling site (2420-2482 m asl, 43°12'12.8″ N and 84°47'47.7″ E) in May 2009, which is located in a mountainous area of the Aikendaban region.
All of tree-ring samples were dried, mounted, surfaced and cross-dated following standard dendrochronological procedures. Based on chronological statistics, chronological curves and correlations in the all-frequency, high-frequency and low-frequency domain among five kinds of tree-ring gray value chronologies (TRG, TEG, TLG, MXG and MIG) and four kinds of tree-ring density chronologies (TED, TLD, MXD and MID), results show that the variation of earlywood average gray values and latewood average gray values can reflect well the variation of earlywood average densities and latewood average densities and the negative correlation between them is significant. However, the variation of tree-ring maximum gray values and tree-ring minimum gray values do not correlate well with variation of tree-ring minimum densities and tree-ring maximum densities.
In addition, meteorological data were used from Bayinbuluke station, which is closet to the sampling site. Correlation function analysis was applied to investigate the relationship between tree-ring gray values and tree-ring densities and climate during the calibrated period (1958-2008). Correlation results between the tree-ring gray value chronologies and tree-ring density chronologies and mean monthly temperatures, mean monthly highest temperatures, mean monthly lowest temperatures and mean monthly precipitations from Bayinbuluke station from previous October to current September (1958-2008). The results showed that the tree-ring gray value chronologies were negatively correlated with the three kinds of temperature. Moreover, the tree-ring density chronologies were positively correlated with the three kinds of temperature. In addition, the relationships between the chronologies (tree-ring gray values and the tree-ring densities) and precipitation were weak. After combinations, the higher correlation coefficient between one kind of tree-ring gray value chronologies (TRG_t) and the mean monthly highest temperatures was found for the period from May to August (r=-0.542, P<0.0001, N=51).
Our results provide basic information for reconstructing climatic variations in the past based on tree-ring gray values, and confirm the potential for the use of tree-ring gray values in the study of dendroclimatology.