The Tree-ring Climate Response Relationships At Different Microtopography In Arid Alpine Region Of Northwestern China

Tree ring is one of the important paleoclimate proxy used for reconstructions, which has been widely used in global change studies. Large numbers of tree-ring based climate reconstructions had been conducted in arid region of northwestern China. However, the tree ring growth climate relationships at different microtopography are still ambiguous.The arid region of northwestern China in this article refers to the region located at the monsoon fringe area, including the Gansu, Ningxia and Qinghai provinces, from the Qaidam basin to the Ningxia plain along the Yellow River. This region is one of the hotspot that the earliest dendroclimatology researches were conducted. In present study, 773 cores from 386 trees in 10 sampling sites were collected, including the Chinese pine trees(Pinus tabuliformis) from the Helan Mountains and Hasi Mountains, and the Qilian juniper trees(Sabina przewalskii) from the Zongwulong Mountains near the Delingha city. After the lab processing, cross dating and detrending, the STD, RES and ARS chronologies for each site were developed. We calculated the statistical characters of these chronologies and did the comparisons among them with correlation and cluster analysis. Then, we investigated the tree ring growth-climate response relationships by the correlation analysis, the first order differential correlation and trend surface analysis between the chronologies and climate factors. Finally, we discussed the different relationships that related to the habitats. The conclusions of this article are as followed:The mean sensitive values of Chinese pine tree rings were more than that of Qilian juniper, while it was also related with the altitude. Qilian juniper trees near the lower treeline were more sensitive than Chinese pines(all the Chinese pine sites were located near the upper treeline). Precipitation was clarified to be the mainly limited factor for trees growth in the 10 study sites. The correlations also indicated that the summer air temperature negatively influenced trees growth in the 5 Chinese pine sites and 4 Qilian juniper sites, while trees at the highest elevation site showed positive correlations with the air temperature. All the 5 juniper chronologies positively correlated with May-June precipitation in the current year, and negatively correlated with June temperature in current year and November precipitation in previous year.Comparisons among the Qilian juniper sites from different elevations indicated that the tree ring-precipitation response relationships became stronger with the altitude decreasing. When the precipitation was scarce, the ring width from the lower treeline showed negative correlations with the air temperature. When the precipitation was enough, the ring width from the upper treeline indicated insignificant correlations with the air temperature.Comparison between the Chinese pine trees from two different sampling sites(Helan Mountains and Hasi Mountains) indicated that all the 5 chronologies positively correlated with precipitation, while the correlation coefficients and the periods with significant correlations were different. 3 chronologies at the Helan Mountains positively correlated with the precipitation from March-June in the current year, and negatively correlated with temperature in the current March and July. However, trees grown in the 2 sites at the eastern mountain indicated higher climatic sensitivity than the trees in western site(BS). Both series from the Hasi Mountains positively correlated with the precipitation in previous July-September and current March-June, with the correlation coefficients were higher in previous year than that in current year. The first order correlation analysis showed increased correlation values with the precipitation in March-June, suggesting that the growth accumulative in previous year might affect the ring width in the current year.Comparison between the original series and high frequencies series indicated gently differences in the correlations for the Chinese pine tree-ring series. However, the differences were obvious for the Qilian juniper tree-ring series. The first order correlation results revealed that tree ring chronologies correlated positively with May-June precipitation and negatively correlated with temperature in June and previous November precipitation. There was significant positive correlation between DLH02, DLH03, which located near the upper treeline, and the monthly mean air temperature and monthly minimum air temperature. The consistent increasing trend in both the tree-ring chronologies and temperature curve caused the high correlations between the Qilian juniper chronologies and temperature, which might be unreal. Therefore, the Qilian juniper trees mainly response precipitation variation on high frequency.We reconstructed the precipitation in the Gansu-Ningxia monsoon fringe region(1761-2012) and Delingha region(1357-2013), respectively, using the PC1 of tree ring chronologies based on the correlation analysis results. The spatial correlation field of reconstructed series and instrument data with the CRU 0.5°*0.5° gridded precipitation data showed the similar patterns. With the empirical mode decomposition program, we decomposed each series into 8 intrinsic mode functions(IMF) from high to low frequencies. Every IMF showed cycles on different frequencies. The two reconstructed series both showed the cycles of 2-4, 6, 11, 23 and 46 years. Besides, the series from the GN monsoon fringe region had a cycle of 92 years and the Delingha region had one of 138 years.By comparing with the recently reconstructed large-scale climate indices, we found that our precipitation reconstruction in Delingha showed the similar trend with the northern hemisphere temperature. The reconstructed precipitation in the GN monsoon fringe region also indicated the similar variation before 1960, but the precipitation decreased in recent 50 years.The running correlations with 30-year windows between the reconstructed precipitation and winter NAO index(wNAO) suggested a positive correlation with the reconstructed air temperature of China. We discussed its mechanism.