Mean relative humidity reconstruction based on the tree-ring width from Chinese pines since 1890 from the Qinglong region, China
Boyang ZHAO,Yu LIU,Huiming SONG,Qiang LI
DOI: https://doi.org/10.7515/JEE201605008
2016-01-01
Abstract:Background, aim, and scope The study of humidity is essential for better understanding the past climatic variations, however, there were few long-term humidity reconstructions based on tree-ring widths worldwide. Tree-rings have been considered as one of the best known archives in the past climate research field with their high resolution in time and reliability in cross-dating. In our study, a mean relative humidity (MRH5-7) reconstruction curve was carried out using the Chinese pine tree (Pinus tabulaeformis Carr.) from the Qinglong region in northern China.Materials and methods Employing the standard methods sponsored by the International Tree-Ring Data Bank (ITRDB), we collected 32 tree cores from 16 trees in Qinglong region (40.49°N, 119.53°E, 380 m a.s.l.) during October, 2013. The site has a discontinuous canopy with the sparse trees and the locally dominated tree species is Chinese pine (Pinus tabulaeformis Carr.). Following the standard dendrochronological procedures, all samples were carefully preprocessed in Laboratory of Tree Ring Research, Institute of Earth Environment, Chinese Academy of Sciences. All 32 cores were cross-dated precisely with the COFECHA program and three tree ring chronologies (STD, RES and ARS chronologies) were carried out with the ARSTAN program. In our case, we reconstruct past relative humidity with RES chronology because RES chronology has a good quality in storing high frequency climate signal. To clarify the climatic conditions in our study area, the meteorological data were extracted from the Qinglong station (40°24′N, 118°57′E, 227.5 m a.s.l., records from 1957 to 2012). PDSI which was extracted from the nearest data grid (41.25°N, 118.75°E) was compared with our RES chronology. To test the spatial representativeness of our reconstruction, spatial correlations between observed relative humidity and reconstructed relative humidity with CRU scPDSI were plotted via KNMI Climate Explorer (http://climexp.knmi.nl).We also further analyzed our reconstruction with the multi-taper method (MTM) to detect the periodicity of relative humidity variation.Results Mean relative humidity reconstruction (May—July) in Qinglong region since 1890 was carried out using tree ring RES chronology. The reconstruction explained 39.1% (38.0%, after adjustment for the loss of degrees of freedom) of the instrumental variance during the calibration period (1957—2012). A split calibration-verification method was employed to examine the stability of our reconstruction and the statistics certiifed that our linear regression model is reliable and could be used to reconstruct the past relative humidity variation. In addition, first-order differential and detrended reconstructed and observed May—July mean relative humidity show a significant correlation, which demonstrates that the reconstructed and observed curves showed a conformance in high frequency. On the decadal scale, there are ifve high-value MRH5—7 intervals (1895—1899, 1906—1914, 1924—1926, 1950—1955, 1984—2000) and five low-value MRH5—7 intervals (1900—1905, 1917—1921, 1927—1949, 1956—1973, 1975—1981). CRU scPDSI was employed as a bridge to connect our reconstruction and observation. The spatial correlation reveals that our reconstructed mean relative humidity variation show a synchronously relationship with the mean relative humidity observation in adjacent area. The MTM of spectral analysis was performed on our reconstruction, and significant high-frequent peaks were found at 5.20-year (95% C.L.), 4.20-year (95% C.L.), 4.03-year (95% C.L.), 3.86-year (95% C.L.), 3.58-year (95% C.L.) and 2.21-year (95% C.L.) interannual cycles.Discussion Relative humidity has a considerable inlfuence on photosynthesis and transpiration by controlling the stomata conductance on plant leaves, and then change the tree growth rate. The signiifcant correlation between relative humidity and tree ring width is reasonable and our reconstruction has a robust physiological base. We compare our reconstruction with drought/lfood index in Tangshan city. Historical records and mean relative humidity reconstruction signiifcantly correlate with Spearman rank correlation coefifcient ofr = ? 0.394 (n=111, p<0.001). The drought (lfood) events often occurred in the low (high) mean relative humidity value year. The drought years: 1900, 1919, 1936, 1944, 1968, 1972, 1981, 1992, 1999 and wet years: 1890, 1893, 1894, 1898, 1938, 1949, 1953, 1973, 1985, 1991, 1995 are recorded by both drought/lfood index and our reconstruction, which demonstrates that both curves showed a conformance in high frequency. Two curves still have tendency correlation after 11-year moving averages method was employed. Combined with spatial correlation with CRU scPDSI, our reconstruction reflects the drought/flood changes to a certain extent around the study area. MTM analysis indicates that the mean relative humidity in study area might also be influenced by ENSO episodes, which demonstrates that our reconstruction shows the large-scale representativeness of sea-land coupling.Conclusions Mean relative humidity during May to July is the dominated limiting factors in Qinglong region. A mean relative humidity (MRH5—7) reconstruction curve was carried out using tree ring width index and linear regression model. The reconstruction explained 39.1% (38.0%, after adjustment for the loss of degrees of freedom) of the instrumental variance during the calibration period (1957 — 2012). During past 123 years, there are ifve high-value MRH5—7 intervals (1895 — 1899, 1906 — 1914, 1924 — 1926, 1950 — 1955, 1984 — 2000) and five low-value MRH5—7 intervals (1900 — 1905, 1917 — 1921, 1927 — 1949, 1956 — 1973, 1975 — 1981). MTM analysis indicates that the reconstruction has six major cycles, 5.20-year (95% C.L.), 4.20-year (95% C.L.), 4.03-year (95% C.L.), 3.86-year (95% C.L.), 3.58-year (95% C.L.) and 2.21-year (95% C.L.).Recommendations and perspectives It is relatively rare to reconstruct past relative humidity in northern China, especially with tree ring width chronology. Here, we carry out a May to July mean relative humidity reconstruction using tree ring RES chronology. Our reconstruction is reliable and has signiifcant spatial representativeness based on the correlation analysis.