Abstract: Based on the one-year (2005) observations with a frequency of half hour on the stem sap flow of  Larix gmelinii plantation trees planted in 1969 and the related environmental factors air humidity (RH), air temperature (T_air), photosynthetic components active radiation (PAR), soil temperature (T_soil), and soil moisture (TDR), principal analysis (PCA) and correction analysis were made on the time lag effect of the stem flow in different seasons (26 days of each season) and in a year via dislocation analysis, with the complexity and its integrative effects of the time lags of environment factors affecting the stem sap flow approached. The results showed that in different seasons and for different environmental factors, the time lag effect varied obviously. In general, the time lag of PAR was 0.5-1 hour ahead of sap flow, that of T_air and RH was 0-2 hours ahead of or behind the sap flow, and the time lags of T_soil and TDR were much longer or sometimes undetectable. Because of the complexity of the time lags, no evident improvements were observed in the linear correlations (R², slope, and intercept) when the time lags based on short-term (20 days) data were used to correct the time lags based on whole year data. However, obvious improvements were found in the standardized and non-standardized correlation coefficients in stepwise multiple regressions, i.e., the time lag corrections could improve the effects of RH, but decreased the effects of PAR, T_air, and T_soil. PCA could be used to simplify the complexity. The first and the second principal components could stand for over 75% information of all the environmental factors in different seasons and in whole year. The time lags of both the first and the second principal components were 1-1.5 hours in advance of the sap flow, except in winter (no time lag effect).

Key words: Larix gmelinii, time lag, principal components analysis, stepwise regression analysis, environmental factor