Abstract

This paper analyzes climate-vegetation correlations over 30 years in the northern Tianshan mountains in Xinjiang Province. The analysis combines the wavelet crosscorrelation technique with trend analysis and spline interpolation. The study areas include the Irtysh River, Bortala Valley, Ili River Valley, the northern slope of the Tianshan mountains and the western Junggar Basin. Daily meteorological data were acquired from 21 eligible weather stations between January 1, 1981 and December 31, 2009, inclusive. The satellite data comprise NOAA/AVHRR NDVI digital images (1981–2001) and SPOT-4 VEGETATION NDVI digital images (1998–2010). The results showed that (a) the correlation between early climate changes and subsequent vegetation had multiple-time-scale and time-lag characteristics. In all regions, the overall correlations were similar on different time scales; (b) The most useful time scale for evaluating precipitation effects was 15 tenday periods, while the optimal time lag was 5 ten-day periods. The exception was the western Junggar Basin (where the optimal time lag was 2 ten-day periods); (c) The most useful time scale for evaluating temperature effects was 15 ten-day periods, except in the western Junggar Basin (where the most useful time scale was 10 ten-day periods) and (d) Early changes in temperature were more correlated with subsequent vegetation than with early changes in precipitation.