Little is known about the spatial variability in tree growth and its responses to climate on the Tibetan Plateau; however, such information is essential for improving predictions of forest ecosystem response to climatic change. A network of 16 ring width chronologies was developed along a latitudinal transect in the Qilian Mountains, northeastern Tibetan Plateau. A principal components analysis revealed that the residual chronologies had a positive loading on the first unrotated principal component (PC1). After rotation, PC1 yielded the highest loadings on the driest sites in the northwest and decreased to the south and to the east. PC2 was negatively correlated with altitude. Moisture availability was a dominant limiting factor for tree growth, and this dominance increased northwards and westwards along the precipitation gradient. Loadings of the first two rotated principal components separated the 16 forest sites into three major groups corresponding to the three regions affected by the East Asian Monsoon, Westerlies, and their interaction. Thus, spatial variability in tree growth is an excellent bioindicator of regional climate.