We present a tree-ring based reconstruction of water-year (October-September) streamflow for the Manasi River in the northern Tien Shan mountains in northwestern China. We developed eight Tien Shan spruce (Picea schrenkiana Fisch. et Mey.) chronologies for this purpose, which showed a common climatic signal. The hydroclimatic forcing driving tree growth variability affected streamflow with a three-to four-year lag. The model used to estimate streamflow is based on the average of three chronologies and reflects the autoregressive structure of the streamflow time series. The model explains 51% of variance in the instrumental data and allowed us to reconstruct streamflow for the period 1629-2000. This preliminary reconstruction could serve as a basis for providing a longer context for evaluating the recent (1995)(1996)(1997)(1998)(1999)(2000) increasing trends in Manasi River streamflow and enables the detection of sustained periods of drought and flood, which are particularly challenging for managing water systems. Several of the reconstructed extended dry (wet) periods of the Manasi River correspond to reconstructed periods of drought (flood) in Central Asia in general and in other Tien Shan mountain locations in particular, suggesting that the analysis of Tien Shan spruce could contribute significantly to the development of regionally explicit streamflow reconstructions.

Understanding plant phenological change is of great concern in the context of global climate change. Phenological models can aid in understanding and predicting growing season changes and can be parameterized with gross primary production (GPP) estimated using the eddy covariance (EC) technique. This study used nine years of EC-derived GPP data from three mature subtropical longleaf pine forests in the southeastern United States with differing soil water holding capacity in combination with site-specific micrometeorological data to parameterize a photosynthesis-based phenological model. We evaluated how weather conditions and prescribed fire led to variation in the ecosystem phenological processes. The results suggest that soil water availability had an effect on phenology, and greater soil water availability was associated with a longer growing season (LOS). We also observed that prescribed fire, a common forest management activity in the region, had a limited impact on phenological processes. Dormant season fire had no significant effect on phenological processes by site, but we observed differences in the start of the growing season (SOS) between fire and non-fire years. Fire delayed SOS by 10 d ± 5 d (SE), and this effect was greater with higher soil water availability, extending SOS by 18 d on average. Fire was also associated with increased sensitivity of spring phenology to radiation and air temperature. We found that interannual climate change and periodic weather anomalies (flood, short-term drought, and long-term drought), controlled annual ecosystem phenological processes more than prescribed fire. When water availability increased following short-term summer drought, the growing season was extended. With future climate change, subtropical areas of the Southeastern US are expected to experience more frequent short-term droughts, which could shorten the region’s growing season and lead to a reduction in the longleaf pine ecosystem’s carbon sequestration capacity.

No abstract