Abrupt climate change has an important impact on sustainable economic and social development, as well as ecosystem. However, it is very difficult to predict abrupt climate changes because the climate system is a complex and nonlinear system. In the present paper, the nonlinear local Lyapunov exponent (NLLE) is proposed as a new early warning signal for an abrupt climate change. The performance of NLLE as an early warning signal is first verified by those simulated abrupt changes based on four folding models. That is, NLLE in all experiments showed an almost monotonous increasing trend as a dynamic system approached its tipping point. For a well-studied abrupt climate change in North Pacific in 1976/1977, it is also found that NLLE shows an almost monotonous increasing trend since 1970 which give up to 6 years warning before the abrupt climate change. The limit of the predictability for a nonlinear dynamic system can be quantitatively estimated by NLLE, and lager NLLE of the system means less predictability. Therefore, the decreasing predictability may be an effective precursor indicator for abrupt climate change.
A study was carried out to investigate the grazing pressure of heterotrophic nanoflagellates(HNF) on bacteria assemblages in the Yellow Sea Cold Water Mass(YSCWM) area in October, 2006. The results show that the HNF abundance ranges from 303 to 1 388 mL-1, with a mean of 884 mL-1. The HNF biomass is equivalent to 10.6%–115.6% of that of the bacteria. The maximum abundance of the HNF generally occurred in the upper 30 m water layer, with a vertical distribution pattern of surface layer abundance greater than middle layer abundance, then bottom layer abundance. The hydrological data show that the YSCWM is located in the northeastern part of the study area, typically 40 m beneath the surface. A weak correlation is found between the abundances of HNF and bacteria in both the YSCWM and its above water layer. One-way ANOVA analysis reveals that the abundance of HNF and bacteria differs between inside the YSCWM and in the above water mass. The ingestion rates of the HNF on bacteria was 8.02±3.43 h-1 in average. The grazing rate only represented 22.75%±6.91% of bacterial biomass or 6.55%+4.24% of bacterial production, implying that the HNF grazing was not the major factor contributing to the bacterial loss in the YSCWM areas. 相似文献