【摘要】 自从 May ( 1 974 )指出即使是简单的种群模型也能揭示混沌动态以来 ,自然种群是否存在混沌一直具有争论 ,如何检测自然种群的混沌行为也成为种群动态研究的一个难点。通过时间序列分析和反应面模型建摸的方法分析了马尾松毛虫的复杂性动态。用自相关函数对马尾松毛虫发生的时间动态分析的结果认为动态是平稳的 ,其周期性不显著 ,而具有一定的复杂性 ,这种类型可以是减幅波动、有限周期或弱混沌 ,波动主要由系统内因引起。进一步采用反应面模型估计全局李雅普若夫指数和局域李雅普若夫指数结果均为负 ,显示马尾松毛虫种群动态不存在混沌现象 ,但是在增加一个小的噪音以后 ,局域李雅普若夫指数变为在 0以上波动 ,说明系统对噪音非常敏感 ,噪音对松毛虫种群动态具有很大的影响 ,可以将其从非混沌状态变为混沌。研究结果认为全局李雅普若夫指数λ是一定时间内两个变动轨迹的总平均偏差 ,而随着种群动态的波动 ,指数也是波动的 ,所以对于检测自然种群的混沌来说不是一个好的指标。局域李雅普若夫指数λM能更好地表示自然种群混沌的存在和产生混沌的条件。对害虫管理来说对种群暴发初期的预测是尤其重要的 ,而此时又最难于预测 ,所以对种群动态的监测就尤为重要。由于马尾松毛虫的代间种群动态为第一级密度相更多还原

【Abstract】 Since May (1974) pointed out that even simple population models had the potential to display chaotic dynamics, ecologists have been debating on existing of chaotic behavior in natural populations. To detect the chaos in natural populations is still difficult. In this paper time series analyses and response surface method were used to analyze the complex dynamics of mason pine caterpillar(MPC), Dendrolimus punctatus Walker. Autocorrelation function (ACF) analyses of the population dynamics of MPC demonstrate the pattern of the oscillatory decay to zero, which reveal the population is stationary, and fluctuations with an endogenous periodic component. The deterministic dynamics may be damped oscillations, a limited cycle, or “weak” chaos. Endogenous component plays more important role than exogenous component It has no significant period. MPC population dynamic pattern may be different at different sites or time and spatial scales. Though both Global Lyapunov Exponents and Local Lyapunov Exponents are negative, adding a small noise the local Lyapunov Exponents fluctuated above zero. The results reveal that noise can play an important role in the population dynamics and can move the system into chaos. The Global Lyapunov Exponents λ is based on average divergence rate which may not be a good indicator for dynamics of real biological system because Lyapunov Exponents will fluctuate with population. So estimation of Local Lyapunov Exponent λ_M is a more appropriate way to detect the chaos. For the management of MPC the most important prediction is at the beginning of the outbreak, so continuous monitoring is needed. Because of the first order density dependent, the prediction for the next generation by the status of present generation should be most reliable.更多还原