科研之友
微信
新浪微博
Facebook
分享链接
摘要
Using cellular automata (CA) model, in this article, a mechanistic approach has been conducted to help to insight into spread process of plant population such as annual weed population. A controllable CA model with 25 neighbourhood cells has been built, in which seed dispersal, as a key process of annual weed population dynamics, has been described by Gaussian distribution. The hypothesis that initial configurations affect annual weed population dynamics and control strategies has been tested by simulating. And the results support strongly the hypothesis. Patch, pattern and configuration are important concepts in weed patch management, which have been mathematically defined in this article. The aggregation effects have been found out while simulating, which is subject to patch size, patch density and distribution function of seed dispersal. Perhaps it could partly explain the reason why weeds often distribute in patch. Accordingly, true-patch and generalised patch have been distinguished from each other, which would be practically applicable to weed patch management. Based on this approach to the particular weed-crop system, a new hypothesis has been proposed: in a real controllable weed-crop system, the mean law might hold. If so, algorithm of patch control for long-term weed management could be simplified to control a true-patch only according to its mean density, it is unnecessary to control it according to the density in its each cell.
