水文多变量相空间重构自记忆模型研究

Study on self-memory hydrologic forecasting model based on multi-dimensional phase-space reconstruction theory

【目的】针对单变量相空间重构自记忆模型存在的不足,研究多变量相空间重构自记忆模型在水文预报中的适用性。【方法】根据多变量相空间重构理论构造多维相空间,并在此基础上,结合自忆性原理,建立多变量相空间重构自记忆模型,最后利用新疆和田绿洲的实测月蒸发能力资料进行验证。【结果】和田绿洲月蒸发能力实测资料检验结果表明,建立多变量相空间重构模型是可行的,可以取得理想的效果。【结论】多变量相空间重构自记忆模型的建立使得相空间重构自记忆模型从一维拓展到了多维,也使得该模型更加符合生产实际。

[Objective] As phase-space self-memory reconstruction model by the single variable have deficiencies, the multi-dimensional phase-space was studied. [Method] A multi-dimensional phase-space was reconstructed according to phase space reconstruction theory with multi factors. Based on this,a selfmemory hydrologic forecasting model was built and verified by monthly evaporation conducted in Hotan oasis,Xinjiang. [Result] Through monthly evaporation capacity forecast in Hotan oasis,the case study indicates that the model is reasonable and ideal. [Conclusion] Multivariable phase space reconstruction selfmemory modeling changes one dimension phase space reconstruction into multi-dimension, which is much more in line with practical production.