The time series of precipitation in flood season (May-September) at WuhanStation, which is set as an example of the kind of time series with chaos characters, is split intotwo parts: One includes macro climatic timesc...The time series of precipitation in flood season (May-September) at WuhanStation, which is set as an example of the kind of time series with chaos characters, is split intotwo parts: One includes macro climatic timescale period waves that are affected by some relativelysteady climatic factors such as astronomical factors (sunspot, etc.), some other known and/orunknown factors, and the other includes micro climatic timescale period waves superimposed on themacro one. The evolutionary modeling (EM), which develops from genetic programming (GP), is supposedto be adept at simulating the former part because it creates the nonlinear ordinary differentialequation (NODE) based upon the data series. The natural fractals (NF) are used to simulate thelatter part. The final prediction is the sum of results from both methods, thus the model canreflect multi-time scale effects of forcing factors in the climate system. The results of thisexample for 2002 and 2003 are satisfactory for climatic prediction operation. The NODE can suggestthat the data vary with time, which is beneficial to think over short-range climatic analysis andprediction. Comparison in principle between evolutionary modeling and linear modeling indicates thatthe evolutionary one is a better way to simulate the complex time series with nonlinearcharacteristics.展开更多
基金Supported by the National Natural Science Foundation of China under Grant No. 42075034.
文摘The time series of precipitation in flood season (May-September) at WuhanStation, which is set as an example of the kind of time series with chaos characters, is split intotwo parts: One includes macro climatic timescale period waves that are affected by some relativelysteady climatic factors such as astronomical factors (sunspot, etc.), some other known and/orunknown factors, and the other includes micro climatic timescale period waves superimposed on themacro one. The evolutionary modeling (EM), which develops from genetic programming (GP), is supposedto be adept at simulating the former part because it creates the nonlinear ordinary differentialequation (NODE) based upon the data series. The natural fractals (NF) are used to simulate thelatter part. The final prediction is the sum of results from both methods, thus the model canreflect multi-time scale effects of forcing factors in the climate system. The results of thisexample for 2002 and 2003 are satisfactory for climatic prediction operation. The NODE can suggestthat the data vary with time, which is beneficial to think over short-range climatic analysis andprediction. Comparison in principle between evolutionary modeling and linear modeling indicates thatthe evolutionary one is a better way to simulate the complex time series with nonlinearcharacteristics.
