A seasonal stochastic simulation method of daily runoff based on conditional dimensionality reduction and reconstruction

Abstract

The invention discloses a daily runoff seasonal random simulation method based on condition dimension reduction reconstruction, and belongs to the field of random hydrology. The daily runoff seasonalrandom simulation method includes the steps: constructing a three-dimensional Copula function by adopting a conditional dimensionality reduction reconstruction theory; according to the two-dimensionaljoint distribution of the daily flow at the time t-1 and the time t-2 and the two-dimensional joint distribution of the daily flow at the time t and the time t-2, establishing three-dimensional jointdistribution of daily flow at the moments t, t-1 and t-2, and converting the three-dimensional Copula function into the condition distribution and the two-dimensional Copula function, so that the parameter solving difficulty caused by directly establishing the high-dimensional Copula function is avoided, and the difficulty of constructing the high-dimensional Copula is reduced; and the parametersof the high-dimensional Copula are easier to estimate, and the calculation becomes easier. According to the daily runoff seasonal random simulation method, the second-order lag time correlation relation is considered in daily runoff random simulation, and the problem that an existing daily runoff random simulation method based on the Copula function cannot consider second-order lag time is solved, and statistical characteristics such as daily runoff sequence mean value, variance and Cs can be well simulated, and the simulation effect is better.

Description

technical field

[0001] The invention belongs to the field of stochastic hydrology, and more specifically relates to a seasonal random simulation method of daily runoff based on conditional dimensionality reduction and reconstruction. Background technique

[0002] As the input of water resource system, runoff sequence plays a vital role in system simulation, and is often used to formulate reservoir operation strategy, assess the risk of water resource system, determine various hydraulic parameters, etc. Since the measured hydrological series are often only several decades or even more than ten years, the series is relatively short, and it is difficult to meet the needs of water resources system analysis and risk assessment. Therefore, runoff stochastic simulation methods are often used to generate long series of hydrological data.

[0003] The commonly used stochastic simulation method is mainly the first-order autoregressive method (AR(1)). This model has a simple structure...

Images