报告人：Alistair G.L. Borthwick  (The University of Edinburgh)

时  间：2019-05-27  15:00- 16:00

地  点：环境大楼B112

Abstract: River processes are notoriously difficult to predict.  Physical parameters including morphology, texture, and flow discharge, are at best approximate.  Model parameters may be inexact and model assumptions place a limit on the physics.  Uncertainty relates to the state of knowledge, and is best expressed using probability functions.  Environmental engineers concerned with river processes need to evaluate how uncertainty in physical and model parameters translates to key outputs such as water levels, flow speed, flood extent, and water quality. This seminar will examine a technique based on the conservation of probability that can give efficient and robust estimates of expected values, variance and higher order statistical moments of model outputs, without recourse to Monte Carlo simulation.   An example of maximum water level estimation for tidal rivers in Bangladesh will illustrate the technique.