| |
CRC PUBLICATIONS |
Joint Probability Approaches to Design Flood Estimation: A Review
Ataur Rahman,
Tam Hoang,
Erwin Weinmann,
Eric Laurenson
Publication Type:
Technical Report
This is a publication of the initial CRC for Catchment Hydrology
CRC Program:
Flood Hydrology (Previous CRC)
|
Publication Keywords:
Rainfall/Runoff Relationship
Floods and Flooding
Design Data
Modelling (Hydrological)
Distribution (Mathematical)
Frequency Analysis
Precipitation (Atmospheric)
|
Abstract / Summary:
Summary
Rainfall-based flood estimation techniques are common in hydrologic practice. The currently used methods are based on the design event approach; they use a probabilistic rainfall depth in combination with representative values of other inputs and then assume that the resulting flood has the same frequency as that of the input rainfall depth. In many cases, this assumption is unreasonable, and the arbitrary treatment of various inputs is likely to introduce significant bias in flood estimates for a given frequency. The report critically examines the limitations of the current design event approach and identifies the potential alternative methods that might lead to an improved rainfall-based design flood estimation technique.
The serious limitations of the current design event approach stem mainly from the simplifying assumption that with a representative set of inputs and model parameter values, the design flood output will preserve the annual exceedance probability of the design rainfall depth input. While there have been recent improvements in defining more representative design values for losses (Hill et al., 1996a, b) and there is some scope for developing more consistent sets of temporal patterns, even improved sets of single-valued design inputs will not be able to adequately allow for the complex interaction of rainfall and loss parameters with other catchment attributes (e.g. catchment size, shape, drainage characteristics).
The most promising alternatives to the design event approach include the continuous simulation approach (possibly also in its simplified form using runoff files) and the joint probability approach. These two approaches are similar in their (deterministic) modelling of the hydrograph formation process (runoff routing), but differ in the form of their basic inputs and how they use these to represent the runoff generation phase. This report focuses on the joint probability approach while the continuous simulation approach is being covered in a paralled report.
The promise of the joint probability approach stems from the fact that it can readily utilise the (deterministic) models and much of the design data used with the current design event approach, but will apply them within an appropriate probability framework. (Laurenson, 1974). This framework also exists and only needs to be adapted to this specific application. Joint probability methods therefore have the potential to lead to significant improvements in flood estimation, with relatively modest effort. The challenge lies in distilling the best elements of the existing methods and design data, and then combining them in a practically useful way to produce the required design tools.
The report presents a review of the previous studies in the area of joint probability approach to design flood estimation. It focuses particularly on the results of previous studies in relation to practical applicability of the methods. It has been found that most of the previous applications employing the joint probability approach were limited to theoretical studies; mathematical complexity, difficulties in parameter estimation and limited flexibility preclude these techniques to be applied under practical situations.
This review indicates that rainfall duration, rainfall temporal pattern and losses are key variables to be treated as random variables in addition to rainfall depth. An initial loss-continuing loss model combined with a semi-distributed non-linear runoff routing model (e.g. RORB, URBS) would be appropriate to use with the joint probability approach. From the consideration of practical applicability and ability to account for dependence between the flood producing variables, Monte Carlo simulation and the application of the Total Probability Theorem to discretized distributions appear to be the most promising methods for determining derived flood frequency distributions. This review concludes with a list of research tasks to develop a practical design tool for flood estimation using the joint probability approach.
This report is not currently available for downloading. Printed copies can be purchased from the Centre Office
 |
1 Download(s) |
|
|
|
|
|
|
|
|
Centre Office:
CRC for Catchment Hydrology
Dept of Civil Engineering
Building 60
Monash University Vic 3800
Tel: +61 3 9905 2704
Fax: +61 3 9905 5033
|
|
|
