Professional Documents
Culture Documents
Hydrographs
10-2
Storm Duration and Depth
Effect of varying storm duration assuming all other variables constant
Intensity
Shorter duration storm
Discharge Q
Time
Intensity
Longer duration storm
Time
Time
Intensity
Greater depth
Discharge Q
Time
Intensity
Smaller depth
Time
10-3
Antecedent Wetness and Temporal Patterns
Effect of changing antecedent wetness assuming all other variables constant
conditions)
Time
Effect of changing storm temporal pattern assuming all other variables constant
Intensity
Peaked pattern
Discharge Q
Time
Intensity
Uniform pattern
Time
10-4
Design Flood Hydrographs
There are 3 general approaches to deriving hydrographs whose
peak flow has an ARI of Y years:
Continuous Simulation
Randomly sample a long sequence of rainfall and convert to a long
sequence of streamflow.
This is considered best approach but capability to simulate high-
resolution long rainfall sequences is just being developed
In future this will be the main method
Empirical
eg Probabilistic rational method is based on frequency analysis of
rainfall and flood data to estimate Y-year peak flow
10-5
Average Variability Method
Average variability method is the main method used in Australian
design practice. This is because rainfall IFD has been regionalised
across Australia and there is limited streamflow gauging.
10-6
Design Storms
Intensity
mm/hr
(a) ARR IFD analysis yields gross
characteristics of design storms: average
i
Time
intensity and storm duration
ts td
Intensity
mm/hr (b)
ARR temporal patterns disaggregate
gross storm into a design storm
Time
ts
Walsh (1991) derived design initial loss rates for eastern NSW
using the following approach:
• Select a gauged catchment
• Estimate Y-year flood using flood frequency analysis
• Calibrate a rainfall-runoff model
• Using Y-year design storm find the design initial loss which yields
Y-year flood peak obtained from the frequency analysis
• Repeat for a number of catchments and regionalize results
Walsh, M. initial loss for design flood estimation in NSW, International Hydrology and Water
resources Symposium, Institution of Engineers, Aust., Perth, 1991
10-9
Initial Loss Results for Eastern NSW
Walsh, M. initial loss for design flood estimation in NSW, International Hydrology and Water
resources Symposium, Institution of Engineers, Aust., Perth, 1991
10-10
Design Loss Parameters for Eastern NSW
Walsh, M. initial loss for design flood estimation in NSW, International Hydrology and Water
resources Symposium, Institution of Engineers, Aust., Perth, 1991
10-11
Reliability of Average Variability Method
The average variability method is approximate in that peak flow does not
necessarily have the same ARI as the design rainfall.
10-13
Default kinDog Parameters
Process Parameter Default value Comment
Channel Channel 6 m0.2s-1 Cr affects the speed of flood wave. Increasing Cs causes
conveyance Cr floodwave to travel faster
Hillslope Overland flow Cs 0.7 s1/2 This is the most sensitive parameter in kinDog. Increasing Cs
reduces and delays peak flow
Infiltration Initial loss IL Highly variable Need to assess antecedent conditions. A value close to 0 mm
means catchment will produce overland almost immediately
after start of rainfall
Continuing loss φ 2.5 mm/hr IL and φ control the storm water balance.
10-14
kinDog Calibration to Observed Events
Calibration involves adjusting kinDog parameters until observed and simulated
hydrographs have best match
10-16
kinDog Calibration to Observed Events
This is an unusually
good fit! Ceff =1 for
a perfect fit.
Variance of errors ∑t (q t − q̂ t )
2
It is defined as 1 − =1−
Variance of observed flow ∑ (q t − q )2
t
where qt is the observed flow at time t, q̂t is the simulated flow at time t
and q is the arithmetic mean of the observed flows. 10-17
kinDog Design Hydrograph
Select ARR design
storm
10-18
kinDog Design Hydrograph
Select design
storm duration
Enter design
initial loss
10-19
kinDog Design Hydrograph