21. Fitting a distribution function (Gumbel)
1000
1500
2000
2500
3000
3500
1 10 100
return period (years)
discharge(m3/s)
data
function
Average Return Interval (ARI)
22. Extrapolation for larger return periods
1000
1500
2000
2500
3000
3500
4000
4500
5000
1 10 100 1000 10000
return period (years)
discharge(m3/s)
data
function
26. Joint probability methods
River system
Takes into account that there are multiple factors may contribute to flood
levels:
Rainfall depth
Spatio-temporal rainfall patterns
Initial soil conditions
Initial reservoir conditions
Ocean water levels
…
27. Example: tidal river
Sea
River discharge
local water level influenced
by river discharge and sea
water level
2.75 3
3
3.25
3.25
3.5
3.5
3.5
3.75
3.75
3.75
3.75
4
4
4
4
4.25
4.25
4.25
4.5
4.5
4.5
4.75
4.75
5
discharge [1000 m
3
/s]
seawaterlevel[m+NAP]
local water level [m+NAP]
1 2 3 4 5 6 7 8 9 10
1
1.5
2
2.5
3
3.5
4
28. example tidal river
Sea
River
2.75 3
3
3.25
3.25
3.5
3.5
3.5
3.75
3.75
3.75
3.75
4
4
4
4
4.25
4.25
4.25
4.5
4.5
4.5
4.75
4.75
5
discharge [1000 m
3
/s]
seawaterlevel[m+NAP]
local water level [m+NAP]
1 2 3 4 5 6 7 8 9 10
1
1.5
2
2.5
3
3.5
4
example: flooding occurs
if h>4.5 m+NAP
29. example tidal river
Sea
River
Flood events
1. identify all events for
which h>4.5 m+ref
2.determine the probability for
each of these events and take the
sum
30. Analogy: game of two dice
opponent has to throw 10 or more to win the game
6 7 8 9 10 11 12
5 6 7 8 9 10 11
4 5 6 7 8 9 10
3 4 5 6 7 8 9
2 3 4 5 6 7 8
1 2 3 4 5 6 7
1 2 3 4 5 6
31. Analogy: game of two dice
opponent has to throw 10 or more to win the game
6 7 8 9 10 11 12
5 6 7 8 9 10 11
4 5 6 7 8 9 10
3 4 5 6 7 8 9
2 3 4 5 6 7 8
1 2 3 4 5 6 7
1 2 3 4 5 6
Six undesired “events”
Each with probability 1/36
Total probability: 6/36 = 1/6
32. Flood risk complexer than dice
number of variables may be larger than 2
variables can be correlated
continuous values instead of discrete values
complex hydrodynamic processes
⇒ no analytical solution available
⇒ estimation techniques required
39. Brisbane River Catchment Flood Study – phase 2
Estimate probabilities of flood discharges and volumes at 23 locations for
‘no dam’ and ‘with dam’ conditions, using the following methods:
1.Flood frequency analysis
2.Design event approach
3.Joint probability (Monte Carlo) approach
40. Essence of Monte Carlo Simulation
1. Generate a large number of synthetic events
2. Simulate each event with a hydrological/hydraulic model
3. At each location, quantify how often flood levels are exceeded
Main challenges:
oSynthetic events have to be realistic and representative
oComputation times
41. Correlations
1. Extreme rainfall likely to co-incide with
increased ocean water levels
2. Reservoir volumes at the beginning of an
event are related (if one reservoir is full,
the other reservoirs are more likeley to be
full as well)
3. Initial soil conditions in the various
subcatchments are related
45. Spatio-temporal rainfall patterns (STEPS)
128-256-512 km 64-128-256 km 32-64-128 km
16-32-64 km 8-16-32 km 2-4-8 km4-8-16 km
Break up the rainfall map into a set of maps for different scales
Parameterize time evolution of each scale, large scales evolve slowly, small scales fast
Multiply maps to get the rainfall forecast.
From: Seed (2013)
50. Current status of BRCFS – phase 2
Finalising draft computation results of the three methods for ‘no
dam’ and ‘with dam’ conditions
Start of the reconciliation proces
Sept 11: workshop with client, IPE and various stakeholders