Flood analysis (cont)

Francois Ilunga PhD (Wits)

1
3.7 Statistical methods
 Basis of statistical methods is historical records
(e.g. observed and or infilled/patched data series)
 30 years or more of records can be used for good
estimates of statistical parameters in design of
hydraulic structures.
 However short records more prevalent in
developing countries than in developed countries
 Return periods T determined in design of water
structures, e.g. culverts could be 25 yrs, bridge
50 yrs; etc.
 Hence associated Qp expected not to be
exceeded more than once in N years
3.7.1 Probable maximum flood (PMF)
 PMF: largest flood expected to occur in a
particular region. T may be assumed as large as 2

possible (infinity). contrary to probability theory

3.7.2 Probability of the n-year event
 1 : 50 year flood either occurring or being
surpassed in any year with a return period of
50 years: P(Q > or = Q50) = 2 %
1
 Generally P = = P(Q > or = Qt)
T
 Hence risk of exceedence
 Probability of exceeding such an event in n
year for a return period T
P(X> or = x)n = 1-(1-1/T)n

Tut (Activity/question 5)
 Determine the probability of a flood equal or
bigger than the 1 : 50 year occurring at least
once in a particular 3 year period
3
3.7.3 Calculating the value of the n-year
event
Depends on probability distributions fitted to
hydrological records (time series)
 Normal distribution (!!!) (symmetrical)
 Log normal distribution (symmetrical)
 Extreme value (EVI, II & III)-Gumbel distribution
 e.g Log normal, EV (skewed distributions)
shown to be common in South Africa
 For normal distrib. QT = Qav + SZ (1), use
 Qav: average of data; s: standard deviation;
 Z: standardised/reduced value(Table 3A.1b&3A1a)
 For log-normal: similar to (1), but use log data
 (Table 3A.3)
 EV1: QT = Qav + S (0.78W T-0.450) (2), 4

where g =1.14 (skewness coefficient), table 3A.4

 3.7.4 Value of n-year using plotting
position
 Plotting position: rank hydrological data, e.g.
Maximum annual series, descending/ascending
 General formula of probability of exceedence:
 T = (An+b)/(cm+d) (2)
 n: number of data points; m: rank order (high-
low); a, b, c, d : constant according to authors
 Commonly use of Weibull Formula T n 1 (3)
r
m
Tut (Activity/ question 6)
 At a gauge of the Vaal dam, annual maximum
flood were recorded between 1936 and 1989.
Determine the probability of exceeding the 5 th
highest annual flood value?
5
 3.8 Flood routing
Estimate stage and discharge at various points
along watercourse for floods, Q = f (H) and vice-
versa. Hydrological routing methods determine
outflow hydrograph from inflow hydrograph
translated into lag vs attenuation
3.8.1. Channel routing
 Storage volume not a simple a function of stage.
I Wedge Storage

Y1

Prism Storage

Y0
6
 For simplified case, linear storage S = KQ (ideal
storage function of discharge)
 Real case for a river reach/channel-> storage
volume not a simple a function of stage, but a non
linear relationship due to storage wedge
 Hence S function of I inflow(I) and outflow (Q) for
channel routing
 S = K[ xI + (1 - x ) D ] (4a)
 S: storage of river reach; I; inflow, Q: outflow; x:
weighting constant factor-geometry of channel; K:
related to travel time of flood wave (storage coeff.)
 Combining continuity Eq.(4b) & (4a), one gets Eq(5)
 ½ (I1 + I2 ) t - ½ (D1 + D2) t = S2 -S 1 (4b)
 D2 = C0 I2 + C1 I1 + C2 D1 (5)
 (Muskingum method for channel/river routing)
 t: routing time, usually as fraction of travel wave 7
time
 Hence coefficients Cs determined as follows:

C0
Kx 0,5t (6a)
K Kx 0,5t

Kx 0,5t
 C1 (6b)
K Kx 0,5t


K Kx 0,5t (6c)
C2
K Kx 0,5t

 C1 + C2 + C3 =1
 Tut (Activity/question 7 (see
Addendum 4.1) 8
3.8.2. Reservoir routing
 Situational case

Inflow [I] Live storage [S]

Non overspill crest

Overspill crest

Stage h

Outflow [D]

9
 Flow over spillway determines reservoir routing
 Flow volume and reservoir storage=f(stage);
and for a weir V = a y b [m3]; y = stage above
overspill, a, b = reservoir constant
 Reservoir routing performed from continuity
equation ( Puls’ method):
 ½ (I1 + I2) t + (S1 - ½ D1 t) = (S2 + ½ D2 t)
 Or
 ½ (I1 + I2) t - ½ (D1 + D2) t = S2 - S1
 I1 , I2, ………,In known from inflow hydrograph
 D1 :initial discharge know across spillway
 Si..n and Di..n unknown at the start

10
 Tut (Activity/Q) 8 (see addendum 4.1)
Highlights
 The surface-elevation relationship known (a)
and (b) needs to be established.
 Flood routing focuses on live storage

 Storage between any two levels is the mean

reservoir surface area between these levels;
e.g. between 54 and 56: 2(4.16 +4.2)/2 = 8,36
106 m3 , hence b/n 56 and 58, 16,84 106 m3
etc up to 72, live storage-water level
relationship is derived

11
 Calculate discharge associated with each
incremental water stage above overspill crest,
using sluice gate orifice (t = 0, water level is
63,5 m, assuming, Cd = 0,8
 Q ORIFICE C d A 2 g H H = 63,5 - 54 = 9,5 m,
 Q=125 m3/s
 For spillway, Qs = Cs L H2/3 ,Cs = 2,2 and L
= 72,5 m; t = 0, water level = 63,5 m, hence
2,5 m below the overspill crest elevation at
66,0.
 Hence total discharge at time t = 0 is 125m3/s.
 Elevation - discharge curve for the reservoir is
now computed starting at 52,65 m (datum)
 Hence Table 11. can be computed 12
 The storage - elevation relationship and the
elevation - discharge relationship are now
combined to develop the routing storage
curve via elevation (Storage-discharge)
 routing period = 6h; storage unit= routing
period 1 m3/s=1 6 3 600 [m3/s h
s/h ] = 21,6 103 m3
 S ± ½ Dt curves are derived
 Curve discharge D versus S can be
constructed
 Routing calculation

13
 3.9 Hydrograph
 Generated from rainfall event(s)

Rainfall
(mm)

x LAG

Peak flow

Total Runoff

(m3/s) Rising limb

Surface

runoff Recession curve

Baseflow

Time (h)
14
 Baseflow and rain storm may contribute
differently to runoff
 Hence these 2 can be separated (surface
runoff, infiltration, etc)
3.9.1 Unit hydrograph (UH)
 UH enables to correlate rainfall and surface
runoff (hence net rainfall and surface runoff)
 Uniform intensity net rainfall on a particular
catchment and for rainfall storms of different
intensities but equal duration Ratio of
rainfall intensities and ratios of runoff are
identical
 Principle of superposition: Hydrograph
results from several rainfall intensities
15
3.9.1.1. Conversion of a short duration unit
hydrograph to a longer duration
hydrograph

1 : 2 hour @ 5 mm/h = 10 mm rainfall

5

i (mm/h) 1 2

0
2 4 Time (h)

Addition of the two 2 hour hydrographs

2 hour unit hydrographs

Q (m3/s)

4 hour unit hydrograph

16
Time (h)
 Example: Two 2 h UH of 0.5 mm/h, with unit
of rainfall = 1mm may be added to obtain 4 h
UH of same rainfall intensity. Hence 2 mm (=
4x0.5) of 4 h UH. According to principle of
superposition.
 Similarly two 2h UH of 5 mm/h, with 1mm as
unit of rainfall, yield 4 mm of 4 h UH.
 Time base of 4 hour unit hydrograph is
longer than that of 2 hour unit hydrograph.
 Hence rainfall occurred at lower intensity for
a longer time
 Ordinates of the hydrograph 2h UH divided
by 2 to obtain 4 hour unit hydrograph

17
 3.9.1.2. Conversion of a long duration unit
hydrograph to a shorter duration unit
hydrograph
 Use of S-curve (continuous uniform-
intensity storms producing the unit of rain
in t1 h on the catchment), hence intensity
to be factored by 1/ t1
i (mm/h)

1/t1
t1 t1 t1 t1

t1 h s-curves; t2 h apart

t2 h unitgraph

Q (m3/s)

S-Curve Difference

18
t2 Time (h)
 If a second S-curve is drawn one unit
period to the right of the first, difference b/n
the two S-curves = runoff of one t hour unit
hydrograph.
 Check at steady state: Qsts = 2.78A/t1

A (km2); t (h)
 A UH of a shorter storm of t2 hours, can be
can obtained from second S-curve.
Intensity will be factored with 1/t2.
 Ordinates of the difference S-Curves graph
are multiplied by the inverse ratio of the
intensities, rain intensity becomes t1/t2
mm/h and total rain is again the unit of
rainfall. 19

Tut (Activity/Q) 9 (see Addendum 4.1)

 Extra readings
 Read only related topics covered (in notes)
from the following
 Drainage manual, 2013: Chap 3: p.3.12-3.16
 Hydrology & hydroclimatic principles and
applications, 2013, 2009: Chap 13 (p.658-670),
Chap 10, 10.3
 Introduction to flood hydrology: Chap 8, 9 &10
 Engineering Hydrology, 1983: Chap 9, Chap 8;
Chap 7
 Notes
 Drainage manual, 2013-APPLICATIONS
GUIDE: workable examples, p.3.5-3.47.(relate
these only to topic covered, see sessions 1-4)
20