May long-term historical hydrological data be misleading for flood frequency analysis in current conditions of climate change?
Alexandra Fedorova1, Nataliia Nesterova1,2,
Olga Makarieva1,3, and Andrey Shikhov4
1Saint-Petersburg University, St. Petersburg, Russia
2State Hydrological Institute, St. Petersburg, Russia
3Melnikov Permafrost Institute SB RAS, Yakutsk, Russia
4Perm State University, Perm, Russia
EGU-2020
1
❑ 25 people died
❑ 8 people are missing
❑ 3.7 thousand homes flooded
❑ 15 bridges destroyed
❑ 70 tons of crop washed away
❑ Economic damage from the
flood in 2019 amounted up to
half a billion Euro
This flood became the
most hazardous one in
the region in 80 years
history of observations.
2
What has happened? – Historical flood
3
Where has it happened? – the Iya River
▪ The South-Eastern part of Siberia, Russia;
▪ The northern slopes of the Eastern Sayan;
▪ The Iya River basin (14500 km2);▪ Maximum height (2789 m);▪ The climate is sharply continental
What did cause the flood?
https://www.irk.ru/news/20190702/dam/
The aim of the study was to analyze the factors that led to the
formation of a catastrophic flood in June 2019, as well as
estimate the maximum discharge at the Iya River.
▪ heavy rains as a result of
climate change?
▪ melting of snow and glaciers
in the mountains of the East
Sayan?
▪ deforestation of river basins
due to clearings and fires?
4
What did cause the flood? – heavy rains
5
▪Melting of snow and glaciers in
the mountains: less than 10% of
the area was covered with snow.
This could not cause flooding of
such magnitude
▪Deforestation: the area of losing
forest in the basin consists of no
more than 4% of the total
catchment area.
▪From June 25 to 27, from 170 to
250 mm of precipitation fell. The
main cause of the flood was a
heavy rain.
Hydrograph model
➢ Parameters: measured properties of
soils and vegetation cover
✓ Applicable to catchments of all sizes
✓ Applicable on basins in the permafrost
zone
➢ Input: temperature, humidity,
precipitation
➢ Output: hydrographs in the last
discharge section line, water balance
characteristics, soil and snow conditions
Distributed deterministic model of hydrological processes
6
Developed by prof. Yu.B. Vinogradov
(SHI, Saint Petersburg)
Precipitation
Rain Snow
Interception Snow cover
formation
Heat energy
Heat dynamics in snow Heat dynamics
in soil
Infiltration and surface flow
Initial
surface
losses
Water dynamics in soil
Slope
transformation
of surface flow
Channel transformation
Runoff at basin outlet
Underground flow
Transformation of
underground flow
Snow melt and water vield
Evaporation
Model verification for 3 basins
7
3 basins:a – the Kirej river, Ujgatb – the Iya river, Arshanc – the Iya river, Tulun.
River a b c
Period 1959-20171963-2017
1941-2017
S (km2) 2950 5140 14500
H (m) 873 1483 979
Flow.obs. 374 540 326
Flow.sim. 402 528 338
Precip. 688 771 586
Evap. 286 243 247
NS (m/av) 0,66/0,57 0,69/0,62 0,72/0,67
For annual maximum water discharge;
Period: 1970-1996;
The difference in values does not exceed 300 m3s-1 (8%).
Model verification
8
Data for modelling catastrophic flood
The assessment of
the maximum water
discharge in June
2019 based on two
types of input:
▪ Observed
weather stations’
data (Arshan,
Ikey, Tulun);
▪ ICON climate
model data
9
Results
1, 2 – the amount of
precipitation for the catchment -
3-hour precipitation according to
the ICON weather model and
daily precipitation based on data
from weather stations;
3 – the observed flow
hydrograph (based on
extrapolation of the dependence
of water flow on the level);
4, 5 – calculated 3-hour and
averaged daily flow hydrograph
according to the ICON weather
model;
6 – calculated daily runoff
hydrograph based on data from
weather stations.
The results of flood modeling at the Iya River – Tulun in June 2019:
10
Results of modeling
▪ Qmax. based on ICON: 4780 m3s-1 (daily) and 5260 m3s-1 (3-hour)
▪ Qmax. weather station data: 6570 m3s-1 (daily)
▪ The maximum discharge based on ICON data is 1400 m3s-1 lower than the
observed, however, its formation coincides in the term. According to
weather station data, the maximum discharge coincides in dimension, but
its formation is delayed by 1 day;
▪ We attempt to show the need to expand the meteorological and hydrological
network. We also demonstrate the capabilities of the modern calculation
methods and forecasts in case of insufficient observed data;
▪ We showed that the ensemble of input meteorological data from various
sources could potentially be used to satisfactorily predict the magnitude and
duration of the catastrophic flood in order to minimize the consequences;
11
Has this flood been observed before?
12
Year Water level, m Discharge, м3/с
1937 8.5 1907
1980 9.0 2520
1984 11.0 4400
2019 13.8 6800 (preliminary assessment)
https://pikabu.ru/story/masshtab_navodneniya_v_gorode_tulun_irkutskaya_oblast_6789293
The level of protective dam is 12 m.Why?
The flood in the Tulun town.
Why was the maximum level of the dam 12 m?
13
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
400 600 800 1000 1200 1400
Max
imu
m w
ate
r d
isch
arge
, m
3/s
Water level, sm
1984
historical maximum
Dam construction:
2006-2009
Q(H) for 1980-2006
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71M
axim
um
wat
er
dis
char
ge, m
3/s
Qmax for 1936-2006
▪ This series of discharge are
not homogeneous;
▪ Probability of the flood
(1984) was underestimated
as historical maximum;
1984
0
1000
2000
3000
4000
5000
6000
7000
8000
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82M
axim
um
wat
er
dis
char
ge, m
3/s
0
2000
4000
6000
8000
10000
12000
400 600 800 1000 1200 1400
Max
imu
m w
ate
r d
isch
arge
, m
3/s
Water level, sm
What will be the new max level of the dam?
14
Q(H) for 1980-2019
Qmax for 1936-2019
▪ This series of discharge
also are not homogeneous;
▪ Will the probability of the
flood (2019) be
underestimated?
1984
historical maximum
2019historical maximum?
1984
2019
Conclusion
15
▪ The estimated discharge has exceeded previously observed one by about
50%.
▪ The results of the study have shown that recent flood damage was
caused mainly by unprepared infrastructure.
▪ The safety dam which was built in the town of Tulun just ten years ago
was 2 meters lower than maximum observed water level in 2019.
▪ This case and many other cases in Russia suggest that the flood
frequency analysis of even long-term historical data may mislead design
engineers to significantly underestimate the probability and magnitude of
flash floods.
Thank you for attention!
16