u. bethers, j.seņņikovs -...

A flood risk assessment and mapping for Riga city

U. Bethers, J.SeņņikovsLaboratory for Mathematical Modelling of Environmental and Technological Processes, Faculty of Physics and Mathematics, University of Latvia

Riga (population ca. 800,000) is the capital of Latvia. The city is located at the

southern coast of Gulf of Riga along the lower stretch of the River Daugava.

The water bodies adjacent or within the territory of the city are the River

Daugava itself, its connection (Bullupe) to another major river – Lielupe,

several minor tributaries of the River Daugava, as well as an interconnected

lake system of estuarine origin. The flow in the lower stretch of the River

Daugava is regulated by Riga hydropower plant which is situated approx. 25

kilometers upstream the river mouth. The lowest 12 kilometers of the river

channel is dredged to ensure the operation of the Riga Freeport.

The aims of the study were (1) the identification of the flood risk situations, (2) the

quantification of the flooding scenarious of different return periods, (3) the building

and calibration of the hydrodynamical mathematical model for the domain potentially

vulnerable for flooding, (4) the calculation of flood events with different return period,

and (5) the detailed (horizontal accuracy around 10 m) mappping of the potentially

flooded areas.

1. Scenarious

The combination of storm surges in the southern part of Gulf of Riga with

unfavorable regime of hydropower plant operation was found as the most

dangerous flooding situation. The time series of water level at the mouth of the

River Daugava was analysed for more than 130 year long time period. The

significant trend was found in the annual peak water level. Five significant

storm events were found in time period 2001-2007 which roughly correspond to

storm surges with return period once in 5, 10, 20, 50 and 100 years. The model

storm events were created by scaling waterlevel and meteorological conditions

during these selected events, and superposing them with hydropowerplant

operation regime.

2. Model

The finite-element based

shallow water model was

built for the area,

potentially vulnerable for

flooding. Heterogeneous

depth/terrain information

from various sources was

integrated in the model.

The linear objects

(watercourses, dams,

etc.) of hydraulic

importance were included

in the model. The typical

spatial resolution of

approx. 50-100 m was

reached with total number

of finite elements around

250,000.

501000 501500 502000 502500 503000 503500

322600

322800

323000

323200

323400

323600

323800

324000

324200

324400

324600

324800

325000

-22.48

-18.77

-15.06

-11.35

-7.636

-3.926

-0.216

3.494

7.204

10.91

14.62

18.33

3. Calibration

The hydrodynamical model was calibrated on the basis of water level

observations in 5 different locations during 5 selected real storm events. Model

forcing was hourly observations of 2 sea levels (Lielupe, Daugavgrīva),

Daugava (at HPP) and Jugla river discharges, and wind speed. Calibration

results are illlustrated for the storm of Jan-2005 comparing hourly observed and

calculated water levels

0,5

1

1,5

2

2,5

3

20

05

.01

.08

9:0

0

20

05

.01

.08

11

:00

20

05

.01

.08

13

:00

20

05

.01

.08

15

:00

20

05

.01

.08

17

:00

20

05

.01

.08

19

:00

20

05

.01

.08

21

:00

20

05

.01

.08

23

:00

20

05

.01

.09

1:0

0

20

05

.01

.09

3:0

0

20

05

.01

.09

5:0

0

20

05

.01

.09

7:0

0

20

05

.01

.09

9:0

0

20

05

.01

.09

11

:00

20

05

.01

.09

13

:00

20

05

.01

.09

15

:00

20

05

.01

.09

17

:00

20

05

.01

.09

19

:00

Wa

terl

ev

el,

m

A ndrejos ta - obs erved

K īš ez ers - obs erved

R īgas HE S - obs erved

A ndrejos ta - c alc ulated

Ķ īš ez ers - c alc ulated

R īgas HE S - c alc ulated

The hydrodynamics of the flood

scenarios were calculated for the

model storm situations. The

importance of the dynamical

modeling of flooded areas was

shown for the domain with a

complex channel system and

typical length of storm event

below 18 hours. Figure: water

level time series for storm with

return period 10 years.

4. Scenario calculation

Time

W

a

t

e

r

l

e

v

e

l

,

m

5. Flood mapping

The method for the mapping of the results of

hydrodynamical calculations on the digital

terrain map of much higher (10 m) spatial

resolution was proposed and applied. See

illustration of flooded area correction.

6. Examples of flooded areas

Traffic junction (left above), river delta

(right above), northern part of Ķīšezers

lake (below). Return period 10 years

(green), 100 years (red).