Assessment of design flood characteristics for ungauged permafrost basin

Olga Semenova 1,2 , Lyudmila Lebedeva 3,4 , Anatoly Zhirkevich 5 1 Gidrotehproekt, 2 St. Petersburg State University, 3 Nansen Centre, 4 Hydrological Institute, St.Petersburg

State 5

Hydroproject Institute, Moscow, Russia

Motivation

 Sparse hydrometric and meteorological network of Eastern Siberia  Hydrologic regime affected by permafrost  Changing climate conditions  Practical needs of reliable assessment of design flood characteristics for present and future

Goal

• Develop parameterization scheme for assessment of possible hydrologic changes in the Timpton River basin • Make preliminary assessment of future runoff characteristics at the Kanku hydropower plant (under construction) by different approaches using the hydrological model Hydrograph • Compare the results of three approaches: PMF (possible maximum flood), traditional frequency analysis and deterministic-stochastic modelling

Approaches

• Frequency analysis of observed runoff characteristics (annual mean, maximum, minimum). It is based on assumption of hydrological processes stationarity, long-term observational series and basin-analogues. • Probable Maximum Flooding (PMF) method includes identification of crucial meteorological factors of maximum flooding and assessment of runoff characteristics according to them (in our case, by hydrological modelling) • Deterministic-stochastic modelling – combined use of stochastic weather generator and deterministic hydrological modelling

1 2 3 № Gauge

Study object – the Timpton river basin

Ust’-Baralas The Kanku hydropower plant Ust’-Timpton

Distance from river mouth, km 337 201 20 Basin area, km 2 13 300 27 300 43 700 Period of runoff measurements 1954 – cont.

1952 – cont.

    Altitude varies from 600 to 1700 m Continental climate Bare rocks, tundra and larch forest Zone of discontinuous and sporadic permafrost

Variety of landscapes

Bare rocks

Deep active layer, Subsurface runoff

Bush tundra Larch forest

Shallow active layer, surface runoff

Riparian vegetation

www.hydrograph-model.ru

The Hydrograph model

 Process-based ( includes all processes ) explicitly  Observable parameters, minimum calibration ( can be obtained apriori )  Common input daily data ( air temperature and moisture, precipitation )  Free of scale problem ( from soil column to large basin ) initially developed by Prof. Yury Vinogradov

www.hydrograph-model.ru

Basin schematization and model parameters

Basin area was divided into three landcover types: 1. Bare rocks 2. Tundra 3. Larch forest

Density, kg/m 3 Porosity, % Water holding capacity, % Infiltration coefficient, mm/min Heat capacity, J/(kg o C) Heat conductivity, W/(m o C) Wilting point, % Moss and lichen

500 90 60 10

Peat

1720 80 20-40 0.0005-0.5

Clay with inclusion of rocks

2610 55 13 0.0005

Bedrock

2610 35 7 0.05-1 1930 0.8

8 1930 0.8

6-8 840 1.2

4 750 1.5

2-3

Q 1 2 0 0 0 1 0 0 0 0 8 0 0 0 6 0 0 0 4 0 0 0 2 0 0 0

Model validation on two adjacent basins

1. Ust’-Timpton (43700 km 2 ), 1970 – 1973 1970 1 0 0 0 0 8 0 0 0 6 0 0 0 4 0 0 0 2 0 0 0 1971 6 0 0 0 5 0 0 0 4 0 0 0 3 0 0 0 2 0 0 0 1 0 0 0 I II III IV V VI VII VIII IX X XI XII 1972 8 0 0 0 6 0 0 0 4 0 0 0 2 0 0 0 I II III IV V VI VII VIII IX X XI XII 1973 I II III IV V VI VII VIII IX X XI XII I II III IV V VI VII VIII IX X XI XII Т

Model validation on two adjacent basins

Q 3 0 0 0 2 5 0 0 2 0 0 0 1 5 0 0 1 0 0 0 5 0 0 2. Ust’-Baralas (13300 km 2 ), 1970 – 1973 1970 2 5 0 0 2 0 0 0 1 5 0 0 1 0 0 0 5 0 0 1971 I II III IV V VI VII VIII IX X XI XII I II III IV V VI VII VIII IX X XI XII 3 0 0 0 2 0 0 0 1 0 0 0 1972 2 5 0 0 2 0 0 0 1 5 0 0 1 0 0 0 5 0 0 1973 I II III IV V VI VII VIII IX X XI XII I II III IV V VI VII VIII IX X XI XII Т

• • • •

Stochastic Weather Generator

Simulation of daily precipitation, temperature and relative humidity Simulation of annual and intra-seasonal variations Temporal correlation of meteorological elements Spatial correlation of meteorological elements • • Parameters are estimated from observed series of meteorological data Parameters may be modified according to climate change projections

Projected scenarios of air temperature (left) and precipitation (right) by changes 2100 (two models, two scenarios)

CanESM2 model 40 1966-1984 rcp 8.5 - CanESM2 20 rcp 2.6 - CanESM2 200

Reference period – 1966-1984

1966-1984 rcp 8.5 - CanESM2 rcp 2.6 - CanESM2 150 0 100 -20 50 -40 I II III IV V VI VII VIII IX NorESM1-M model 40 1966-1984 rcp 8.5 - NorESM1-M 20 X month XI XII rcp 2.6 - NorESM1-M 150 100 0 0 50 -20 I II III IV V VI VII VIII IX X XI XII 1966-1984 rcp 8.5 - NorESM1-M rcp 2.6 - NorESM1-M -40 I II III IV V VI VII VIII IX X month XI XII 0 I II III IV V VI VII VIII IX X XI XII

Probable Maximum Flooding approach

1. Main factors of maximum flooding: • •

Snow water equivalent

•

Rain on snow during snowmelt

• Intensity of air temperature increase during snowmelt • Date of the temperature transition from positive to negative values in autumn antecedent year Precipitation of the last warm month of antecedent year 2. The values of 1, 0.1, and 0.01% probability of chosen factors were used to generate artificial meteorological series 3. Generated meteorological data were used as the forcing for the Hydrograph model to simulate probable maximum flood

Results – the Kanku hydropower gauge

3

PMF

2. 5 2 1. 5 1 0. 5 0 - 0. 5 - 1 - 1. 5 - 2 2 5 0 0 0 2 0 0 0 0 1 5 0 0 0 1 0 0 0 0 5 0 0 0

FA

0 0 . 1 0 . 4 1 2 h ist oric 4 6 8 1 2 2 0 3 0 E xc . pr o b. , % 4 5 Can Esm rsp 2. 6 6 0 7 5 8 5 9 2 9 6 9 8 Can Esm rcp 8. 5

Frequency analysis Probable Maximum Flooding

Conclusions

Three different approaches were used to access extreme flood characteristics for an ungauged basin in permafrost area.

The results of three approaches are highly inconsistent.

Probable Maximum Flooding approach relies on estimation of crucial meteorological factors values of low probability. Combination of the latter in a forecast results in significantly overestimated discharge.

Conventional frequency analysis is based on the assumption of hydrological processes stationarity, which is undermined by the ongoing environmental change in high latitudes.

DS-modelling approach is seen as a preferable tool for the assessment of flood characteristics in ungauged basins. It allows explicitly accounting for the factors causing extreme runoff events.

Thank you for attention!

15