Water-centric Inter-linkages Some Case Studies in Sri Lanka S.B - - PowerPoint PPT Presentation

S.B Weerakoon1

• Water with Floods, Climate, Energy

Cho Thanda Nyunt2

• Water with Climate Change

Yoshimitsu Tajima3

• Water with Coastal Environment

1University of Peradeniya, Sri Lanka 2University of Tokyo, Japan 3University of Tokyo, Japan

Some Case Studies in Sri Lanka

Water-centric Inter-linkages

S.B Weerakoon1, Srikantha Herath2 , Gouri De Silva1

1Department of Civil Engineering,

University of Peradeniya, Peradeniya, Sri Lanka

2United Nations University, Shibuya-ku, Tokyo, Japan

Rainfall Forecasting and Flood Inundation along the Lower Reach of Kelani River Basin under Changing Climate

Flood inundation in Colombo and suburbs create heavy economic damages

4

Elevation distribution (DEM)

 Region

– Wet Zone

 Total Basin Area – 2,230 km2  Uppermost Elevation – 2,250 m  Length of the River

– 192 km

 Average Annual Rainfall

– 2,400 mm

 Peak flow

– 800-1500 m3/s

 Vegetation cover  Upper basin

– Tea, rubber, grass and forest

 Lower basin

– heavily urbanized

Kelani River Basin

• 1. Weather forecasting for flood warning

Rainfall forecasting using downscaling of 72 hr climate model data by Weather Research and Forecasting (WRF) model

• to provide early warning on rainfall and floods

Application of Weather Research and Forecasting model (WRF model)

Nesting option – 135/45/15/5 𝑙𝑛 (4050 ×4050 𝑙𝑛 / 1530 ×1395 𝑙𝑛 / 465 ×465 𝑙𝑛 / 245 ×260 𝑙𝑛) with input resolution of 10 minutes

Input data downloaded from NCAR web site

WRF – short term rainfall forecasting

 Calibration – 21st November 2005

• Validation – 27th and 28th April 2008, and 30th, 31st May and

1st June 2008

Justification

• Mean Absolute Model Error percentage

𝑵𝑵𝑵𝑵 % = 𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻 𝑺𝑻𝑻𝑺𝑺𝑻𝑻𝑻 − 𝑷𝑷𝑷𝑻𝑷𝑷𝑻𝑻 𝑺𝑻𝑻𝑺𝑺𝑻𝑻𝑻 𝑷𝑷𝑷𝑻𝑷𝑷𝑻𝑻 𝑺𝑻𝑻𝑺𝑺𝑻𝑻𝑻 × 𝟐𝟐𝟐𝟐

Calibration and validation

Difference between WRF predictions and observed rainfall for the date 21st November 2005

Calibration Validation

Difference between WRF predictions and observed rainfall for the date 27th April 2008

Validation

Difference between WRF predictions and observed rainfall for the date 31st May 2008 Difference between WRF predictions and observed rainfall for the date 1st June 2008

• 2. Flood inundation analysis under

changing climate

• Climate pattern up to 2099 under A2 and B2 Emission

Scenarios of AR4 by – Statistical Downscaling Model

• Inundation modeling in the lower Kelani basin using

FLO-2D model

Study area for rainfall analysis (about 2200 km2) Study area for flood analysis (about 500 km2) Study area for rainfall – runoff simulation (about 1700 km2)

Source: Department of Irrigation, Sri Lanka

Hanwella

14

Data collection Rainfall forecast for future under both A2 and B2 scenarios (SDSM) Generate inflow at the upstream for future (FLO-2D) Forecasting future flood conditions according to future rainfall Meteorological data Topographic data Hydrological data Calibration and verification of HEC-HMS Analyze forecasted rainfall data

Grid and boundary of catchment

Flood inundation model (FLO-2D)

Parameters

 Channel and catchment characteristics such as,

• Infiltration
• Manning’s

coefficient

• Channel

roughness

• Channel

shape and dimensions

Grid size – 250 m Elevation Distribution (DEM) Manning's coefficient according to land use

Return Peri riod / d / (yr yr) 3 da day ra rain infall / / (mm) mm) A 2 B 2 50 50 391 383 100 100 429 420 Return Peri riod / d / (yr yr) Dail ily ra rain infall / l / (mm) mm) A 2 B 2 50 50 425 377 100 100 476 417 3 day rainfall for upper basin Daily rainfall for lower basin

Calibration and validation – with respect to discharge hydrograph at the D/S gauging station (Nagalagam Street) and flood inundation maps

17

Justification

• Normalized Objective Function (NOF)
• Nash – Sutcliffe efficiency R2

NS

• Percentage bias (δb)
• Fraction of the domain classified correctly by the simulations (F)

𝐺 = 𝑇𝑝𝑝𝑝 ∩ 𝑇𝑛𝑝𝑛 𝑇𝑝𝑝𝑝 ∪ 𝑇𝑛𝑝𝑛 × 100

Calibration

Observed and simulated flow during 2005 flood

18

Validation

Observed and simulated flow at during April-May 2008 flood

19

Event 𝑂𝑂𝐺 𝑆2

𝑂𝑇

𝜀𝑐 November 2005 0.09 0.98 6.98% April-May 2008 0.14 0.97 10.37%

Validation

Time series of observed and simulated flow at the D/S gauging station during May 2010 flood Observed Inundation extent (from DMC data) Simulated Inundation extent

F = 73%

HEC–HMS was used to compute inflow into the lower basin at the upstream

Inundation extents due to 50 year return period rainfall under A2 scenario Inundation extent correspond to 50 year return period rainfall under B2 scenario Inundation extents due to 100 year return period rainfall under A2 scenario Inundation extent correspond to 100 year return period rainfall under B2 scenario

Results

Water Resources in Sri Lanka

2012 Source: CEB Kelani Basin MahaweliBasin

About 140 MHPs with 350MW capacity at present contribute 6.2% Goal- Renewable energy to supply 10% by 2016

Hydro with Reservoir Hydro Run-of- River Diesel Coal Heavy Oil

S1 15 1 778 974 778

200 400 600 800 1000

in kt eq. CO2/TWh

Source: CEB

 Run-of-river MHP/SHP provides lowest

contribution during dry period

 Operation of multi-purpose large reservoirs has

great impact on energy generation

Source: CEB

 For the Kelani River Basin

 Forecasting of weather for early warning systems  Inundation extents and high risk areas of inundation

by rainfalls of 50, 100 year return periods under both A2 and B2 scenarios in Colombo were investigated

 For hydro-energy



Integrated water management under changing

climate is important.

7th July 2008, Weerakoon

Devon waterfall ( Upper Kotmale Subbasin)

Adaptation strategies

Levee and detention basin

• A levee of 1.0 m height and 10 km long from the

downstream

• Detention reservoirs; several marshy lands were

identified from land use maps and converted in to detention reservoirs.

10km long Levee Developed marshy lands as detention reservoirs

50 year return period rainfall under A2 scenario

Results – Inundation extents under (c)

50 year return period rainfall under B2 scenario

Reduces average risk about 65% Reduces the average risk about 40%

100 year return period rainfall under A2 scenario 100 year return period rainfall under B2 scenario

Results – Inundation extents under (c)

Reduces the average risk about 32% Reduces the average risk about 25%