DSD-INT 2015 - Developing an operational hydrologic forecast system using EPS and Satellite Data - Gökçen Uysal
2 likes721 views
This document summarizes the development of an operational hydrologic forecast system in mountainous basins in Turkey using ensemble prediction systems (EPS) and satellite data. The system was developed for the Karasu basin using the HBV and SRM hydrologic models forced by ECMWF-EPS weather predictions. Probabilistic streamflow forecasts were produced out to 10 days lead time and evaluated against observations. The system was later expanded to include the Murat and Seyhan basins.
![SRM DEVELOPMENT IN RTC-TOOLS
Model Inputs : for each zone,
1. Daily total precipitation
2. Daily average temperature
3. Daily snow covered area (Depletion curve)
Model parameters (cr, cs, tcrit, RCA…)
Qn+1= [cSn* an(Tn+ Δ Tn) Sn+ cRnPn] * (A · 10000 / 86400) * (1-kn+1) + Qnkn+1](https://image.slidesharecdn.com/dsd-int2015-developinganoperationalhydrologicforecastsystemusingepsandsatellitedata-gkenuysal-151110093004-lva1-app6891/85/DSD-INT-2015-Developing-an-operational-hydrologic-forecast-system-using-EPS-and-Satellite-Data-Gokcen-Uysal-30-320.jpg)
DSD-INT 2015 - Developing an operational hydrologic forecast system using EPS and Satellite Data - Gökçen Uysal
- 1. Gökçen Uysal1, Bulut Akkol1, Cansaran Ertaş1, Cihan Çoşkun1, Arda Şorman1, Aynur Şensoy1, Dirk Schwanenberg2 1 Anadol u Uni versi ty , Eski ş ehi r , Turkey 2 Uni versi ty of Dui sburg - E s s en , Germany DEVELOPING AN OPERATIONAL HYDROLOGIC FORECAST SYSTEM USING EPS AND SATELLITE DATA IN MOUNTAINOUS BASINS OF TURKEY
- 2. Between Asia and Europe, Turkey is a high-altitude country with over 25 river basins, including the trans-boundary Tigris and Euphrates rivers. GEOLOGY AND TOPOGRAPHY OF TURKEY (1)
- 3. GEOLOGY AND TOPOGRAPHY OF TURKEY (2)
- 4. SNOW POTENTIAL OF TURKEY NOAA Web Site – www.noaa.gov Hypsometric mean elevation is about 1130 m. Snow is major component in the mountanious Eastern Part.
- 5. HYDROPOWER POTENTIAL OF TURKEY DSİ, 2012 report
- 6. HYDROPOWER DEVELOPMENT OF TURKEY (1) * https://www.hydropower.org/country-profiles/turkey ** IHA 2015 Hydropower status report, https://www.hydropower.org/ Turkey remains one of Europe’s leading markets for future hydropower development… Turkey has plans for hydropower development over the coming decade with a total installed electric power capacity of 2002 32 GW 2014 64 GW 2023 100 GW (with 30 per cent of electricity generation coming from renewables*). In early 2015, the Turkish Government announced it would allocate USD 16 billion to hydro development until 2018 as part of its Tenth Development Plan**.
- 7. Hydropower development In 2003, In recent years, HYDROPOWER DEVELOPMENT OF TURKEY (2) will be further supported by Turkey’s interconnections into the European grid and potential for further linkages east into Asia. Turkey liberalized its energy market and embarked on a process of privatizing existing assets. international private companies have made major investments in the country.
- 8. Considering the geographic location of Turkey and hydroelectric power potential, our R&D activities as well as the future operational implementation will provide; scientific outcomes, economic benefits, water resources benefits also for downstream stakeholders: flood mitigation, irrigation, hydropower… MOTIVATION FOR RESEARCH
- 9. KARASU (A PILOT BASIN) Keban, Karakaya, Atatürk etc. large dams are located at the downstream of the basin 60–70% of the total runoff volume originates during snowmelt period….
- 10. Keban Dam Karakaya Dam Atatürk Dam INSTALLED CAPACITY: 1 330 MW ENERGY PRODUCTION: 6 000 GWh/year INSTALLED CAPACITY : 1 800 MW ENERGY PRODUCTION : 7 354 GWh/year INSTALLED CAPACITY : 2 400 MW ENERGY PRODUCTION : 8 900 GWh/year Birecik Dam Karkamış Dam INSTALLED CAPACITY : 189 MW ENERGY PRODUCTION : 652 GWh/year CASCADE DAMS IN EUPHRATES RIVER INSTALLED CAPACITY : 672 MW ENERGY PRODUCTION: 2 516 GWh/year
- 11. In 2004, flood risk occurred due to sudden snow melt. At the time, there was no real-time forecasting system. Governmental agencies decided to (pre-)release additional 2,000 m3/s over 10 days from the most upstream reservoir through the cascade. Post-event evaluation showed that too much water was released and lost for other objectives such as irrigation or hydropower. FLOOD EVENT IN 2004??
- 12. Karasu Basin is one of the pilot basins selected for various; NATIONAL PROJECTS (DPT – State Planning Organization, TUBITAK – The Scientific and Technological Research Council of Turkey), INTERNATIONAL PROJECTS (NATO SfP-NATO Science for Peace, EUMETSAT H-SAF – European Organization for the Exploitation of Meteorological Satellites, Satellite Application Facilities on Support to Operational Hydrology and Water Management). PREVIOUS EFFORTS
- 13. The ongoing projects and execution periods; - Developing an Operational Hydrologic Forecast System using EPS and Satellite Data in Mountainous Euphrates and Seyhan Basins, TUBITAK Project, Oct 2013 - Dec 2015. - Implementation of Hydrological Ensemble Prediction System, Anadolu University Scientific Research Project (1307F284), Dec 2013 – Dec 2015. - Monitoring and Validation of Snow Components using Satellite and Automated Ground Stations, Anadolu University Scientific Research Project, Apr 2014 – Apr 2015. ONGOING NATIONAL EFFORTS
- 14. THROUGH INTERNATIONAL PROJECTS http://www.ecmwf.int/en/h-saf-and-hepex-workshops-coupled-hydrology • TUBITAK project is extended through COST (European Cooperation in Science and Technology) Project (ES1404) "A European network for a harmonised monitoring of snow for the benefit of climate change scenarios, hydrology and numerical weather prediction" action, • Moreover, preliminary DA results of H-SAF products (H-10, H-12 SCA and SWE) in cooperation with Duisburg-Essen University for Karasu pilot results are presented at the HSAF HEPEX WORKSHOP.
- 15. SCOPE & INNOVATION It is aimed to deploy an operational hydrologic forecast system in snow dominated mountainous headwater basins in Turkey. For the deterministic and probabilistic flow forecasts, numerical weather predictions (WRF, ECMWF-EPS, etc.) force hydrological models. The main motivation and most novel part of this work is the exploitation of ensemble forecasts for the first time in Turkey within an operational hydrologic forecasting system.
- 17. Elevation zones The Upper Euphrates Basin (Karasu) The main land-cover types are pasture, cultivated land and bareland. Elevation range (m) Area (km2 ) Hypsometric Mean Elevation (m) Slope (%) Karasu 1125-3487 10275 1983 19.24
- 18. RECENT EXTENTIONS TO NEW BASINS Seyhan Basin Euphrates Basin MODIS satellite image February 2012 (white areas indicate snow cover)
- 19. SEYHAN BASIN HAS DEVELOPING HYDROPOWER POTENTIAL ex: water structures located in Seyhan Basin, Göksu River
- 20. CONTRIBUTIONS Anadolu University University of Duisburg- Essen Institut Royal Météorologique de Belgique State Hydraulic Works Turkish State Meteorologi cal Service (TSMS)
- 21. MONITORING Runoff for Upper Euphrates Basin 0 50 100 150 200 250 300 350 400 1-Feb 15-Feb 29-Feb 14-Mar 28-Mar 11-Apr 25-Apr 9-May 23-May 6-Jun 20-Jun 4-Jul 18-Jul 1-Aug Discharge(m3/s) 2008 2009 2010 2011 2012 2013
- 22. Meteorological & SGS Point & SnoTel Stations EUPHRATES OBSERVATION NETWORK FP ID River & SGS Elevation (m) Basin Area(km2) Obs. Period E21A019 Karasu - Kemah 1123 10350 1953 - E21A022 Murat - Tutak 1552 5882 1953 -
- 23. FIELD WORK
- 24. 1-DAY LEAD TIME ACCURACY OF WRF RELATED TO GROUND STATIONS
- 25. EPS DATA FOR PROJECT CATCHMENTS ECMWF produced 28 km/grid resolution (T639/N320) EPS data Total prec and mean air temp EPS data 6 x 10 pixels for Karasu Catc., 4 x 6 pixels for Murat Catc. and 4 x 5 pixels for Seyhan Catch.
- 26. ENSEMBLE VERIFICATION SYSTEM (EVS) RESULTS
- 27. SNOW DEPLETION CURVES (SDC) Upper Euphrates SDC (MODIS after cloud removal)
- 28. HBV (Hydrologiska Byråns Vattenbalans) SRM (Snowmelt-Runoff Model ) HYDROLOGICAL MODELS (HBV & SRM) (Lindströmet al., 1997) (Martinec et al., 2008)
- 29. HBV Good at continuous modeling More parameters Low response to instantaneous melting Does not require SDC (output) SRM Good at melting period Less parameters Fast response to instantaneous melting Requires SDC (input) WHY MULTI-MODELS ? Through SRM Development
- 30. SRM DEVELOPMENT IN RTC-TOOLS Model Inputs : for each zone, 1. Daily total precipitation 2. Daily average temperature 3. Daily snow covered area (Depletion curve) Model parameters (cr, cs, tcrit, RCA…) Qn+1= [cSn* an(Tn+ Δ Tn) Sn+ cRnPn] * (A · 10000 / 86400) * (1-kn+1) + Qnkn+1
- 32. Ensemble Streamflows are developed for three different basins (Karasu, Murat, Çukurkışla, Kilicmehmet), using two different models for the 2011 melting period. The next results are presented every 10 days although the forecast is available daily. Ensemble Streamflows are compared with simulations. PROBABILISTIC FORECASTS
- 33. UNDER DELFT-FEWS PLATFORM (WITHOUT DATA ASSIMILATION AND OUTPUT CORRECTION) HBV SRM Spagetti plot Quartiles
- 35. ADDED VALUE OF FEWS PLATFORM Import Data Module (in particular for EPS and satellite data), Grid processing (basin averages etc.), Forecast Handling (pre- and post processing, administration, storage, analysis, export) Easy link with EVS for ensemble analysis, Data Assimilation (embedding of MHE and several Kalman Filters via OpenDA and RTC-Tools in progress), Potential future use as operational platform (suitable to transfer knowhow to governmental agencies)…
- 36. CONCLUSIONS AND OUTLOOK Streamflow forecasting accounting snow accumulation and melt is a major topic for Turkey due to flooding and hydropower Flood early warning may decrease inundation risk and related damages based on forecasts to take the right action Flow forecasts assist the short-term operation of reservoirs and HEPP and increase their economic value Potential for improvement in Numerical Weather Predictions (higher resolution, bias correction, etc.) Data Assimilation and probabilistic forecasting are necessary tool to issue accurate and robust forecasts (further collaboration in H-SAF with Deltares and University of Duisburg-Essen on various DA approaches)
- 37. CONCLUSIONS AND OUTLOOK (2) HBV and SRM models perform OK, but may have potential for improvement (further collaboration with Federal Institute of Hydrology, Germany and University of Duisburg-Essen). Further work will also focus on the short-term management of the HEPP by the use of forecasts and predictive control techniques Interdisciplinary work brings together staff from the government, private sector, academia providing an environment for sharing knowledge and creating beneficial results for all members.
- 38. THANK YOU FOR YOUR ATTENTION gokcenuysal@gmail.com This study is funded by 113Y075 TUBITAK (The Scientific and Technological Research Council of Turkey) and BAP 113Y075 (Scientific Research Project) of Anadolu University, Turkey.