Modeling Effects of Kizildere Geothermal Plant on Great Menderes River Water Quality

1. NUMERICAL MODELING OF THE EFFECTS OF KIZILDERE GEOTHERMAL POWER PLANT ON WATER QUALITY OF THE GREAT MENDERES RIVER, TURKEY Assoc.Prof.Dr.Şebnem ELÇİ, Adem Erdem, Assoc.Prof.Dr.Gülden Gökçen Izmir Institute of Technology, Urla, Izmir, Turkey

2. Outline • Scope of theStudy • Introduction – GeothermalEnergy • Numerical Model – WASP 7.2 • Study Site - Great Menderes River • Implementation of theNumerical Model • DiscussionandConclusions

3. Scope of the Study • Motivated by highboron concentrations observed in crops and soils of a river basin located in the south-west of Turkey. • Part of a projectfundedbyTheScientificandTechnologicalResearchCouncil of Turkey (Tübitak -104M301) which had an objective of ‘Assessment of environmentaleffects of geothermalapplications’. • Duringtheproject, soil, airandwatersampleswerecollectedfromKizildereGeothermal Power Plant site andanalysiswereconducted. • Thispaperfocused on the effects of the plant on water quality of the Great Menderes River.

4. Introduction - GeothermalEnergy • Considered as one of the cleanerforms of energy • Lowcostelectricalenergy (80% cheaper as comparedtofossilefuelsandnuclearenergy) • Turkey is rich in potential (feasiblegeothermalcapacityforelectricity is about 4500 MW, whereascurrentproduction is 27.7 MW) • Importantgeothermalfields in Turkey • Denizli-Kızıldere(242°C) (20 MW) • Aydın-Germencik (232°C) • Çanakkale- Tuzla (171°C) • Aydın-Salavatlı (174°C) • Kütahya-Simav (162°C) • İzmir-Seferihisar (150°C) • Caferbey-Manisa (150 °C)

5. Introduction • When the geothermal fluid (composed of water, steamandgasessuch as CO2, H2S, NH3, andBoronreaches to the wellhead, it is directed to a separator where steam and liquid phases are separated. • Then steam is sent to the turbine where the electrical power is maintained, whileliquid, which is 88.5% of the total flow rate, is dischargedintonearbystream (Great Menderes River)by a channel. • Whenthisliquid is not re-injectedtothewell, it can causeincrease of Li, As, H2S, Pb, salinity, Boron, NH3 in streams

6. Numerical Model – WASP 7.2 Water Quality Analysis Simulation Program • WASP 7.2, developed by US Environmental Protection Agency (EPA), is designed for aquatic systems and solves for the advection, dispersion, point and diffuse mass loading and boundary exchange processes in 1, 2, and 3-D systems. • WASP utilizes the kinematic wave model for hydrodynamic calculations but, when needed, it can also be linked with other hydrodynamic and sediment transport models (EFDC, CE-QUAL) . • Although the model can not execute hydrologic computations, it can be linked to hydrologic models such as HSPF and SWIMM.

7. Numerical Model – WASP 7.2 ModelingFramework WASP Modeling Framework Model Preprocessor/Data Server Binary Wasp Input File (wif) CSV, ASCII Output WASP Input Messages Models Hydrodynamic Interface BinaryModelOutput Stored Data BMD Eutrophication Hydro Conservative Toxicant ExportedModelResults Organic Toxicants MOVEM Mercury Graphical Post Processor Heat TOXI :toxicants; EUTRO : ammonia, nitrate, dissolvedoxygen, salinity, organic/inorganicphosphorus; MERCURY :mercury; HEAT :temperatures

8. Study Site – G. Menderes River

9. The modeled segments of the Great Menderes River and the locations of monitoring stations

10. Implementation of theWater Quality Model • Segmentation of theriver, derivation of thelength and width of the segments from the 1/25000 scalemapsandcalculation of the surface area and volume for each segment from the measured water levelsat the monitoring stations • Calculation of theslopes of the segments by dividing the elevationdifferences by the segment length • Estimation of the roughness coefficientusingformulationbyMeyer-Peter ve Müller (1948) for G. Menderes River: d90: 0,0008 n: 0,0117 • Acquiring daily flow rates measuredat the stations and the flow rate of the water withdrawnat the two water intakes (Yenice and Feslek) for irrigation from DSI and EIEI during the simulation period (water year 2006).

11. Implementation of theWater Quality Model • Preparation of input data - daily average air temperature (°C), wind speed (m/s), and solar radiation (cal/cm2) – for the numerical model • Derivation of daily water temperatures and dissolved oxygen values from monthly measured observations • Validation of the numerical model with the observations

12. Validation of the numerical model RMSE: 0.45

13. Results of the Water Quality Model (EUTROPHICATION)

14. Results of the Water Quality Model (TOXI)

15. DiscussionandConclusions • This studyinvestigated the environmentaleffects of KızıldereGeothermal Plant on Great Menderes River. • Legal arrangements for forcing ‘Environmental Impact Assessment’ beforeestablishing new geothermal power plants isrequired and the assessmentshouldinvolveeffects on bothsoil and water resources. • Water qualityparametersweresimulated by EUTRO and TOXI modules and the resultswerediscussed for different scenarios with and without the discharge of effluent.

16. DiscussionandConclusions • Resultsindicateda decrease in dissolvedoxygen concentration (duetowarmereffluent of the plant),an increase in salinity and ammonia concentrations. • Resultsalsoindicatedincrease in Boronconcentrationsespeciallyduringwinterwhenflowsarelow. • Accumulation of Boron at sedimentswasalsoobserved (140 ppm in soilsamplestaken at thechannel, morethan 3 timestheothersamplingstations), because it can be easilyabsorbedbysuspendedparticles.

17. Acknowledgements • Funding for thisstudywasprovided by TUBITAK (Project no: 104M301). Wewouldlike to thank to project team for their contributions. Wealsowouldlike to extendourthanks to State Hydraulic Works (DSI) and Electrical Power Resources Survey and Development Administration (EIEI) for providing water quality and flow rate data.