Graduation Semester and Year
2019
Language
English
Document Type
Dissertation
Degree Name
Doctor of Philosophy in Civil Engineering
Department
Civil Engineering
First Advisor
Dong-Jun Seo
Abstract
The objective of this research is to improve the cost-effectiveness, reliability and resilience of water supply system operation through the utilization of ensemble forecasting and optimization of the operating policies. For the effective operation of a multi-reservoir water supply system skillful inflow and outflow forecasts are required. To that end, this research first assesses the value of medium-range ensemble precipitation forecasts generated with the HEFS developed by the US NWS in increasing skill and lead time of ensemble inflow forecasts. The skill of reservoir outflow forecasts is then assessed using three different models with varying complexity. Also assessed in the above is the relative importance of meteorological, hydrologic and reservoir modeling uncertainty in outflow forecasts. Lastly, the proposed policy modifications were optimized via global sensitivity analysis using the variance-based Sobol method to improve cost-effectiveness of the system operation further. The analysis results identify influential policies, assess their impact on the cost response of the system operation, and determine policy modifications. Main conclusions of this research follow below. Compared to using only the 72-hr RFC QPF, the use of the medium-range precipitation forecasts from GEFS increases skill and lead time of mean daily inflow forecasts from HEFS by up to 3 days for significant events. The HEFS-produced multi-daily inflow forecasts are significantly more skillful than the daily inflow forecasts, and extends the lead time of skillful forecasting further. It is demonstrated that the use of the HEFS-produced ensemble inflow forecasts results in significant savings in mean annual pumping cost, compared to the TRWD’s current practice of inflow forecasting. Among the three reservoir models considered in this work, RiverWare provides the largest skill in MEFP-GEFS-forced outflow forecasts. It is shown that meteorological, hydrologic and reservoir modeling uncertainties are comparably importance in improving skill of reservoir outflow forecasts. However, the decomposition of total predictive uncertainty into the above three indicates that the relative importance varies significantly with lead time and among different reservoir models. Evaluation of reservoir outflow forecasts for specific large outflow events shows that, although the reservoir outflow forecasts forced by the HEFS inflow forecasts are not probabilistically unbiased for very large to extreme events, the ensemble spread of the outflow forecasts is generally able to encompass the observed pool elevation and outflow. As such, the HEFS inflow forecasts provide additional critical information not available from single-valued forecasts for risk-based decision making in reservoir operations. Among the five operating policies selected for considered for modification, only one or two exert large influence on the cost for a given year. The influential polices, however, vary very significantly from year to year. The cost response of the system to policy modifications is complex and shows large interannual variations. Hydroclimatic conditions, storage states, inflow, and demand largely determine the influential policies and their modifications. It is shown that annually-varying, or dynamic, determination of policy modifications offer significantly larger potential for cost savings than using optimized, but fixed, policy modifications.
Keywords
Water forecasts, Forecast verification, Multi-reservoir system, Water supply, Reservoir modeling, Reservoir operation optimization
Disciplines
Civil and Environmental Engineering | Civil Engineering | Engineering
License
This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.
Recommended Citation
Limon, Reza, "IMPROVING MULTI-RESERVOIR WATER SUPPLY SYSTEM OPERATION USING ENSEMBLE FORECASTING AND GLOBAL SENSITIVITY ANALYSIS" (2019). Civil Engineering Dissertations. 396.
https://mavmatrix.uta.edu/civilengineering_dissertations/396
Comments
Degree granted by The University of Texas at Arlington