Filling out missing daily streamflow data using fuzzy rule-based models

Date

2020

Author

Akgün, Ömer Burak

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

230
views

0
downloads

Daily streamflow observations are used for many purposes including analysis of current water-resources conditions in a basin, development of water-resources planning and management strategies and climate change adaptation measures. Streamgages are used to collect streamflow data; however, many streamgages suffer from a common problem: data-gaps. In this study, a Takagi-Sugeno Fuzzy RuleBased (TS_FRB) Model that uses Subtractive Clustering (SC) for rule generation is developed to fill out missing daily streamflow data due to a streamgage becoming inoperative for a long period. Fuzzy Rule-Based (FRB) model uses only daily streamflow data of neighboring streamgages, thus is very advantageous in terms of data requirement. Ergene Basin, Turkey is used as the case study and FRB models are developed to fill out missing daily streamflow data at four streamgages found in this basin. Numerous models are built to investigate the effect of the SC parameters (i.e., the number of cluster centers and the cluster radius) by which the rule-base of the FRB is identified, and the number of input variables on the performance of the models. Small cluster radius results in similar fuzzy rules to be devised, which vi reveals the needs for more rules. On the other hand, as the number of cluster centers increases, the risk of overfitting increases. Thus, selection of the best cluster radius and number of cluster centers combination is a challenging task and requires a trialand-error procedure. FRB models developed in this study provides good and robust ({u1D441}{u1D446}{u1D438} values around 0.67) estimations for the closely spaced streamgages located on the same tributary. On the other hand, FRB model performance is poor for the streamgage that is located far away from its neighboring streamgages and for the streamgage that is located on a different tributary than its neighbors. Moreover, anthropogenic effects in the Ergene Basin, makes the training of the FRB challenging and influences the model performance negatively.

Subject Keywords

Missing data, Subtractive Clustering, Takagi-Sugeno Fuzzy Rule-Based Models, Missing Data, Infilling, Daily Streamflow

URI

http://etd.lib.metu.edu.tr/upload/12625527/index.pdf
https://hdl.handle.net/11511/45756

Collections

Graduate School of Natural and Applied Sciences, Thesis

Suggestions

OpenMETU
Core

Estimation of Streamflow Using Takagi-Sugeno Fuzzy Rule-Based Model Akgün, Ömer Burak; Kentel Erdoğan, Elçin (2018-06-01) In this study, a tool is developed to estimate streamflow at Guvenc Basin, Ankara by using Takagi-Sugeno (TS) Fuzzy Rule-Based (RB) model. The model takes precipitation and runoff at time � as predictor (input) and estimates the runoff at time � + 1. The approach used to generate the TS RB model is based on density based clustering. Each cluster center is used to generate a fuzzy rule that represents the system behaviour. Satisfactory results are obtained especially after including the seasonal behaviour of...
Estimation of predictive hydrologic uncertainty using the quantile regression and UNEEC methods and their comparison on contrasting catchments Dogulu, N.; López López, P.; Solomatine, D. P.; Weerts, A. H.; Shrestha, D. L. (Copernicus GmbH, 2015-7-23) In operational hydrology, estimation of the predictive uncertainty of hydrological models used for flood modelling is essential for risk-based decision making for flood warning and emergency management. In the literature, there exists a variety of methods analysing and predicting uncertainty. However, studies devoted to comparing the performance of the methods in predicting uncertainty are limited. This paper focuses on the methods predicting model residual uncertainty that differ in methodological complexi...
Regional intensity-duration-frequency analysis in the Eastern Black Sea Basin, Turkey, by using L-moments and regression analysis Ghiaei, Farhad; Kankal, Murat; ANILAN, TUĞÇE; YÜKSEK, ÖMER (2018-01-01) The analysis of rainfall frequency is an important step in hydrology and water resources engineering. However, a lack of measuring stations, short duration of statistical periods, and unreliable outliers are among the most important problems when designing hydrology projects. In this study, regional rainfall analysis based on L-moments was used to overcome these problems in the Eastern Black Sea Basin (EBSB) of Turkey. The L-moments technique was applied at all stages of the regional analysis, including det...
Calibration and evaluation of a flood forecasting system: Utility of numerical weather prediction model, data assimilation and satellite-based rainfall Yücel, İsmail; Yılmaz, Koray Kamil (2015-04-01) A fully-distributed, multi-physics, multi-scale hydrologic and hydraulic modeling system, WRF-Hydro, is used to assess the potential for skillful flood forecasting based on precipitation inputs derived from the Weather Research and Forecasting (WRF) model and the EUMETSAT Multi-sensor Precipitation Estimates (MPEs). Similar to past studies it was found that WRF model precipitation forecast errors related to model initial conditions are reduced when the three dimensional atmospheric data assimilation (3DVAR)...
Comparison of Frequency Analysis with Density Based Clustering in Evaluating Extreme Streamflows Kentel Erdoğan, Elçin(2018-12-31) Analysis of streamflow data has significant importance concerning sustainable use of water resources, planning supply of short and long term demands. Streamflow, unlike many other completely random variables, is a dependent variable which is closely related to both meteorological (such as precipitation, temperature) and catchment characteristics (such as land use and soil type). Thus inter-dependencies should be taken into consideration while analyzing streamflow. Besides this, while analyzing streamflow da...

Citation Formats

Ö. B. Akgün, “Filling out missing daily streamflow data using fuzzy rule-based models,” Thesis (M.S.) -- Graduate School of Natural and Applied Sciences. Civil Engineering., Middle East Technical University, 2020.