Year Established: 2016 Start Date: 2016-03-01 End Date: 2017-02-28
Total Federal Funds: $16,000 Total Non-Federal Funds: $32,000
Principal Investigators: Ramesh Teegavarapu
Abstract: The main objective of the proposed research study is to develop and evaluate several variants of statistical, rule-based and data-mining based algorithms for identification and detection of outliers and data anomalies in hydro-meteorological data, especially stage data. This study is continuation of work that was completed in in year 2015. In the first phase of the work, algorithms for identification stage data anomalies were developed and evaluated and a prototype tool was developed. The third phase (this proposed study) will focus on refinement of tool by inclusions of new features and algorithms and also application and testing of the same for stage observations at different hydraulic structures in South Florida Water Management District (SDFWMD) region. New neighborhood value-based methods developed in the previous study will be evaluated in this study. The methods currently existing in the tool will be evaluated using a number of indices, skill scores and performance measures based on input from modelers and data management personnel at SFWMD. The study will use data from a network of stage monitoring sensors automatically uploaded into the databases of SFWMD. The proposed research is highly relevant and critical to a number of water resources management agencies (e.g. South Florida Water Management District (SFWMD)) that currently stage data for modeling and management of day-to-day operations of water resources systems and development of protocols for flood control warnings.