Comparison of Monte Carlo Schemes in the Modeling of Extreme Flood in Tropical Rain Forest Basins

• Published in: KSCE journal of civil engineering 2023, 27(12), , pp.5175-5189
• Main Authors: Ekwueme, Benjamin Nnamdi, Ibeje, Andy Obinna
• Format: Article
• Language: English
• Subjects: Autoregressive models; Autoregressive processes; Catchment area; Civil Engineering; Coefficients; Correlation coefficient; Correlation coefficients; Discharge measurement; Engineering; Gaging stations; Geotechnical Engineering & Applied Earth Sciences; History; Hydraulic Engineering; Hydrologic data; Industrial Pollution Prevention; Kurtosis; Probability distribution; Probability theory; Rainforests; River catchments; Rivers; Root-mean-square errors; Stream discharge; Stream flow; System effectiveness; Systems design; Tropical forests; Water resources; 토목공학
• ISSN: 1226-7988; 1976-3808
• DOI: 10.1007/s12205-023-0980-5

Insufficient length of streamflow record poses challenges in capturing extreme events, frequencies, and sequences crucial for effective water resources systems design. Consequently, simulation and application of statistically similar streamflow values becomes imperative. In this study, two Markov-based models, namely the Thomas-Fiering (T-F) model and the Autoregressive (AR) model, were developed based on a log-normal probability distribution to synthesize monthly and annual flows using a decade of historical data (1980–1989) collected from the Umulokpa gauging station in the Adada River catchment of Enugu State, Nigeria. Except for the month of December (p = 0.03), no significant difference (p > 0.05) was observed between the monthly and annual model-predicted streamflows. However, both models demonstrated limitations in fully capturing the characteristics of the observed flows, particularly regarding Kurtosis. Further accuracy assessment from computed correlation coefficients (0.018 for T-F and 0.122 for AR), along with root mean square error (RMSE) values (12.84 for T-F and 11.95 for AR), indicate that while both models deviated from the observed flows, the AR model demonstrated superior performance compared to the T-F model. However, to apply these models reliably, for future flow prediction in the Adada catchment, streamflow data from rivers with comparable catchment characteristics is essential.