Dam Failure Analysis for the Lago de Matrullas Dam, Orocovis, Puerto Rico

Abstract

The U.S. Geological Survey, in cooperation with the Puerto Rico Electric Power Authority, completed a hydrologic and hydraulic study to assess the potential hazard to human life and property associated with the hypothetical failure of the Lago de Matrullas Dam, located within the headwaters of the Río Grande de Manatí. The hydrologic study yielded outflow hydrographs and peak discharges for Lago de Matrullas and other subbasins in the Río Grande de Manatí hydrographic basin for three extreme rainfall events: (1) a 6-hour probable maximum precipitation (PMP) event, (2) a 24-hour PMP event, and (3) a 100-year-recurrence, 24-hour rainfall event. The hydraulic study simulated the hypothetical dam failure of Lago de Matrullas using hypothetical flood hydrographs generated from the hydrologic study and selected dam breach parameters. The flood wave resulting from the failure was downstream-routed through the lower reaches of the Río Matrullas, the Río Toro Negro, and the Río Grande de Manatí for determination of water-surface profiles developed from the event-based hydrologic scenarios and “sunny day” (no precipitation) conditions. The Hydrologic Modeling System (HEC–HMS) and the River Analysis System (HEC–RAS) computer programs, developed by the Hydrologic Engineering Center (HEC) of the U.S. Army Corps of Engineers, were used for the hydrologic and hydraulic modeling, respectively. The flow routing in the hydraulic analyses was performed using the unsteady-state flow module available in the HEC–RAS model.

At the Lago de Matrullas Dam, inflow peak discharges of about 1,104 and 1,032 cubic meters per second (m³/s) were estimated with HEC–HMS for the 6- and 24-hour PMP events, respectively. The 100-year recurrence, 24-hour rainfall event simulation resulted in a peak discharge of about 418 m³/s. For the hydrologic analysis, no dam failure conditions were considered with the model. The results of the hydrologic simulations indicated, however, that the dam would be overtopped by the simulated 6- and 24-hour PMP events. Unlike the 6- and 24-hour PMP events, no overtopping was observed during the simulated 100-year recurrence, 24-hour rainfall event.

For the dam-breach hydraulic analysis, the hypothetical failures of the Lago de Matrullas Dam included two possible failure modes: overtopping and piping. Overtopping failure was evaluated in this study for the 6- and 24- hour probable-maximum-precipitation breach scenarios. Piping dam failure was simulated for sunny day conditions and for the 100-year-recurrence, 24-hour rainfall scenario.

Results from the simulated dam failure of the Lago de Matrullas Dam using the HEC–RAS model for the 6- and 24-hour PMP events showed peak discharges at the dam of 3,149.33 and 3,604.70 m³/s, respectively. Dam failure during the 100-year-recurrence, 24-hour rainfall event resulted in a peak discharge of 2,103.12 m³/s directly downstream from the dam. Dam failure under sunny day conditions produced a peak discharge of 1,695.91 m³/s at the dam assuming the antecedent lake level was at the morning-glory spillway invert elevation. Flood-inundation maps prepared as part of the study depict the flood extent and provide valuable information for preparing an Emergency Action Plan. Results of the failure analysis indicate that a failure of the Lago de Matrullas Dam could cause flooding to many of the inhabited areas along stream banks from the Lago de Matrullas Dam to the mouth of the Río Grande de Manatí. Among the areas most affected are the low-lying regions in the vicinity of the towns of Ciales, Manatí, and Barceloneta. The delineation of the flood boundaries near the town of Barceloneta considered the effects of a levee constructed during 2000 at Barceloneta in the flood plain of the Río Grande de Manatí to provide protection against flooding to the near-by low-lying populated areas. The results showed overtopping can be expected in the aforementioned levee during 6- and 24-hour probable-maximum-precipitation dam failure scenarios. No overtopping of the levee was simulated, however, during dam failure scenarios under the 100-year recurrence, 24-hour rainfall event or sunny day conditions.

First posted June 11, 2015