A 500-Year Storm, a Question of Preparedness
Over the course of a seven-day period in August 2017, Hurricane Harvey dropped more than 50 inches of rainfall over the city of Houston, Texas. The storm has subsequently been dubbed “the most signiﬁcant rainfall event in U.S. history” by the United States Geological Survey. The estimated costs from the storm ranged from $125 to $180 billion USD – making it the second most costly storm in U.S. history.
Can we Model Hurricane Harvey over San Antonio?
In the days leading up to Hurricane Harvey, the San Antonio River Authority (SARA) relied on its ﬂood warning system (FWS) developed by HDR Engineering, Inc. (HDR) in InfoWorks RS to get a predictive analysis and inform a preparedness plan with the City of San Antonio, should Harvey’s path lead it in that direction. SARA’s Wayne Tschirhart, technical lead for the FWS, evaluated several anticipated rain events based on the estimated rainfall rates being reported by the Weather Service so the city could have an idea of what to expect. The storm’s path turned and stalled over Houston, never making it to San Antonio.
One week later, Texas Floodplain Management Association (TFMA) hosted hydrologic and hydraulic experts from around the state for the 2017 TFMA Fall Technical Summit. When the extent of Harvey’s damage became understood it sparked the question: would San Antonio have been prepared if the storm headed in another direction? San Antonio City Council asked Steve Graham, SARA Executive, if it were possible to model the impacts of Hurricane Harvey. SARA approached Brandon Hilbrich, PE, Project Manager, and Anthony Henry, Senior Modeler, at HDR with this very question. SARA and City Council wanted to know how a storm of Harvey’s magnitude would have impacted their city, and they wanted to know posthaste.
A Model Built Efficiently, an Accurate Depiction of a Worst-Case Scenario
With a four-week deadline to provide meaningful results for the Downtown area, the HDR team used InfoWorks ICM to conduct the analysis. HDR had already developed the ﬂood warning model in the legacy Innovyze product InfoWorks RS, which was easily converted to ICM. HDR had also already conducted a pilot study converting the Upper San Antonio River (USAR) watershed in InfoWorks ICM. The USAR watershed covers San Antonio’s downtown and the majority of inner Loop 410.
Working with this baseline, HDR could import operational data from ICMLive, steady state HEC-RAS & HEC-HMS model data, as well as XPSWMM 1D/2D data to get the model up to the necessary level of detail for simulating Harvey. With ICM’s integrated catchment modeling engine, running a model of this level of complexity was made possible and time-eﬀective to meet SARA’s deadline.
InfoWorks ICM helped HDR to accurately model the unique USAR watershed characteristics, including underground tunnels using pressure ﬂow conduits, and operational control for a dam and two ﬂood gates.
The ICM model allowed HDR to produce results for the entire USAR watershed in a short period of time. This analysis would have taken several months longer with traditional HEC-HMS and HEC-RAS modeling.” – Brandon Hilbrich, PE, HDR Inc.
Gauge Adjusted Radar Rainfall (GARR) from NEXRAD data was provided by SARA and applied to the model. In addition, each subbasin parameter was incorporated to develop ﬂow hydrographs. 2D areas were incorporated to allow major storm drain systems to surcharge and capture overland ﬂow characteristics.
Due to the amount of runoﬀ volume generated by Harvey, several 2D areas were also added where overland ﬂow volume could transfer between major tributaries and through urban areas. “The ICM model allowed HDR to produce results for the entire USAR watershed in a short period of time, which would have taken several months longer to do with traditional HEC-HMS and HEC-RAS modeling.” Hilbrich noted.
By utilizing ICM as a complete catchment solution, SARA and HDR had a holistic model capable of advanced simulations on a single engine, without having to stitch together models from various sources. Additionally, the GARR was re-positioned directly over the USAR watershed, providing a truly worst-case scenario assessment for the USAR watershed if the same 50 inches had fallen over San Antonio.
Insight into Impact, and a City Reassured
As Harvey moved over the Harris County area, there were 14 deﬁnable storm waves based on the GARR data. In replicating those weather patterns over San Antonio, each wave would have been considered a near 100-year event. The most intense instance, wave 11, would have lasted 24 hours and would have produced 19 inches of rain in 10 hours. As a result, the City’s Olmos Dam would have received 79% of TCEQ 72-hour Probable Maximum Precipitation (PMP), been overtopped by nearly six feet, and have spilled onto San Antonio’s major highway, HWY 281. Assuming no rainfall following the event, the highway would have been inundated for 11 days and the dam would have taken 12 days total to empty.
By referring to the model results, SARA’s Graham was able to explain to City Council how Harvey ﬂooding would come in pulses. “The danger is being caught out in the middle of a storm,” Graham said.
The study results also helped identify three critical facilities that would have been susceptible to ﬂooding.
The good news? San Antonio would not have been as severely impacted by this type of storm as Houston.
“It’s terrible. It would be an event of record, but I don’t think it would have been quite like what Harris County saw,” Graham stated to City Council.
The geography of the San Antonio area – 1,500 feet of fall across Bexar County vs. 300 feet in Harris County, to be exact – would have produced “ﬂashy” inundation as opposed to prolonged ﬂooding. The city’s tunnel system would have helped mitigate impacts, however, street and residential areas would have likely seen more extensive ﬂooding.
In summarizing key takeaways from the simulation, HDR noted, “Overtopping velocities at crossings resulting in high water rescues would be the overwhelming concern.” And because of the analysis done by SARA and the HDR team, the City is now aware of these impacts. As Hilbrich put it, “The study illustrated the usefulness of the ﬂood warning model in the ICM modeling software and its ability to model not only real-time data, but hypothetical extreme events to provide a better understanding of the City’s preparedness.”