Theme for 2014:
Estimating Probable Maximum Precipitation in Mountainous Watersheds
University of California, Davis
|On this page:|
Why this theme?
Dam failures are considered "low-probability, high-loss" events because of the potential for massive death and destruction of communities located downstream. One major focus in designing a dam is performing a risk assessment of the potential causes of failure. All dams must be designed to prevent water overtopping the dam. The primary cause of dams failing from overtopping has been inadequate capacity of the dam's spillway, that is, more water coming into the reservoir behind the dam than the spillway can release. The water has to go somewhere so it is over the top.
Designing the dam's spillway must include the ability to pass the theoretical maximum flood, which is termed the "probable maximum flood" or PMF. To estimate the PMF requires an estimation of the theoretical maximum precipitation, which is called "probable maximum precipitation" or PMP. The hydrologists and meteorologists responsible for estimating PMP are working to update their tools and information used for this work. There are two main types of approaches for estimating PMP: Hydrometeorological and Statistical. The 2014 Symposium is focused on hydrometeorological approaches.
Special Recognition Award
The 2014 Special Recognition Award was presented to Hydrologic Engineering Center (HEC) of the U.S. Army Corps of Engineers for their history of of providing excellent public service and exemplifying the best in government agencies. See the award language and hear the presentation on the Symposium's Hydrologic Engineering Center award page.
2014 Photo Gallery
We took pictures at the 2014 Symposium — our 20th year! — to memorialize the event held for the last time at Freeborn Hall, UC Davis. See the pictures …
Comparing American River PMP Estimates to Historical Floods
- Download presentation [PDF, 0.9 MB]
The California-Nevada River Forecast Center (CNRFC) hydrologic models were used to estimate Probable Maximum Precipitation (PMP) values using the historical precipitation and temperature patterns from the 1986 and 1997 floods. The historical precipitation patterns were uniformly scaled up to match the U.S. Army Corps of Engineers (USACE) 3-day volume of the Probable Maximum Flood (PMF). Total precipitation required for the 1997 pattern was similar to the USACE PMP total. The 1986 pattern required more precipitation than the USACE PMP primarily because it was a colder event, and the entire watershed was not producing runoff. The 1986 precipitation totals were very similar to the USACE PMP when the precipitation pattern was combined with warmer temperatures.
Peak and 1-day flows at Folsom Reservoir were quite different. The 1997 pattern had a much higher 1-day flow than the 1986 pattern and the USACE PMF. It appears that this is primarily due to precipitation timing; most notably on the South Fork of the American River. Antecedent conditions and precipitation distribution also played minor roles.
Maximization of Historical Severe Precipitation Events over American, Yuba and Feather River Basins
- Download presentation [PDF, 3.9 MB]
A new approach to the maximization of precipitation over American, Yuba and Feather river basins in California was developed. The first step in this approach is to select all available historical severe storms over the California region that encompasses the three target watersheds, and then to reconstruct the detailed atmospheric conditions of those selected severe storms over the modeling domain. A byproduct of this exercise is the validation of the regional atmospheric modeling over the modeling domain by means of historical observations. Once this exercise was performed, it became clear that all of the historical severe storms over the target watersheds were due to atmospheric rivers.
Consequently, the new approach to the maximization of historical precipitation over each of the target watersheds is to estimate by numerical simulations by a regional atmospheric model (MM5 regional model) the optimal location and the optimal orientation of the historical atmospheric river event, corresponding to each selected historical storm event, with respect to each watershed in such a way to produce the maximum basin average precipitation for a selected duration over the target watershed for the particular historical event. After the determination of the optimal location and orientation of the atmospheric river (with respect to each target watershed) for a particular historical severe storm event, the moisture flux from the atmospheric river to the target watershed is maximized in order to obtain an estimate of the basin average maximum precipitation for a particular duration over the target watershed for the particular historical storm event.
For each target watershed, the above procedure was repeated for 72-hour storm duration for 61 severe historical storms during the 1950-2011 period in order to develop an ensemble of 61 72-hour basin average maximum precipitation estimates for each watershed. The maximum of those 61 estimates is taken as the historical 72-hour basin average maximum precipitation estimate for a particular watershed.
The Meteorology of Extreme Orographic Precipitation in California — A Synthesis as of 2014
- Download presentation [PDF, 2.1 MB]
During the past decade, a wide range of insights about the character and causes of extreme orographic precipitation in California has emerged, based on our growing understanding of the presence, mechanisms and impacts of "atmospheric rivers" in the extratropical atmosphere.
Atmospheric rivers (ARs) are constantly moving and evolving pre-cold-frontal jets of intense low-level water vapor transport thousands of kilometers long but only about 500 km wide, and are responsible for roughly 90% of all atmospheric vapor transport outside the tropics. When an AR reaches and encounters the Coastal Ranges and Sierra Nevada of California, the fast moving, moisture-laden air contained therein generally flows up and over the ranges, cooling, condensing and yielding almost ideal conditions for intense and sustained orographic precipitation. The resulting orographically driven storms are key players in many important weather, hydrologic and ecological processes in the state. The intensities, storm totals, geographical distributions and impacts of AR storms in California are determined by many factors, including among the most straightforward:
- Numbers of ARs making landfall each year
- Amounts of vapor being transported by the ARs
- Direction of transport in the AR relative to perpendiculars to mountain ranges (for maximum uplift)
- Duration of AR passage overhead of a given location
- Temperature of an AR as a determinant of snowline altitudes
- Stability of the atmosphere within which the AR is embedded
- Closeness of the air in the AR to saturation (how much uplift is needed to drive intense precipitation)
- Presence or absence of a resulting Sierra Barrier Jet, and
- Antecedent soil-moisture conditions
Much research has been conducted in recent years to expand and exploit our knowledge of these AR storms, and much more is in progress. Given the developing understanding of the AR phenomenon and its impacts in California, several noteworthy new activities are underway.
USACE Extreme Storm Team and Hypothetical Storm Analysis
- Download presentation (with notes) [PDF, 2.2 MB]
Extreme precipitation (and the resulting runoff) is an important component for the USACE Dam and Levee Safety program. Historic and hypothetical floods are used for design of critical infrastructure. The U.S. Army Corps of Engineers Extreme Storm team has been tasked with cataloging major historic precipitation events (those that have occurred after current Probable Maximum Precipitation (PMP) reports were developed), performing site specific PMP studies, and updating Corps guidance for hypothetical storm analysis.
The historic storm database is being populated using information from existing historic precipitation studies, radar rainfall data, and interpolated rain gage data. Information for the storm event, the raw data, depth-area duration information, and geographical information, is included in the database.
These storms will be available for site specific PMP studies and they could be used for other analyses, area-reduction curves, whatif scenarios, or for calibration events for rainfall runoff modeling. The Extreme Storm team has completed three site specific PMP studies and four studies are planned or in progress.
New tools are under development at the Hydrologic Engineering Center (HEC) to support the USACE Dam and Levee Safety program.
HEC-MetVue is a new tool that can be used to visualize and manipulate historic or hypothetical storms; increasing/decreasing precipitation depths and moving/rotating the storm.For dam safety studies, a required product from the hydrologic modeling analysis is a reservoir stage frequency curve that extends out to the stage resulting from a Probable Maximum Flood (PMF) event (with uncertainty included). The reservoir stage frequency curve is part of the information used to determine whether a reservoir is hydrologically deficient and needs remediation. In most cases, the only readily available information to define the reservoir stage frequency curve is historic measurements of reservoir stage and there is not enough data to define the curve out to the PMF event.
New Monte Carlo simulation capabilities are being added to a number of HEC software that will support a stochastic simulation where hundreds of thousands of events are generated. The Monte Carlo simulation could be used to define the reservoir stage frequency curve using meteorologic, rainfall-runoff, and reservoir operation simulations.
The Watershed Analysis Tool, HEC-WAT, is a developing piece of software that can be used to "link" HEC software together in a user-defined compute sequence. Specifically for the case of generating reservoir stage frequency curves, an uncertainty analysis simulation is possible where HEC-WAT samples precipitation events, passes the precipitation to the Hydrologic Modeling System (HEC-HMS) for rainfall-runoff simulation, and then flows are passed to the Reservoir System Simulation program (HEC-ResSim) for simulation of the reservoir operations.
Probable Maximum Flood Using HMR 59 for Piru Basin
- Download presentation (with notes) [PDF, 1.2 MB]
Pyramid Dam was constructed as part of the State Water Project, a system of dams reservoirs, pumping plants, power plants, aqueducts, and tunnels that conveys water through California. Pyramid Dam and Lake are located within Los Angeles County in Southern California, adjacent to Interstate 5, about 50 miles northwest of Los Angeles. The dam is 400 feet high with a crest length of 1,090 feet. The lake holds approximately 170,000 acre feet and provides for recreation, emergency storage, acts as a forebay to Castaic Power Plant, and is an afterbay for Warne Power Plant. The water surface elevation generally ranges between elevations 2,578 (maximum) and 2,570 (minimum) feet and is not drawn down for flood control.
The dam sits within the Piru Creek drainage, a 423 square mile watershed that is tributary to the Santa Clara River. The reservoir itself collects runoff from about 295 square miles with the balance of the water shed downstream of the dam. All the water stored in Pyramid Lake is "project water" and is delivered through Warne Power Plant. Runoff is not stored within the lake.
Freeboard and spillway size were based on the Probable Maximum Flood study performed prior to construction (DWR, 1968). Since then, there have been revisions to the PMF to reflect updated design criteria — HMR 36 and HMR 58. HMR 58 calculation procedures for the general storm were the latest to be used for Pyramid Reservoir. Probable Maximum Precipitation (PMP) estimated from HMR 36 totaled 28.1 inches for a 72-hour storm and 31.5 inches using HMR 58. The PMP total from original design was 26.1 inches. For total runoff, the original PMF study resulted in estimated inflow of 180,000 cubic feet per second (cfs). Using HMR 36, also results in a peak inflow of 180,000 cfs, while the peak inflow from HMR 58 is 278,000 cfs.
These changes in the PMP and subsequent PMF have resulted in the estimated freeboard originally starting at four (original design) to five feet using HMR 36 to overtopping by 1 to 2 feet using HMR 58. These latest results direct resources to address spillway changes or changes in operations to accommodate the increased runoff demand.
Calculating Probable Maximum Precipitation (PMP) in Complex Terrain — Updating HMR 59 for Piru Basin in Southern California
- Download presentation [PDF, 3.9 MB]
Applied Weather Associates (AWA) has completed a site-specific PMP study for the Piru Creek drainage basin in southern California. The study was supported by California DWR for Pyramid Dam and United Water for Santa Felicia Dam. The site-specific PMP study provided refined and updated PMP values from those published in Hydrometeorology Report 59 (HMR 59).
HMR 59 is the newest of the HMRs published in 1999. The site-specific PMP study incorporated a better understanding of meteorology, updated storm datasets, improved spatial analysis and introduced an updated method to quantify topographic effects. NOAA Atlas 2 was used in HMR 59. NOAA Atlas 14 was used in the site-specific study. Major issues with HMR 59 are:
- The Storm Separation Method (SMM) used is highly subjective and not reproducible
- Inconsistencies and calculations errors
- It is unclear how actual storm data were used to develop PMP values
This presentation will discuss how the site-specific PMP study was performed with an explicit discussion of how the Orographic Transposition Factor (OTF) was developed and implemented. A description of the Storm Precipitation Analysis System (SPAS) will be provided along with the use of the HYSPLIT trajectory model used to determine moisture inflow vectors from the Pacific Ocean.
PMP, Extreme Storm Probabilities, and Dam Safety: New Data, New Methods, Time to Update?
- Download presentation [PDF, 5 MB]
Abstract:Probable Maximum Precipitation (PMP) has been used as the storm rainfall design standard for estimating the Probable Maximum Flood (PMF) at critical facilities, including high-hazard dams and non-coastal nuclear power plants in the United States. However, this PMP design standard is being questioned and challenged, with some design modifications being based on risk, with cost savings realized.
The Bureau of Reclamation, Army Corps of Engineers, and Federal Energy Regulatory Commission are moving towards or are now fully utilizing risk-informed decision making to evaluate the safety of high-hazard dams. The Bureau of Reclamation has recently developed risk-based Inflow Design Flood (IDF) standards that rely in-part on extreme precipitation probability estimates for watersheds. Some states have embraced risk analysis, while most rely on PMP estimates or percentages of the PMP and PMF. FEMA recently published IDF guidelines that discourage the use of PMP or PMF percentages.
The current state of practice on PMP and risk is in a state of flux. Data and methods used to estimate PMP, as published in National Weather Service generalized Hydrometeorological Reports, are severely outdated. There are also diverse needs for extreme storm rainfall and probabilities, up to and including PMP. States may be open to risk-based products if they are clearly defined, simple to use, and in Federal guidelines. Guidance and training is clearly needed. The time is ripe for an update of extreme storm products. Several factors point in this direction:
- Questions asked today by decision makers and regulators about risk, climate change, and extremes
- Increased use of risk analysis for dam safety by major Federal agencies
- Improved knowledge on extreme storm mechanisms, data, and advances in numerical modeling and statistics
This presentation highlights some ongoing work and issues on extremes and PMP: activities of the Extreme Storm Events Work Group; the current industry state-of-flux on extreme precipitation needs; some collaborative research efforts; and example Reclamation extreme precipitation frequency studies in the Trinity Alps and Sierra Nevada, California.