A Vision of Future Observations for Western U.S. Extreme Precipitation Events and
Flooding: Monitoring, Prediction and Climate
M. Ralph1, M. Dettinger2, A. White1, D. Reynolds3, D. Cayan2, T. Schneider4, R. Cifelli5, K. Redmond6, M.
Anderson7, F. Gherke7, K. Mahoney8, L. Johnson1, S. Gutman9, V. Chandrasekar10, A. Rost11, J. Lundquist12, N.
Molotch13, L. Brekke14, R. Pulwarty15, J. Horel16, L. Schick17, A. Edman18, P. Mote19, J. Abatzoglou20, R. Pierce21,
G. Wick1, S. Matrosov8
1
NOAA/Earth System Research Laboratory/Physical Sciences Division, Boulder, Colorado
2
U.S. Geological Survey, Scripps Institution of Oceanography, La Jolla, California
3
NOAA/National Weather Service/Monterey Weather Forecast Office, Monterey, CA
4
NOAA/NWS/Office of Hydrologic Development, Boulder, Colorado
5
Cooperative Institute for Research in the Atmosphere, Fort Collins, Colorado
6
NOAA/Western Region Climate Center, Reno Nevada
7
California Department of Water Resources, Sacramento, California
8
Cooperative Institute for Research in Environmental Sciences, Boulder, Colorado
9
10
NOAA/Earth System Research Laboratory/Global Systems Division, Boulder, Colorado
Colorado State University, Department of Electrical and Computer Engineering, Fort Collins, Colorado
11
NOAA/NWS/ National Operational Hydrologic Remote Sensing Center, Chanhassen, Minnesota
12
University of Washington/Dept. of Civil and Environmental Engineering, Seattle, Washington
13
University of Colorado at Boulder, Geography Department, Boulder, Colorado
14
U.S. Bureau of Reclamation, Technical Services Center, Denver, Colorado
15
NOAA/OAR/Climate Program Office, Physical Sciences Division, Boulder, Colorado
16
University of Utah, Department of Meteorology, Salt Lake City, Utah
17
.S. Army Corps of Engineers, Seattle, Washington
18
NOAA/NWS Western Region Headquarters, Salt Lake City, Utah
19
Oregon State University, Oregon Climate Change Research Institute, Corvallis, Oregon
20
University of Idaho, Department of Geography, Moscow, Idaho
21
NOAA/NWS/San Diego Weather Forecast Office, San Diego, California
Contact: F. Martin (Marty) Ralph at marty.ralph@noaa.gov, 303-497-7099 (office)
Provided to the Western States Water Council on October 5th, 2011 in Idaho Falls, Idaho
Capsule Summary
This document provides a vision of 21st Century observations for tracking, predicting and ultimately
managing the occurrence and impacts of major storms in the west. The vision recommends innovations
and enhancements to existing monitoring networks for rain, snow, snowmelt, flood and their
hydrometeorological precursor conditions over land and ocean. These include new radars to monitor
winds aloft and precipitation, streamgages, Snotel enhancements, and more, as well as entirely new and
only recently-possible observational tools. The document presents the scientific and management
motivations and contexts for this vision, describes its key components and an implementation strategy,
along with expected benefits. This document supports a Resolution of the Western States Water Council
and an MOU between NOAA and the Western Governors Association, both addressing extreme events.
Monitoring Atmospheric Conditions
That Fuel Extreme Precipitation & Flood
New Snow and Streamflow Monitoring
for Better Snow Melt Forecasts
Offshore Monitoring to Extend Forecast
Lead Times for Extreme Precipitation
Atmospheric River Observatories to Fill
Largest Single Monitoring Gap
st
Fig. E1. Four broad conceptual elements of the vision for 21 Century monitoring in the Western US. Each of the
frames will appear again later in the document at full scale and will be explained and motivated in detail.
2
1.
The Challenge
This document was requested for inclusion in a report from the March 2011 workshop on extreme events
sponsored by California’s Department of Water Resources, Western States Water Council (WSWC), and
the Western Governors’ Association (WGA).
Key elements of this plan were presented at the July 2011 WSWC meeting in Bend Oregon. The
following day, July 29, 2011, a Resolution was approved by the WSWC as an official “position” for
consideration by the WGA. Key elements of this Resolution are directly related to this Vision document,
and provides the following overarching guidance quoted from the Resolution titled
“RESOLUTION of the WESTERN STATES WATER COUNCIL supporting FEDERAL
RESEARCH AND DEVELOPMENT OF UPDATED HYDROCLIMATE GUIDANCE FOR
EXTREME METEROLOGICAL EVENTS”
“WHEREAS, advances in weather forecasting research, such as that of NOAA’s Hydrometeorological
testbed program on West Coast atmospheric rivers, demonstrate the potential for improving extreme event
forecasting at the operational time scale; and”
“WHEREAS, WGA and NOAA signed a memorandum of understanding on June 30, 2011, regarding
state adaptation to climate variability and change that focuses on climate extremes, variability and future
trends as they relate to disaster risk reduction and improved science for coastal and marine resources
management”
“BE IT FURTHER RESOLVED, that the Western States Water Council supports development of an
improved observing system for Western extreme precipitation events, to aid in monitoring, prediction,
and climate trend analysis associated with extreme weather events; and
BE IT FURTHER RESOLVED, that the Western States Water Council urges the federal government to
support and place a priority on research related to extreme events, including research on better
understanding of hydroclimate processes, paleoflood analysis, design of monitoring and change detection
networks, and probabilistic outlooks for climate extremes.”
A copy of the full Resolution is provided at the end of this document. The purpose of this White Paper
is to provide a vision of next generation observations that could aid in monitoring, prediction and
climate understanding associated with extreme weather events that affect either water supply or
flooding in the semi-arid Western United States. The Vision expressed in this document represents
a next step toward implementation of the guidance contained within the WSWC Resolution noted
above. At the request of the WSWC, a half-day tutorial will be provided to WSWC members and staff in
Idaho Falls, Idaho on 5 October 2011 on the uses of, and needs for, current and future observations in
weather and hydrology prediction, as well as for climate monitoring, along with a synopsis of the
Observing System Vision described here. The presenters are Dave Reynolds (NWS-Weather), Harold
Optiz (NWS-Hydrology), Jessica Lundquist (Univ. of Washington), and Marty Ralph (NOAA-Research).
3
Key societal imperatives related to extreme precipitation, flooding and water supply:
-
-
-
-
Provide the necessary flood protection while ensuring adequate water supply in an environment
characterized by extreme events
Provide enough warning lead time with quantified forecast uncertainty to enable preemptive
actions by emergency preparedness officials, including out to 10 days where feasible
Address risks due to aging flood control infrastructure (e.g., the Howard Hanson Dam crisis is an
example as are levees across the west)
Avoid a “Katrina-of-the-West” scenario in which an extreme event disrupts or overwhelms
existing operations or aging infrastructures catastrophically, which studies such as ARkStorm
have identified as a significant risk (with projected damages that could exceed $500 Billion)
Address risks associated with the effects of warming on the water cycle and atmospheric
processes, which may include:
o decreasing overall snowpack (i.e., the summer water supply)
o shortening the snow season and thus extending the flood season earlier into spring
o possible expansion of flood season on west coast to earlier in the fall or later into spring
o increasing flood risk with warmer, and possibly more intense, storms
Mitigate potential impacts of climate change by providing a better information base for developing
adaptation strategies such as forecast-based reservoir operations that can enable greater water
supply while maintaining maximum flood control using existing structures (Fig. 1)
Provide low-cost alternatives for consideration in planning major new flood control and water
supply infrastructure
Protect endangered species, such as salmon, aided by potential benefits to water supply
Improve next-generation meteorological and hydrologic models using revolutionary observations
Address increase wild fire activity, which can lead to flash flooding and debris flows
During a 17-year period studied by Pielke et al. (2002), the Western States of WA, OR, CA, ID, NV, UT,
AZ, MT, WY, CO, NM, ND, SD, NE, KS, OK, and TX experienced $24.7 Billion in flood damages, an
average of $1.5 Billion annually. California, Washington and Oregon combined to account for $10.6 B
(46%) of this regional total (see Appendix A).
A key challenge for the next decade is "lead time -- lead time -- lead time". The bar has been raised. It
used to be adequate to provide a few hours of lead time. Today, however, community leaders not only
have to be prepared from a safety standpoint, but they need to be able to minimize the costs of taking
preparedness actions, e.g. by shifting work schedules so that preparatory work can be done on regular
time versus overtime. Increasingly community leaders need lead times out to 7, 10- even 14 days.
To address these major challenges will require innovative solutions that, in turn, will require a strong
scientific enterprise of monitoring, observations and science. Solutions will depend upon better
understanding, tracking and prediction of the causes of extreme events and how they might change in the
future. Solutions will also depend on innovative engineering efforts to develop capabilities that can costeffectively fill gaps in observations, forecasts and related services that to support vibrant economies,
healthy ecosystems and reliable water and living resources.
4
Fig. 1. Schematic illustration of possible outcomes associated with early spring runoff depending upon whether the decision is
made to release water to preserve flood control space for use in a potential late season flooding storm or to store water in
expectation of summertime water-supply demands.
2.
The Context
a) Lessons learned from recent projects
The main drivers for a future observing network are needs related to real-time monitoring, predictions
(from minutes to days for flooding, and to seasons for water supply), and climate trend analysis. The
network described here draws upon many existing documents describing needs and requirements,
including
-
NRC reports (Flooding on Complex Terrain; Network of Networks; GPM Satellite system),
Interagency needs assessments, e.g., Workshop on non-stationarity…(2010), USBR Science and
Technology Program (2011), USWRP Workshop on QPF (2005)
Planning for Integrated Water Resources Science and Services (IWRSS)
Planning for a National Water Center,
Experience from NOAA’s Hydrometeorological Testbed (HMT; hmt.noaa.gov),
Understanding of the joint roles of atmospheric rivers in extreme events and water supplies in the
west (Ralph and Dettinger 2011; Dettinger et al. 2011)
The ARkStorm emergency preparedness exercise in California
Ongoing NOAA Regional Integrated Science and Assessment studies
Experience from State Climatologists
5
-
Observing system gap analysis for atmospheric rivers (ARs) by NOAA’s Unmanned Aircraft
Systems program, and
NOAA’s radar network planning.
This paper takes advantage of a series of significant recent advances in observing system technology,
research findings, experience gained by testing prototype observing systems in NOAA’s HMT, the
National Integrated Drought Information System (NIDIS) and the North American Monsoon Experiment
(NAME; Higgins et al. 2006), as well as advances in numerical weather prediction, hydrologic forecasting
and real-time communications. Interagency needs assessments have also been produced relating to these
issues, including a report focused on extreme precipitation forecasting (Ralph et al. 2005), and more
recently a major report on dealing with issues related to quantifying extreme event probabilities
(Workshop on Nonstationarity, Hydrologic Frequency Analysis, and Water Management; Olsen et al.
2011, http://www.cwi.colostate.edu/publications/is/109.pdf).
One broad conclusion of observing system research has been that monitoring of the atmospheric column,
and not just the surface meteorology, and monitoring the atmosphere over the Pacific (where key weather
features take shape before moving ashore) is vital to understanding/predicting precipitation intensity and
form (rain/.snow) in the western states.
Another vital requirement is that data from key elements of this network, especially those that provide
observations aloft and/or offshore, need to be assimilated into numerical weather prediction models to
reap maximum rewards. Existing models either already are able to assimilate key observations, e.g., wind
profiler, GPS-met and dropsondes data, or methods to assimilate other data can be developed. Many direct
uses of these same data exist, even without this data assimilation but maximum benefits will accrue from
the combination of both the direct uses and model assimilation. These several findings are at the root of
the vision presented here.
This vision was developed by a team of experts from federal, state, and local agencies, the academic
community in hydrometeorology, hydrology, and climate, and representatives of some of the western
water-related systems that require information on extreme precipitation, flooding and water supply.
Lessons learned from the recent flood control crisis in Washington State, where the flood protection from
Howard Hanson Dam above Seattle was seriously compromised by seepage that developed after a record
storm, are incorporated (White et al. 2011; Appendix D). Similarly, experience from an emergency
preparedness scenario, named ARkStorm (Appendix M), that explored the potential impacts of a
devastating series of atmospheric rivers hitting California, also informed this report. The ARkStorm
exercise concluded that over $500 billion in economic impacts could occur from a single major storm
sequence in California through damages, business disruption, and other dimensions, and that significant
loss of life could occur (Porter et al. 2010). Historical flood damages provide much the same perspective.
The vision presented here was developed within, and recognizing, the context of several recent and
ongoing efforts to enhance observing capacities and networks related to extreme precipitation events in
the West:
6
Enhanced Flood Response and Emergency Preparedness (EFREP; led by CA Dept. of Water
Resources, NOAA and Scripps Institution of Oceanography; Appendix E): Development and deployment
of statewide monitoring, modeling and decision-support programs drawing from, and making operational,
key findings from HMT-West, for better detection, monitoring and prediction of ARs and their impacts.
A key component is a “picket fence” of four coastal atmospheric river observatories, a statewide soilmoisture network, snow-level radars, and land-based monitoring of vertically integrated water vapor
(IWV), with associated decision support capabilities. Out of a total of four “tiers” (i.e., levels of
complexity, investment and protection), the ongoing implementation covers tiers 1 and 2 for California
(93 field sites in all), and will be complete in 2013. A brochure from CA DWR is available online at
http://www.esrl.noaa.gov/psd/atmrivers/projects/pdf/Advanced%20Monitoring%20Network%20FloodER
%20Prgrm.pdf. The lessons there apply directly to Washington and Oregon. The top two tiers, 3 and 4,
are broader in scope and cost, with benefits stretching across much of the Western U.S., and are included
in this white paper.
Deployment of a NEXRAD on the Washington Coast (led by NOAA/NWS): The Next-Generation
Radar (NEXRAD) network consists of over 150 radars across the continental U.S. and territories.
Recently a major gap in NEXRAD radar coverage off of the Washington coast has been identified. This
has led to the installation of a NEXRAD radar on the Washington coast. The site will be operational late
in 2011.
The Winter Storm Reconnaissance Program (led by NOAA/NWS): Using NOAA’s G-IV and military
C-130s, for the last few years an annual monitoring effort has been conducted for several weeks over the
Pacific Ocean to improve forecasts across the US. These efforts have demonstrated techniques that could
be refined and extended to provide early characterizations of many more extreme events affecting the
Western U.S.
Enhancements of the SNOTEL high altitude network (led by NRCS): This includes the addition of
soil moisture at a number of SNOTEL sites, as well as selected other instrument upgrades.
Enhancement of the Climate Reference Network (USCRN) to include soil and humidity
measurements (led by NOAA): Supported by the National Integrated Drought Information System
(NIDIS), by September 2011, all 114 CRN locations had soil probes and associated data loggers and
relative humidity instruments installed. By the end of FY 2011, subject to funding, all 114 USCRN
stations in the conterminous United States will have received soil moisture probes and RH sensors. (See
FY 2010 USCRN Annual Report at http://www.ncdc.noaa.gov/crn/annual-reports.html, including its Fig.
6)
Deployment of Regional Climate Reference Network (RCRN; led by NOAA): The USRCRN
consists of about 430 new stations nationally that meet siting and instrumentation standards of CRN.
(These sites do not include soil moisture or relative humidity sensors.) A pilot project involving 71 grid
points in the Four Corner states is wrapping up, and another 76 grid points are now being surveyed by
WRCC in the five westernmost continental Pacific states. Initial emphasis is on western states to serve
NIDIS needs. See http:www.ncdc.noaa.gov/crn/usrcrn/.
7
Addition of dual-polarimetric capability to NEXRAD (led by NOAA): Beginning in 2011, NEXRAD
radars will be upgraded to include dual polarization capability. Dual polarization offers several
advantages compared to current NEXRAD single-polarization radar systems, providing additional
information about the size, shape, and orientation of precipitation particles. This information can be used
to more accurately identify the type of precipitation (e.g., hail vs. rain), correct for signal loss
(attenuation) in heavy precipitation, and more easily identify and remove non-meteorological radar
echoes. Dual polarization phase measurements allow for rainfall estimation that is much less affected by
problems related to absolute calibration of the radar system, to signal attenuation effects, as well as to
partial beam blocking.
Testing of Unmanned Aircraft Systems (UAS) for offshore weather data collection (led by NOAA, in
close partnership with NASA): The “Winter Storms and Pacific Atmospheric Rivers” (WISPAR)
experiment was conducted in early 2011 over the eastern Pacific Ocean. The ability of the unmanned
NASA Global Hawk aircraft to carry a NOAA dropsonde system that directly measured atmospheric
profiles within atmospheric rivers over the ocean was demonstrated. Flights were up to 25 hours long,
and were coordinated with NOAA’s G-IV reconnaissance aircraft, which also sampled an AR near
Hawaii (Appendix H).
b) Meteorological context
Extreme precipitation is the primary cause of flooding in the region, and when these events occur in
winter they can substantially increase snowpack and thus water supply. Another major cause of extreme
runoff and flooding in some regions is snow melt in the spring and summer, which can be triggered by
unusually warm and sunny conditions, and/or warm, heavy rainfall.
The meteorological phenomena responsible for extreme precipitation varies across the region, as
illustrated in Fig. 2. Four primary precipitation-causing phenomena are highlighted here:
Atmospheric rivers (cool-season storms that come ashore from the Pacific; Appendix C)
Southwest monsoon (involving summer thunderstorms and remnant tropical storms)
Great Plains convective storms, i.e., thunderstorms (in spring and summer)
Upslope storms along the “front Range” of the Rocky Mountains (in spring)
8
Atmospheric
Rivers
(fall and winter)
Great Plains Deep
Convection
(spring and summer)
Spring Front Range
Upslope (rain/snow)
Southwest
Monsoon
(summer & fall)
Fig. 2. Schematic illustration of regional variations in the primary weather phenomena that lead to extreme precipitation and
flooding and contribute to water supply in the Western U.S. This has been developed with input from experts on these
phenomena who contributed directly to this document.
Although there is some overlap in the seasonality of these phenomena, they are relatively distinct, with
hydrometeorologically significant atmospheric rivers (ARs) occurring primarily from October through
March, spring upslope storms occurring from April to June, the southwest monsoon being almost
exclusively in July-October, and Great Plains deep convection from April to August.
The availability of COOP daily precipitation data from thousands of sites in the region going back 30
years or more, helps to clarify the geographic domains for each of these phenomena. For each COOP site,
the 10 historical days with the largest daily precipitation totals were identified (i.e., the top 10 out of
roughly 10,000 dates or more days in a site’s period of record). The season that had the largest number of
these “top-10” extreme precipitation events at each COOP site was then plotted on a map (Fig. 3a). While
the seasonal and geographic boundaries can blur or overlap to a degree (as in the Colorado Front Range),
the overall patterns are clear enough to illustrate the regionally and seasonally varying causes of extreme
precipitation across the West. Peak-streamflow dates (Fig. 3b) help to corroborate these patterns.
9
Fig. 3. (a) Seasonality of extreme precipitation events across the Western U.S. based on daily precipitation totals from
thousands of COOP observations (dots) covering at least 30 years each. The color of each site corresponds to the season of
the year when more of the top-10 daily precipitation events occurred than any other season. (b) Seasonality of annual peak
daily stream flows highlighting the geographic distributions of AR-, snowmelt and monsoon dominated regions.
c) Challenges associated with the regional geography
The geography of the west severely complicates both the monitoring and prediction of these extreme
events, including weather and climate time scales for the following reasons.
Many storms originate over the Pacific Ocean where there are major limitations in weather
observations, e.g., severely limited measurements of vertical profiles of winds, temperature and
moisture in ARs and the larger-scale weather systems within which they are embedded.
The presence of large and complex mountain ranges creates large differences in important weather
conditions over short distances, making adequate monitoring and forecasting much more
challenging.
Large areas of the mountainous west are uniquely vulnerable to future increases in flooding during
extreme events because over half of the largest (historical) daily precipitation totals occurred when
daily mean temperatures were between -3 and 0 deg C, meaning that a warming of 3 deg C would
greatly exacerbate flooding (Fig. 4; Bales et al. 2006)
Much of the weather observing infrastructure in place today is simply too sparse, outdated (e.g.,
most vertical profiling today depends on infrequent in-situ balloon measurements when higher
observation frequencies are needed), or is not deployed in an optimal manner for western US
10
applications (see Fig. F1 in Appendix F for an example of NEXRAD coverage gaps in the western
U.S.).
d) Selected requirements
-
-
-
Accurate hydrometeorological forcings for the next generation flood-forecasting and streamflow
models, including those being developed as part of IWRSS and the National Water Center, including
soil moisture conditions, snow pack, existing stream flow, base flow, precipitation inputs,
temperature, evapotraspiration, etc.
Accurate Quantitative Precipitation Estimates (QPE) in complex terrain
Accurate Quantitative Precipitation Forecasts (QPF) for extreme events
Nowcasts and shortShort-term high-resolution QPF over urban areas for water, stormwater, and
sewage management
6-h forecasts of the end of heavy rain over key watersheds for reservoir operations
1-5 day guidance/forecasts of the location and intensity of extreme events to support future forecastbased reservoir operations to optimize flood control and water supply
Improved situational awareness that provides enough lead time with quantified forecast uncertainty to
enable preemptive actions by emergency preparedness officials, including out to 10 days where
feasible
Seasonal guidance of the potential for both flood risk and water supply that can enable decision
makers to select optimal policy options
Accurate QPE and QPF for fire risk management and fire fighting
short term (minutes to a few hours) forecasts of rain rates that initiate debris flows over recent burn
scars
Increase in the use of GIS to pinpoint areas of concern to help local forecasters keep track of multiple
impact related events
Guidance on the potential risk of extreme events in a changing climate so as to inform major
infrastructure planning (e.g., Fig. 4 shows those areas for which warming of 3 degrees C would
transform snow accumulation in extreme events into liquid, and thus change the stream flow in ways
that could lead to great cool-season peak flows and possibly flooding at times of year currently not as
vulnerable at those times)
These requirements correspond closely with the information needs as outlined in the U.S. Bureau
Reclamation Science and Technology Program’s in the climate change and variability priority area,
specifically addressing the following needs related to Shorter Term Climate Variability (USBR 2011,
section 5.A.2) on time scales of days to years:
• Improved use of existing-quality weather, climate and/or hydrologic predictions in the development of
operations outlooks (e.g., improved use of forecast uncertainty through novel methods or tool
development)
• Development of superior-quality weather and climate and predictions relative to current information
products from Reclamation's forecast providers
11
• Enhanced communication of uncertainties and risks associated with weather, climate predictions in the
development of Reclamation's operations outlook.
Fig. 4. Areas for which an increase in cool-season extreme precipitation and flooding would most likely occur in the case of a
3 deg C warming of the regional climate during storms. From Bales et al. 2006.
12
3.
The Vision
a) Overview
This section outlines a strategy that would lead, over a 6-yr period, to a fully functional modern
monitoring system optimized for the west over a 6-year period. It would utilize existing organizations
with the suitable expertise, infrastructure and missions to perform the implementation and then to carry on
long-term operation, maintenance and ongoing improvements. While many details remain to be worked
out, in principle, capabilities (i.e., knowledge, skills, tools) currently exist to execute this strategy if
adequate funding becomes available. The vision is an expert opinion on what can be done to address these
major societal challenges based on existing and emerging technologies and techniques, but does not
identify funding sources.
The observational tools and related efforts are based on
A set of requirements (often overlapping or complementary to one another) associated with
detecting the phenomena that cause extreme precipitation as well predicting them hours to days
beforehand and monitoring signatures of climate change
Progress in numerical modeling of both the atmosphere and of hydrologic conditions, and the need
for high resolution observations and predictions of key hydrometeorological conditions that drive
them
Several years of testing and prototyping new tools and methods in the NOAA HMT-West project
and in other experimental settings and campaigns, such as CalWater and the Yosemite HighAltitude Hydroclimate Network, have provided many lessons about what works and does not
work, and how much effort is involved
Emergence of new observational technologies (summarized below), information systems (e.g.,
GIS), real-time and reliable communication from remote sites, and better understanding of the
physical phenomena themselves.
Our vision consists of
A broad land-based network (Fig. 5) customized to monitor conditions in the extreme precipitation
regimes identified in Fig. 2, with observations focused on water vapor content and its transport,
QPE, snow level, soil moisture and dust- and rain-on-snow effects. A key strategy is monitoring
conditions aloft, below roughly 3 km altitude. This is the layer most difficult to observe using
satellite or scanning radars, and yet is where some of the most important meteorological
conditions, in terms of extreme precipitation, reside. It is where most of the water vapor and
clouds are concentrated exists, where airflows interact with terrain, and where the atmospheric
“boundary layer” conditions set the stage for heavy precipitation.
An enhanced mountain array (Fig. 6), also land based, focused on snow conditions on the ground
for tracking melt (Appendix K & L)
Offshore monitoring (Fig. 7), with an emphasis on AR conditions and assimilation of data into
weather models from the critical upwind region over the eastern Pacific from which so many
13
western storms approach. States and communities need time to prepare, to minimize the cost
impact of preparations/recovery from major storms in any ways they can. Offshore monitoring can
help to provide such lead time before storms and flooding in many parts of the west. This need
not be a perfect forecast...but a heads up that a big storm is coming and an initial statement of the
characteristics of storm, e.g., cold, wet, windy, a once in 10 year event, or a typical big winter
storm.
Fig. 5. Schematic network of new sensors (land-based) to improve monitoring, prediction and climate trend detection for
hydrometeorological conditions that create extreme precipitation and flooding.
The land-based network (Figs. 5 and 6) described here would ideally include
100 new low-mid altitude soil moisture observing sites (Fig. 5, Appendix J)
125 existing high-altitude sites with new snow-related data (Fig. 6, Appendix K)
100 new GPS-met observing sites (Appendix I)
25 snow-level radars
25 wind profiling/ARO sites (section 3c and Appendix E)
14 C-band scanning radars (Appendix F)
10 X-band scanning radars or mini CASA networks (Appendix F)
14
Fig. 6. Existing SNOTEL sites color coded to their altitude range. Red ovals highlight regions where a subset of the existing
SNOTEL sites would have additional sensors emplaced to support better spring snow melt monitoring and prediction, or where
new sites would be needed to broaden the altitude range of coverage. For more information, see Appendix K.
Fig. 7. Conceptual elements of an offshore observing network to improve monitoring, prediction and climate trend detection
for hydrometeorological conditions that create extreme precipitation and flooding.
15
b) Overarching gap: Monitoring water vapor transport – fuel for extreme precipitation
An overarching gap in observations associated with extreme precipitation events is monitoring of water
transport, i.e., the “fuel” for the precipitation. Existing observations typically do not measure winds and
water vapor aloft except in a handful of locations twice per day, and yet these are the key variables that
determine the water vapor transports that fuels precipitation. The observing system that is proposed here
is based on filling this gap, and designed in many settings on the specific characteristics of ARs that are
most important to severe storms and that have now been well documented, including their width, depth,
snow level, water vapor transport and orientation. Studies have shown that half of the hour-to-hour
variance on coastal mountain rainfall is due to variance in the upslope wind speed, direction and water
vapor transport at about 1 km above sea level, both within AR and beyond as well. In addition, it has
been shown that the direction of the low-altitude wind in ARs has a major controlling effect on exactly
which watersheds will receive the greatest rainfall in a given event (Ralph et al. 2003, Neiman et al.
2011), as is illustrated in Fig. 8.
When atmospheric rivers strike coastal mountains (Ralph et al. 2003)
Air ascends coastal mountains, water vapor condenses, heavy rainfall occurs
Details of the atmospheric river determine which watersheds flood
Fig. 8. Illustration of the role of coastal mountains in controlling the spatial distribution of heavy rainfall when an
atmospheric river strikes the coast. The orientation of the winds in the AR creates rain shadows that can shelter some regions
from the heaviest rainfall. In the case, where the “dividing streamline” crossed the Santa Cruz Mountains, it separated floodproducing rain to its west from moderate rain to its east that did not create flooding.
16
Fig. 9. Schematic of atmospheric river conditions and an associated observing strategy (known as an ARO) combining wind
profiler, GPS met, snow-level radar, and surface data. Because an average AR is about 400 km wide, the optimal horizontal
spacing of AROs is 200 km.
Based on a decade of development at NOAA and in HMT, a method has been invented that deploys
unattended sensors coupled with numerical weather model output to fill this gap (Fig. 9). Four
atmospheric river observatories (AROs) are now being deployed along the California coast as part of its
EFREP program and the technique was part of NOAA’s rapid response to the Howard Hanson Dam
crisis.
The present vision involves completing a west coast array of AROs to capture water vapor transport from
over the Pacific Ocean in ARs, along with a few more AROs farther inland to monitor water vapor
transports through preferred pathways into the intermountain west. Similar instrumental arrays (called
TARS) already exist along the border with Mexico, to support drug interdiction efforts by providing wind
profile information aloft for operation of tethered balloons carrying surveillance radars to detect lowflying aircraft. The present plan would develop tools to take advantage of this border array to help
monitor water vapor influx into the Southwest from the South, especially as part of the summer monsoon
which derives its moisture through transport from either Mexico or the Gulf of Mexico. The plan also
applies this technology along a line roughly 100 km east of the Front Range of the Rockies to monitor
water vapor influxes from the Great Plains towards the Rockies in upslope storms, and would be coupled
with snow-level radars in the foothills to monitor, in realtime, the fraction of vulnerable watersheds that
would be exposed to rain versus snow. This fraction a critical factor determining the mix of flooding
versus snowpack water storage from any given storm. Because water vapor transport from north to south
across the Canadian border tends to be very weak (due to the prevalence of dry cold continental air there),
no such array is proposed for that area.
17
c) Phenomenological and regional requirements and relevant observations
This section summarizes specific observational enhancements addressing gaps related to the phenomena
primarily responsible for extreme precipitation and streamflow.
•
Broad area coverage components (addressing all four meteorological phenomena that cause
extreme precipitation, plus spring/summer rapid snowmelt)
These observations represent low-cost sensors that would be of benefit across the entire Western
U.S.
Soil moisture (Appendix J)
GPS met for IWV (Appendix I)
“Bulk” snow measurements (depth, SWE, precipitation; Appendix K)
Air-snow interface (snow albedo, wind, dew point temperature, solar radiation)
Snowpack internal conditions (e.g., chemistry of dust, snow densities, Appendix L)
Additional stream gages
•
Atmospheric River-focused components
Another element of the network emphasizes horizontal water vapor transport in ARs, along with
snow level monitoring and enhanced QPE through gap-filling scanning radars over major urban
areas. The design also requires offshore monitoring to extend the range of predictions out to
several days.
Atmospheric river observatories (ARO) in the AR-impacted regions
Snow level radars
Scanning polarimetric radars over major cities (e.g., Seattle, Bay Area, Phoenix,Vegas)
Offshore monitoring (coastal scanning radars - C-band, aircraft, buoy-mounted wind
profilers, all taking advantage of satellites)
•
Monsoon-focused components
As indicated in Fig. 2, the southwestern U.S., including Arizona, Nevada, Utah, Colorado, and
New Mexico, are impacted by the precipitation associated with the North American Monsoon
(NAM). In particular, parts of southern Arizona and New Mexico receive over 50% of their
annual precipitation during the NAM season (Douglas et al. 1993; Adams and Comrie 1997). The
key gaps for monsoon monitoring include both water vapor transport from Mexico or the Gulf of
Mexico, plus scanning radars to fill very large gaps in current radar coverage.
ARO-like installations, but focused on Monsoon-related water vapor transport and taking
advantage of the existing TARS profiler network
Gap-filling scanning radars with polarimetric capability over major urban areas
•
Great Plains convection components
The Great Plains have one of the best existing radar networks for monitoring broad area storms
(although gaps for monitoring tornadoes and severe thunderstorms remain). The present network
vision emphasizes water vapor distribution and transport, as well as soil moisture.
18
-
Gap filling polarimetric radars in flood-prone front-range watersheds
Enhanced lightning network
•
Front Range upslope storms
Upslope storms primarily occur along the east slopes of the Rocky Mountains, extending from
Montana to New Mexico (Fig. 2). These storms usually consist of a 200-300 km wide area of
easterly or northeasterly winds north of a cold front and surface low pressure system. The passage
of the cold front is often complicated by local terrain, and uncertainties regarding the specific
location of the low pressure center development typically confounds forecasts. The monitoring
system thus emphasizes monitoring these features and the conditions that modulate them, with
A network of 449 MHz wind profilers 100 km east of the base of the foothills to monitor cold
fronts and upslope winds, plus another array of 915 MHz wind profilers 10 km east of the base of
the foothills to monitor both barrier jet winds and snow level (449 HMz profilers are more
expensive, but can see much higher into the atmosphere than 915 MHz systems)
Snow level radars at the base of the foothills as an alternative to the 915 MHz wind profilers
•
Snow melt
As highlighted in Fig. 3b, spring melt creates most of the highest flows seen on many rivers,
especially in the intermountain region. For this reason special emphasis is given here to snowmelt
conditions. Most of the proposed actions for this purpose require new sensors at high altitudes,
but could also include sensors at altitudes where cold storms deposit snow that can contribute to
flooding in somewhat rare events. The most effective approach would leverage the existing
SNOTEL network, which already targets regions of significant snowpack, and has infrastructure
that could facilitate implementation of new tools. In fact, some of this is already happening within
the NRCS, and with California’s EFREP, in terms of adding soil moisture measurements to some
sites. Figure 6 highlights the locations where system enhancements are recommended, and lists
the key sensors that would be used, while Appendix K describes novel needs and methods
recommended here.
d) Some excluded elements
While additional NEXRAD systems have potential uses here (such as illustrated by the planned
deployment of a new NEXRAD radar on the Washington coast) the NEXRAD network is optimized more
for Great Plains convection and detection of tornadoes and hail. For example, the NEXRADs do not scan
at elevations angles below 0.5 above the horizon. In the west, where radars are commonly sited in areas
of high terrain to reduce blockage, the ability of the radar to observe at or below the horizon is critical for
accurate QPE. Moreover, the NEXRAD network is already far more dense in the Great Plains than over
the mountainous part of the west (see Fig. F1 in Appendix F) and individual new NEXRAD radars are
prohibitively expensive relative to the scope of the present vision.
Also, although new satellite techniques (i.e., new sensors or next generation satellites) hold potential to
support the objectives of this white paper, their costs are beyond the scope of what is being considered
19
here for regional applications. Nonetheless, the proposed network will integrate with existing and future
radar and satellite observations, e.g., by using existing satellite sensors and NEXRAD data in new ways.
For example, a) passive microwave and GPS occultation satellite observations are crucial over the ocean
for AR monitoring; however, the passive microwave method does not work over land. A land-based
GPS-met land-based network is proposed here specifically to help fill this gap. Also, no satellite methods
currently monitor AR winds and water vapor transports aloft over the oceans or land, so airborne
reconnaissance is proposed here to fill this gap. Finally, some key satellite footprints/spatial resolutions
are simply too coarse to resolve conditions associated with the especially complex terrain of the west.
e) Additional actions related to achieving the objectives of this modernization:
-
-
-
Measure the performance of precipitation forecasts that use measures of accuracy focused specifically
on extreme precipitation events
Measure snow-level and soil moisture forecast performance
Create a scale for assessing the strength of ARs offshore (e.g., compare water vapor transport in an
AR to the average flow of the Mississippi river; as in “this AR transports in one day a amount of
water as vapor similar to what 7 to 15 Mississippis transport in one day as liquid)
Create a simple scale for communicating and comparing magnitudes of extreme precipitation and
stream flow over land
Develop predictive modeling and decision support tools/systems to optimize use of the data
Use of new GIS-based tools for information management, display and dissemination
Ensure that data from key elements of this network, especially those that provide observations aloft
and/or offshore, are assimilated into operational weather prediction models. Note that models either
are already able to assimilate many of the observations (e.g., wind profiler, GPS-met and dropsondes
data) proposed here when and if they are made, and that methods to assimilate other data types could
be developed.
Sustaining an improved western observing network requires a stronger collaboration across several
agencies
4.
Transition to Operations
This new observational network would be established in the midst of significant new capabilities in the
nation’s water resources science and services, which would facilitate the operational transition and
increase the operational impacts of these new information streams. This section discusses this operational
framework as well as related existing infrastructure.
IWRSS
The Integrated Water Resources Science and Services (IWRSS) is a new multi-agency framework to meet
the nation’s growing water resource challenges. In May 2011, the National Oceanic and Atmospheric
Administration, the US Army Corps of Engineers, and the US Geological Survey signed a memorandum
of understanding establishing IWRSS. In the future, additional federal agencies (24 in all) that have
water resource responsibilities will join the consortium. Through IWRSS, these agencies will engage in
20
collaborative science and develop services and tools to support integrated and adaptive water resources
management. IWRSS consortium members will operate within a common operating picture, employing
shared modeling and information services frameworks, allowing their individual decision support systems
to work in concert towards meeting the water challenges faced by our Nation.
IWRSS considers a broad range of time scales – from historical “analyses of record” through forecasts
and projections spanning weather and climate. Three crosscutting themes on (i) Human Dimensions
(including stakeholder interactions and communications), (ii) Information Technology (agency-spanning,
service oriented architecture), and (iii) Operational Science (a summit to sea modeling and prediction
framework) inform IWRSS efforts. The IWRSS common operating picture will provide decision support
system interoperability and data synchronization services in which these new information streams can be
shared enhancing their value and impact, across multiple federal agencies.
Another key element of the IWRSS strategy, are a series of regional watershed demonstrations (or pilots).
A number of candidate pilot studies have been identified in the Western US, but the first to be explored
will be a regional demonstration centered on Northern California, leveraging ongoing and developing
efforts by HMT, NIDIS and other key programs. A series of IWRSS-led workshops is anticipated
beginning in the fall of 2011, that will identify gaps and stakeholder requirements, explore technical and
scientific solutions and approaches and produce recommendations to inform a regional pilot plan.
More information on IWRSS is available at:
http://www.nohrsc.noaa.gov/~cline/IWRSS/IWRSS_ROADMAP_v1.0.pdf
The NOAA NWC
An important development within IWRSS is the establishment of the National Water Center (NWC). The
NWC will be housed in a new NOAA facility being constructed in Tuscaloosa, Alabama and will serve as
a focal point for IWRSS developments and NOAA’s water resource information services. The NWC is
expected to be operational by the Winter of 2014. The NWC will provide an end-to-end, summit-to-sea
data and modeling framework for all water related processes at and near the Earth’s surface. As noted
earlier, these information services will cover a historical period, an analysis of current conditions, and
forecasts and projections on weather and climate time scales. As such they will span too much and too
little water conditions. Within NOAA, the NWC will support the NWS River Forecast Centers with endto-end, high-resolution gridded water resource information, much like NCEP provides for the NWS
Weather Forecast Offices.
The NWC will provide capabilities to integrate the new observations discussed herein, through
assimilation into the IWRSS forcing data engine and operational modeling framework, and by integrating
them into a data model that supports both NWS operations, as well as those of the IWRSS consortium.
The NWC then becomes a vehicle for widespread dissemination and application to operational centers
and stakeholders and customers of water resource information.
21
NCEP/WFOs
IWRSS/NWC will complement and enhance the NWS’s long-established and highly efficient framework
for disseminating weather and water information, forecasts and warnings. The NOAA National Centers
for Environmental Prediction (NCEP) provide 24x7x365 large scale modeling forecasts and guidance.
This information supports a network of 122 Weather Forecast Offices (WFOs) and 13 River Forecast
Centers (RFCs), which serve as focal points of local expertise and knowledge, as well as for engagement
with local communities and stakeholders. These data will feed into WFOs, supporting their role in
issuance of locall watches and warnings.
NCEP leads NOAA’s efforts in assimilating observations into operational numerical models, including
wind profiler and dropsondes data from aircraft. GSD innovates on assimilating observations from many
sources in high resolution models. The existing capabilities in these arenas will be utilized, including use
of current assimilation methods and development of new ones.
5.
Implementation, costs and timelines
A tiered approach to implementation was used to develop the 21st Century vision for monitoring being
implemented by California’s EFREP program. A similar approach could be used to organize, schedule
and prioritize the westwide vision presented here (Fig. 10).
A tiered approach for new obs to help address CA’s water resource issues
IV:
Offshore
recon.
Tier III:
Newer technology
Ex: Gap-filling radars,
Buoy-mounted WPs
Tier II: Expand on well-defined
needs with proven technology
Ex: Wind profilers, Coastal
Atmospheric river observatory
Tier I: Address well-defined needs with
proven technology
Ex: Soil moisture sensors at CIMIS sites, GPS
receivers of opportunity, snow-level radars
Fig. 10. A tiered approach to implementation could be used. Lower tiers are the least costly and most readily implemented.
The upper tiers are more expensive and require greater effort to deploy. (From the California-DWR sponsored HMT-West
Legacy/Enhanced Flood response and Emergency Preparedness project; Appendix E).
22
The costs to operate, maintain and optimize the network can be reduced considerably by deploying
automated monitoring systems and methods for monitoring performance, diagnosing faults, emphasizing
reliability during critical meteorological conditions, training local “hosts” who can perform less technical
tasks (e.g., restarting a computer), depending upon existing sensor network communications and quality
control tools (such as MADIS - http://madis.noaa.gov/ and MESOWEST http://mesowest.utah.edu/index.html), maintaining a small, technical and engineering staff that can be on
call to perform repairs, and using methods developed partly in HMT for maintaining regional networks in
a cost-effective manner. The proposed network is also structured to also take advantage of existing,
underutilized networks, such as the TARS profiler network along the southern border with Mexico, preexisting GPS sensors currently deployed for seismic monitoring, and extensive existing SNOTEL and
streamgage networks, and provides resources to ensure key stream gauges are supported long-term.
Approximate cost for elements of this major improvement in western U.S. infrastructure are presented
below but are very rough, order of magnitude (VROM) estimates, assuming optimal use of existing
network facilities and co-location with pre-existing power supplies and telemetry streams where possible.
The following list summarizes the VROM cost estimates for key components of this modern system:
Land-based network
Acquire and deploy land-based in situ and small remote sensors: $30 M ($10 M for in-situ sensors
and GPS-met; $20 M for profiling wind and precipitation radars)
Acquire and deploy land-based scanning radars: $35 M (14 C-band radars + 10 X-band radars)
Operate, maintain and optimize use of the land-based network (including one central group of
technical experts, three regional support hubs and one optimization R&D program); this would
include a “Rapid Response” capability that would be on call for deployment of a scanning
polarimetric radar, 2 mobile atmospheric river observatories and supporting surface
meteorological data – consider the Howard Hanson Dam issue as an example; Operate a snow and
rain sample collection network (and diagnostic systems) for isotope, chemistry and aerosol
monitoring, including for “dust on snow” and Asian dust: $13 M/year
Ensure long-term support for key existing stream gauges in the region: $5 M/year
Offshore network and tool development
Develop an optimized offshore monitoring system and concept of operations using aircraft, buoy
and island-mounted systems, integrated with existing and planned satellite observations and
numerical modeling tools: $15 M (includes $5 M for buoy-mounted ARO development)
Acquire and outfit necessary aircraft (likely including the no-cost transfer of outdated military
aircraft, such as Unmanned Aircaft Systems -UAS, to either NOAA or NASA): $25 M
Operate, maintain and optimize use of offshore network (1 aircraft including expendables such as
dropsondes, 4 buoy-mounted AROs, 8 island observing sites, optimization R&D): $17 M/year
Total integrated network (land-based + offshore)
Develop, acquire and deploy: $105 M (one-time costs)
Operate, maintain and optimize: $35 M (annually)
23
Timeline
Initially, emphasis could be placed on targeted development tasks, acquisition of existing technologies
and deployment of systems. As time progresses, the need to operate, maintain and optimize uses
increases, as does the opportunity to develop new applications and tools. The fully implemented system
can be in place within 6 years, including $105 M of one-time costs. The final ongoing costs are $33
M/year, and could be carried out by a core facility plus three regional hubs.
Develop, Acquire,
Deploy ($M)
Operate, Maintain,
Optimize ($M)
Total for the
year ($M)
1
27
8
35
2
24
11
35
3
21
14
35
4
18
17
35
5
15
20
35
6
0
35
35
$105 M
$105 M
$210 M
Year
Total
Highlights/notes
Expand EFREP network to OR,
WA, ID, NV; develop offshore tools;
Order long-lead equipment
Start Front range and snow
networks; deploy soil moisture and
GPS-met to intermountain area;
First gap-filling radars deployed;
continue snow network deployment;
Expanded EFREP network done
Front-range network complete;
Airborne assets in place; core OMO
staff and facilities established
Gap-filling radar network complete;
snow network complete; offshore
system completed
Full network in place; full OMO
system in place; ongoing service
improvement system in place
Ongoing costs $35 M/year
Note: These estimates represent rough order of magnitude costs based on a concept of operations that
highly leverages existing facilities, organizations and applies cost-saving strategies tested in HMT.
Table 1. Strawman implementation strategy to establish long-term capability for the Western U.S.
We would also seek to coordinate the development of the enhanced observation system with the
information systems and administrative mandates of the primary federal, state and related water data and
management agencies, and to leverage existing data wherever possible.
6.
Expected Outcomes
Based on past experience with implementing major new observational infrastructure (e.g., NEXRAD,
satellites, EFREP, and HMT-West), the time frame for implementation is roughly 5-10 years. During this
time the methods and tools to use the new data would also be being enhanced, and benefits would begin
to accrue. This is not a “quick fix” strategy. It is a thorough, scientifically based, strategy to modernize
the region’s observational infrastructure for dealing with the extremes of too much, or too little water.
24
This 21st century observing system for the West will meet many of the 21st Century needs for solid, realtime information based on the most up-to-date science, to support decision making affecting millions of
people, businesses and the environment of the west.
Benefits will include:
Mitigating risks of >$100 B Katrina-like disasters on the U.S. West coast (see NRC reports from
before Katrina and ARkStorm activities since then)
Reducing flood risks and damages through development of modern decision support tools that
allow for pre-storm releases from flood control reservoirs that are refilled by the storm, thereby
enhancing also water supply for the summer
Enabling forecast-based reservoir operations for combined benefits of improved flood control and
enhanced water supplies, possibly offseting some requirements for new reservoir space being
considered in the west ($Billions in savings)
Improving drought monitoring and associated benefits in support of NIDIS
Improving the capacity for early detection of climate change impacts on water supply and flooding
in the west and to inform adaptation and mitigation strategies optimized for regional impacts and
needs
Tighter collaborations across federal, state, local agencies to assure effective implementation into
existing water information and management systems
Strengthening of science and technology job sectors in the region
From a policy perspective, it is perhaps most notable that improved water-year-type forecasts in
California alone have estimated values that could exceed $100 M in a single year (Simpson et al. 2004).
In terms of flood risks, enhanced preparedness for flooding events that mitigates loss of life and property
for flooding – the phenomena that is responsible for the greatest impacts as defined through Presidential
disaster declarations (e.g., average flood losses in the region are $1.5 Billion per year-Appendix A) – is a
primary strategy for controlling and reducing those risks. With improved monitoring, improved forecasts
and longer lead times, community leaders will be better able to prepare for extreme events not only from a
safety standpoint, but can minimize the costs of taking preparedness actions, e.g. by shifting work
schedules so that preparatory work can be done on regular time versus overtime. Thus we believe that
prudence dictates a proactive approach to monitoring for extreme precipitation events and flooding across
the western states, and we hope that the vision presented here can be implemented as a proactive strategy.
25
7.
Full text of the Resolution of the Western States Water Council passed on July 29, 2011
Position No. 332
RESOLUTION
of the
WESTERN STATES WATER COUNCIL
supporting
FEDERAL RESEARCH AND DEVELOPMENT OF UPDATED HYDROCLIMATE
GUIDANCE FOR EXTREME METEOROLOGICAL EVENTS
Bend, Oregon
July 29, 2011
WHEREAS, Western states have recently been experiencing near-record flooding, droughts, or
wildfires that threaten public safety, tax aging water infrastructure, and/or have significant economic
consequences; and
WHEREAS, before the first half of 2011 was over, the year had already set records for extreme
weather events, with the nation having experienced eight $1 billion-plus disasters, according to the
National Oceanic and Atmospheric Administration (NOAA); and
WHEREAS, extreme weather events have grown more frequent in the U.S. since 1980,
according to NOAA, and are believed to be caused by a changing climate; and
WHEREAS, the top twelve warmest years on record globally all have occurred since 1997, and
climate change is expected to result in future increases in the frequency, extent, and/or severity of
floods, coastal inundation, and droughts.
WHEREAS, some of NOAA’s probable maximum precipitation estimates used by water
agencies for dam safety analyses have not been updated since the 1960s and the federal Guidelines for
Determining Flood Flow Frequency Analysis (published as Bulletin 17B) have not been revised since
1981, and neither of these guidance documents address hydroclimate non-stationarity; and
WHEREAS, flood frequency analyses are used by public agencies at all levels of government
to design and manage flood control and stormwater infrastructure, with Bulletin 17B still representing a
default standard of engineering practice; and
WHEREAS, federal funding for hydrology research has waned since the 1970s-1980s, and
alternative statisticalmethodologies for flood frequency analyses or deterministic analytical procedures
are not being supported and transitioned to common engineering practice; and
WHEREAS, the Federal Emergency Management Agency has adopted a process for local
communities to explicitly incorporate “future conditions hydrology” in the national flood insurance
program’s flood hazards mapping; and
WHEREAS, a federal agency committee composed of the U.S. Army Corps of Engineers, U.S.
Bureau of Reclamation, NOAA, U.S. Geological Survey, and U.S. Environmental Protection Agency
held a 2010 national science workshop on non-stationarity, hydrologic frequency analysis, and water
management, to identify information gaps and the state of the science for handling hydroclimate
uncertainty; and
26
Position No. 332
WHEREAS, the Council co-sponsored a 2011 workshop on hydroclimate non-stationarity and
extreme events, to identify actions that could be taken at planning to operational time scales to improve
readiness for extreme events; and
WHEREAS, the federal and the Council workshops identified multiple approaches that could
be employed at the planning time scale, including ensembles of global circulation models, paleoclimate
analyses, and alternative techniques for flood frequency analysis; and
WHEREAS, advances in weather forecasting research, such as that of NOAA’s
Hydrometeorological testbed program onWest Coast atmospheric rivers, demonstrate the potential for
improving extreme event forecasting at the operational time scale; and
WHEREAS, the 2006Western Governors’ Association (WGA) report on Water Needs and
Strategies for a Sustainable Future and the follow-up 2008WGA Next Steps report identify addressing
climate change impacts as a priority for moving forward, and make specific recommendations for
actions that the federal government and the states should take to support adaptation, including detailing
research and planning needs.
WHEREAS, WGA and NOAA signed a memorandum of understanding on June 30, 2011,
regarding state adaptation to climate variability and change that focuses on climate extremes, variability
and future trends as they relate to disaster risk reduction and improved science for coastal and marine
resource management; and
WHEREAS, the Draft Vision and Strategic Framework for a Climate Service in NOAA
includes changes in extremes of weather and climate as one of the four key societal challenges that will
initially be a focus of the climate service.
NOW, THEREFORE, BE IT RESOLVED, that the federal government should update and
revise its guidance documents for hydrologic data and methodologies – among them precipitationfrequency
estimates, flood frequency analyses, and probable maximum precipitation – to include
subsequently observed data and new analytical approaches; and
BE IT FURTHER RESOLVED, that theWestern States Water Council supports development
of an improved observing system forWestern extreme precipitation events, to aid in monitoring,
prediction, and climate trend analysis associated with extreme weather events; and
BE IT FURTHER RESOLVED, that theWestern States Water Council urges the federal
government to support and place a priority on research related to extreme events, including research on
better understanding of hydroclimate processes, paleoflood analysis, design of monitoring and change
detection networks, and probabilistic outlooks of climate extremes.
BE IT FURTHER RESOLVED, that theWestern States Water Councilwill work with NOAA
in supporting efforts on climate extremes, variability, and future trends as called for in theWGA-NOAA
memorandum of understanding.
F:\POSITION\2011\Bend, OR\-#332 WSWC Position on Extreme Meteorological Events 2011July 29.doc
27
Appendices
A: Flood damages in the western U.S.
B: Comparison of extreme precipitation in the western US with other regions
C: Storms, floods, and the science of atmospheric rivers
D: Howard Hanson Dam crisis Rapid Response
E: California’s Enhanced Flood Response and Emergency Preparedness (EFREP) project
F: Scanning radar options
G: A possible scaling for extreme 3-day precipitation
H: WISPAR experiment with the Global Hawk Unmanned Aircraft System (UAS)
I: Global Positioning System – Meteorology (GPS-Met)
J: Soil moisture observing network design and instrumentation
K: An Approach to Enhancement of the Western SNOTEL and Streamgage Networks
L: Dust-on-snow sampling
M: ARkStorm: Emergency Preparedness Scenario for An Extreme Atmospheric River Event
N: Gap-filling radar example for Colorado Front Range flood from a burn area
References
Adams, D. K., A. C. Comrie, 1997: The North American Monsoon. Bull. Amer. Meteor. Soc., 78, 2197–
2213.
Bales, R., Molotch, N., Painter, T., Dettinger, M., Rice, R., and Dozier, J., 2006, Mountain hydrology of
the western US: Water Resources Research, 42, W08432, doi:10.1029/2005WR004387, 13 p.
Douglas, M.W., R. Maddox, K. Howard, and S. Reyes, 1993: The Mexican monsoon. J. Climate, 6,
1665–1667.
Dettinger, M.D., Ralph, F.M., Das, T., Neiman, P.J., and Cayan, D., 2011: Atmospheric rivers, floods,
and the water resources of California. Water, 3 (Special Issue on Managing Water Resources and
Development in a Changing Climate), 455-478.
Higgins, W., D. Ahijevych, J. Amador, A. Barros, E.H. Berbery, E. Caetano, R. Carbone, P. Ciesielski, R.
Cifelli, M.C. Vazquez, A. Douglas, M. Douglas, G. Emmanuel, C. Fairall, D. Gochis, D. Gutzler, T.
Jackson, R. Johnson, C. King, T. Lang, M.I. Lee, D. Lettenmaier, R. Lobato, V. Magana, J. Meiten, K.
28
Mo, S. Nesbitt, F.O. Torres, E. Pytlak, P. Rogers, S. Rutledge, J. Schemm, S. Schubert, A. White, C.
Williams, A. Wood, R. Zamora, and C. Zhang. 2006: The NAME 2004 field campaign and modeling
strategy. Bull. Amer. Meteor. Soc., 87, 79-94.
IWRSS Roadmap: http://www.nohrsc.noaa.gov/~cline/IWRSS/IWRSS_ROADMAP_v1.0.pdf
Neiman, P. J., L. J. Schick. F. M. Ralph, Mi. Hughes, and G. Wick, 2011: Flooding in Western
Washington: The connection to atmospheric rivers. J. Hydrometeor., (In press).
Olsen, J. Rolf, Julie Kiang and Reagan Waskom, (editors). 2010. Workshop on Nonstationarity,
Hydrologic Frequency Analysis, and Water Management. Colorado Water Institute Information Series
No. 109. www.cwi.colostate.edu.
Porter, Keith, Wein, Anne, Alpers, Charles, Baez, Allan, Barnard, Patrick, Carter, James, Corsi,
Alessandra, Costner, James, Cox, Dale, Das, Tapash, Dettinger, Michael, Done, James, Eadie, Charles,
Eymann, Marcia, Ferris, Justin, Gunturi, Prasad, Hughes, Mimi, Jarrett, Robert, Johnson, Laurie, Dam LeGriffin, Hanh, Mitchell, David, Morman, Suzette, Neiman, Paul, Olsen, Anna, Perry, Suzanne, Plumlee,
Geoffrey, Ralph, Martin, Reynolds, David, Rose, Adam, Schaefer, Kathleen, Serakos, Julie, Siembieda,
William, Stock, Jonathon, Strong, David, Sue Wing, Ian, Tang, Alex, Thomas, Pete, Topping, Ken, and
Wills, Chris; Jones, Lucile, Chief Scientist, Cox, Dale, Project Manager, 2010: Overview of the
ARkStorm scenario: U.S. Geological Survey Open-File Report 2010-1312, 183 p. and appendixes.
Ralph, F. M., Neiman, P. J., D. E. Kingsmill, P. O. G. Persson, A. B. White, E. T. Strem, E. D. Andrews,
and R. C. Antweiler, 2003: The impact of a prominent rain shadow on flooding in California’s Santa Cruz
mountains: A CALJET case study and sensitivity to the ENSO cycle. J. Hydrometeor., 4, 1243-1264.
Ralph, F. M., R. M. Rauber, B. F. Jewett, D. E. Kingsmill, P. Pisano, P. Pugner, R. M. Rassmussen, D.
W. Reynolds, T. W. Schlatter, R. E. Stewart and J. S. Waldstreicher, 2005: Improving short-term (0-48
hour) Cool-season quantitative precipitation forecasting: Recommendations from a USWRP Workshop.
Bull. Amer. Meteor. Soc., 86, 1619-1632.
Ralph, F.M., and M.D. Dettinger, 2011: Storms, Floods and the Science of Atmospheric Rivers. EOS,
Transactions, Amer. Geophys. Union., 92, 265-266.
Simpson, J. J., M. D. Dettinger, F. Gehrke, T. J. McIntire, G. L. Hufford, 2004: Hydrologic Scales, Cloud
Variability, Remote Sensing, and Models: Implications for Forecasting Snowmelt and Streamflow. Wea.
Forecast., 19, 251-276.
U.S. Bureau of Reclamation (USBR) 2011: FY 2011 Science and Technology Program, Research and
Development Office, Denver, CO. www.usbr.gov/research/docs/S&T_2011_research_abstracts.pdf
White, A. B., B. Colman, G. M. Carter, F. M. Ralph, R. S. Webb, D. G. Brandon, C. W. King, P. J.
Neiman, D. J. Gottas, I. Jankov, K. F. Brill, Y. Zhu, K. Cook, H. E. Buehner, H. Opitz, D. W. Reynolds,
L. J. Schick, 2011: NOAA's Rapid Response to the Howard A. Hanson Dam Flood Risk Management
Crisis. Bull. Amer. Meteorol. Soc. (in press July 2011), doi: 10.1175/BAMS-D-11-00103.1.
29
A Vision of Future Observations for Western U.S. Extreme
Precipitation Events and Flooding: Monitoring, Prediction and Climate
Appendices
Marty Ralph (NOAA), Mike Dettinger (USGS), et al.
(See author list in main text, and authors listed for each Appendix)
A: Flood damages in the western U.S.
B: Comparison of extreme precipitation in the western US with other regions
C: Storms, floods, and the science of atmospheric rivers
D: Howard Hanson Dam crisis Rapid Response
E: California’s Enhanced Flood Response and Emergency Preparedness (EFREP) project
F: Scanning radar options
G: A possible scaling for extreme 3-day precipitation
H: WISPAR experiment with the Global Hawk Unmanned Aircraft System (UAS)
I: Global Positioning System – Meteorology (GPS-Met)
J: Soil moisture observing network design and instrumentation
K: An Approach to Enhancement of the Western SNOTEL and Streamgage Networks
L: Dust-on-snow sampling
M: ARkStorm: Emergency Preparedness Scenario for An Extreme Atmospheric River
N: Gap-filling radar example for Colorado Front Range flood from a burn area
Provided to the Western States Water Council on October 5, 2011
1
Appendix A: Historical flood damages in the Western U.S.
M. D. Detttinger (USGS) and F.M. Ralph (NOAA)
The following figure was published by Pielke et al. in 2002 and used an analysis of NWS
data for the period 1983-1999. During these 17 years, the Western States of WA, OR,
CA, ID, NV, UT, AZ, MT, WY, CO, NM, ND, SD, NE, KS, OK, and TX experienced
$24.7 Billion in flood damages, an average of $1.5 Billion annually. California,
Washington and Oregon combined to account for $10.6 B of this (i.e., 43% of the
regional total), with California alone accounting for 25% of the western total.
Fig. A1. Flood damages nationally, and for the 17 states in the contiguous western US
(red).
2
For comparison, if the strawman modernization of Western US observations for extreme
precipitation and flooding, achieved even a 2% reduction in flood damages, i.e., $30
M/year in savings, then it would be cost effective even without considering the benefits
to water supply that would also accrue. A general idea of how large the water-supply
benefits might be is provided by an estimate from Simpson et al. (2006) that a better
water-year status forecast for California, in a single year, could save the State (not
including businesses in the State) about $150 million.
Still other benefits would include reduced loss of life. For example, in 1998, experimental
offshore reconnaissance enabled NWS forecasters in coastal California to issue a warning
with enough lead time so that emergency responders were able to be pre-position
equipment and supplies, and thus affected hundreds of rescues with only 1 fatality. A
similar storm 15 years earlier caused many flood fatalities (Ralph. et al. 2003).
Finally, the vulnerability of the region is highlighted by recent conditions at Howard
Hanson Dam in Washington that put $4 B in infrastructure at risk, and where the most
effective management response proved to be installation and operation of many of the
observational assets described in this vision in order to provide highest quality forecast
capabilities to the reservoir managers who had to managed those risks (White et al. 2011;
Appendix D).
-----------------Pielke RA Jr., Downton MW, Barnard Miller JZ (2002) Flood damage in the United
States, 1926-2000--A reanalysis of National Weather Service Estimates: UCAR, Boulder,
CO, 86 p.
Ralph FM, Neiman PJ, Kingsmill DE, Persson POG, White AB, Strem ET, Andrews ED
and Antweiler RC, 2003, The impact of a prominent rain shadow on flooding in
California’s Santa Cruz Mountains—A CALJET case study and sensitivity to the ENSO
cycle: J. Hydrometeorology, 4, 1243-1264.
Simpson, J.J., Dettinger, M.D., Gerhke, F., McIntyre, T.J., and Hufford, G.I., 2004,
Hydrologic scales, cloud variability, remote sensing and models—Implications for
forecasting snowmelt and streamflow: Weather and Forecasting, 19, 251-276.
3
Extreme Precipitation and Atmospheric
Rivers on the U. S. West Coast
BY F. M. RALPH1 AND M. D. DETTINGER2
Physical Sciences Division, Boulder, Colorado
2U.S. Geological Survey, Scripps Institution of Oceanography, La Jolla, California
1NOAA/ESRL,
Submitted on 22 Aug 2011 to Bull. Amer. Meteor. Soc. as a contribution to “Map Room”
INTRODUCTION. Strong winter storms
battered the U.S. West Coast from western
Washington to southern California in
December 2010, producing from 250 to over
670 mm (10 to 26 inches of rain) in preferred
areas (Fig. 1). A common denominator among
these events is that the synoptic weather
patterns produced a series of strong
atmospheric rivers (AR) that transported large
amounts of water vapor from over the Pacific
Ocean to the U.S. West Coast (Fig. 2), which
fueled the heavy rain and flooding, and
provided beneficial increases in snowpack.
For example, by December 22 – the first full
day of winter, the southern Sierra had already
received 75% of its annual average snowpack
and was well on its way to one the deepest
annual snowpacks recorded.
Just how “extreme” were these events
relative to other atmospheric river cases in the
region? More generally, how does West Coast
AR-fed precipitation compare with extreme
precipitation in other parts of the U.S., such as
from landfalling hurricanes and tropical
storms?
This report uses decades of
Cooperative
Observer
(COOP)
daily
precipitation reports from over 5800 stations
across the U.S. to address these questions, and
then summarizes the West Coast events and
forecasts of December 2010.
A
BRIEF
BACKGROUND
ON
ATMOSPHERIC RIVERS. Atmospheric
rivers (AR) are long, narrow zones within
extratropical cyclones that contain large water
vapor contents and strong winds and that are
responsible for > 90% of all atmospheric water
vapor transport in midlatitudes (Zhu and
Newell, 1998). They are 1000s of km long
and, on average, only 400 km wide (Ralph et
al. 2004); 75% of the water vapor transport
occurs below 2.25 km altitude (Ralph et al.
2005).
ARs regularly produce extreme
precipitation in coastal regions because they
transport large quantities of water vapor and
comprise almost ideal conditions for
producing heavy orographic rains and flooding
when they encounter mountains (Ralph et al.
2005, 2006, 2011; Neiman et al. 2008, 2011).
The west coast of North America is
particularly vulnerable to ARs, as are the west
coasts of other midlatitude continents, such as
Fig. 1. 14-day observed precipitation in
the Western U.S. (Courtesy, NWS
Advanced
Hydrologic
Prediction
Service).
4
12 Dec 2010
18 Dec 2010
Fig. 2. Polar orbiting satellite observations of vertically integrated water vapor
from SSM/I and SSM/IS showing atmospheric river conditions associated with two
of the extreme precipitation events in December 2010 on the U.S. West Coast.
These images represent two separate and independent ARs. (Courtesy of G. Wick
and D. Jackson.)
Europe (Stohl et al. 2008). Although they are
linked to extreme precipitation and flooding,
ARs also produce 25-50% of the annual
precipitation on the U.S. West Coast (Guan et
al. 2010, Dettinger et al. 2011), and thus are
important in generating water resources in the
region.
as well as cloud liquid water and rain rate
(e.g., Wick et al. 2008).
For more information on ARs, including a
list
of
publications,
see
http://www.esrl.noaa.gov/psd/atmrivers/.
ANALYSIS OF EXTREME PRECIPITATION USING COOP DATA. Although
much is now known about key geophysical
characteristics of ARs, a systematic
comparison of extreme AR rainfall on the U.S.
West Coast with extreme precipitation
elsewhere has not been made. To provide such
a comparison, long-term (>30 yr) COOP
precipitation records of 3-day precipitation
totals were used to determine where and how
often storms in each of several simple rainfall
categories (R-Cats) have been reported across
the contiguous U.S. The categories used are
listed in Table 1.
a) Methodology
The categories developed and used here were
based on daily accumulated precipitation totals
reported in the Summary of the Day
observations from cooperative weather
stations across the United States [National
Weather Service (NWS), (1989)]. Missing
data were excluded, as were accumulations
from multiple days reported only as multi-day
totals. While daily totals are a common
measure of precipitation, multiple-day
precipitation totals can be of more practical
importance with respect to progressive
hazards, like floods on main-stem rivers and
some landslides.
A “station event” is defined here as an
occasion when a single COOP station reports a
precipitation total within one of the R-Cats.
An “episode” is defined as a 3-day period
during which at least one COOP station
reports precipitation within one of the R-Cats.
A single episode can, and usually does,
include multiple station events. Multiple-day
precipitation totals are considered because
multi-day totals are commonly most relevant
for landslides, flooding and other hydrological
impacts on main-stem rivers. Although 2- and
4-day totals were also considered, 3-day
windows were used because: (a) when 2-day
windows were considered, roughly half of the
5
Table 1. Rainfall categories used in this study, and national frequencies of
occurrence. Note that an “episode” is defined as a single 3-day period for which
one or more stations observed at least 200 mm of precipitation in the same
general area.
75 W
Fig. 3. Maximum 3-day precipitation totals at 5877 COOP stations in the
conterminous US.
major storms (by 3-day standards) were
missed, and (b) when 4-day periods were used,
one of the four days typically contributed little
to the multiple-day totals (on average,
nationally, the driest day of 4-day site-events
contributed only 4% of the 4-day totals,
whereas, on average across all events, the
driest day in a 3-day window still accounts for
10% of the total).
b) Results
Historical patterns of extreme precipitation,
labeled by R-Cats, are shown in Fig. 3, which
reveals that, although R-Cat 1–2 events are
reported in most states, nearly all R-Cat 3–4
events occurred in California, Texas, or the
southeastern states.
Thus, extremeprecipitation events in the mountains of
California are found to be comparable with
those in the southeastern U.S. (including
Texas) and are only equaled by storms in that
6
Fig. 4. Seasonality of extreme precipitation events in the Eastern
U.S. versus the Western U.S. Number of 3-day episodes achieving
the highest rainfall categories, east (pink) and west (blue) of 105ºW,
by month of year, normalized to the number of COOP sites in each
region. Two thresholds are used, light shading for R-Cat 2 (i.e., >300
mm), and dark shading for R-Cat 3-4 (i.e., >400 mm).
region.
Extreme precipitation events in
California are further notable because, unlike
those in Texas and the Southeast, it is not
uncommon for California stations to have
experienced multiple R-Cat 3-4 episodes
during their periods of record.
By defining R-Cat “episodes,” i.e., 3-day
periods during which at least some stations
exceeded a given R-Cat threshold, we verified
that the more extreme the episode - the larger
its areal extent (Table 1). Episodes were then
binned by month of year for stations east and
west of 105°W. The resulting episode counts
(Fig. 4) reveal that most eastern episodes
occurred during the spring, summer, and fall
seasons, whereas the western episodes
occurred almost exclusively during the cool
season (Nov-Apr) when strong ARs typically
impact the region (Neiman et al. 2008).
Despite the timing difference, for a wide range
of 3-day precipitation totals the fraction of wet
days exceeding those totals are nearly the
same for the cool-season (Nov-Apr) in the
western U.S. as for the warm season (MayOct) in the eastern U.S. (Fig. 5).
Evaluation of all R-Cat episodes from
1997-2005 in terms of meteorological
conditions showed that in all 17 episodes west
of 115°W that met or exceeded the R-Cat 2
threshold, satellite data showed that an AR
had struck the west coast during the 3-day
episode (Neiman et al. 2008). During the
same period, roughly half of the major eastern
events were associated with tropical storms
and hurricanes. Also, by using daily NCEPNCAR Reanalysis estimates of vertically
integrated water vapor transports from 19502008, it was found that 44 of 48 R-Cat 3-4
episodes in the western US coincided with
landfalling ARs there. Upon normalizing longterm eastern and western counts by number of
stations in each region, the annual-averaged
frequencies of extreme R-Cat episodes east
and west of 105°W are identical, i.e., with 44
R-Cat 1-4 episodes per year. Similarly, the
normalized annual-averaged frequencies of RCat 3-4 episodes are 2.1 - 2.2 per year in both
areas. Thus, 3-day precipitation extremes
associated with landfalling ARs on the U.S.
West Coast are heavier than extreme storms
anywhere else in the country outside the
southeast U.S. (including those related to
landfalling tropical storms and hurricanes).
Also, they yield comparable precipitation
totals with the southeastern storms, and occur
station-by-station just as frequently as the
extreme-precipitation episodes there.
7
earthen Trees Branch Dam along the Virgin
River near Springdale, Utah. In the case of the
southern California and Utah areas, a strong AR
stalled in the region for several days, providing a
persistent supply of tropical water vapor from
near Hawaii (which also experienced flooding).
As part of NOAA's standard procedures for
issuing precipitation and stream flow forecasts,
quantitative precipitation forecasts (QPF) were
produced in a collaborative effort between the
Hydrometeorological Prediction Center (HPC),
the Northwest River Forecast Center (NWRFC),
the California-Nevada River Forecast Center
(CNRFC), and the local NWS Weather Forecast
Offices (WFOs) in the region. These QPFs were
then transformed into quantitative stream flow
forecasts for key watersheds by the CNRFC and
NWRFC.
By the time the first storms hit the
Washington Coast from 1200 UTC 10 Dec-
Fig. 5.
Frequency of occurrence of
extreme precipitation east and west of
105 deg W (normalized to number of
COOP stations in each region).
EXTREME PRECIPITATION STRIKES
THE U.S. WEST COAST IN DECEMBER
2010. As an example, the extreme nature of the
precipitation associated with the ARs of
December 2010 is described below, along with a
synopsis of the associated forecasts. The first
major storm in the series produced 292 mm
(11.5 inches) of rain at Quinault on the western
side of Washington’s Olympic Mountains and
localized flooding from 10-12 December. Thus,
this was an R-Cat 1 event, although it nearly
achieved an R-Cat 2 rating. The second set of
storms struck California from 17 – 22 December
2010, producing more than 670 mm (26 inches)
of rain in the San Bernardino Mountains of
southern California over those 6 days, and
upwards of 10-15 feet of snow in the southern
Sierra Nevada Mountains. Within this period of
heavy rain the 3-day total at Lytle Creek in
southern California was 440 mm (17.32 in) from
19-21 December. Thus it was an R-Cat 3 event.
In addition to flooding in Washington and
California, these storms produced 432 mm (17
inches) of rainfall in the mountains of southern
Utah over 5 days between 18-23 December at
the “Little Grassy” SNOTEL site. The 3-day
maximum accumulation at that site was 351 mm
(13.8 in) from 20-22 December, and thus was an
R-Cat 2 event. Flooding in southern Utah caused
serious property damage and damage to the
Fig. 6. HPC 3-day precipitation forecasts
(in inches). These forecasts were issued
a) 9 Dec 2010 at 1350 PST, b) 17 Dec
2010 at 0151 PST, and c) 20 Dec 2010 at
0202 PST.
8
precipitation from ARs, such as the water vapor
“AR flux tool” (Neiman et al. 2009).
In practical terms, the R-Cat categorization
applied here is simple enough to facilitate
communication in both technical and public
arenas, could be used to standardize research
analyses and forecasts, and could improve
reporting
of
climate-change
projections
regarding
extreme
precipitation.
The
categorization complements standard returnperiod
methods,
straightforwardly
accommodating the nonstationarities of climate
change (Milly et al. 2008) and short record
lengths that can make estimation of returnperiods very difficult. Such categorizations will
be all the more important as research into the
potential impacts of a changing climate on ARs
and extreme precipitation begins (e.g., Dettinger
2011). The extent to which the R-Cat episodes
(from all mechanisms) are identifying truly
extreme precipitation episodes may be judged
either economically or physically. For example,
nationally, six of the R-Cat 3–4 episodes from
1997-2005 were associated with damages
exceeding $1 billion each. Meanwhile, R-Cat 2–
4 episodes (combined) occur historically roughly
just as frequently as do hurricanes (Atlantic and
Eastern Pacific combined, measured by the
Safir-Simpson scale) and as do violent tornadoes
(measured by the Fujita scale), indicating that
the simple R-Cat scale used here is identifying
extreme precipitation that is just as rare and
extreme,
nationally,
as
are
standard
categorizations of these other mechanisms of
extreme weather. Thus the R-Cat description of
the nation’s most extreme precipitation episodes
could
be
an
exceptionally
useful
communications and research tool, and in the
present analysis provided a clear, objective
perspective on the severity of AR storms in the
western US.
ember to 1200 UTC 13 December 2010, QPFs
produced by HPC (Fig 6a) provided valuable
guidance that heavy rain was on its way, and the
river hydrographs forecasted by NWRFC gave
ample warning for the flooding that ensued.
HPC, NOAA's Environmental Modeling Center
(EMC), NWRFC, and the Seattle WFO
produced
specialized
precipitation
and
hydrologic forecasts for many western
Washington river basins including the Green
River Basin near Seattle because of the limited
flood prevention capability provided by the
damaged Howard A. Hanson Dam (White et al.
2011). On December 17, when the next series of
powerful storms began to make landfall in
central and southern California, weather
forecasters were armed with a similar set of
forecasts from HPC, EMC and CNRFC,
including remarkable precipitation forecasts
(Figs. 6b and 6c) of more than 250 mm (10
inches) in 72 hours for both December 17-20
(1200 UTC to 1200 UTC) and December 20-23
(1200 UTC to 1200 UTC).
IMPLICATIONS.
Dave
Reynolds,
meteorologist-in-charge of the National Weather
Service's (NWS) San Francisco WFO, said
NWS operational forecasters were well prepared
for the storms in December 2010 in part because
of enhanced awareness and improved
understanding of the role of ARs. Reynolds
recently presented an online briefing to NWS
Western Region staff on the AR phenomenon
and related scientific advances. These advances,
led by NOAA's Physical Sciences Division
(PSD) in Boulder, Colorado, were conducted
under NOAA’s Hydrometeorology Testbed
(HMT), and used modern satellite and other
observational tools to reveal the importance of
ARs to both flooding and water supply in the
region. Evaluations of QPF products by HMT
during the cool season of 2005/06 (Ralph et al.
2010) also revealed that 18 of the 20 most
extreme precipitation events that season were
associated with ARs, and that QPF was biased
low by roughly 50% in the events with greater
than 127 mm (5 in) of precipitation in one day.
Ongoing research and prototyping is underway
to further assess QPF and to explore potential
new tools to assist in prediction of extreme
FOR FURTHER READING
Dettinger, M.D., F. M. Ralph, T. Das, P.J.
Neiman, and D. Cayan, 2011: Atmospheric
rivers, floods, and the water resources of
California: Water, 3, 455-478.
-----, 2011: Climate change, atmospheric rivers
and floods in California—A multimodel
analysis of storm frequency and magnitude
9
profile and atmospheric-river characteristics.
Mon. Wea. Rev., 133, 889-910.
-----, ----- G. A. Wick, S. I. Gutman, M. D.
Dettinger, D. R. Cayan, and A. B. White,
2006: Flooding on California’s Russian
River:
Role of atmospheric rivers.
Geophys. Res. Lett., 33, L13801,
doi:10.1029/2006GL026689.
-----, E. Sukovich, D. Reynolds, M. Dettinger, S.
Weagle, W. Clark, P.J. Neiman, 2010:
Assessment
of
extreme
quantitative
precipitation forecasts and development of
regional extreme event thresholds using data
from HMT-2006 and COOP observers. J.
Hydrometeor., 11, 1288-1306.
-----, and M.D. Dettinger, 2011: Storms, Floods
and
the
Science
of
Atmospheric
Rivers. EOS, Transactions, Amer. Geophys.
Union., 92, 265-266.
Stohl, A., C. Forster, and H. Sodemann, 2008:
Remote sources of water vapor forming
precipitation on the Norwegian west
coast at 60º N - A tale of hurricanes and an
atmospheric river. J. Geophys. Res.113,
D05102, doi:10.1029/2007JD009006.
White, A. B., B. Colman, G. M. Carter, F. M.
Ralph, R. S. Webb, D. G. Brandon, C. W.
King, P. J. Neiman, D. J. Gottas, I. Jankov,
K. F. Brill, Y. Zhu, K. Cook, H. E. Buehner,
H. Opitz, D. W. Reynolds, and L. J. Schick,
2011: NOAA's Rapid Response to the
Howard A. Hanson Dam Flood Risk
Management Crisis. Bull. Amer. Meteorol.
Soc. (In Press), doi: 10.1175/BAMS-D-1100103.1.
Wick, G. A., Y. Kuo, F. M. Ralph, T. Wee, and
P. J. Neiman, 2008: Intercomparison of
integrated water vapor retrievals from
SSM/I and COSMIC, Geophys. Res. Lett.,
35, L21805, doi:10.1029/2008GL035126.
Zhu, Y, and R. E. Newell, 1998: A proposed
algorithm for moisture fluxes from
atmospheric rivers. Mon. Wea. Rev., 126,
725-735.
changes: J. Amer. Water Resour. Assoc., 47,
514-523.
Guan, B., N. P. Molotch, D. E. Waliser, E. J.
Fetzer, and P. J. Neiman, 2010: Extreme
snowfall events linked to atmospheric rivers
and surface air temperature via satellite
measurements. Geophys. Res. Lett., 37,
L20401, doi:10.1029/2010GL044696.
Milly, P.C.D., J. Betancourt, M. Falkenmark,
R.M. Hirsch, Z.W Kundzewicz, D.P.
Lettenmaier, and R.J. Stouffer, 2008:
Stationarity is dead: Whither Water
Management?, Science, 139, 573-574.
National Weather Service (NWS), (1989),
National Weather Service Observing
Handbook Number 2. Cooperative Station
Observations: Observing Systems Branch,
Office of Systems Operations, 94 pp.
Neiman, P. J., F. M. Ralph, G. A. Wick, J. D.
Lundquist and M. D. Dettinger, 2008:
Meteorological characteristics and overland
precipitation impacts of atmospheric rivers
affecting the West Coast of North America
based on eight years of SSM/I satellite
observations. J. Hydrometeor., 9, 22-47.
-----, A.B. White, F.M. Ralph, D.J. Gottas, and
S.I. Gutman, 2009: A Water Vapor Flux
Tool for Precipitation Forecasting. U.K.
Journal of Water Management, 162, 83-94.
-----, L. J. Schick, F. M. Ralph, M. Hughes, G.
A. Wick, 2011:
Flooding in Western
Washington:
The
Connection
to
Atmospheric Rivers. J. Hydrometeor. (In
Press), doi: 10.1175/2011JHM1358.1.
Ralph, F.M., P.J. Neiman, and G.A. Wick, 2004:
Satellite and CALJET aircraft observations
of atmospheric rivers over the eastern NorthPacific Ocean during the El Niño winter of
1997/98. Mon. Wea. Rev., 132, 1721-1745.
-----, -----, and R. Rotunno, 2005: Dropsonde
observations in low-level jets over the
Northeastern Pacific Ocean from CALJET1998 and PACJET-2001: Mean vertical-
10
Appendix C: Storms, Floods and the Science of Atmospheric Rivers
Published as the Cover Story in EOS, a weekly, peer reviewed journal of the
American Geophysical Union. August 2011. (Ralph, F.M., and M.D. Dettinger,
2011: Storms, Floods and the Science of Atmospheric Rivers. EOS, Transactions,
Amer. Geophys. Union., 92, 265-266.)
11
12
13
14
Appendix D: NOAA’s Rapid Response to the Howard A. Hanson Dam Flood
Risk Management Crisis
ALLEN B. WHITE1, BRAD COLMAN2, GARY M. CARTER3, F. MARTIN RALPH1,
ROBERT S. WEBB1,
4
DAVID G. BRANDON , CLARK W. KING1, PAUL J. NEIMAN1, DANIEL J. GOTTAS1,
ISIDORA JANKOV5, KEITH F. BRILL6, YUEJIAN ZHU7, KIRBY COOK2, HENRY E.
BUEHNER2,
8
HAROLD OPITZ , DAVID W. REYNOLDS9, and LAWRENCE J. SCHICK10
1
NOAA/Earth System Research Laboratory/Physical Sciences Division, Boulder, CO
2
NOAA/National Weather Service/WFO Seattle, Seattle, WA
3
NOAA/National Weather Service/Office of Hydrologic Development, Silver Spring, MD
4
NOAA/National Weather Service/Western Region Hydrology and Climate Services,
Salt Lake City, UT
5
Cooperative Institute for Research in the Atmosphere, Colorado State University,
Fort Collins, CO, and NOAA/Earth System Research Laboratory/Global Systems Division,
Boulder, CO
6
NOAA/National Weather Service/Hydrometeorological Prediction Center, Suitland, MD
7
NOAA/NWS/National Centers for Environmental Prediction/Environmental Modeling Center,
Camp Springs, MD
8
NOAA/National Weather Service/Pacific Northwest RFC, Portland, OR
9
NOAA/National Weather Service/WFO San Francisco Bay Area, Monterey, CA
10
U.S. Army Corps of Engineers, Seattle, WA
In Press at Bulletin of the American Meteorological Society (July 2011)
Capsule Summary:
NOAA Operations and Research join forces to implement recent scientific and technical
advances to better predict a possible flood in western Washington and help calm public fears
regarding reduced flood protection from Howard A. Hanson Dam.
ABSTRACT
The Howard A. Hanson Dam (HHD) has brought flood protection to Washington’s Green
River Valley for more than 40 years and opened the way for increased valley development near
Seattle. However, following a record high level of water behind the dam in January 2009 and the
discovery of elevated seepage through the dam’s abutment, the U.S. Army Corps of
Engineers declared the dam “unsafe.” NOAA’s Office of Oceanic and Atmospheric Research
(OAR) and National Weather Service (NWS) worked together to respond rapidly to this crisis for
the 2009/10 winter season, drawing from innovations developed in NWS offices and in
NOAA’s Hydrometeorology Testbed (HMT).
15
New data telemetry was added to 14 existing surface raingages allowing the gauge data to
be ingested into the NWS rainfall database. The NWS Seattle Weather Forecast Office produced
customized daily forecasts, including longer lead-time hydrologic outlooks and new decision
support services tailored for emergency managers and the public – new capabilities enabled by
specialized products from NOAA’s National Centers for Environmental Prediction (NCEP) and
from HMT. The NOAA Physical Sciences Division (PSD) deployed a group of specialized
instruments on the Washington Coast and near the HHD that constituted two atmospheric river
observatories (AROs) and conducted special HMT numerical model forecast runs. Atmospheric
rivers (ARs) are narrow corridors of enhanced water vapor transport in extratropical oceanic
storms that can produce heavy orographic precipitation and anomalously high snow levels, and
thus can trigger flooding. The AROs gave forecasters detailed vertical profile observations of AR
conditions aloft, including monitoring of real-time water vapor transport and comparison with
model runs.
Fig. D1. Base map indicating the locations of the newly telemetered raingages provided by the
NWS (open black diamonds), the atmospheric river observatory equipment deployed by PSD (three bull’s eyes), and
a few additional observing system assets in the region (see key). The wedge depicting the prime range of low-level
wind directions in ARs for Green River flooding is based on a recent study by Neiman et al. (2011).
16
Appendix E: HMT-West Legacy Network supporting California Department
of Water Resources’ Enhanced Flood Response and Emergency Preparedness
Effort within the Flood Safe Program
Allen B. White1, F. M. Ralph1, Mike Dettinger2, Dan Cayan2, Mike Anderson3, Art Hinojosa3
1
NOAA/Earth System Research Laboratory/Physical Sciences Division, Boulder, CO
2
USGS/Scripps Institution of Oceanography, La Jolla, CA
3
California Department of Water resources, Sacramento, CA
A partnership between California’s DWR, NOAA HMT and USGS/Scripps Institution of
Oceanography has been established to implement a 21st Century Observing System for
California. It includes over 90 field sites, 43 of which are for soil moisture (and rainfall), 36 for
water vapor aloft (using GPS-Met technology), 10 for newly designed snow-level-radars, and 4
atmospheric river observatories. Tools are being developed to combine these observations with
high-resolution modeling techniques into decision support tools for use in monitoring and
predicting extreme precipitation and flooding. The network is optimized for monitoring
atmospheric river conditions as storms strike the coast and move inland. It complements existing
observations, including NEXRAD and satellite systems. Implementation is over 5 years, and is
supported primarily through $10.5 M of Bond funding from California DWR. Expertise from
NOAA, USGS, DWR, and Scripps is being used to develop and deploy the systems.
A brochure from CA DWR is available online at
http://www.esrl.noaa.gov/psd/atmrivers/projects/pdf/Advanced%20Monitoring%20Network%20
FloodER%20Prgrm.pdf.
Fig. E1. Basemap showing new observations being deployed as part of the HMT-West Legacy/EFREP project
between 2009-2013.
17
Appendix F: Scanning Radar Requirements and Alternatives for QPE in the
Western U.S.
Rob Cifelli1, V. Chandrasekar2, Marty Ralph1
1
NOAA/Earth System Research Laboratory, Boulder, Colorado
2
Colorado State University, Fort Collins, Colorado
Providing accurate quantitative precipitation estimation (QPE) in the western U.S. is difficult due to the
combined challenges of low density of observations, sharp, topographically-forced precipitation
gradients, complex microphysical processes resulting from the interaction of precipitation producing
processes with orography, and limitations in the existing operational NEXRAD radar network design
(including areal coverage – see Fig. X). Moreover, the operational NEXRAD network radars are not able
to scan below 0.5 elevation angle. The result is that, due to earth’s curvature and the increasing size of
the radar beam with range, the useful range of NEXRAD for QPE is limited to < 150 km (see Fig. Y).
Figure F1. NEXRAD coverage at (a) 3 km AGL and (b) 1 km AGL using 0.5 deg beam elevation angle (from
McLaughlin et al. 2009)
18
Figure F2. Schematic diagram showing the Seattle NEXRAD (KATX) radar beam view in the Green River Basin
(south east of Seattle, WA).
Given regional and localized variability in terrain and precipitation processes, it is likely that there is no
“one size fits all” solution for monitoring extreme precipitation and QPE in complex terrain. What is
needed is an adaptive strategy to optimize the coverage (through a combination of gauges and radar) for a
given location.
The strategy for HMT’s QPE activities is called AdaPtive Precipitation Estimation Network Design
(APPEND). The APPEND approach is to evaluate the QPE systems in different topographic/regional
settings so that improved understanding of advantages, disadvantages and uncertainty of the various
sensor networks, data assimilation and modeling tools incorporated into the current QPE systems can be
obtained. The APPEND approach will include a design platform for developing next generation QPE
systems that builds upon the best aspects of the current systems and take advantage of future sensor
technologies and data processing algorithms.
Although much remains to be determined regarding optimal mix of observations, Fig. Z summarizes
characteristics and costs of 4 types of radar that could be considered. Based on the points made above
regarding QPE, and on the costs of the largest radar type, the strawman network in this white paper
focuses on C-band and X-band radar systems, both of which have been tested in the complex terrain of
the western U.S. in HMT. An additional radar system, called CASA (Collaborative Adaptive Scanning of
the Atmosphere - a sophisticated networking approach to scanning with X-band radars), could be an
option in the future, although its design has primarily targeted severe weather (i.e., tornadoes, and severe
thunderstorm detection in the Great Plains) that is less prevalent in much of the mountainous west.
CASA has demonstrated very promising QPE results in the Oklahoma region and is therefore listed in the
table below; however, its potential for QPE in mountainous terrain remains untested.
19
Fig. F3. Comparison of four types of scanning radar systems, including three that have been tested in HMT-West.
Based on the complex terrain of the west, the importance of low altitude conditions, and costs, this white paper
focuses on C-band and X-band type radars, both including polarimetric capability.
20
Appendix G: Categories for Comparison and Communication of ExtremePrecipitation Storm Totals across the US: Potential Application to NWS
Hydrologic Operations
David Reynolds (NWS Monterey), F. Martin Ralph (NOAA/ESRL/PSD),
and Michael Dettinger (US Geological Survey)
The magnitude and frequency of occurrence of extreme precipitation events 1 are documented
geographically and seasonally across the contiguous U.S using a simple categorization scheme [1] with
potential for use in technical and public arenas (Fig. 1).
Fig. G1. Three-day accumulations of precipitation (P) from 59 years of Cooperative Observer (COOP)
precipitation measurements [2] at 5877 sites were used to construct the categorization scheme for extreme
precipitation used in the figures and table, according to:
1
Extreme rainfall in this context refers to large area wide rainfall events, most often associated with land falling
tropical cyclones or extratropical winter cyclones, with the potential to cause main stem river flooding as opposed
to smaller scale convectively driven events associated with flash floods or urban flooding. The categories here are
determined by maximum 3 day point rainfall.
21
Rainfall Category 1 (R-CAT 1) storms totaling 200<P<300 mm (~8-12 in)
Rainfall Category 2 (R-CAT 2) storms totaling 300<P<400 mm (~12-16 in)
Rainfall Category 3 (R-CAT 3) storms totaling 400<P<500 mm (~16-20 in)
Rainfall Category 4 (R-CAT 4) storms totaling P>500 mm (~>20 in)
Storms in these categories are of similar magnitude and frequency east and west of 105ºW, where
a natural dividing line between western extreme precipitation events (predominately cool season events
associated with land-falling atmospheric rivers [3, 4]) and eastern events (predominately warm season and
mostly associated with land-falling tropical cyclones) was identified. This is shown in the insert to Fig. 1,
and in Fig. 2.
Fig. G2 Illustration of the comparability of frequency of occurrence of extreme precipitation west of 105ºW from
Nov-Apr relative to east of 105ºW from May-Oct.
To put the extremity and frequency of these new storm categories into perspective, consider that
the combined frequency of Category 2, 3 and 4 extreme precipitation episodes, averaging 12/yr, is nearly
the same as the number of hurricanes in the Atlantic and Eastern Pacific ocean basins, 15/yr, and as the
number of violent tornadoes (F4-F5) annually, 10-20/yr. The more extreme storms, Category 3 and 4
storms occur 3 times per year whereas “major” hurricanes (Saffir-Simpson CAT 3, 4 or 5) occur 6 times
per year somewhere in the Atlantic Basin [5], and F5 tornadoes occur 1-2 times per year [6]. Thus the
extreme-precipitation scaling system above plays much the same role in 1) focusing and standardizing
research analyses and forecasted rainfall intensities, 2) in evaluating climate change projections of
extreme precipitation, and 3) in communicating storm risks to non-specialists, as do the Saffir-Simpson
22
scale of hurricane intensities and modified Fujita scale for tornadoes (both based primarily on wind
speeds rather than precipitation). The lack of a similarly recognized and simple categorization scale for
extreme-precipitation events has made communicating and comparing storm magnitudes complicated.
This scaling helps address gaps in current quantitative precipitation information identified in scientific,
forecast and forecast-user community planning efforts in the past [7], and is an outcome of NOAA’s
Hydrometeorology Testbed [7; hmt.noaa.gov].
The rainfall categorization above differs from many earlier studies that describe rainfall in terms
of “return periods”. While return periods are useful in many applications and naturally scale for each
locality, they are difficult to compare inter-regionally and are confusing to the public. Return periods also
have significant uncertainty when return periods exceed historical record lengths and, under 21st Century
conditions, when they are likely to be nonstationary [8] and thus very difficult to estimate from real-world
precipitation series.
The fact that the most extreme precipitation events in one region of the country are not as extreme
as in others could be considered a disadvantage. However, NWS makes extensive use of categorizations
that are invaluable for those regions of the country impacted by these phenomena, e.g., hurricanes,
blizzards, and in some respects, tornadoes as well (Fig. 3), yet are rarely, if ever, used in other major parts
of the country.
23
Approximate Regional Distribution of
Major Meteorological Hazards in the U.S.
Blizzard
“Tornado
Tropical Storms
Fig G3. Schematic illustration of the inherently regional nature of several major meteorological hazards across the
U.S. The domains are not intended to be exact, but generally highlight areas most often affected by these hazards.
For operational purposes, a “site event” can be defined as an occasion when a given cooperative
station reports a precipitation total within the range of one of the new extreme-precipitation categories. In
contrast, an “episode” or “Rainfall Category-n storm” (i.e., R-CAT-1, 2, 3 or 4) can be defined as a period
during which at least one station reports precipitation within one of these ranges, and the episode or storm
is then labeled in terms of the largest 3-day precipitation total reported. A single episode thus can (and
usually does) include multiple site events. To avoid double counting, for wet spells longer than 3 days,
only one 3-day site event is cataloged (in the figures above), with that event being identified by the 3-day
window with the largest precipitation total. Missing data were excluded, as were accumulations from
multiple days reported only as a multi-day total. While daily totals are a common measure of
precipitation, multiple-day precipitation totals can be more useful with respect to progressive hazards, like
floods and some landslides. Three-day periods are used in the present categorization because, although 2and 4-day windows were also considered, roughly 1/2 of the major storms (by 3-day standards) were
missed when 2-day totals were used instead and, when using 4-day periods, one of the four days typically
contributed little to the multiple-day totals (e.g., on average, nationally, the driest day of 4-day events
24
contributed only 4% of the 4-day totals). These same definitions and standards can be applied, as needed,
to gridded QPEs and model outputs.
Fig. 4 provides an example of the application of the R-Cat methodology to the USGS ARkStorm
[9]. This multiday event consisting of two major precipitation episodes, one striking southern California
and the second impacting northern California about a week later, is a realistic characterization of a very
rare but highly probable event similar but not as intense as the 1861-62 California mega-flood.
Fig. G4 R-Cat levels for the USGS ARkStorm flood scenario.
1. Ralph, F.M., and M.D. Dettinger, submitted, How extreme is precipitation in atmospheric rivers?:
Science, 4 p.
2. National Weather Service (NWS), 1989, National Weather Service Observing Handbook Number
2—Cooperative Station Observations: Observing Systems Branch, Office of Systems Operations,
94 pp.
3. Zhu, Y., and R.E. Newell, 1998, A proposed algorithm for moisture fluxes from atmospheric
rivers: Mon. Wea. Rev. 126, 725-735.
4. Neiman, P.J., F.M. Ralph, G.A. Wick, J. Lundquist, and M.D. Dettinger, 2008, Meteorological
characteristics and overland precipitation impacts of atmospheric rivers affecting the West Coast
of North America based on eight years of SSM/I satellite observations: J. Hydrometeor. 9, 22-47.
5. Jarvinen, B.R., C.J. Neumann, and M.A.S. Davis, 1988, A tropical cyclone data tape for the North
Atlantic basin, 1886 1983 Contents, limitations,
and
uses:
NOAA
Technical
Memorandum NWS NHC 22, 21 p., http://www.nhc.noaa.gov/pdf/NWS NHC 1988 22.pdf.
6. Schaefer, J.T. and R. Edwards, 1999. The SPC tornado/severe thunderstorm database. Preprints,
11th Conference of Applied Climatology, Amer. Meteor. Soc., 215 220.
25
7. Ralph, F. M., R. M. Rauber, B. F. Jewett, D. E. Kingsmill, P. Pisano, P. Pugner, R. M.
Rassmussen, D. W. Reynolds, T. W. Schlatter, R. E. Stewart and J. S. Waldstreicher, 2005:
Improving short-term (0-48 hour) Cool-season quantitative precipitation forecasting:
Recommendations from a USWRP Workshop. Bull. Amer. Meteor. Soc., 86, 1619-1632.
8. Milly, P.C.D., J. Betancourt, M. Falkenmark, R.M. Hirsch, Z.W Kundzewicz, D.P.
Lettenmaier, and R.J. Stouffer, 2008, Stationarity is dead: Whither Water Management?:
Science, 139, 573 574.
9. Porter et al, 2011: An over view of the ArKStorm scenario. To be released in January,
2011. USGS Multi hazards Demonstration Project.
26
27
28
Appendix I: Global Positioning System-Meteorology networks (GPS-Met) –
leveraging solid earth monitoring for hydrometeorology applications
Seth Gutman1
1
NOAA/Earth System Research Laboratory, Glbal Systems Division, Boulder, Colorado
Dual-use Observations from GPS networks
The West Coast of North America is one of the most seismically active areas on Earth. To
mitigate the impact of these inevitable and occasionally catastrophic events, several U.S.
Government Agencies (notably the National Science Foundation, U.S. Geological Survey and
NASA) in collaboration with public and private academic institutions in California and other
states, have deployed an extensive network of geophysical monitoring systems and sensors in
critical areas along the North American Plate Boundary. The purpose of this network is to
improve our understanding of the tectonics of the region, characterize areas susceptible to
movements in response to geological forces, and identify regions along the plate boundary where
strain is accumulating and fault motions are most likely to occur.
One of the most widely deployed instruments is the satellite Global Positioning System receiver,
deployed in extensive networks throughout the region. These unattended instruments
continuously record the radio signals broadcast by the GPS satellites in Earth orbit, and the
information is transmitted to data processing centers where it is used to monitor the positions of
the GPS antennas in near real-time with extraordinary (millimeter) accuracy. Figure I1 (courtesy
of the California Real-Time Network headquartered at Scripps Institution of Oceanography
http://sopac.ucsd.edu/projects/realtime/) shows the locations of many of the continuously
operating GPS reference stations in California. Red and blue symbols denote the locations
providing data to CRTN in real-time; light blue and pink symbols identify the locations of
planned real-time systems. Similar, though less dense, networks also exist across many of the
Western States for solid earth monitoring (Fig. I2).
One source of error in monitoring the positions of these systems comes from changes in the
atmosphere associated with weather events. The scientists and engineers who developed the
techniques used to monitor site positions with this level of precision also recognized the impact
of the atmosphere on GPS measurement accuracy and they developed clever ways to reduce
atmospherically-induced errors by treating the atmosphere as a nuisance parameter and removing
its effects. At about the same time this was happening, atmospheric scientists in Boulder, CO
were looking for improved (i.e. more accurate, cost effective and all-weather) ways to monitor
water vapor in the atmosphere. Why did they want to do this? Water vapor is the source or raw
materials of clouds and precipitation, and an ingredient in most major weather events. The
problem is that water vapor at that time was very difficult to observe because it varies rapidly on
the scale of clouds. So to monitor water vapor effectively, they needed lots of closely spaced
measurements made rapidly. The most operational common ways to observe water vapor at that
time were to use weather balloons and/or the information provided by Earth orbiting
environmental satellites. The problems with these methods were that weather balloons were (are
are) expensive and launching them is labor intensive so they are only launched twice daily at
sites spaced about 400 km (250 mi) apart. Satellites are incredibly expensive, the moisture
29
observations over land are not reliable in the presence of clouds, and satellite techniques to
monitor water vapor under all weather conditions are effective only over the open oceans.
When it was realized that the errors in position caused by weather were extremely useful to
meteorologists, new applications rapidly emerged. Dual use of GPS observations for
simultaneously monitoring the solid Earth and the Earth’s atmosphere is not only cost effective,
it’s almost free. GPS techniques are now used routinely by NOAA for weather forecasting and
climate monitoring. One of the most dramatic and successful applications of GPS technology in
weather forecasting is the use of CRTN GPS data to monitor land-falling atmospheric rivers and
make improvements in heavy precipitation and flash flood predictions for the people of
California, as is being done with the implementation of the HMT-West Legacy network through
the California DWR-sponsored EFREP project (Appendix E). In key ways it is the best inland
version of the very useful SSM/I satellite data offshore (SSM/I cannot derive water vapor content
over land due to technical constraints on the remote sensing method it uses).
Fig I1. Locations of continuously operating GPS receivers in CA. Red and blue symbols are sites
currently delivering real-time data to the California Real-Time Network (CRTN) for use in earthquake
monitoring and other applications. Light blue and pink symbols are other potential sites for the network.
30
Fig. I2. Network of GPS Solid Earth monitoring stations around the Western U.S. that could potentially
be modified t include real-time communications and surface meteorology sensors and could thus provide
continuous monitoring of vertically integrated water vapor (IWV) that is a key factor in extreme events.
31
Appendix J: Soil moisture observing network design and instrumentation
R. J. Zamora1, F. M. Ralph1, E. Clark2, and T. Schneider2
1
NOAA/Earth System Research Laboratory/Physical Sciences Division, Boulder, CO
2
NOAA/National Weather Service
The National Oceanic and Atmospheric Administration (NOAA) Hydrometeorology Testbed
(HMT) program has deployed soil moisture observing networks in the watersheds of the Russian
River and North Fork (NF) of the American River in Northern California, and the San Pedro
River in southeastern Arizona. These networks were designed to serve the combined needs of the
hydrological, meteorological, agricultural, and climatological communities for observations of
soil moisture on time scales that range from minutes to decades.
The networks are a major component of the HMT program that has been developed to accelerate
the development and infusion of new observing technologies, modeling methods and recent
scientific research into the National Weather Service (NWS) offices and to help focus research
and development efforts on key hydrological and meteorological forecast problems. These
forecast problems are not only of interest to the NWS, but they also play a crucial role in
providing input to water managers who work at the national, state, and local government levels
to provide water for human consumption, agricultural, and other needs.
The HMT soil moisture networks have been specifically designed to capture the changes in soil
moisture that are associated with heavy precipitation events and runoff from snowpack during
the melt season. This paper describes the strategies used to site the networks and sensors as well
as the selection, testing, and calibration of the soil moisture probes. Figure J1 illustrates the
design of a soil moisture monitoring site.
In addition two illustrative examples of the data gathered by the networks are shown. The first
example (Fig. J2) shows changes in soil moisture observed before and during a flood event on
the Babocamari River tributary of the San Pedro River near Sierra Vista, AZ, on 23 July 2008.
The second example (Fig. J3) examines a 5-year continuous time series of soil moisture gathered
at Healdsburg, CA. The time series illustrates the transition from a multiyear wet period to
exceptionally dry conditions (drought) from a soil moisture perspective.
Zamora, R. J., F. M. Ralph, E. Clark and T. S. Schneider, 2011: The NOAA Hydrometeorology
Testbed soil moisture observing networks: Design, instrumentation and preliminary results. J.
Atmos. Oceanic Technol., 28, 1129-1140.
32
Fig. J1. Schematic diagram showing the orientation of the soil probes and surface
meteorological observations used at a typical soil moisture observing station.
Fig. J2. Evolution of the cumulative precipitation, and soil wetness fractions measured at 10, 20,
and 50 cm depth for the period 18-26 July 2008 at Whetstone, AZ that included a flash flood.
33
Fig. J3. Observations from coastal California. Five year soil wetness fraction climatology
(black), daily total precipitation (red) and atmospheric rive events (blue). Dashed line indicates a
soil wetness fraction of 0.25. Palmer Drought Indices are shown in bold type.
34
Appendix K: An Approach to Enhancement of the Western SNOTEL and
Streamgage Networks for Snowmelt, Rain-on-Snow and other Flood
Mechanisms
Mike Dettinger (USGS, Scripps Inst. Oceanography, La Jolla, California)
The snow fields of the mountains of the western US are important contributors to high flows and
flooding in the region, either during especially rapid and voluminous snowmelt seasons (e.g.,
2011) or when they melt rapidly or facilitate rapid rainfall runoff during warm storms. McCabe
et al. (2007) recently found a wide range of trending rain-on-snow frequencies around the
western US. Lower altitudes areas have tended to have fewer such episodes in recent decades, as
snowlines have retreated upward with regional warming so that low snowfields have been less
available to be rained on; higher altitudes have tended to experience more rain-on-snow episodes
as warming has brought rainfall to snowfields at ever higher altitudes. Marks et al (2001) have
shown that, during some well monitored and simulated rain-on-snow events in Idaho, the
delivery of heat into snowpacks by the rain itself was a small contribution compared to the heat
deposited on snow by rapid condensation of vapor onto the surface. The heat released by this
condensation was the primary cause of melting. Even when melting snow is a minor contribution
(as often happens at middle and higher altitiudes), rain falling onto snowy surfaces can run off
quite efficiently and rapidly so that the rain itself can yield significant flooding. In warm storms,
very large parts of river basins that normally would receive snowfall and generate little
immediate runoff can instead receive rainfall so that the areas contributing rapid runoff to
downstream flows and flooding are greatly enhanced (Detinger et al. 2009). For example, White
et al. (2002) showed that for each of four watersheds studied in California, there existed an
altitude range for which a 2000 ft rise in the snow level of a storm would triple the runoff.
Thus a vision of observations for extreme storms and flooding in the West necessarily includes
consideration of the region’s snowfields (Bales et al., 2006) and, in particular for the purposes of
this plan for extreme storms and flooding, a focus on surficial and meteorological observations
of those snowfields. At present, high-altitude precipitation, snow-water contents, snow depths,
and air temperatures are continually monitored at almost 1000 SNOTEL (or equivalent) sites in
the western US, by the USDA Natural Resources Conservation Service and by the California
Cooperative Snow Surveys. These observations are used to track the seasonal storage and release
of water for water-supply forecasts and management purposes across the region.
As these sites are already established and operated, for the purposes of the present vision for
storm and flood monitoring in the region, several enhancements at selected SNOTEL sites are
proposed. Thus, at about 20 SNOTEL sites in each of the ovals in Fig. K1, enhanced
observational regimes would be instituted to track more closely the fate of the snow waters in the
soil and on the landscape and to monitor several conditions of flood generation associated with
warm storms.
35
Fig. K1. Locations of existing SNOTEL (or equivalent) stations in the West, with subregions
within which enhancements might be undertaken indicated by red ovals.
Immediate enhancements to monitoring of the fate of snowmelt in the soil could be achieved by
addition of soil-moisture and temperature monitors at several depths below the ground surface at
the selected SNOTEL sites. With new interests in monitoring and forecasting floods from rapid
snowmelt and rainfall runoff in the mountain catchments of the West, especially as new areas are
exposed to rainfall and snowmelt accelerations in response to warming trends, the region’s
streamflow networks need to be preserved and, if possible, enhanced along with the SNOTEL
network (e.g., Kenney et al. 2011). Enhanced monitoring of the fate of snowmelt (and flooding)
across the western landscape could be accomplished by some combination of insuring the
funding sources for selected existing streamgages (thus ensuring their permanent operation), by
restarting some long-term gages (where appropriate) that have been discontinued in recent years,
and by establishing new gages, particularly in increasingly important high-altitude locations.
Next, with snowmelt- and warm-storm contributions from the snowfields to flooding in the
region in mind, a new regime of monitoring of conditions at the surface of the snowpack could
also be added to the selected SNOTEL sites. These snow-surface observations would include
36
monitoring of wet-bulb temperatures to complement the existing dry-air temperature
measurements. Wet-bulb temperatures is a measure of humidity and availability/readiness of
vapor for condensation, and would reflect the conditions under which large amounts of water
vapor are available and likely to condense onto and into snowpacks, which—as noted
previously—can result in rapid melting of snow. The mass of vapor that condenses under such
conditions also depends on how much moist air is passing over the snow surface, so that winds
would also be monitored. Under other conditions (when condensation is notw the immediate
cause of snowmelt), the amount of solar radiation absorbed by the snowpack is critical, and thus
these sites would also be enhanced to include continual measurements of solar insolation and
surface albedo.
The measurements proposed thus far draw on well-established technologies and methods, and
would leverage the existing power and communications at SNOTEL sites. For the future, an
important direction for new developments in monitoring of snowpacks of the West might focus
more on conditions in the interior of the snowpacks (i.e., between the ground surface and the top
of the accumulated snow). There is an increasing understanding that dust and soot that falls on
snow and that gets incorporated into snowpacks can greatly affect snowmelt timing and runoff
amounts (Painter et al. 2011) and that aerosols in the atmosphere can influence the forms and
amounts of precipitation, including snowfall, in the first place (Ault et al. 2011). Furthermore it
is well established that structures like ice layers inside of snowpacks influence rates, locations,
and amounts of snowmelt and release of melt waters into streams (Pfeffer and Humphrey, 1996).
It would be useful to know more about what is happening in the interior of snowpacks to
improve our understanding and predictions of the consequences of extreme storms and fairweather melt conditions.
Thus we also envision a future phase of monitoring enhancements at the selected SNOTEL sites
to add measurements of aerosols in the near-surface atmosphere or falling onto the snow surfaces
and regular (albeit probably only once-per-year once the bulk of the year’s snowpack has been
formed) sampling of snow for measurements of dust, black-carbon, and aerosol contents. The
strategies and some of the technologies to make these measurements in long-term operational
modes have not yet been developed and so implementation will be preceded by some research
and development. Monitoring of physical conditions within snowpacks are also somewhat
speculative, but could include continual measurements of snow temperatures and densities at
various heights above the soil surface in the snowpack and might (with more development)
include monitoring of liquid-water contents at the same heights.
Ault A.P., Williams C.R., White A.B., Neiman P.J., Creamean J.C., Gaston C.J., Ralph F.M.,
Prather K.A., 2011, Detection of Asian dust in California orographic precipitation: Journal of
Geophysical Research - Atmospheres.
37
Bales, R., Molotch, N., Painter, T., Dettinger, M., Rice, R., and Dozier, J., 2006, Mountain
hydrology of the western US: Water Resources Research, 42, W08432,
doi:10.1029/2005WR004387, 13 p.
Dettinger, M.D., Hidalgo, H., Das, T., Cayan, D., and Knowles, N., 2009, Projections of
potential flood regime changes in California: California Energy Commission Report CEC-5002009-050-D, 68 p.
Kenney, T.A., Buto, S.G., and Susong, D.D., 2011, Analysis of watersheds monitored by the US
Geological Survey streamflow-gaging station network in the Upper Colorado River basin: US
Geological Survey Scientific Investigations Report 2011-5081, 102 p.
Marks, D.G., Link, T., Winstral, A.H., and Garen, D., 2001, Simulating snowmelt processes
during rain-on-snow over a semi-arid mountain basin: Annals of Glaciology, 32, 195-202.
McCabe, G.J., Clark, M.P., and Hay, L.E., 2007, Rain-on-snow events in the western United
States: Bull., American Meteorological Society, 88, 319-328.
Painter, T.H., Deems, J.S., Belnap, J., Hamlet, A.F., Landry, C.C., and Udall, B., 2011,
Responses of wind erosion to climate-induced vegetation changes on the Colorado Plateau: Proc.
Natl. Acad. Sci., 108, 3854-3859.
Pfeffer, W.T., and Humphrey, N.F., 1996, Determination of timing and location of water
movement and ice-layer formation by temperature measurements in sub-freezing snow: Journal
of Glaciology, 42, 292-304.
White, Allen B., Daniel J. Gottas, Eric T. Strem, F. Martin Ralph, Paul J. Neiman, 2002: An
automated brightband height detection algorithm for use with Doppler radar spectral moments. J.
Atmos. Oceanic Technol., 19, 687-697.
38
Appendix L: Dust-on-snow sampling
Jayne Belnap (USGS) and
Mike Dettinger (USGS, Scripps Inst. Oceanography, La Jolla, California)
Dust production is increased by drought, disturbance of the soil surface, the invasion of exotic
annual grasses and fire. Droughts reduce soil moisture, reducing plant cover and thus soil surface
protection. Disturbance reduces or removes components that naturally stabilize soils and soildisturbing activities—energy exploration and development, grazing, and recreation—are
increasing substantially in the western U.S. Surface disturbance enhances the invasion of exotic
annual grasses. Their presence increases fires that, in turn, increase dust production. Also, in
drought years, annual grasses do not germinate, leaving soils barren and vulnerable to erosion.
Future increases in temperatures will slow the recovery of vegetation and soil from land use
disturbance, further increasing the frequency and magnitude of dust storms.
Large dust storms have both local and regional effects on human and natural resources. Small
particles can be lodged in the lungs, causing respiratory disease, and dust can carry organisms
that cause diseases such as Valley Fever. Fatalities have resulted when visibility has been
reduced on highways. Soil fertility is lost with the dust because nutrients are often attached to
dust particles.
Much of the dust produced from low-elevation land is deposited on the snowpack of nearby
mountains. The dark-colored dust absorbs heat, increasing the melt rate of the underlying
snowpack. Earlier or more rapid melt may contribute to, or decrease, flood risks in the melt
seasons. Early melting leaves soils exposed longer to solar radiation and gives plants a longer
growing season, resulting in more water being lost to the atmosphere and thus less water entering
streams and rivers. In addition, late season water supplies are reduced (Painter et al., 2010).
Water quality is reduced when windblown material is deposited in ephemeral washes or when
less quantity concentrates sediment, pollutants, and results for higher temperatures.
A reduction in water quantity and quality affects most aspects of resource management in the
western U.S., as water is needed for wildlife, agriculture, energy production, household
consumption, and recreation. Early snowmelt and reduced surface flow will produce additional
challenges for those responsible for delivering this water, as climate changes and increasing
human populations will increase the demand for water just as supplies are diminishing. Future
energy production alternatives, including solar thermal energy, oil shale development, and
nuclear power, also use large amounts of water; hence a decline in water availability may limit
39
their siting and production levels as well. There is a multi-billion dollar recreation and tourism
industry in this region is also dependent upon current stream flows and lake levels.
In order to more effectively manage our water resources, we need a better understanding of 1)
what factors affect dust production, including vegetation, soils, and the timing, type, and
intensity of activities; 2) what areas consistently produce dust and what are the relative total dust
contribution of low intensity diffuse sources (e.g., livestock grazing) versus acute point sources;
3) the major dust deposition zones and where are they relative to sources and 4) how does dust
impact snow melt under different conditions? Answering these questions will require a
monitoring network that continues through time, with dust outputs and inputs measured.
In the present vision, monitoring to better understand and track (from year to year) the role of
dust in snowmelt across the region, snow samples will be collected late (approximately May) in
the snowpack development season, but prior to the onset of major melting, and the samples will
be analyzed for total dust contents. Sample sites will most likely be associated with either snow
courses or SNOTEL installations, and snow from the same sites will be sampled each year.
These dust determinations will provide a baseline complement to ongoing and developing
research efforts to monitor the conditions that encourage dust production and the radiative effects
of dust in snow by other groups.
Painter, T. H., J. S. Deems, J. Belnap Jayne, A.F. Hamlet, C. Landry and D. Udall, 2010:
Response of Colorado River runoff to dust radiative forcing in snow. Proc. Nat. Academy
Sciences U.S.A. Vol. 107, pp 17125-17130, DOI: 10.1073/pnas.0913139107
40
Appendix M: ARkStorm Emergency Preparedness Scenario
Mike Dettinger (USGS. Scripps Inst. Oceanography), F. M. Ralph (NOAA), M. Hughes
(CIRES), T. Das (Scripps Inst. Oceanography), P. J. Neiman (NOAA), D. Cox (USGS), G. Estes
(Extreme Precip Symposium), D. Reynolds (NOAA), R. Hartman (NOAA), D. Cayan (USGS.
Scripps Inst. Oceanography), and L. Jones (USGS)
The USGS Multihazards Project is working with numerous agencies to evaluate and plan for
hazards and damages that could be caused by extreme winter storms impacting California.
Atmospheric and hydrological aspects of a hypothetical storm scenario have been quantified as a
basis for estimation of human, infrastructure, economic, and environmental impacts for
emergency-preparedness and flood-planning exercises. In order to ensure scientific defensibility
and necessary levels of detail in the scenario description, selected historical storm episodes were
concatentated to describe a rapid arrival of several major storms over the state, yielding
precipitation totals and runoff rates beyond those occurring during the individual historical
storms. This concatenation allowed the scenario designers to avoid arbitrary scalings and is
based on historical occasions from the 19th and 20th Centuries when storms have stalled over the
state and when extreme storms have arrived in rapid succession. Dynamically consistent, hourly
precipitation, temperatures, barometric pressures (for consideration of storm surges and coastal
erosion), and winds over California were developed for the so-called ARkStorm scenario by
downscaling the concatenated global records of the historical storm sequences onto 6-and 2-km
grids using a regional weather model of January 1969 and February 1986 storm conditions. The
weather model outputs were then used to force a hydrologic model to simulate ARkStorm runoff,
to better understand resulting flooding risks. Methods used to build this scenario can be applied
to other emergency, nonemergency and non-California applications.
Dettinger, M. D., F. M. Ralph, M. Hughes, T. Das, P. J. Neiman, D. Cox, G. Estes, D. Reynolds,
R. Hartman, D. Cayan, and L. Jones, 2011: Design and quantification of an extreme winter
storm scenario for emergency preparedness and planning exercises in California., Nat. Hazards,
(in press July 2011).
Porter, Keith, Wein, Anne, Alpers, Charles, Baez, Allan, Barnard, Patrick, Carter, James, Corsi,
Alessandra, Costner, James, Cox, Dale, Das, Tapash, Dettinger, Michael, Done, James, Eadie,
Charles, Eymann, Marcia, Ferris, Justin, Gunturi, Prasad, Hughes, Mimi, Jarrett, Robert,
Johnson, Laurie, Dam Le-Griffin, Hanh, Mitchell, David, Morman, Suzette, Neiman, Paul,
Olsen, Anna, Perry, Suzanne, Plumlee, Geoffrey, Ralph, Martin, Reynolds, David, Rose, Adam,
Schaefer, Kathleen, Serakos, Julie, Siembieda, William, Stock, Jonathon, Strong, David, Sue
Wing, Ian, Tang, Alex, Thomas, Pete, Topping, Ken, and Wills, Chris; Jones, Lucile, Chief
41
Scientist, Cox, Dale, Project Manager, 2011: Overview of the ARkStorm scenario: U.S.
Geological Survey Open-File Report 2010-1312, 183 p. and appendixes.
Fig. M1. Summary map of some extreme weather conditions resulting from the ARkStorm
scenario associated with two stormy periods impacting the state over 23 days. The conditions
are based on real storms from 1986 and 1969, but assuming they occurred back-to-back, and that
the 1969 storm stalled for an additional day. Each storm included a strong atmospheric river.
42
Appendix N: Example of Utility of Gap-Fill Radar in a Colorado
Flash Flood Event
Rob Cifelli1, S. Matrosov2
1
2
NOAA/Earth System Research Laboratory, Boulder, Colorado
Cooperative Institute for Research in Environmental Sciences, Colorado
In late summer 2010, a wildfire burned over 6,000 acres and 160 homes in the Fourmile Canyon
area, immediately west of Boulder, Colorado (Fig. N1-top). As a result of the fire, much of the
steep terrain within the canyon area became vulnerable to debris flows and flash flooding. The
FourMile Canyon is area is ~80 km from the nearest NEXRAD radar (KFTG) and is not well
monitored by the operational National Weather Service (NWS) radar network.
As a result of the fire and in response to
concerns from the NWS regarding the increased
disaster potential, NOAA/Physical Sciences
Division (PSD) deployed a X-band polarimetric
radar in Erie, CO during the summer of 2011 to
supplement the NEXRAD coverage in the area
and determine the added value of gap fill radar
data in this region. On July 13th 2011, a series
of slow moving storms produced heavy rain and
hail over parts of FourMile Canyon, resulting in
debris flows and flash flooding in FourMile
Canyon (Fig. N1-bottom).
Figure N1. (top) Smoke plume from the FourMile
Canyon fire in September 2010 (image courtesy of
NASA). The locations of Boulder and Denver are
indicated. (bottom) Flood waters in FourMile
Canyon on July 14, 2011.
An example of the difference in radar coverage between the NOAA-PSD X-band and NEXRAD
KFTG for the July 13, 2011 event is shown in Fig. N2. The X-band radar, located only 30 km
from the FourMile Canyon region, provides higher resolution and is able to capture the spatial
gradients in precipitation much better compared to the NEXRAD. Although the X-band system
operates at a higher frequency and is more prone to signal loss in heavy rain than NEXRAD, the
polarimetric phase information can be used to correct for the attenuation and allow for improved
rainfall estimation compared to NEXRAD.
43
Table N1 lists the total rainfall accumulation from a rain gauge deployed by the National Center
for Atmospheric Research in FourMile Canyon as well as the corresponding NEXRAD and
NOAA X-band for an approximate one hour period on July 13. The NEXRAD total is biased
high compared to the rain gauge due to the existence of hail mixed with rainfall. The NEXRAD
radar cannot discriminate between the hail and rain and overestimates the actual rainfall total. In
contrast, the X-band radar shows better agreement with the gauge due to the polarimetric phase
information, which is relatively immune to hail and is able to isolate the rainfall contribution.
Figure N2. (top) NEXRAD radar reflectivity at 00:13 UTC on 14 July, 2011 (18:13 MDT 13 July, 2011).
(bottom) Corresponding attenuation-corrected image from the NOAA-PSD X-band radar. The FourMile
Canyon burn perimeter is indicated by the irregular boundary in each image. The location of the rain
gauge in FourMile Canyon used for rainfall comparisons is indicated by the white circle in each image.
The X-band radar data shown have not been quality controlled and should be considered preliminary.
44
The combination of high resolution and polarimetric information demonstrates the utility of gap
fill radar for rainfall estimation in watersheds where NEXRAD coverage is poor.
Rain Gauge (inches)
0.9
NOAA-PSD X-band radar (inches)
0.7
NEXRAD radar (inches)
1.5
Table N1. Rainfall total for the period 23:45 (13 July) - 01:00 (14 July) UTC (17:45-19:00 MDT 13 July).
Rainfall totals are based on preliminary data and are approximate.
45