RAIN-ON-SNOW OCCURRENCE & MAGNITUDE ACROSS ELEVATIONS�IN MARITIME WESTERN WASHINGTON �����Monte Carlo Simulation of Large Storms Under Recent & Projected Climatic Conditions

Matt Brunengo G S O C Meet-Up

mbruneng@pdx.edu 22 April 2023

(2013 YouTube)

What is rain-on-snow?

• Snowmelt during rainfall
• commonly / importantly in big storms
• temperatures typically warmer than seasonal

​

• combined rain + melt 🡪 major water inputs

​

• rain doesn’t cause most of the melting

major energy sources = sensible & latent heat, long-wave radiation

BUT – rain contributes some heat and most WATER

Where does ROS occur?

• Most common in maritime–temperate climates

​

​

Western North America

• winter storms from Pacific
• snow on the mountains

🡪 common in middle elevations

​

​

(see: McCabe, Clark & Hay, March 2007, Bull. AMS)

​

Atmospheric rivers:

• Wikipedia article
• NASA: 1 week global precipitation
• NASA: video

5

local topographic effects:

Puget Sound convergence zone,

on lee side of Olympic Mtns; flow thru Chehalis Gap

orographic lifting 🡪 rainy / snowy on windward side, dry (rain shadow) on lee side (chinook winds)

When does ROS occur?

• Seasonal storm occurrence
• frontal–cyclonic storms riding jet stream
• especially in atmospheric rivers / pineapple express

(warm air holds more moisture)

• most occur in fall & winter rain > melt
• some in spring melt > rain
• influenced by atmospheric circulation patterns

​

• no specifically “ROS storms” – many factors must combine

​

​

1^st recorded ROS in the PNW ?

8 January: “ …. we would observe that the almost constant mild, warm rains of the last two weeks have completely annihilated the snow on the various prairies, and it is fast disappearing from the woodland.”

​

15 January: “The heavy rains which have followed in the wake of the unprecedented depth of snow, have almost flooded the country between [Olympia] and the Columbia. The Cowlitz, we understand, has become as wild as a cataract — completely forbidding navigation .... The Newaukum and Skookum Chuck .… are said to be higher than ever known in the recollection of the oldest [Euro-American] inhabitant .… “

The Columbian

Olympia, Northern Oregon Territory, 1853

Major ROS events in the PNW

• Wellington avalanche, 1910
• Vanport flood, Memorial Day 1948
• Christmas 1964
• January 1965
• 1975–76
• 1977
• 1980
• 1983

​

Major ROS events in the PNW

​

• 1989–91
• Feb 1996
• 1996–97
• (November 2006 – not really ROS – little snow yet)
• December 2007
• January 2009

(most aren’t catastrophic)

​

Should we care about ROS?

​

“Major floods in the [White–Puyallup], as on all streams of the Puget Sound area, occur almost without exception during November to March, and are augmented to a marked degree by the rapid melting of the heavy mantle of snow that normally covers the higher portions of the area during those months. Obviously, therefore, any investigation into the flood hydrology [or LS, or …] of the basin must include the determination not only of the precipitation that falls during the flood period, but also of the antecedent precipitation stored in the form of snow and of the factors that accelerate or retard its melting.”

Geo. F. Hopkins, 1940, Transactions AGU

Hydrologic investigations for Mud Mtn Reservoir

Why does ROS matter?

• Important water–input process in PNW
• changes snowpack storage vs runoff relations

🡪 less water in spring & summer

​

• ROS affects people (and our stuff)
• high runoff, floods, channel erosion, habitat
• landslides, avalanches
• roof loads, transportation problems, etc.

​

• People affect the process
• effects of land use (forest mgt ), climate change

Why does ROS matter?

• Hydrologic significance
• ROS relative to simple precipitation

* what is measured in the gauge?

​

​

​

* what goes into the ground?

​

Why does ROS matter?

• Hydrologic significance
• ROS relative to simple precipitation

trivial * rainwater alters the snowpack

* some water leaves the pack (WAR )

significant * WAR > storm precipitation

​

• magnitude: how much?
• frequency: how often?
• geography: where?

Use the record to study ROS?

• Limited sites, instruments, record length
• National Weather Service
• usually temp, precip, snow depth (some wind)
• typically at lower elevations
• Cooperative Snow Survey (NRCS)
• snow courses to SNOTEL
• snow–water equivalent (SWE) (some depth, temp)
• typically at higher elevations
• few stations of any kind in middle elevations
• Most sites ~70 years maximum (as of 2005)
• Record is too short and limited to get a large & broad sample of events

West–central Washington Cascades

• Seattle & Tacoma municipal watersheds, flood control projects
• National Weather Service: 7 COOP stations + Stampede Pass
• Cooperative Snow Survey: 28 snow courses and/or SNOTEL
• Records from 1940 – 2005+

Cascades terrain

• average elevations

(Mitchell & Montgomery,

Quaternary Research, 2005)

• but a bit of variation ….

upper Cedar River

Stampede Pass

​

• NWS airways observation site
• 3958 ft ≈ 1206 m
• T, P, W, snow, etc.
• hourly or better
• staffed: high-quality observations
• almost continuous since mid-1940s
• snow course & pillow downhill

(3860 ft / 1175 m)

Sidebar: snow measurement

• Several ways to quantify snow
• snowfall: daily accumulation
• depth, measured on a snow board
• snow on the ground = accumulation to that point
• also as depth, for a given time or day
• measured manually or instrumentally
• occasional (at a station or by snow survey) or continuous (weather sta’s, sensors at Snotel sites)
• snow water equivalent (swe)
• amount of liquid water the snowpack would yield
• also occasional (weather stations) or continuous (snow pillows at Snotel sites)

Meadows Pass SNOTEL site, Cedar River basin��NRCS snow information

• 20 snow courses
• 8 SNOTEL sites
• many co-located
• ≥ 10 yr record

Snow measurements

The problem

ROS events vary with climate, weather, terrain and vegetation, all of which affect their magnitude, frequency & spatial effects over time

Observational record is too short & limited to get a large & broad sample of events

​

Project: use probabilistic modeling to quantify aspects of ROS over the long term

Probabilistic modeling

• Monte Carlo methods
• use of random sampling techniques

(commonly in computer simulations)

to obtain approximate solutions

for mathematical & physical problems

• Virtual experiments – compensating for sporadic events & scattered stations

Model structure

• Computer model of storm events
• Excel workbooks, VisualBasic code

​

• M–C simulation for probabilistic factors
• sample from frequency distributions 🡪

initial & hourly weather & snow conditions

​

• deterministic components
• calculate snow accumulation and melt
• track percolation of rain ± meltwater 🡪 ground

​

Model architecture

• Monte Carlo mode (the full Monte )
• specify parameters & frequency distributions
• random numbers 🡪 generate event values
• 1000 yr = ~4400 events 🡪 long realization series

​

• climate “constant”
• simple code
• precip events, sites

just moving on ….

M–C Results: for StpP

M–C Results: for StpP

Results: Elevation

ROS variations with elevation

ROS–susceptible elevations

rain-on-snow zone

Rain-on-snow in the W Washington Cascades

Model: statistically, WAR ≠ P at most elevations

Most important in mid-elevations maximum frequency & hydrologic significance of ROS

preferred ROS zone ~ 500–1100 m

peak at ~ 800 m

Why should we care, again?

• Climate change: what’s happening?
• temperatures rising
• precipitation – uncertain (some +, some –)
• snow volumes declining, snowlines rising

​

• The Monte Carlo model can be used to examine climate-change scenarios

Climatic shifts

• Climate change: what’s happening?
• temperatures rising
• snow volumes declining (indexed to Δ 1 Apr swe)

​

• Scenarios
• cc 1: storm T_i + 1° C, swe less 10%
• cc 2: storm T_i + 2° C, swe = f [ elev ]
• cc 5: storm T_i + 5° C, swe = f [ elev ] / 2

Previous results: elevation

Rain-on-snow in the Northwest in the 21^st century

Still – most important in mid-elevations

Model: for T↑ and SWE↓

🡪 more ROS events in most areas

🡪 preferred ROS zone moves up

​

​

Effects of ROS also likely to shift & expand

Conclusions

• Monte Carlo simulation is a useful tool for studying highly variable phenomena such as the big storms producing ROS

• M–C model seems to work
• reproduces the input series
• generates reasonable snow & perc responses
• creates interesting frequency distributions
• not perfect or universal: limitations, oddities

​

Conclusions

• For the west–central Cascades, the model results indicate:

​

• Differences during many events over long time periods 🡪 WAR is statistically different from precipitation, at all elevations examined

​

• Results suggest a maximum in the frequency & hydrologic significance of ROS between ~ 500 – 1200 m in the study region, with peak at ~ 800 m

So – why should we care?

• Frequency analysis: WAR ≠ precip
• be careful when using rain–gauge data to calculate magnitude–frequency relations
• in many events, the ground feels something different from gauged precipitation
• differences increase from lower 🡪 higher elevations

Climatic shifts

• Climate change:
• temperatures rising, snowpack shrinking

​

• Model can examine change scenarios
• greater % of ROS events, and the ROS zone is moving slightly uphill
• changes in hillslope & fluvial processes ?

So – why should we care?

• Land use: forest management
• removed from this project early on (or I’d still be working on it … )
• in ROS, differences in:
• snow accumulation
• melt rate

​

​

• does large-scale harvest affect ROS processes and rates enough to change hillslope and fluvial processes?

Questions ?

Web sites in this presentation

• MJB presentation, 2013 PNW Climate Science Conf

https://www.youtube.com/watch?v=ruNuhGe12mY

​

• Atmospheric rivers

Wikipedia https://en.wikipedia.org/wiki/Atmospheric_river

NASA global satellite loop https://svs.gsfc.nasa.gov/vis/a000000/a004200/a004285/imergert_1080p_30.mp4

NASA video (4:47) https://svs.gsfc.nasa.gov/4960

​

• USDA NRCS snow survey and Snotel information

Nat’l Water & Climate Ctr www.wcc.nrcs.usda.gov

​

Weather & climate web sites

• NOAA, National Weather Service

National www.weather.gov

Portland, NW OR & SW WA www.weather.gov/pqr

​

• Climate data

Western Reg’l Clim Ctr www.wrcc.dri.edu

Nat’l Clim Data Ctr www.ncdc.noaa.gov

NRCS Nat’l Water & Clim Ctr www.wcc.nrcs.usda.gov

Oregon Clim Service (OSU ) ocs.oregonstate.edu

UW Atmos Sci Dept www.atmos.washington.edu

UW Climate Impacts Group www.cses.washington.edu

​

53

Hydrology & streamflow sites

• USGS Water Resources Division

Oregon www.usgs.gov/centers/or-water

Washington www.usgs.gov/centers/wa-water

Waterwatch water.usgs.gov/waterwatch

USGS real-time data waterdata.usgs.gov/nwis/rt

​

• Flood-related sites

FEMA (FIRM maps, etc.) www.fema.gov

NWS river obs & forecasting www.weather.gov/ahps/

NWS flood damage data www.fema.gov/data-visualization/historical-flood-risk-and-costs

​

​

​

54