Transcript Statistics of Extreme Events Pacific Northwest Climate Change

Climate Change ???
How Is it Detected?
Difficulties of Detecting Climate Change
Climate systems are intrinsically noisy
 The climate system is not well sampled
 Insufficient temporal baseline to determine
proper amplitude of “natural” variations
 Which recording sties are best tracers of
climate change?
 What indicators are best tracers of climate
change?

Statistics of Extreme Events
Pacific Northwest Climate Change
Real or is it just our imagination?
1998 Official Climatology Statement
It appears that Oregon is poised for a relatively warm, dry
winter, after three years of cooler, wetter winters. That
does not mean a return to drought conditions. In fact, the
evidence suggests the PNW is in for a long period of cold
and wet.

Reduced snow pack effects regional power
generation; 2001 drought resulted in 46%
price hike by BPA
Reduced soil moisture in the Cascades
elevates the risk of catastrophic forest fires
Central Question: Is the PNW
getting warmer and drier?
Proposition to Prove: The PNW is
getting warmer and drier
The Physics of Climate Change
Astrophysics and Climatology?

Digging signal out of noisy data constitutes
most of observational astrophysics. Why
not apply similar techniques (e.g. large
scale smoothing, waveform analysis, etc)
to climate data.
The PNW Climate Waveform?
Y-axis is in units of standard deviation
PNI Questions:
Are the “features” significant? (remember
Y-axis units are in standard deviations)
Is the PNI a fluke?
Is their bias due to incomplete sampling?
PNI and Fish Counts
Mid Century World Wide Cooling
May be reflected in the PNI?
A Tale of Two Jets
Decadal Jet Stream Variations?





Physical Cause?
ENSO (El Nino)
PDO (Pacific Decadal Oscillilation)
NPI (North Pacific Index)
Other?
ENSO


Really has the wrong waveform
Amplitude variations occur on more rapid
timescales (3-5 years)
PDO



First Discovered in 1996
6-18 month persistence of temperature
anomalies
Unknown cause
NPI

Area weighted sea level pressure over the Gulf
of Alaska
Combined PDO + ESNO Does
Matter
The basic issue is whether or not
the cyclical behavior seen in the
last century will continue into this
one. We will show later that the
PNI shows no sign of returning to
the cool/wet cycle, even though its
statistical time has arrived to do so.
Waveform Example
Testing the Robustness of the PNI
• Can the PNI be broadly reproduced by
randomly selecting sites?
• Does the PNI have a primary driver – rain,
snow, temperature?
• If the PNI can be reproduced (and is
therefore not a fluke of these 3 sites) can
we fit the time series to make a prediction
about future climate?
Calibration Check
Using Annual Snowfall
On Robustness
Several thousand trials were run to randomly
select sites and construct individual site specific
indices to compare against the original PNI.
In all cases one can recover the basic PNI
waveform, although there are amplitude
differences
Conclude that the PNI waveform is real and
does represent climate West of the Cascades
Rain, Temp or Snow?
Temp shows most variation; followed closely by snow;
Cool/wet periods are best defined by temp + snow;
Reconstructing the Time Series
Fundamental Period = 53 years
Fundamental (53)+Secondary (13)
3 component Fit
In 1998, we should have started back into the
cold and wet cycle
Recent Data inconsistent with return
to cool/wet cycle




Eugene Rainfall 19601998
50.1 (mean)
9.4 (sd)
1.3 (m.e.)




Eugene Rainfall 19992004
36.7
5.6
2.5
Z-statistic is 4.8 (1 in 1,000,000 change of being random)
2001 Annual rain lowest in last 60 years
2005 Rain to date lower than 2001
Reduced Snowfall
Local Manifestation of Reduced
Snow Pack
http://climate.ntsg.umt.edu/html/science_western
_water.pdf
Conclusions:



The PNI is robust and recoverable from
analysis of many random sites
PNI amplitude is exaggerated by use of March
15 snowfall instead of annual snowfall
Statistical pattern suggests we should now be
in the middle of a cool/wet cycle; empirical
data (still limited of course) argues otherwise
Human Implications of This Climate
Change
Decreased power production from
hydroelectric dams due to lower snow
packs. RATES will rise.
 Don’t own a ski resort
 Longer and likely more severe forest fire
seasons
 Continued milder winters might produce
less weather related accidents

The End