Climate Indexing

1
Climate Indexing

• Application to
• the Pacific Northwest
• G. Bothun and S. Ostrander
• Dept. of Physics, University of Oregon

2
Premise

• Climate represents seasonal weather patterns
  averaged over time.
• Such a representation comprises a quantitative
  climate waveform
• Departures from that waveform, if statistically
  significant, therefore comprise evidence of real
  climate change

3
(PNI), developed by at the UW by Ebbesmeyer and
Strickland (1995) is a composite index that
characterizes Pacific Northwest climate. The PNI
uses three parameters 1) air temperature at
Olga 2) annual precipitation measured at Cedar
Lake WA 3) snowpack depth at Paradise on Mount
Rainier on March 15 of each year. Statistical
Z-scores are computed for each year and the data
is smoothed with a 5 year boxcar to yield a
waveform that indicates multi-decadal periods of
cool/wet (blue) and warm dry (red) climates.
4
Our study uses hundreds of climate sites in
Western Washington to reform the PNI. Our new
PNI, hereafter called the SHI (Stephanies
Harmonic Index), is formed from the 20 best sites
in the data. The index generated from that data
(GD in the figure below) compares quite favorably
with the PNI indicating that the PNI is a robust
waveform and is not a fluke of using only 3 data
points.
5
The original PNI is un-weighted. Below are
example departures from the SHI that occur when
either temperature data (weight 1) or
precipitation data (weight 2) Is given more
weight. The detailed waveform seems more
sensitive to precipitation than it is to
temperature. Work is in progress on weighted
seasonal representations of the SHI.
Weight 2 Rainfall
6
The yellow waveform below represents an attempt
to game the system by selecting those sites
(starting in 1935) that will yield the most
amount of positive signal (i.e. to bias the data
towards warm/dry periods). As can be seen, the
only significant departure occurs in the
1960-1970 period. No attempt to maximize the
post 1975 signal by preferential site selection
produced a waveform that is any different than
the SHI. This suggests the SHI is
representative.
7
The final step now involves fitting an
approximate spherical harmonic function to the
data in the last century (in this case 1920-2000)
in order to predict future climate based on last
centuries wave form. While the model is not a
perfect fit to the data by any means, it is a
reasonable approximation which clearly shows
that, as of 2000, the PNW should be entering
another extended cool/wet period. The data are
in sharp disagreement with this prediction and
therefore provide an indicator that the current
period of warm/dry weather is likely to extend
into the indefinite future.
8
Conclusions

• The PNI indexing system is robust
• The detailed climate waveform is more sensitive
  to precipitation than temperature
• Preliminary fits to the waveform suggest that the
  current period of warm/dry Western Washington
  climate that started around 1975 will continue
  into the indefinite future.
• This implies a) reduced hydropower, b) truncated
  ski season, c) possibly extended forest fire
  seasons.