remedial science: the problems with proxies
(too old to reply)
2018-08-07 11:07:15 UTC
Raw Message
Executive summary:

* Tree rings give an estimate of both local temp and precip.

* The estimates are noisy -- approx +-100% error for temps
and +-70% for precip in the Mongolia proxy used below.

* If temp and precip changes are correlated -- e.g. temps go up and
precip goes down -- tree rings growths don't register the change
robustly, and maybe not at all.

Tree ring proxies give a rough guide to past temperatures and rainfall
patterns. But even a "trsgi" standardised growth rate is (therefore)

As an example we can check some tree ring data for 16th-20th cent
Mongolia (Jacoby 2006).

Let's look at the error bars when we try to interpret tree ring growth
data as temperatures or precipitation.

Running the Mongolia tree ring data against annual av temps for the
region gives us the model:

y = 0.312647*x + -1.41843
beta in 0.312647 +- 0.316268 95% CI
P(beta>0.000000) = 0.973688
r2 = 0.0406499

Which indicates TRSGI captures something about temperatures in
Mongolia, but the low R2 warns us the error bars are quite wide. The
95% bounds are essentially a +- 100% error. Somewhat more than
satellite surveys or ground thermometers.

The precip relationship is similar.

y = 1.20691*x + 18.9548
beta in 1.20691 +- 0.825839 95% CI
P(beta>0.000000) = 0.997676
r2 = 0.0865449

IOW the conversion of a tree ring growth to an annual precip has a 95%
error bar of +- 68% of the central estimate. The R2 warned us the
conversion was noisy even if there is a statistically significant link
between tree ring growth and annual precipitation.

Being a "double proxy" has its downside. Correlated changes can become
invisible to the proxy. If temperatures and precipitation change in
"the right way" there is an attenuated -- maybe 0 -- change in the
TRSGI proxy.

If we run the TRSGI from the Mongolia tree ring data against modern
data (which partly overlaps the proxy data for the early 20th cent)
for regional temperatures (C) and precipitation (mm/mo) we find the model:


temp 0.10839 0.05236 2.07005 0.04131
precip 0.02931 0.01891 1.54956 0.12476
CONSTANT 0.48698 0.39519 1.23226 0.22106

THE F-VALUE = 3.797 (WITH 2 AND 90 DF) P - VALUE = 0.02611

Which suggests the fit is still very noisy.

And we note the temp and precip coefficients.

We can deduce that if temps and precip change so that

.10839*deltatemp approx== -.02931*deltaprecip
deltatemp approx== -.2704 *deltaprecip

then there would be essentially no change in the TRSGI.

Example. Supposed monthly precip reduced by 10mm over some period.
Our equation shows if temps over the same period rose ~2.7C then the
tree rings would appear to grow at the same -- the change would not
register as strongly, maybe not at all.

Unfortunately, reducing rainfall and increasing temperature is a
pattern that is more likely to happen than the opposite. We suspect
tree ring data will not very well track some region that is tending to
become hotter and drier.

In a strikingly ignorant tweet, Trump gets almost everything about
California wildfires wrong
Los Angeles Times, 06 Aug 2018 18:19Z

European nations keep melting as temperatures hit near-record highs
Kuwait Times, 06 Aug 2018 18:24Z

Assaad Razzouk @AssaadRazzouk 06 Aug 2018 18:27Z
Human activity is causing cancer in many species of #wildlife - and this
effect is greatly underestimated: Study reveals that 27% of Beluga whales
from the highly polluted Saint Lawrence Estuary have cancer
buff.ly/2LXbe9v pic.twitter.com/APRmCTie9a
<Loading Image...>

Europe bakes again in near-record temperatures
Daily Times, 07 Aug 2018 04:24Z
Europe baked in near-record temperatures on Mon but some respite was on
the ... initially expressed optimism they were getting the better of the
raging fire.

'Race and gender bias' behind alleged Liberal Party kebab shop brawl: police
ABC News, 07 Aug 2018 05:03Z
A man is facing charges stemming from an alleged brawl which erupted
during a Liberal Party branch meeting at a Kebab shop in Sydney's
south in June.

Water bills to soar as climate change takes hold
Increasingly frequent droughts, floods and superstorms are set to
drive water utility costs higher, according to a new report from the
Institute of Mechanism Engineers, the Telegraph reports. The
engineering authority said the longer, hotter summers - an expected
outcome of climate change -will mean water treatment plants will need
to run at peak flow rates for longer. This would raise maintenance and
running costs, energy consumption and the amount of chemicals needed
to clean the water, the report said. "The report's author Jenifer
Baxter, head of engineering at the institution, warned that water
companies will need to adapt to climate change by preparing for both
drought and flood risks," the Telegraph says. Energy Live News also
has the story. -- Jillian Ambrose, The Telegraph

The next major innovation in batteries might be here
Quartz has a feature about a new battery developed by Pellion
Technologies, a Massachusetts-based startup. It claims it "has made
the leap beyond lithium-ion that will bring the battery industry to
the next stage of technological disruption. [CEO Dave Eaglesham] and
his colleagues have accomplished something researchers have been
struggling with for decades: they've built a reliable rechargeable
lithium-metal battery." Quartz explains why this could be significant:
"Pellion's battery can pack nearly double the energy of a conventional
lithium-ion battery, making it able to, for example, double the time a
drone can spend in the air. That 100% increase in energy density is a
step change compared to the annual 10% or so improvement the battery
industry currently averages. If Pellion overcomes early limitations,
its batteries have the potential to power a Tesla car for 800 km (500
miles) on a single charge, rather than today's upper limits of 400 km."
-- Akshat Rathi, Quartz

Seasonally dependent responses of subtropical highs and tropical
rainfall to anthropogenic warming
Global warming could delay the arrival of rains in the "dry" season in
the tropical northern hemisphere, a new study suggests. "These changes
scale linearly with warming, with increasing implications for
projecting climate changes in the tropics and subtropics as warming
continues," the researchers say.
-- Nature Climate Change

Australia - 'Big call': minister refuses to link drought to climate
change on Q&A
Calla Wahlquist, The Guardian

France approves bids for 720 MW solar power projects
Bate Felix, Reuters

Wetter weather caused by climate change reduces forest soils'
absorption of greenhouse gases: study

Healthcare not very green compared to other industries
Linda Carroll, Reuters

Australia: Abbott policies reduced effectiveness of foreign aid to
combat climate change
Paul Karp, The Guardian

Anders Adamsen @anderspreben 06 Aug 2018 14:29Z
The scorching #heatwave in Europe continues ## This #map shows all-time
European heat records by country. The #Danish record will be challenged this
week according to the Met Office # # @JakubMarian | #climate
<Loading Image...>

Paul Dawson @PaulEDawson 07 Aug 2018 04:30Z
"The heatwave we now have in Europe is not something that was expected with
just 1C of warming," "Several positive feedback loops are already in
operation, but they are still weak. We need studies to show when they might
cause a runaway effect" #Climate theguardian.com/environment/20#

Sony Corp
TYO: 6758 - 7 Aug., 3:00 pm GMT+9
6,163 JPY +89 (1.47%) *** up 1.5% ***

Royal Dutch Shell Plc Class A
ETR: R6C - 7 Aug., 9:52 am GMT+2
28.81 EUR +0.11 (0.40%) up

Panasonic Corporation
TYO: 6752 - 7 Aug., 3:00 pm GMT+9
1,450 JPY +1 (0.069%) even

Liquefied Natural Gas Ltd
ASX: LNG - 7 Aug., 4:10 pm AEST
0.70 AUD -0.0100 (1.40%) *** down 1.4% ***

BHP Billiton Limited
ASX: BHP - 7 Aug., 4:10 pm AEST
33.62 AUD -0.48 (1.41%) *** down 1.4% ***
Wally W.
2018-08-07 11:42:33 UTC
Raw Message
Post by M***@kymhorsell.com
* Tree rings give an estimate of both local temp and precip.
* The estimates are noisy -- approx +-100% error for temps
and +-70% for precip in the Mongolia proxy used below.
* If temp and precip changes are correlated -- e.g. temps go up and
precip goes down -- tree rings growths don't register the change
robustly, and maybe not at all.
Global Warming's Tree Ring Circus Brings Us The Costliest Show On

"If global warming isn’t the Greatest Show on Earth, it’s certainly
the costliest and most bizarre. An early act featuring a hockey stick
–shaped graph published by the UN Intergovernmental Panel on Climate
Change (IPCC) in 2001 profoundly influenced world energy and
environmental policies.

Based heavily upon data taken from tree growth rings on the Yamal
Peninsula in Siberia, it indicated that world temperatures which had
been stable for 900 years until the 20th century suddenly soared due
to human fossil fuel-burning greenhouse gas emissions – at least that
was the IPCC’s story.

... it's way past time for the carnival we have witnessed to fold up
its tent and get science out of show business."