I received the following message from a reader of this web-log:
"You have previously exposed the drawing of shonky conclusions by government agencies from time to time. Here is a developing case. The Queensland DNRM (Department of Natural Resources) web site makes the following claim about a recently released CSIRO report titled 'Climate Change in Queensland under Enhanced Greenhouse Conditions', projected climate changes for Queensland can be summarized as follows: Annual rainfall over Queensland is projected to decline over most of the State, although projections of rainfall change are less certain than for temperature.
The full report link is at http://www.longpaddock.qld.gov.au/ClimateChanges/pub/FullReportLowRes.pdf .The issue is, no where in the report does it state that average annual rainfall is projected to decline, quite the opposite. See statements in Section 2.2.2, page 13 "...the Mark 3 model projects slight increases in annual rainfall in Queensland by 2050 (about 7%; see section 3.2)." Other references to rainfall increases are made on pages 48 and 73. Page 74 states "average rainfall in Queensland is projected to change little by Mark 4." Mark 4 is the next version of the Mark 3 climate model.
DNRM (Department of Natural Resources) spokespersons have recently been taking every opportunity at workshops and seminars to play up large projected declines in Queensland rainfall and I understand that the Minister was making statements to that effect today on radio too."
In fact the report is not recent. It was published in January 2002. However, this web-log reader and others probably heard the Minister on radio on Wednesday morning and it sounded like he was talking about a new report. The ABC radio news was actually quoting the Minister answering Estimates Committee questions in the Queensland Parliament on Tuesday. This is what was said in Parliament on 12th July 2005:
"Mrs DESLEY SCOTT: Minister, page 9 of the MPS mentions the department's role in researching the potential impacts of climate change. Can the minister outline the findings of any government
research into the possible effects of future climate change on Queensland rainfall patterns?Mr ROBERTSON: For four years my department has been researching the possible effects of future greenhouse concentrations on Queensland's climate. Our studies into past climate conditions have been central in covering the link between greenhouse gas concentrations, natural variability, ozone depletion and declining rainfall across much of eastern Australia. The results indicate that growing greenhouse gas concentrations and ozone depletion have contributed significantly to Queensland's rainfall decline over the past 20 years.
In collaboration with the CSIRO's Atmospheric Research Division we are researching whether this declining trend will continue. We expect rainfall to continue to decline over the next 70 years. Our modelling experiments suggest that over much of the state annual rainfall may decline by as much as 13 per cent by 2030 compared to conditions in the 1990s. By 2070 the decline may be as much as 40 per cent compared to conditions in the 1990s."
Was not the Minister referring to that 2002 report that was a four-year contract between the Queensland Government and the CSIRO and that is now 3 years old?
Roger Stone contributed to that 2002 report.
Graham Young (Editor of e-journal Online Opinion and Blogger) has been remembering Roger Stone's predictions. This is what Graham wrote on 1st July 2005:
"Another entry for my yet to be constructed database of predictions is climatologist Dr Roger Stone of the University of Southern Queensland and the Queensland Department of Primary Industries.
On the first of June ABC Rural carried these pars:The prospects for normal winter rainfall and crops have deteriorated, with news today Australia is officially in a borderline El Nino.
The southern oscillation index is in the negative.
Dr Roger Stone, a climatologist with the Queensland Department of Primary Industries, says the outlook is bleak for winter crops.
He says three eastern states have almost no chance of a normal crop, with the outlook worst in New South Wales.
"For the state as a whole, less than 10 per cent chance of getting normal winter crop," he said.
"This is normal yield, so it doesn't miss out altogether on getting what we call median yields.
"For most of those shires to the west and south-west of Parkes and Dubbo, stretching down towards the Victorian border, in fact for most of New South Wales, Victoria and South Australia, the chances of getting a normal wheat crop are about 10 per cent to 20 per cent at most."Then, on the 30th June, a mere 29 days later, the Courier Mail carried this sentence about the sometimes torrential rain South-East Queensland has been receiving since mid-June:
Climatologist Roger Stone said the rain was likely to continue at least through winter due to a one-in-10-year climate phenomenon."
Yeah, we did have a bit of rain including through the NSW wheat belt -and contrary to the original prediction!
I note that the Queensland Government in their summary of that 2002 report claims that:
"Maximum tropical cyclone wind speeds are likely to increase by 5 to 10%, by 2050. This will be accompanied by increases of 20 to 30% in peak tropical cyclone precipitation rates. However, little change is expected in the regions of tropical cyclone formation, and there is no evidence that tropical cyclones travel further south."
I don't think cyclone-expert Chris Landsea would approved, see
http://sciencepolicy.colorado.edu/prometheus/archives/science_policy_general/000318chris_landsea_leaves.html .
And I was recently sent this information in relation to my blog about the Wyaralong Dam (my blog of 7th July titled Government Commits $2.3 million to Unknown Quantity of Water):
"I suspect DNR (Queensland Department of Natural Resources) is fudging on projected volumes because the decision to go with Wyaralong rather than Glendower as the next dam site was taken late in the piece. The Government has bought up almost all the land it needs for Gledower, not so for the dam it is now shackled to. Old timers tell me that there is no way the water volumes from the Teviot Brook catchment will produce the sort of dam the Government is talking about, but I suspect there is a bit of politics in that, too. Locally, the dam is seen as water storage for metropolitan use located smack dab in an area crying out for more water for rural use.
DNR has always maintained that the Mary River valley is where the next big dam will be located for seq's urban water needs. I believe all this talk of Wyaralong is an attempt to control the policy agenda until the Government sniffs how the political breeze will blow in relation to a bigger dam on the Mary."
I think this message is saying that the Queensland Government already knows that the dam they have committed to, will catch no water.
In making this determination I suspect the Queensland Government did not even consider the predicted 40 percent decline in rainfall over the next few decades as per the Minister's comments in state Parliament on Tuesday?
ANSWERS to JOHN MCLEAN
Glad to see John has done some more selective reading …
Comments with asterisks to John’s comments.
Sorry but I've been busy elsewhere for a few days.
To respond to Fletcher's comments of July 18....
1. Thanks for the correction of "CO2 equivalents" but your corrections
only increases the incorrectness of the graph. The influence of water
vapour has been omitted yet according to a conversation last year with one
of the CSIRO's top atmospheric physicists water vapour is the most
significant gas when it comes to global warming ("but we concentrate on CO2
because humans can do something about that".)
So let's add water vapour at a Global Warming Potential (GWP) of about 2
and we'll assume across humid tropical Queensland and drier central
Queensland it averages 1.2% of the atmosphere (ie. 12000 ppm). That means
24000 ppm should be added to the Y-axis of the graph, making it
24860. Further, the estimated increase (doubtless based on many
assumptions) over the 110 years of 960 ppm, or to about 300% of the
starting figure with water vapour excluded, now becomes an increase of
only 6% when water vapour is included.
By the way, assuming that the GWP of carbon dioxide is half that of water
vapour for the sake of argument, then it follows that the average annual
increase in carbon dioxide of 1.5ppm would be matched by an increase in
water vapour of just 0.0006%.
**ll water vapour is a funny thing – it has both negative and positive effects.
**ter vapour is inherent in the model feedbacks.
** it is modelled. And the yep – small differences in gas forcings appear to have a significant effect. Do you think if we pull the small amount of CO2that is actually in the atmosphere out that the Earth might be a tad cold ??
2. My understanding of GCMs is that they work by an iterative process from
a given base year and both a set of assumptions at that time and a set of
assumptions about increases.
** – you’re telling the story here. Go on and enlighten us. What do THINK oes on …
They also appear to use 12 months as their
standard cycle period and so I suggest that one could expect that annual
"intermediate" results should be reasonably close to the mark.
** Whoa – medication wearing off here …. 3 hourly actually …
Further, if what you say about the inclusion of sea surface temperatures
(SSTs) is even vaguely correct - and I am not conceding that it is - then
why would one create a report that does not include historical SSTs but
claim all kinds of accuracy for the modelling?
** Read last response. They’re just examining how CO2 evolves alone. They have provided other info on goodness of fit.
In section 5, "Future Work", we find "... Further model development is
envisaged, however, as it is fair to say that while current climate models
now have a reasonable simulation of the general characteristics of ENSO,
they still do not simulate all of the characteristics of ENSO in detail.
For instance, most state-of-the-art climate models generate a region of
anomalously cool water along the equator that is not observed." I don't
accept the first part of this statement as fact but what bothers me more is
that the CSIRO asdmit that the models are producing results that don't
accord with reality.
** Oh – gee they aren’t very good are they. This gets back to the performance of models issue.
** Unlike the anti-AGW crew at least they provide full disclosure on limitations…
** I think we’ll wait for your answer as what to do from your PC.
3. I also see that the "Future Work" includes the statement "Because of
the inevitable uncertainty in future greenhouse gas emissions, projections
of climate change impacts will always be accompanied by a range of
predictions." The CSIRO is ignoring a substantial body of research into
variations in solar radiation and other solar emissions, and into
considerable research into aerosols, predominantly carbon in the form of
soot, and the influence these have on climate. To those we might add that
El Nino conditions have a dramatic influence on Queensland climate, as do
water vapour (as we saw above) and wind. I dare say that there is an
inevitable uncertainty about these unconsidered influences too.
** mmmm… starting to do some systems analysis. You’ll be modelling before you know it.
** Yep CSIRO and other examine solar forcing. La de dah.
** Try http://www.iges.org/c20c/
** And you need to model solar + volcanoes + aerosols + greenhouse to represent the temperature warming properly.
4. No Mr Christian, I am not implying that the scientists have an
integrity problem - please read what I said! The output from models is
not verified by real world conditions and therefore (a) the models should
be rejected and (b) any statement based on the output of those models is
without credibility.
** Jeepers creepers. Didn’t we go over that in the last post. If you don’t like the answer given – you win then – uncle !!! (P.S. can you ring CSIRO up and tell them they have goofed… someone had better tell them … better you than us).
5. It is true that uncertainties and problems are explicitly brought to
the readers attention - on the preface page and on the final page (as
mentioned in 3 and 4 above). These are nowhere near as frequent in the
body of the report where more assertive and definite statements are usually
found.
** Oh do go on ….
6. In minor contradiction to the above but laughable for the wrong reasons
are, on page 9, the predicted hot days for Brisbane (ie. days over 35C),
currently averaging about 3 but which in 2070 could be anywhere between 4
and 35. In other words it might be x or it might be 9 times x.
**Yep. From a range of models and range of emission pathways. Please inform us what the exact future emission pathway will be. If we listen to your goods self and ignore the need to stabilise CO2 pick 35 !!! But then heat stress can cause manic laughter ….
7. I am not entirely sure what Mr Christian is referring to when he states
in reference to temperature, rainfall and teh validity of models "There is
a major amount of work on that very issue." and that I am kidding.
**Yes there is a very large amount of work internationally on verifying model behaviour. Jeezzz !!!
**Try http://www.iges.org/c20c/
(this a is a web link – you click on it and other information that is not in your head is then presented on the screen !! – you then read it and decide if it has any value… if you can’t do this nor have any time to do this … simply continue to do what you’re doing now …)
he
means Bureau of Meteorology temperature data then yes, the data is
available but at a price for each weather station and have neither the time
nor the money to be inclined to purchase all that to investigate in more
detail.
** OK then you don’t have the time. We can’t trust the scientists – who will do the work then ? Alan Jones. Kerry Packer perhaps….
I was in fact referring to the free data that is readily available
from the BoM's website. If his comemnts are about the validity of climate
models then we go around in a circle again and I point out that the model
outputs correlate poorly with reality and that is the only verification
that matters.
Despite the claims for accuracy, the information I have from the USA,
Canada and UK is that models are no more correct than they were a few years
ago but what has changed is that we hear/read far more claims that they are
improved and that the science has advanced.
** DO TELL THEN !!! So the implementation of a dynamic world vegetation feedbacks model has had not impact on the vegetation results. (previous models have had a static vegetation feedback models e.g. fixed albedos….) and improvements in convective rainfall and cubic conformal grids.
This no clear proof of such
statements and one suspects that they could be dismissed as "marketing" by
scientists who, lacking clear and conclusive evidence, are becoming
increasingly desperate to convince the public.
** oh come on – most of the public are not interested nor convinced. Up the Bulldogs !!
*** it is the scientists who are by far the underdogs vs public vs big business vs many state, federal and international govts including George Bush ….(and even versus overstatements by greens)
8. Mr Christian talks of consensus between models as if it has merit. It
simply none at all - it is the real world verification that matters.
** Only up to a point – we don’t have a replicate planet earth to play with – you need to be able isolate and explore factors. And in any case a HUGE amount of work is undertaken internationally to verify models against obs … jeeez….. was it even necessary to tell you this – a duh !
Let me explain by example: Suppose you created a computer program
containing a very complex function and when you entered 5 + 5 it gave the
answer 9. Someone else copies your function and again finds that 5+5 =
9. A third person looks at the basis for your complex function and creates
his own but again 5 + 5 = 9. Would you say that on the basis of consensus
between these models we should all assume that 5 + 5 does equal 9? This is
what Mr Christian would have us believe.
** You assume too much. 2 + 2 is 5 for exceptionally well calculated values of 2 !!!
** They do share some core subroutines. Some very well tested and reviewed core routines.
** They are largely very different.
** They are different models with a range of methods of coupling ocean and atmosphere and a range of technical details. However what is interesting is that often they produce similar patterns (reassuringly even) and depending on the severe degree of testing of each of the models allow you to represent their “overall opinion” (like a jury) on a variety of evidence.
9. Finally governments do appear to be formulating policy on the basis of
these reports because no other reports are particularly relevant to the
respective states. It's hard to have faith in global predictions, after
all, Australia's temperatures have been rising since about 1945 or 50
whereas the global average declined from 1945 to about 1975. (Yes, okay,
this shows that temperature changes are regional but that's rather in
contradiction to the theory that carbon dioxide, a relatively evenly spread
gas, causes warming.)
** John AGW does not state that we will have “even” effects from GHG increases. That is WAY to simplistic. The atmosphere has certain thresholds and step changes can occur. You may also have noticed a fair bit of non-linear behaviour, circulations, persistent features like El Nino and chaos ! For example a recent phenomenon appears to be due to changes in southern annular mode consistent with declines of stratospheric ozone (whoops another human effect) and increases in tropospheric greenhouse gases. Some models (and oh we hate those don’t we !!) suggest that this may be responsible for major circulation changes in the southern hemisphere and Antarctic circumpolar vortex and dare we even hypothesise a mechanism behind Australian drying trends and tropical cyclone patterns of recent decades.
NSW Premier Bob Carr has shown how state politicians use these reports as
the basis of public statements. He has widely proclaimed that the CSIRO
told him that NSW rainfall shows a declining trend since 1950 of 14.1
mm/decade but he fails to mention - or perhaps the CSIRO report failed to
tell him - that 1950 rainfall was the highest recorded in the state and
about 50% above the average from 1900 to 1949. The claim was correct as
far as it went but the circumstances were clearly exceptional and comments
based on that year are misleading.
** John have a look at http://www.bom.gov.au/cgi-bin/silo/reg/cli_chg/trendmaps.cgi
** Tell us you want to stick with it – there has been some persistent trends for some time.
** And the climate drying trend “correlates” with CO2, global changes and warming. Well many climatologist would say something happened in 1976 on a wide range of fronts.
** OK let’s say you are right John McLean – we know nothing. We cannot prove nor disprove climate change. So all we have is 100 years of rainfall measurements which we hope those loyal and faithful volunteers measured correctly. And let’s also assume that you are the Premier of NSW. You notice things have been getting drier, you read the papers about climate change. And as an avid reader of this blog you are now utterly confused with all the rabble (including mine). You have to decide what to do… So you can stick with your 100 years and know nothing else. Or you can take some risk assessment and make some proactive provisions for the future. But that will cost money and the public want another hospital. What is your advice to the Premier John McLean … (it’s not as easy when you actually are responsible for a decision and can’t just turn your Internet off and go to bed).
** Dear Premier,
**………………….. Fill in here …….
**Over and out from Fletch
**Time for another Bex and a lie down.
FROM JOHN McLEAN
Sorry but I've been busy elsewhere for a few days.
To respond to Fletcher's comments of July 18....
1. Thanks for the correction of "CO2 equivalents" but your corrections
only increases the incorrectness of the graph. The influence of water
vapour has been omitted yet according to a conversation last year with one
of the CSIRO's top atmospheric physicists water vapour is the most
significant gas when it comes to global warming ("but we concentrate on CO2
because humans can do something about that".)
So let's add water vapour at a Global Warming Potential (GWP) of about 2
and we'll assume across humid tropical Queensland and drier central
Queensland it averages 1.2% of the atmosphere (ie. 12000 ppm). That means
24000 ppm should be added to the Y-axis of the graph, making it
24860. Further, the estimated increase (doubtless based on many
assumptions) over the 110 years of 960 ppm, or to about 300% of the
starting figure with water vapour excluded, now becomes an increase of
only 6% when water vapour is included.
By the way, assuming that the GWP of carbon dioxide is half that of water
vapour for the sake of argument, then it follows that the average annual
increase in carbon dioxide of 1.5ppm would be matched by an increase in
water vapour of just 0.0006%.
2. My understanding of GCMs is that they work by an iterative process from
a given base year and both a set of assumptions at that time and a set of
assumptions about increases. They also appear to use 12 months as their
standard cycle period and so I suggest that one could expect that annual
"intermediate" results should be reasonably close to the mark.
Further, if what you say about the inclusion of sea surface temperatures
(SSTs) is even vaguely correct - and I am not conceding that it is - then
why would one create a report that does not include historical SSTs but
claim all kinds of accuracy for the modelling?
In section 5, "Future Work", we find "... Further model development is
envisaged, however, as it is fair to say that while current climate models
now have a reasonable simulation of the general characteristics of ENSO,
they still do not simulate all of the characteristics of ENSO in detail.
For instance, most state-of-the-art climate models generate a region of
anomalously cool water along the equator that is not observed." I don't
accept the first part of this statement as fact but what bothers me more is
that the CSIRO asdmit that the models are producing results that don't
accord with reality.
3. I also see that the "Future Work" includes the statement "Because of
the inevitable uncertainty in future greenhouse gas emissions, projections
of climate change impacts will always be accompanied by a range of
predictions." The CSIRO is ignoring a substantial body of research into
variations in solar radiation and other solar emissions, and into
considerable research into aerosols, predominantly carbon in the form of
soot, and the influence these have on climate. To those we might add that
El Nino conditions have a dramatic influence on Queensland climate, as do
water vapour (as we saw above) and wind. I dare say that there is an
inevitable uncertainty about these unconsidered influences too.
4. No Mr Christian, I am not implying that the scientists have an
integrity problem - please read what I said! The output from models is
not verified by real world conditions and therefore (a) the models should
be rejected and (b) any statement based on the output of those models is
without credibility.
5. It is true that uncertainties and problems are explicitly brought to
the readers attention - on the preface page and on the final page (as
mentioned in 3 and 4 above). These are nowhere near as frequent in the
body of the report where more assertive and definite statements are usually
found.
6. In minor contradiction to the above but laughable for the wrong reasons
are, on page 9, the predicted hot days for Brisbane (ie. days over 35C),
currently averaging about 3 but which in 2070 could be anywhere between 4
and 35. In other words it might be x or it might be 9 times x.
7. I am not entirely sure what Mr Christian is referring to when he states
in reference to temperature, rainfall and teh validity of models "There is
a major amount of work on that very issue." and that I am kidding. If he
means Bureau of Meteorology temperature data then yes, the data is
available but at a price for each weather station and have neither the time
nor the money to be inclined to purchase all that to investigate in more
detail. I was in fact referring to the free data that is readily available
from the BoM's website. If his comemnts are about the validity of climate
models then we go around in a circle again and I point out that the model
outputs correlate poorly with reality and that is the only verification
that matters.
Despite the claims for accuracy, the information I have from the USA,
Canada and UK is that models are no more correct than they were a few years
ago but what has changed is that we hear/read far more claims that they are
improved and that the science has advanced. This no clear proof of such
statements and one suspects that they could be dismissed as "marketing" by
scientists who, lacking clear and conclusive evidence, are becoming
increasingly desperate to convince the public.
8. Mr Christian talks of consensus between models as if it has merit. It
simply none at all - it is the real world verification that matters.
Let me explain by example: Suppose you created a computer program
containing a very complex function and when you entered 5 + 5 it gave the
answer 9. Someone else copies your function and again finds that 5+5 =
9. A third person looks at the basis for your complex function and creates
his own but again 5 + 5 = 9. Would you say that on the basis of consensus
between these models we should all assume that 5 + 5 does equal 9? This is
what Mr Christian would have us believe.
9. Finally governments do appear to be formulating policy on the basis of
these reports because no other reports are particularly relevant to the
respective states. It's hard to have faith in global predictions, after
all, Australia's temperatures have been rising since about 1945 or 50
whereas the global average declined from 1945 to about 1975. (Yes, okay,
this shows that temperature changes are regional but that's rather in
contradiction to the theory that carbon dioxide, a relatively evenly spread
gas, causes warming.)
NSW Premier Bob Carr has shown how state politicians use these reports as
the basis of public statements. He has widely proclaimed that the CSIRO
told him that NSW rainfall shows a declining trend since 1950 of 14.1
mm/decade but he fails to mention - or perhaps the CSIRO report failed to
tell him - that 1950 rainfall was the highest recorded in the state and
about 50% above the average from 1900 to 1949. The claim was correct as
far as it went but the circumstances were clearly exceptional and comments
based on that year are misleading.
cheers
END OF COMMENT FROM JOHN McLEAN.
I did bother to read your post. I addressed your main points and then made another on climate change derivative of that (sorry if that's not OK with you - we here call it "the flow of conversation")
"At odds with other forecasters" - which ones by the way? Do tell....
Sorry you don't like probabilities - that's all anyone has got ... so looks like it's back to the tea leaves.
And yep - good idea - keep a record of his "forecasts" - but don't forget to record both "wrongs" and "rights". How do you think you do the assessment .. and leave out persistence and random throws etc ?
If you want to get right into it - get a copy of Qld DPI's RAINMAN - you can then see explicitly what skill his system has from the raw data up. (well what it would have over the last 100 years).
And oh - you're a journalist are you - the guys that consistently trivialise and mock climate science (both climate change and forecasting). You'll run all sorts of pro and con stories a week apart. We have a box of clipping on journalists stuffing up with rampant misrepresentation and sensationalising on climate topics. Public interest not well served but controversy sure sells papers.
Christian, Stone can't hide behind the fact that his system allocates "probabilities". If those probabilities are not better than pure chance then he has no business calling them forecasts.
If you had bothered to read my post properly you'd know that my primary concern isn't with global warming - I don't have a problem with the basic science - but with those people who make claims on the basis of the science which are either unsupportable, or exaggerated. I have similar concerns in other areas of human endeavour.
You'd also know that I wasn't judging Stone on the basis of just one prediction. The last par was an invitation to him to respond and the basis of the article was my suggestion that we should build a database of "experts" and see how many of their predictions actually proved correct.
We journalists are always deferring to, and often reporting uncritically, people who claim to be experts, but frequently these experts are not really expert at all. How do you find out who is reliable? My answer is that we should maintain a database and over time build up a picture of reliability.
I'd noticed Roger Stone in the past because his forecasts are often at odds with other forecasters, and they frequently seemed to be wrong. It appears to be a widespread problem in the climate science community, yet we are being asked to take their predictions seriously. So, if I had a database he'd be in there, but then so would many others.