Bishop Hill Diary dates
Aug 2, 2012 Climate: Statistics Royal Society The Royal Society is organising a conference on uncertainty in weather and climate:
This meeting follows on from the 2010 Anniversary Discussion Meeting on “Handling Uncertainty in Science” but with a focus on weather and climate prediction and downstream applications. How is uncertainty represented in weather and climate prediction? How reliable are representations of uncertainty? How can decision makers in weather and climate sensitive sectors make useful decisions in the light of uncertain input? Are current ensemble weather and climate prediction systems useful for decision making across a variety of application sectors? How should probability forecasts be presented to the public?
Speakers include Judith Curry and Peter Webster.
Details here.
Reader Comments (24)
'uncertainty in weather and climate:' surely this is an oxymoron in the latter case given 'the science is settled '
Anyone want to bet the main messages coming out of this will be , we need more funding for research and 'climate doom ' is real but we can't tell you why just trust us on this .
I thought weather and climate were different things? Or so I've been told...
James P, if weather and climate were not different things, then they wouldn't need to focus on both.
Uncertainty is one of my real bugbears that seems to be appearing more and more as they seem to mix together real uncertainty due to a variable outcome and the model actually being wrong.
This is my understanding. I have a model of rolling a die that says it will be a six one sixth of the time. My model says that if I roll a die 600 times it will be a six 100 times. The uncertainty is that in repeated experiments it is a six 92,103,111 etc times. With many experiments and applying statistics you may find on average a six 110 times and prove statistically that the die is loaded and therefore my model was wrong. That was the hypothesis, experiment and failed test in that case.
The problem with GCM's is that the model is wrong in being able to judge whether increased CO2 will increase warming. They don't have the capability to do it and just give you the answer that you've programmed in. That isn't uncertainty at all it is just being wrong or it telling you what you told it to tell you. The models are great for things like weather forecasting and many other studies but for forecasting the climate in 2100 they are completely useless (with certainty)
" How should probability forecasts be presented to the public?"
Truthfully would be a good start.
According to Julia Slingo, the models on climate change are tested daily
- some interesting discussion of computer code - start watching at approx 2:40 (Julia Slingo at approx 2:41)
(Interesting question from Graham Stringer asking if there is a problem with the code)
http://www.parliamentlive.tv/Main/Player.aspx?meetingId=5979&player=windowsmedia
She claims they're 'tested' twice a day for 'robustness' as they are the same codes used for the weather forecasts, they don't always get the weather forecasts right but that's the nature of the chaotic system and doesn't prove that the codes are wrong!
Re: Marion
The transcribed version is here: http://www.publications.parliament.uk/pa/cm200910/cmselect/cmsctech/387b/38724.htm
We make a (data) loss on every (weather prediction) sale, but we make it up with volume (climate prediction)!
Re: Aug 2, 2012 at 2:07 PM | TerryS
"The transcribed version is here:
http://www.publications.parliament.uk/pa/cm200910/cmselect/cmsctech/387b/38724.htm
"
Thanks
Judith Curry will give them an earful. They should invite Jonty Rougier. If they will take seriously what Curry and Rougier can tell them then they will conclude that climate science is in its infancy. In a few decades, climate science might be capable of making predictions about climate.
With Watts American work, the UK work-in-progress (sorry, forget researcher), Treadgold's NIWA and possibly Australian BOM data in dispute, will the Uncertainty Debate include the reliability of adjusted and homogenized data?
A catatenation of problems not randomly or normally distributed. Interesting implication for the legacy of Al Gore, don't you think? "Sir: it is not an Abyssian, 2-year old brown-and-bray cat, it is a dog. A mutt with mange, actually."
steve
I realise, but they do seem to be conflating them ("weather and climate prediction"). When sceptics do that, they receive a load of condescension.
Who is Peter Webster?
Regards
Mailman
Re: Aug 2, 2012 at 5:56 PM | Mailman
See the Bishop's link, the descriptions can be expanded -
"Professor Peter J Webster, Georgia Institute of Technology, AtlantaProbabilistic forecasting of environmental and health hazards in the developing world
Biography
Professor Webster’s career may be divided into two parts: Exploring the dynamics of low-frequency atmospheric circulations such as monsoons and El Nino and determining the predictability of the precipitation associated with these phenomena. More recently, he has attempted to use this theoretical insight to the generation of forecasting models of hazards (floods and tropical cyclones) and the transfer of these modules and their output to the people of the developing world, from government organizations to the villagers, especially in South Asia.
Webster is a professor in Earth and Atmospheric Sciences at the Georgia Institute of Technology. During his academic career he has mentored and graduated 28 doctoral students. He has published over 160 peer reviewed articles and two text books. He has been recognized widely receiving the Mason Gold Medal (2012) and the Adrian Gill Award (2003) from the Royal Meteorological Society (2003). In addition he received the Carl-Gustav Rossby Gold medal (2004) and the Jule G. Charney Research Award from the American Meteorological Society. Webster is a fellow of the Royal Meteorological Society, the American Geophysical Society, the American Meteorological Society and the American Association for the Advancement of Science."
Hmm, I wonder if he would be the one to ask about the 'environmental and health hazard' being caused by the push for biofuel crops in the developing world.
Wonder if that was considered at all in these models of theirs!!
Sooner or later, some twit at the conference will bring up the 'precautionary principle.'
I wonder if they'll pre-book alternative accommodation for if adverse weather disrupts their proceedings?
Mailman, you may also know Peter Webster in the contexts of the FOI requests for
A nice man: while he always respected CRU's requests to keep the datset they sent him confidential, his confirmation that he had received it played a crucial role.
"The Royal Society is organising a conference on uncertainty in weather and climate"
JP has beaten me to it. I was going to say: "In the church hall if wet."
This is a good time to remember the brief spell when Paxman was told to wrap up Newsnight by reading the weather forecast. I remember his reading out blah, blah, showers, and then, exasperated, saying words to the effect of "What do you expect? It's April."
Sorry to be totally boring, but there's nothing wrong with the GCMs apart from the mistakes in the IR physics, heat transfer theory and basic heat transfer measurement. So, in my view no IPCC climate model can predict climate [the weather models are acknowledged to have a 'warming bias'].
Now, I could be wrong. However, so far no-one has told my why. I offer the challenge and I'm prepared to humble myself if proven wrong - the most basic mistakes in heat transfer cannot be justified by any objective scientist yet without them there is no CO2-AGW and a fixed GHE for a water plant.Over to you folks...
Can I remind your Grace of Proverbs 24:14?
Jonathan Jones,
Ah yes...that set a bell off alright!
Thanks guys.
Regards
Mailman
Who is Peter Webster?
Regards
Mailman
I'm pretty sure that Peter Webster is Judy Curry's husband. A handsome pair, as they say.
No CRU representative? Wow!