COLUMBIA, Mo. – The world’s oceans cover more than 72 percent of the earth’s surface, impact a major part of the carbon cycle, and contribute to variability in global climate and weather patterns. However, accurately predicting the condition of the ocean is limited by current methods. Now, researchers at the University of Missouri have applied complex statistical models to increase the accuracy of ocean forecasting that can influence the ways in which forecasters predict long-range events such as El Nińo and the lower levels of the ocean food chain—one of the world’s largest ecosystems.
“The ocean really is the most important part of the world’s environmental system because of its potential to store carbon and heat, but also because of its ability to influence major atmospheric weather events such as droughts, hurricanes and tornados,” said Chris Wikle, professor of statistics in the MU College of Arts and Science. “At the same time, it is essential in producing a food chain that is a critical part of the world’s fisheries.”
The vastness of the world’s oceans makes predicting its changes a daunting task for oceanographers and climate scientists. Scientists must use direct observations from a limited network of ocean buoys and ships combined with satellite images of various qualities to create physical and biological models of the ocean. Wikle and Ralph Milliff, a senior research associate at the University of Colorado, adopted a statistical “Bayesian hierarchical model” that allows them to combine various sources of information as well as previous scientific knowledge. Their method helped improve the prediction of sea surface temperature extremes and wind fields over the ocean, which impact important features such as the frequency of tornadoes in tornado alley and the distribution of plankton in coastal regions—a critical first stage of the ocean food chain.
“Nate Silver of The New York Times combined various sources of information to understand and better predict the uncertainty associated with elections,” Wikle said. “So much like that, we developed more sophisticated statistical methods to combine various sources of data—satellite images, data from ocean buoys and ships, and scientific experience—to better understand the atmosphere over the ocean and the ocean itself. This led to models that help to better predict the state of the Mediterranean Sea, and the long-lead time prediction of El Nińo and La Nińa. Missouri, like most of the world, is affected by El Nińo and La Nińa (through droughts, floods and tornadoes) and the lowest levels of the food chain affect us all through its effect on Marine fisheries.”
El Nińo is a band of warm ocean water temperatures that periodically develops off the western coast of South America and can cause climatic changes across the Pacific Ocean and the U.S. La Nińa is the counterpart that also affects atmospheric changes throughout the country. Wikle and his fellow researchers feel that, through better statistical methods and models currently in development, a greater understanding of these phenomena and their associated impacts will help forecasters better predict potentially catastrophic events, which will likely be increasingly important as our climate changes.
Wikle’s study, “Uncertainty management in coupled physical-biological lower trophic level ocean ecosystem models,” was funded in part by the National Science Foundation and was published in Oceanography and Statistical Science.
Steven Mosher says:
March 22, 2014 at 1:41 pm
Note the difference between “unravelling” and improving a prediction…
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
So I read this hypothetical stuff after I come in from a day of dodging snakes and rats while trimming dense tropical plants in my backyard.
While I sometimes enjoy (i.e.: laugh) at Steven’s stuff, he NEVER deals with fundamental explanation of atmospheric physics, let alone presenting REAL (i.e.: Mother Nature, not model) data to demonstrate his point…just troll-speak. Statistics are great when you have a scientific theory (hypothesis, whatever) and real data…however, statistics as a substitute for a scientific theory is simply “cooking the books” or
“torturing the data”.
But, this evening I’m pretty tired and I did get a good laugh.
Isht der Klimate un zee Vvvveeeerrrrmintsht.
Ha ha. 😀
‘Now, researchers at the University of Missouri have applied complex statistical models to increase the accuracy of ocean forecasting…’
And what was the outcome? When will we get the next el Nino? Is this a falsifiable projection?
After a fun career in the computing and Finance world (notice which I capitalized), I was quite sick of the inelegant clichés codgers and fools liked to spout when discussing technology.
e.g.
“Let’s stick with well tested leading technology and stay away from ‘bleeding edge’ technology.”
“Let’s think outside of the box.”
“I do not want to re-invent the wheel.”
It’s always interesting that every fool and old coot thought that they were amongst the first to spout these inanities and that they hated to have someone point out that their statements contradicted themselves.
I used to throw rejoinders at the speakers hoping to minimize trouble yet stop the fount from continuing to flow. In the same order;
“If it’s ‘well tested’ by your standards it is no longer leading technology.”
“I’d be happy if people would ‘just think’.”
“I don’t mind when people ‘re-invent’ the wheel. One of these days they’ll invent round wheels.”
Apt readers of the above press release will notice that all three of these rejoinders and their causes are in play.
There is no recognition nor admission that the current models are abject failures; just a statement that the new models improve or will improve many things, including:
Those are some incredibly amazing statistics and here I thought that food chains were inherent in themselves.
There is no better faster way to build straight to the bleeding edge, don’t think, don’t assess and absolutely do not peek to see what is wrong or necessary before building ‘new shiny glittering’ models. The previous group went and spent decades morphing ‘weather’ models in pretense of a well designed comprehensive ‘whole earth’ climate scenarios. The new group sees their 28% false surface emulations and raises them by 72% new false sea surface emulations.
Just what the world needs! A ‘best thing since sliced bread’ glamorous press announcement long before code, formulas or , ||’shudder’||, models can be tested certified and found useful.
“””””…..1sky1 says:
March 22, 2014 at 1:17 pm
If only an intrinsically complex physical problem could be unravelled by mere statistics……”””””
Statistics (any sort of statistics, either dumb, or “sophisticated), is performed on discrete sets of already known numbers with actual numerical values.
The results of such analysis are exact, in that statistical mathematics is a rigorous discipline of mathematics, with no uncertainty, in the outcome of correct application of its algorithms.
The numbers can be plotted (on suitable axes) as a scatter plot (of discrete dots).
The results of such analysis tell you nothing about ANY datum, not in the given set.. You cannot predict the value of ANY possible data point that might by some means be introduced before the first element of the set; nor can you predict the value of ANY possible data point, that might by some means be obtained after the last element of the given set. You can’t even predict, whether such extrapolations of the set, will result in higher, or lower or identical values, than the end points of the set; the very direction of any change is quite indeterminate.
Moreover, this information vacuum, includes the space between the individual elements of the set. You cannot know what happens between the dots; EXCEPT in the unique case, where the dots represent instantaneous samples of a continuous band limited function, properly sampled in accordance with the Nyquist sampling theorem.
In that case, it is possible (in theory) to completely recover the continuous band limited function, and determine properly interpolated values. And for that matter, correctly obtain an average of the continuous function over that data set range.
So drawing scatter plots, and then joining the dots sequentially, is an exercise in deception.
Think, science lovers. Think. You really believe what you are espousing? Defend it here, then, if it’s easy for you:
http://johnwilsonbach.com/2014/03/22/kidnapped-by-spring/
Lawrie Ayres says:
March 22, 2014 at 3:35 pm
Mac the Knife says:
March 22, 2014 at 2:11 pm
…..When your batting average is ZERO, improvement should be achievable…….
The Farmers Almanac has done better.
Jennifer Maharosy recently requested of our Minister for the Environment to have the Bureau of Meteorology explain why they have such a poor record of predictions. They use models whereas our more accurate long range forecasters use historical data and observation. The Farmers Almanac is a similar beast and proves that history is a better predictor than computers.
As Dr Phil says ” The best predictor of future actions is past performance”.
Lawrie,
Thanks for the ‘come back’! I searched for and found a ‘Jennifer Marohasy’ blog that is Australian climate oriented. Is that the person you were referencing?
http://jennifermarohasy.com/
Interesting blog and provides insights into the Australian climate cycles that I’ve had little exposure to! Thank you for that!
Aus is a looooog ways from Seattle WA USA… If I may ask, where in Australia are you located?
Mac
I have been unable to find any statistical oceanic biomass data on the effect of the almost total removal of the top of the planktonic food chain, the krill eating whales. Hundreds of thousands of these animals were slaughtered from the mid eighteenth to the mid twentieth century. I am not positing a causal relationship between the right whales decline and increasing CO2, but it would be an interesting exercise to do the sums. There are reasonably good records for landed tonnages of oil. Estimates of the amount of krill eaten could be calculated. Perhaps the biomass of krill grazing on the worlds largest carbon sink, phytoplankton, was much smaller in the past because of the whales. This would indicate a large increase in the phytoplankton biomass and increasing CO2 consumption. Just a thought for the anti-whaling lobby.
In reply to:
“The world’s oceans cover more than 72 percent of the earth’s surface, impact a major part of the carbon cycle, and contribute to variability in global climate and weather patterns. However, accurately predicting the condition of the ocean is limited by current methods. Now, researchers at the University of Missouri have applied complex statistical models to increase the accuracy of ocean forecasting that can influence the ways in which forecasters predict long-range events such as El Niño and the lower levels of the ocean food chain—one of the world’s largest ecosystems.”
William:
Talk is cheap. Use the ‘complex’ statistical model to make a prediction. Statistical models fail when there is a step change in a primary (the primary) forcing mechanism that has not occurred in the past modeled statistical period.
I notice there has been a series of ‘predictions’ concerning a El Niño this year. Good luck with that prediction.
Why is there suddenly a significant increase in sea ice both poles?
Why the increase in cold surface ocean temperature anomalies?
William: The planet is cooling due to the sudden decrease in solar magnetic cycle activity. It will be interesting to watch the creative first attempts to hand wave away global cooling.
http://www.ospo.noaa.gov/data/sst/anomaly/2014/anomnight.3.20.2014.gif
http://nsidc.org/data/seaice_index/images/daily_images/S_stddev_timeseries.png
http://arctic.atmos.uiuc.edu/cryosphere/iphone/images/iphone.anomaly.global.png
C’mon folks I’m a skeptic too, but Employing Baysian statistical methods dis a step forward for every branch of science. Electrical eng signal processing and information n theory have been using and refining this method for 30 years. It has enabled all the ADSL, 4G-LTE data, and spread spec CDMA comms we take for granted now in 2014 for broadband data streaming.
IMHO, Baysian stats can only help the failed climate models to find more reasons why they are failing. The modelers are good people, and they want their models to actually replicate the real climate. So let’s be smart here, but still skeptical too.
Joel O’Bryan:
A distinction should be made between Bayes’s theorem and Bayesian methods. As Bayes’s theorem is logically correct, conformity to it is required else conclusions may not logically be drawn from the associated argument. Though Bayes’s theorem correct, there are methods drawn from it that are illogical from their violation of the law of non-contradiction in the selection of the prior probability density function.The violation of non-contradiction stems from the multiplicity of uninformative priors.
Among the applications violating non-contradiction in this way is the procedure (called “Bayesian parameter estimation) by which climatologists extract a posterior probability density function for the equilibrium climate sensitivity from a global temperature time series. My guess is that the University of Missouri work employs this procedure.
nicholas tesdorf says:
March 22, 2014 at 2:58 pm
Steven Mosher says:
March 22, 2014 at 1:31 pm
“NO THEY CANNOT !!!!!!!!!!!
############## ”
Mosher is shouting again. The strain of watching Mother Nature in action again controlling the climate is getting to him.
More settled science from Warmistas.
__________________________
Sorry, but you mis- attributed that quote to Mosher. What he actually said was even dumber than what you think he said- you misunderestimated him, (or something.)
1sky1 says:
“If only an intrinsically complex physical problem could be unravelled by mere statistics.”
Indeed, It’s well overdue that that climate is analysed as system (ie an engineering problem) rather than a box of random number generators.
Oceans, evaporation, heat transfer are physically inter-linked, not random processes.
Starting with a foregone conclusion that the system can be represented as CO2 rise + “noise” is the fundamental error of the last 30 years of climate science.
PS. It hasn’t worked.
Greg:
Right on. Climatologists err in appropriating the ideas of “signal” and “noise” from electrical engineering and applying them to the problem of controlling the climate. Under Einsteinian relativity, a “signal” is capable of propagating from the present to the future but incapable for propagating from the future to the present for to do the latter it would have to travel at superluminal speed. To control the climate the control system needs information about the conditional outcomes of events but this information cannot be carried by a signal. Violating relativity is one of the many fundamental errors that climatologists have made in constructing their field.
Gixxerboy says:
March 22, 2014 at 4:59 pm
Nicholas Tesdorf
That was not Mosher shouting. It was Stephen Richards. Mosh was parodying it with his line about ‘settled science from skeptics’
Then you and Mosher need to explain how any useless model can PREDICT anything, dickhead.
There isn’t a climate or weather model out there that can predict 100% the future. You need to go back to school an learn to understand the limitations of models and all models have them. The best engineering models do help but they are modeling systems that are moderately predictable. The climate and weather are not in that categorie.
“The world’s oceans cover more than 72 percent of the earth’s surface, impact a major part of the carbon cycle,”
And does the impact have any effect?
“through better statistical methods and models currently in development, a greater understanding of [El Niño and la Niña] and their associated impacts will help forecasters better predict potentially catastrophic events, which will likely be increasingly important as our climate changes.“. They are getting things backwards. First they need to look at the major climate factors – things like solar cycles and ocean oscillations, ie. the things that change climate. Then they can look at the more minor or short term things like El Niño and la Niña. There’s little value in working on El Niño and la Niña on the basis that they will become “increasingly important as our climate changes” if they don’t know what changes climate in the first place. How do I know that they don’t know? They don’t know what caused the MWP, LIA, or any of the climate cycles of the Holocene, so obviously they don’t know what’s causing what looks like a continuation of those cycles that we’re in right now.
Dr Burns Mar 22 3:28pm “When models can forecast with any accuracy whether it will rain in two days time, I may start to have some faith in them.“. You are talking about weather models. Weather models are, and always will be, useless for prediction on any climate scale. For that, climate models are needed, and currently there are none. The models used by the IPCC and others are not climate models, they are weather models.
Loosely translated: ‘we don’t really know whats going on – but we need more money’.
What, where, when and magnitude would be the useful (im)possible predictions. Statistical modelling will be a waste of money in a chaotic, coupled, non-linear system.
son of mulder:
Despite the difficulties that you describe, statistically validated models have already been developed that predict average air temperatures over the western states of the United States over a forecasting horizon of 6 months. This has been made possible by the existence of 328 statistically independent observed events extending back to the year 1850. An impediment to successful development of a model that forecasts over a horizon of 30 years is that there are only between 5 and 6 statistically independent observed events. To gather 328 30-year observed events, one would need to observe the climate for 30 years/event * 328 events = 9840 years.
I would be interested how statistics which are collections of data from past measurements can forecast the behavior of a coupled complex, chaotic, non-linear, dynamic, system of systems. As everyone appears to accept that the climate system is ‘chaotic’ …. from http://mathworld.wolfram.com/Chaos.html
“chaotic systems are distinguished by sensitive dependence on initial conditions and by having evolution through phase space that appears to be quite random.”
Statistics by its very nature deals with averaging the initial conditions (or Bayesian priors) and those are the known initial conditions there are many “unknown unknowns” that may also lead to rapid evolution into different states. Statistics is precisely the incorrect tool for forecasting the behavior of a ‘coupled complex, chaotic, non-linear, dynamic, system of systems’ that is Earth’s climate as a consequence of its unpredictability.
Ian W
The term “statistics” has several meanings. One is “mathematical statistics.” The latter makes it possible for one to deal with a situation that one usually faces in the construction of a model: missing information for a deductive conclusion. That information is missing forces the model builder to replace the rule of classical logic in which every proposition has a truth-value by the rule in which every proposition has a probability of being true, giving rise to mathematical statistics.
The term “statistical model” is sometimes used in reference to a model that is constructed solely from observational data as is the case, for example, under regression analysis. However, it is possible under available methods of mathematical statistics to supplement observational data by natural laws in the construction of a model. To do so results in a better model.
Interesting, Mr. Mosher, You try to dismantle a perceived house of cards by building one.
Bob Tisdale says:
March 22, 2014 at 5:26 pm
Why are they wasting everyone’s time with a press release about something in development?
========================
Because they can take credit for what they are working on. In case what they are working on doesn’t pan out. I see it as a sign that they really have no faith in what they are doing. “We better publish now, while we can.”
Geez! He makes it sound like it’s so hard to model and predict climate and atmospheric events. Hasn’t he read the latest IPCC model confidence indicators? They don’t need his complex data driven statistics to predict nearly 3C of temp rise with extreme confidence. All this mumbo-jumbo makes it sound a little unsettled and difficult to do still. What a denier!
Steven Mosher: note they didnt say unravelled.
note they said
“Their method helped improve the prediction of sea surface temperature extremes and wind fields over the ocean,”
Note the difference between “unravelling” and improving a prediction.
Well said. That post was good.
If it works, good.
If it doesn’t work, when does the effort get abandoned and the funding ceased. If it doesn’t work, when do they try to find out what part is not working and try to fix it.
In climate science, we NEVER get to this point. It just goes on and on and on.
Stephen Richards: There isn’t a climate or weather model out there that can predict 100% the future.
So what? There isn’t a perfect paper clip or light bulb either.
We have models to the effect that the weather distribution of December in the NH is shifted toward the cold end compared to the weather distribution of October; which in turn is shifted to cold end compared to the weather distribution of July. We have models ranking days vs nights, and some regions vs. others. A model of climate that was predicted the correct rank orderings of the climate summaries 50 years hence might be worth a lot. We don’t have one, but that does not mean that we never will.
There isn’t a model for the lift capacity of an aircraft wing that has 0 error. Models do not need 0 error to be useful. What they need is a public record of being accurate enough to achieve the purposes for which they were designed.