Skip to content

COVID19 models – a lesson for those who trust climate scientists-Joe D’Aleo

April 19, 2020

By Paul Homewood


Reposted from ICECAP:

COVID19 models – a lesson for those who trust climate scientists

The models used to estimate U.S. deaths from COVID19 had projections that ranged from over 2 million a few weeks back to 100,000-240,000 a week ago down to 60,000 this week. Hopefully the projections and reality will converge on an even lower number.

The media seemed perplexed about the large changes in the models and questioned their concept and value. Welcome to the world of modeling (see this comparison).

We in the weather business use models as a tool and they present special challenges. We have a plethora of models to choose from which run 2 to 4 times a day or even on a smaller scale hourly. Operational model forecasts go out to as far 16 days into the future. The models are subject to large errors especially in the latter periods when storms are coming inland from data sparse regions like the Pacific. We have a favorite phrase – garbage in, garbage out.

The climate model story is even worse. The climate models overstate the warming from greenhouse gases by a factor of two or more.


Climate model forecasts versus the satellite and balloon observations

The models projected the greatest warming in the tropical high atmosphere (called the Tropical Hot Spot) where air in the mean rises due to convergence of air from both hemispheres. But the models warming results mostly from the release of heat from condensation of water vapor (95% of the greenhouse effect). CO2 is a trace gas, just 0.04% of the atmosphere by volume.

See no warming trend in the upper atmosphere since 1979 where models predict it.


The lack of warming also holds for the tropical Pacific ocean down to 300 meters depth from 160E to 80W.



The climate cabal portray the output from their models as gospel, and the believers confuse the model projections with measured data. The data centers make unsubstantiated claims that a given month or year is the warmest back to the beginning of the record (1880 or even 1850) often by the tiniest of margins (0.05C for example).

But the data is just not there to make those claims.

71% of the earth is oceans, and data before the satellite data became available 40 years ago, ocean data was reliant on ships which travelled along specific routes mainly in the northern hemisphere near the land.

A large percentage of land surface was erratically covered with observing sites and the data too often spare and intermittent.

To create an apparent agreement with their sacred models, the data centers manipulated real data or even generated with their models data for 95% of the planet that had poor coverage before the satellite era.

MIT’s climate scientist Dr. Mototaka here exposed the phoney claims that most years are the warmest ever since the 1850s of 1880s.

“The supposed measuring of global average temperatures from 1890 has been based on thermometer readouts barely covering 5 per cent of the globe until the satellite era began 40-50 years ago. We do not know how global climate has changed in the past century, all we know is some limited regional climate changes, such as in Europe, North America and parts of Asia.”

The world’s greatest scientists in the 1970s knew that and created the first view of global trends by using land temperatures where available in the Northern Hemisphere. It showed a 2F warming from 1880s to around 1940 and then a cooling that by around 1970s cut that by more than half. Further cooling until the late 19870s virtually eliminated the rest.


These early measurements were erased when models and global land (and ocean) data was ‘created’ in the following decades. The models cooled the early data and enhanced the warming to create an apparent steady warming during the eras when fossil fuel use increased. Our conclusion is that there is man made global warming but the men are in NOAA, NASA and Hadley.


By the way, the ocean data coverage and accuracy did not really become reliable until the implementation of the ARGO buoys in 2000 designed to accurately measure temperatures and ocean heat content. Like the satellite used to measure sea level changes (which showed no changes – until artificial adjustments were made), the early results were disappointing – showing no warming. The ARGO data supported the inconvenient near two decade pause in the warming that started in the late 1990s.


Eventually in desperation ahead of the Paris Accord, they made this go away by adjusting the buoy data to match to inferior ship data.


The models are tuned to manipulated (fraudulent) data. In addition to what we have shown above, this can be seen when you examine all the extremes of weather that these models and the theories predict. See here how each of the claims have failed.


Read the full post here.

  1. It doesn't add up... permalink
    April 19, 2020 11:51 am

    Seen elsewhere:

    The modern version is garbage in, gospel out.

  2. Thomas Carr permalink
    April 19, 2020 12:04 pm

    This should be essential reading for the BBC’s phalanx of experts . Time to draw up a list of who they are and send on this and all related reports. They cannot say that they are too busy now.

  3. April 19, 2020 12:44 pm

    So glad to see this article. I have made this point for years. The COVID19 models have served to educate the public on the “reliability” of models in a way which those of us objecting to the climate situation were not able.

    I believe one of the early models was proposed by a British “scientist” from Imperial College, Neil Ferguson. He has kept reducing his predictions. However, he did great damage (or helped the left agenda) as these models were used to scare people and push a panic.

    Current US models are all over the place, but generally are rapidly trending downward. From what I have learned, most of these models are flawed in that they do not take culture into account. In other words, what we do in our lives in West Virginia in the Southern Appalachians is treated no differently from those in New York City.

    Last week a paper by MIT economics professor and physician Jeffrey Harris states: “New York City’s multitentacled subway system was a major disseminator — if not the principal transmission vehicle — of coronavirus infection during the initial takeoff of the massive epidemic.”

    WOW, who could have predicted that? Two days prior to this bit of breaking blockbuster news, I was on I-79, having gone 25 miles south to procure a new chain for my chain saw. A digital sign over the interstate challenged my right to be on the road unless it was “essential.” My trip, in my car alone, is treated no differently than riding the NYC subway system by those who model. Interesting.

    While the climate change models gets quickly into the weeds, this COVID19 model business is on the putting green for all to see and understand.

  4. April 19, 2020 1:20 pm

    The Northern Hemisphere Ice Core records from Greenland show:

    the last millennium 1000AD – 2000AD has been the coldest millennium of the entire Holocene interglacial.
    each of the notable high points in the Holocene temperature record, (Holocene Climate Optimum – Minoan – Roman – Medieval – Modern), have been progressively colder than the previous high point.
    for its first 7-8000 years the early Holocene, including its high point “climate optimum”, had virtually flat temperatures, an average drop of only ~0.007 °C per millennium.
    but the more recent Holocene, since a “tipping point” at ~1000BC, has seen a temperature diminution at more than 20 times that earlier rate at about 0.14 °C per millennium.
    the Holocene interglacial is already 10 – 11,000 years old and judging from the length of previous interglacials the Holocene epoch should be drawing to its close: in this century, the next century or this millennium.
    the beneficial warming at the end of the 20th century to the Modern high point has been transmuted into the “Great Man-made Global Warming Alarm”.
    eventually this late 20th century temperature blip will come to be seen as just noise in the system in the longer term progress of comparatively rapid cooling over the last 3000+ years.
    other published Greenland Ice Core records as well as GISP2, (NGRIP1, GRIP) corroborate this finding. They also exhibit the same pattern of a prolonged relatively stable early Holocene period followed by a subsequent much more rapid decline in the more recent (3000 year) past.


  5. Broadlands permalink
    April 19, 2020 1:24 pm

    James Hansen while at NASA in 1999 said about the U.S. temperature record “In the U.S. the warmest decade was the 1930s and the warmest year was 1934”.

    True for that decade. It’s worth adding that 1917 was the coldest year (and still is), but only four years later 1921 was the warmest. And 1921 remained the warmest year on record until NOAA lowered the US Weather Bureau data to make 2012 the warmest.

    Indeed, NCDC has lowered the official US Weather Bureau monthly temperatures for the majority of US states in a seasonal fashion for 1921, 1934, 1938 and 1940…if not all the years in between.

    As one example (May, 1934). Compare the current NOAA values (Climate at a Glance, Statewide Time Series) for the states listed… averages and extremes. US Weather Bureau Climatological tables…

  6. April 19, 2020 1:42 pm

    Our early COVID19 models were based on those coming from Imperial College scientist, Neil Ferguson, who has kept reducing them.

    Our models, with dire predictions, are also being cut in half at every turn. Those seeking to protect them, claim that they did not take mitigation into account. BUT, they did. These early models had the desired effect of creating panic and fear. Even Drs. Fauci and Birx, have been amazed at how easy it has been to use this to control people and shut down countries. Take note of those observations.

    The models also failed to take culture into account. The models treat those of us in West Virginia in the Southern Appalachians the same as those in New York City. There is a big difference in how we conduct our lives.

    The day after Easter I was on I-79 returning from a trip to Home Depot, 30 miles south of my home, to get a new chain for my chain saw. A digital overhead sign challenged my right to be on the interstate for other than “essential” reasons. I was more than annoyed.

    A study, published 2 days later by MIT (Massachusetts Institute of Technology–and my late father’s PhD school) economics professor and physician Jeffrey Harris, concerned a parallel between ridership on the New York City subway system with some 5 million riders per day and the rapid surge in infections. He states: “New York City’s multitentacled subway system was a major disseminator — if not the principal transmission vehicle — of coronavirus infection during the initial takeoff of the massive epidemic.”

    WOW, who could have predicted that? Took a study. Models which assume that we in West Virginia are subject to the same outcomes as riders of the NYC subway system are set to be unreliable. One size does not fit all.

  7. C Lynch permalink
    April 19, 2020 6:08 pm

    You’d wonder how they get away with it. But then you see the ideological brainwashing that is the norm in Western education systems (3rd Level Institutions in particular), the media and the gutlessness of most politicians and you don’t wonder at all.

  8. Harry Passfield permalink
    April 19, 2020 6:26 pm

    Twenty some-odd years ago I built a PC app that kept six mainframe system (same-app) databases in sync. Nowadays, I might venture, a ‘model’. Coding it was fairly easy: a couple of thousand lines of code.
    The difficulty came with coding the ‘what-ifs’: (it had to run autonomously as far as possible). That took another 20,000+ lines of code.
    It’s always the ‘what-Ifs’ that catch you out. I wonder, did the Covid19 modellers allow for them? (As if the programmer could think of all the possible ‘what-ifs’. )

    • Phoenix44 permalink
      April 20, 2020 8:29 am

      I haven’t looked at the code of the Imperial model, but I have read their paper of 16th March. I’ve been building and overseeing models for 30 years, I can confidently say I would have just shrugged at Imoerials COVID model. It’s a bottom up attempt which gives the impression of accuracy and precision without having either – maybe that’s what fooled government. They did not – could not – know key parameters, so modelled how those key parameters are made up. But they didn’t know those parameters either. This is a key point in ALL modeling – making guesses at lower levels of calculations does not make your top level numbers any more accurate than just guessing them in the first place.

      But then the absolute key numbers – e.g. deaths rates – are just plugged in. So we get absurd numbers – 70.9% of over 80s in ICU – and then just a blanket 50% die. I would have thrown it out and told them to make a simple model that could be used to run a Monte Carlo analysis to produce a probability analysis of most likely outcomes. Instead we have a big model (not actually complex, it just looks like it is) that uses one set of assumptions.

  9. April 19, 2020 7:09 pm

    Models produce scenarios, which is OK up to a point. It’s what the media and many of the so-called ‘experts’ make out of these scenarios in order to alarm the public that’s the problem.

    • April 19, 2020 8:10 pm

      I remember the first Apple computers, with the original spreadsheet, Visicalc!

      What on Earth do you want one of those for, says the boss.

      So we can do “What if” calculations, says I

      Now they call them “models”!

    • Duker permalink
      April 19, 2020 9:08 pm

      We already have two countries in Europe that took 2 different routes.
      Greece and Sweden both just over 10 mill people, one went into lockdown early while the other continues with lower level measures
      the IHME modelling from the University of Washington, and which the White House is using,
      [] projects these results of total deaths by 1 Jun ( with error ranges)
      Sweden 5650
      Greece 119

      It can be clearly see from ‘live runs’ that those 2 countries based on current values will have hugely different outcomes based on what the governments did and how quickly

      The UK is projected by 1 Jun with 37,500 deaths, but it has a much bigger population than just over 10 mill, being almost 69 mill

    • Phoenix44 permalink
      April 20, 2020 8:32 am

      I think the problem with COVID is our government took as Gosoel one set of assumptions run through one model. That’s just crazy. There are 20-60 key assumptions in the Imperual model, each with perhaps 10 plausible figures/ranges. That’s a lot of possible combinations. Imperial should have run a simple model and varied the assumptions to produce probabilistic forecasts. The chance the Imperial model was right was Lottery winning chance.

      • Adam Gallon permalink
        April 20, 2020 8:46 am

        The government used a plan designed with pandemic flu in mind.

      • dave permalink
        April 20, 2020 9:55 am

        “…The University of Washington model…”

        for the United Kingdom,

        as updated on April 17,

        shows a prediction, for deaths the next day, April 18, of 1,239.

        The range of possibility given is 157 to 4,203.

        On April19 at 9:00 AM, the UK Government made its daily announcement of new deaths in hospital of persons with the virus – 596.

        There are nuances in everything. My neighbour, whom I talk to garden to garden, and who is a nurse in a major hospital, says that five Phillipino nurses on just one ward contracted the virus. And yet, in the Phillipines itself, the illness hardly spreads. There is something about being an immigrant from an always sunny country to a country with a gloomy winter that makes you more vulnerable. Vitamin D status?

        The nurses all recovered within ten days, by the way. The hospital itself has gone very quiet – only a few active cases there.

      • Duker permalink
        April 20, 2020 10:03 am

        The basic are fairly well known , as epidemics are reasonably common and of course the yearly flu is minor-major epidemic every year .
        For a new virus the unique parameters arent exactly known, but the progress is . The difference between Greece and Sweden is very predictable , the exact differences not so much.
        Clearly the Greek epidemic has almost petered out , while Sweden has some months to run.
        Thats the real difference to climate models , apart from 1000s of lives, is time span. Its not decades of maybe small changes of temperature with random and seasonal changes thrown in.
        In 6 months its likely to be mostly over with exponential change down to
        reducing linear change.
        Hopefully the huge sums being spend in climate actions will find a better use now

  10. gosportmike permalink
    April 19, 2020 7:32 pm

    Is it not the case that before you can produce a reputable model you must fully understand ALL the constituent parts of that problem. Once you fully understand those parts – why do you need a model?

    • Phoenix44 permalink
      April 20, 2020 8:36 am

      If we really understand how something works, models will accurately forecast what will happen if we change something. And once something has happened, and we have good data, a model can help to understand more fully what has happened, by simulating at different levels an epidemic say. I think the Imperial model would be useful for that – once we gave reliable data.

  11. April 20, 2020 5:43 am

    Wow! What an amazing post. Many thanks to both of you.

    • April 20, 2020 6:02 am

      One minor point. This sentence, “CO2 is a trace gas, just 0.04% of the atmosphere by volume” is really irrelevant. There are many good critiques of the climate science ghg model but “CO2 is a trace gas, just 0.04% of the atmosphere by volume” isn’t one of them.

      • dave permalink
        April 20, 2020 2:07 pm

        In geochemistry, ‘trace of X’ means ‘less than 0.1% of X.’ The ‘importance’ of such a statement depends on what X is.

      • April 20, 2020 2:15 pm


  12. April 21, 2020 10:46 pm

    More than 10 years ago I was still in my comfy corporate job and every year we had an exercise in which the leading brass of the company tried to determine where the journey goes. I remember rosy projections of an ever-growing market. When I was part of this meeting for the first time I was too sure what to think of it. I meekly asked what if the growth prospects would not materialize or god-forbid we would go into a shrinking market. Everyone looked at me as if horns were growing out of my head but the topic was shelved quickly. I was told that those numbers came from a very prestigious consultant and had cost a pretty penny. I understood that I was to shut up and I did. Models are opinions, projections are opinions. Anything that cannot be proven beyond a reasonable doubt is an opinion. All those COVID-19 models fail and there is certainly no lack of motivation to get it right. Same for Climate Change but as long as there is no Climate Change lockdown nobody gives a damn.

  13. saparonia permalink
    April 22, 2020 10:57 pm

    I read today on one of the BBC’s propaganda pages that some people show NO SYMPTOMS of this fantasy deadly lurgi. We are warned that people kept in those hot damp old peoples homes are dying.
    Therefore we have to allow our economy to fail, watch unemployment rise and be kept like animals in our cages until Bates, who is not a doctor or scientist and didn’t even write MsDos has got ready his virus killer injection that will cost the EARTH, Literally.

  14. Spurwing Plover permalink
    April 24, 2020 3:39 pm

    So according to the Communists Broadcasting System(CBS)People are not needed what kind of idiotic hogwash is this? and as for Covid 19 we can look to China as the main culpret

Comments are closed.

%d bloggers like this: