Dagsvik silent on the detection limit of their instrument

The suitability of their method to investigate the presence or absence of global warming has not been demonstrated.

Dagsvik silent on the detection limit of their instrument
In forskning.no, Dagsvik offers smoke and mirrors but fails to respond to the core my criticisms of their controversial discussion note. 

This blog post was publihed in Norwegian in forskersonen.no

In his response to  my critique of Dagsvik and Moen’s discussion note, John K. Dagsvik fails to address my fundamental points:

What is the sensitivity of the model they used?

I asked about the sensitivity of the method they had employed: How much would the climate need to change for them to detect an impact that could be ascribed to CO2? JD did not respond to that point.

JD confirmed elsewhere that their study had not ascertained that CO2 had not influenced the climate, it only found that the analyzed temperatures had not changed enough to be detected by the employed method. This is why I would like to repeat my question: What is the sensitivity of the method?

In his response, Dagsvik explained:

A purely statistical (stochastic) model was used to analyze a selection of temperature series, each by itself. Using more weather stations does not make the analysis more or less precise.

This statement confirms my suspicion that they tried to detect climate change in the temperature records of an individual weather station.

Seeking to detect global climate change in the records of an individual weather station is like trying to ascertain a pandemic by observing one individual's body temperature.

In that context, I would like to point to the discussion note:

From these estimates and the formula above it follows [..] that the upper bound of the temperatures in July in Oslo becomes 51.0 degrees Celsius. The highest temperature ever recorded in Oslo is 35 degrees (21 July 1901), which we note is substantially lower than the upper bound. [..] Thus, even in a stationary model with normally distributed variations, such as FGN [the model employed by JD], quite extreme temperature realizations are possible.

That means, in other words, that the temperature in Oslo would need to rise well above 51 degrees to be detected as abnormally warm. And JD would need several such extreme temperatures to ascertain that the climate had changed. If deviations by tens of degrees, or even several degrees from the historical record are required for the method to indicate there is something going on, we could not expect it to confirm climate change, given that the IPCC reports an increase in global average temperature by just over one degree compared to the preindustrial level. The quotes strengthen me in my suspicion that the method used was really not suitable to investigate the research question that was posed. In that case, the failure of the method to detect a systematic temperature rise would hence indicate the method's insufficient precision rather than the absence of climate change.

What has fluid mechanics to do with temperature analysis?

I criticized the discussion note for a literature review that was off-topic, i.e. it was irrelevant to the research they presented. JD did not respond to this critique. I also showed that their literature review repeated standard climate-denialist talking points based on outdated references and had ignored newer evidence that showed the assertions to be false. While JD provided some arguments with newer references that try to salvage their earlier assertions, this does not address the issue of the irrelevance of these points to the study they had conducted.

I shared Dagsvik's texts with my colleague Simen Åndøy Ellingsen, professor of fluid mechanics. He sums up a lengthy email to me with:

"The argument is irrelevant. These properties of Navier-Stokes' equations, which are inherently correct in principle, do not change the fact that climate models provide predictions good enough to inform best response decisions. It's a common misconception. The flow is chaotic locally, but statistically it is quite predictable; Just think of the flow in a stream – if you look from a few centimetres away second by second, it's chaos, but if you take a step back, you'll see a calm current that changes slowly. Call it weather vs. climate, if you will. Navier-Stokes describes the first image, but it's the latter that's interesting here. Take-home message: Solving Navier-Stokes' equation in full detail is here completely uninteresting, nor necessary to make useful predictions."

Ellingsen also points out that similar approaches to Navier-Stokes' equations are used in almost all practical applications, and form the basis for today's design of aircraft, ships, cars, processor fans, oil pipes, wind turbines, hydropower and much, much more.

Since Dagsvik and Moen clearly do not understand fluid mechanics, why do they see themselves competent to assess these complex arguments about atmospheric physics which they reproduce in a one-sided manner in their note? To argue even that atmospheric physicists are wrong in their use of fluid mechanics?

Given the lack of understanding of the science they are discussing, and the lack of relevance to the question they are investigating, Dagsvik and Moen should remove paragraphs 3-6 and parts of paragraph 1 of the discussion note.