“But Modulz”

All about climate models, aka the modulz. The stoopid ones.


[PEDRO] Interesting. When you use an ML-style method to estimate the effect on climate of CO2 emissions, it comes out fairly small.

[SHOSHIDO] Dear climate scientists and epidemiologists. Your models are garbage. Stop imposing them on us.

[EUAN] Maybe people just aren’t buying the decades of failed predictions and wildly inaccurate models anymore.

[BLACKJACK] If scientists can’t model imminent threats, there’s no chance they can accurately model impact of climate change over decades or centuries.

[Janitor] Climate models are useful for studying climate mechanisms but utterly useless for long term forecasting.

Objections and Replies

(Derivation) I’d like an engineering-quality exposition of sensitivity
Reply. There is no direct calculation to accurately prove this.

(Freedom) Don’t impose modulz on us
Reply. Nobody does. They’re mostly useful tools to explore the future.

(Observation) Should we trust modulz or observations
Reply. Unfortunately observations of the future are not available at this time.

(Science) Modulz are not science
Reply. Have you ever used a thermometer?

(Stoopid) Since modulz are stoopid shows that AGW is stoopid too
Reply. We don’t need no damn models to support AGW.

(Sensitivity) Establishing the true climate sensitivity will save us trillions
Reply. The ratio of temperature change to cumulative carbon emissions, is approximately independent of both the atmospheric CO2 concentration and its rate of change on these timescales.

(Tweaked) Modulz are tweaked to get the results scientists want—
Reply. If only it was that easy, we would not need models in the first place. Many properties emerge from runs. They still need to hindcast properly.

(Until) We need better modulz until—
Reply. They’re what we got. I’ll keep my eyes open even if hindsight isn’t 20/20.


{Cf.} #ButPredictions, #ButObservations, #ButEvidence.

{You Need One} No model, no measurement. No model, no data. No model, no implementation of any theory whatsoever. Unuseable science.

{Simulation} To know more about computer simulations in science, confer to Eric’s entry.

{Spherical cows} All modellers have skeletons of spherical cows in their closets.

Further Readings

2018-01. Timeline: The History of Climate Modelling.

2020-2. But RCPs.

2020-07. An Assessment of Earth’s Climate Sensitivity Using Multiple Lines of Evidence. https://doi.org/10.1029/2019RG000678