Your ectotherms example is case in point, of choosing to view a system through a Gaussian lens, rather than reflecting some true set of dynamics. Lizards do not experience uniform amounts of heat as they move. There are nonlinear interactions between the animal's metabolic rate, hydration and various heat exchange mechanisms, paired with countless external environmental factors such as wind speed, humidity, and ground temperature. As such, the resulting body temperature distribution that is measured over time is nothing like a simplistic average of heat inputs.

Also consider the timing and duration of heat exposure, whereby it matters how fast or slow heat is absorbed or dissipated. The lizard's body does not instantaneously respond to changes in temperature; averaging ignores these temporal dynamics. I could go on about thresholds and tipping points, where a certain critical temperature changes the lizard's behavior in a dramatic burst. This "bursting" makes systems sensitive to extremes, and thus ill-fitted to a scenario where the behavior is best described by a simple average over time.

There is also the influences of landscape, predators, etc. In short, there are countless complex spatial and behavioural interactions that produce distributions of behavior that most definitely are not normally distributed (i.e. very heavy tailed). The normal distribution is a choice, and a bad one at that. It does not reflect the reality of nature.

The same applies to your other examples, as they do with all examples in nature.

Your last point does not make sense. Something being an artifact does not enforce a condition whereby only that instance appears in nature. There is no instance where this is the case.

The Gaussian is a lens, and a poor one at that.