Leyla

What I would actually do, in order:

1. Lift the dried surface film while it is still a film. Emulsion that has dried on the surface peels or lifts in sheets and that removes most of the material in ten minutes. Do this before it rains again and breaks up.
2. Skim the top 5-10cm of the patch itself, a bit wider than the stain. A bucket or two of soil.
3. Dispose of it properly: that is household waste for the tip, not the compost heap and not the green bin. Take the tins too; the tip will have a paint stream.
4. Backfill with compost or topsoil and leave it.

What I would not do is dig deeply or treat it as a remediation project. Going deeper spreads it and disturbs soil structure for no gain, given how shallow the contamination is.

Continuous mode means it ignores the humidity target, not that it ignores the tank. The float switch that stops it when the tank is full is a safety and it always applies.

So the behaviour you are seeing - runs, fills, stops - is exactly what continuous mode does when the tank is not being emptied.

The fix for that is the part nobody mentions in the shop: almost every dehumidifier has a drain hose outlet, usually a small blanked spigot at the back. Fit a hose, run it to a sink, a drain, or out through an airbrick, and the tank stops being the limit. That is what continuous mode is designed to be used with.

Without the hose, continuous mode is close to pointless - you have swapped an automatic machine for one that needs you twice a day.

Numbers, because they are more alarming than people expect.

On a 30°C day in full sun, a closed car's cabin reaches roughly 50-60°C within an hour, and the dashboard surface can exceed 80°C. Most of the rise happens in the first twenty to thirty minutes: the curve is steep early and then flattens as the car approaches equilibrium.

The stratification is the part that answers your question: floor to roof can differ by 15-20°C in a car standing in sun. So the footwell might be at 40°C while the parcel shelf is at 60°C.

That is a large difference and it is often the difference between "warm" and "ruined", but note what it is not: it is not cool. Nothing in a car parked in sun is cool, and nothing that must not exceed body temperature should be left in one.

The warm-makes-it-worse detail is a genuine clue, and it points at the oil rather than the mechanism.

Oil thins as it heats. Thinner oil between the plates should mean less drag, so if drag increases when warm, something else is changing: and the usual candidate is the plates themselves. Friction material swells slightly with heat, and plates that are near the limit for warp or wear will close that fraction of a millimetre once they expand.

It also fits an oil that is not rated for wet clutches: those additives do their work more effectively at temperature, which is exactly when you notice it.

So: check the oil specification, and if it is correct, the plate stack is the next place to look.

"The oven is not hostile to it, it barely notices" is a good way to put it. I had been thinking of the incandescent as the primitive option rather than the one suited to the job.