[BUGFIX] Assumption that Ttop of Lineartemp and HalfspaceCoolingTemp is at z=0km

[ajvanamerongen]

Whats the purpose of this PR?

• Bug fix
• New feature
• Documentation update
• Other, please explain

Describe it in more detail below:
The temperature of LinearTemp and HalfspaceCoolingTemp are bugged as they assume that Ttop coincides with z=0, rather than z=z_top_polygon. This results in a shift of temperatures.

Now: T=gradT * z + T0 whereas it should be T=gradT(z - z0) + T0.

A polygon starting at 10 km depth with a thermal gradient of 10C/km and 0C top temperature will have 100C as top temperature rather than 0C.
A representative example of the issue and fix are shown in the two figures below.

Checklist

• The PR title is descriptive and starts with the appropriate tag: [BUGFIX], [ADDITION], [DOC], etc.
• New tests (either assessing the correct behaviour of new internal functions or the correctness of a tutorial) were added, or old tests were updated
• Affected tutorials have also been updated
• [?] The new feature was added in a way that does not break public API
• New documentation related to the new feature was added
• The new code follows the contributor guidelines, in particular the Runic Style

Figure 1.
A subducting slab (left) described as a polygon from z=-12.5 km, with a halfspace cooling thermal profile with Tsurface=0 and thermal age of 80Myr.
A continental plate (right) described as a polygon from z=-12.5 / -8 km (the tapering sedimentary wedge in yellow is part of the plate). It has a linear geotherm from the top at -8 km to -112.5 km of ~ 10C/km.
The white isotherms are 100C, 150C, 350C, 450C and 900C. The 100C and 150C isotherms are practically at the surface although they should be at least several km's lower.

Figure 2.
The improved thermal model using the bugfix. All isotherms are deeper and the 900C isotherm is below the model domain of 80 km (appropriate for a realistic gradient above 10C/km).