Removing fixed c2m and c2m Issue #1055

[vitorlavor]

As pointed out in Issue #1055, $c_{2m}$ and $c_{2h}$ from Eqs. 28a, b (HF08) are being being set to 0.5 - several tests were performed for multiple neighbourhoods and there were no apparent issues. This PR reverts the fixed values and uses the $c_{2m}$ and $c_{2h}$ calculated directly from Eqs. 28a, b (HF08)

[sunt05]

Thanks for working on Issue #1055. Removing the hardcoded c2 = 0.5 makes sense if the physics-based formula produces better results.

Clarification needed: Is the goal to always use the formula, or to use it except when numerically unsafe? If the latter, some guard logic should remain (see inline comments).

Test update required: This change affects model outputs but does not update the test baseline. Please:

# 1. Rebuild and run tests
source .venv/bin/activate
make dev
pytest test/core/test_sample_output.py -v

# 2. Verify outputs are physically reasonable (no NaN/Inf, sensible magnitudes)

# 3. Regenerate baseline
python scripts/gen_sample_output.py

# 4. Commit updated baseline with your changes
git add test/fixtures/data_test/sample_output.csv.gz
git add src/suews/src/suews_phys_rslprof.f95

[sunt05]

The old IF (phi_hatmZh >= 1.) block was not just forcing c2 = 0.5 — it also short-circuited the formula evaluation. By removing it, we expose this calculation to edge cases:

• Denominator (2.*beta*phim_zh - kappa) could be zero or near-zero
• Large c2 values will overflow EXP(c2/2.)
• Even when (1.-phi_hatmZh) = 0, multiplying by NaN/Inf still produces NaN

Consider keeping a guard:

IF (phi_hatmZh >= 1.) THEN
   cm = 0.
ELSE
   c2 = (kappa*(3.-(2.*beta**2.*Lc/phim_zh*dphi)))/(2.*beta*phim_zh - kappa)
   cm = (1.-phi_hatmZh)*EXP(c2/2.)
END IF

[sunt05]

Same numerical stability concern as c2/cm above — the denominator (2.*beta*phih_zh - kappa*Scc) can cause division errors, and EXP(c2h/2.) can overflow.

[sunt05]

Please delete this commented-out block. Git preserves history, so there is no need to keep old logic as comments — it adds noise and maintenance burden.

[sunt05]

These comments now contradict the code behaviour. Please remove or update them.