fix(propagation): Rothermel wind exponent E uses 3.59e-4, not 10E-4

[3dfirelab]

Rothermel (1972) gives E = 0.715 exp(-3.59e-4 sigma) with sigma in 1/ft. In C++ "10E-4" is 10 x 10^-4 = 1e-3, so E was ~10x too small (SH5: 0.058 instead of 0.556) and the wind factor phi_w = C U^B (beta/beta_op)^-E ~4x too small for shrub fuels. RothermelAndrews2018.cpp and FarsitePropagationModel.cpp already use 3.59e-4.

Calibrations made with the previous value (propagationSpeedAdjustmentFactor, windReductionFactor) need to be revisited.

[antonio-leblanc]

Thanks @3dfirelab, confirmed: sd is already in 1/ft there, and RothermelAndrews2018 and Farsite both use 3.59e-4.

The red CI is expected: tests/runff uses Rothermel and the fire front now has ~630 points vs 408 in real_case.kml.ref. Since windReductionFactor=0.4 may have been tuned with the old value, I'll leave that call to filippi before regenerating the reference.