So it is constructed once and then given as parameter. A Meteo object in Snowpack.cc has been renamed for clarity (and ease of searching for it by name).

In late 2021, the cloudiness computation in MeteoIO was improved in
order to better handle the interpolation over the nights (since it is
based on the measured ISWR being compared with the potential ISWR, this
can not be computed at night). This lead to strong regressions for many
simulations (as reported by Fabiano Monti). In early 2022, a fix was
committed (MeteoIO dc347ac75e528b1ed1c1fdfea4841d4afdaefaec): the
terrain shading had to be taken into account in order to prevent
interpolating a cloudiness computed while in the shade of a neighbouring
mountain. This shading can be provided either as a horizon file or
through a DEM (and the computed horizon can then be saved in a horizon
file for quicker reuse later on).

Unfortunatelly, this was not enough for the IMIS stations that don't
measure ISWR but RSWR instead (although there are less impacted since
for most of the season SNowpack will rely on their measured TSS as
boundary condition). This problem had ...

COUPLEDPHASECHANGES cannot be used when using Richards equation for soil, since both treat soil freezing in an inconsistent way.

Slope angle of snow profile location considered for reading and writing CAAML profiles

See merge request !6

Redoing deleted commits. Original messages: Replacing abs with std::abs, Fixing the indentation, Fix to reduce differences between versions, Small fix, Improving on the commits from last Friday (2 September), Error correction, Renamed MS_SURFACE_RUNOFF into MS_SURFACE_MASS_FLUX, MS_SURFACE_RUNOFF correction, Correction for soil water content, New output variable MS_ICE_SOIL

Virtual lysimeter

See merge request !7

Merge branch 'virtual_lysimeter' of https://gitlabext.wsl.ch/snow-models/snowpack into virtual_lysimeter

Added new parameter MS_SURFACE_RUNOFF, which is MS_SNOW if SWE > 0, and MS_RAIN - MS_EVAPORATION otherwise, i.e. the true mass flux through the soil surface all year long

In some cases (this happened with a very short dataset filled by data
generators), MeteoIO returns an empty data vector (in this case, when exhausting
its buffer). This was not tested for in Snowpack, leading to segfault (trying to
access a given station in the vector that is empty).

Bulk and brine salinity (0540, 0541) were written before 0535, in contrast with what was declared in the file header.

Slope angle of snow profile location considered for reading and writing CAAML files (layer heights and thicknesses are affected)

Correcting a typo, and testing explicitly for density option "FIXED" and throwing an error otherwise (i.e., not assuming that after evaluating the other options, everything else must be "FIXED").

Re-introducing: "Adding the possibility to suppress writing *.haz files (controlled by the key HAZ_WRITE in [Output] section)", which was overwritten with the previous commit. Note that CaaMLIO.h had inconsistent end-of-line characters.