|Oct 14, 2015|
|4 years 7 months ||small fixes for the MPI version
|Oct 28, 2015|
|4 years 7 months ||When enabling a memory use optimization for skipped pixels, a memory
management bug was introduced. In the runoff, a minor optimization
|4 years 7 months ||Documentation fixes, mostly related to the parallel version. Better
documentation of the run.sh as well as osX export that had been lost
in the last few versions.
|Nov 2, 2015|
|4 years 6 months ||This fixes a bug with the MPI version of SimpleTerrainRadiation
(uninitialized values summed through all_reduce_sum as well as
possibly negative sky view factors and the usage of the diffuse
radiation without multiplying it by the sky view factor whe
|Nov 3, 2015|
|4 years 6 months ||Now the sky view factors work properly... The previous versions were
only valid on slope coordinates and were used on horizontal
coordinates (the paper actually 'forgot' to mention it).
|4 years 6 months ||Woops... I swapped two fields
|Nov 4, 2015|
|4 years 6 months ||Added the possibility to write additional data (on top of the runoff
components) in the SMET files generated by the Runoff module. In
particular, it is now possible to write the soil temperature at a
given soil depth.
|Nov 5, 2015|
|4 years 6 months ||An invalid name in the headers has been fixed. Some documentation
has been added.
The runoff module now has a call back to SnowpackInterface (to be
used in an upcoming version).
|Nov 6, 2015|
|4 years 6 months ||And there was still another bug in the sky view factor... Now this
should really match the paper.
|Nov 9, 2015|
|4 years 6 months ||More consistent handling of getGrid(): if the requested parameter is
not available, an empty grid is returned. Then it is possible to
simply test for "grid.empty()" in order to check if the
requested parameter could be retrieved.
|4 years 6 months ||Improvements for the getGrid() method. It now works properly with
OpenMP and properly handles meteo forcing grids.
|Nov 10, 2015|
|4 years 6 months ||Modified class Runoff so that the required grids are directly
obtained from the internal reference to the SnowpackInterface
instance (-> method fillHydroTables() no longer needs to be
|Nov 11, 2015|
|4 years 6 months ||Small adjustments: invalid POIs are now only written out on the
master node. The TSOIL output is exposed in SnowpackInterface and
documented. This gives ideas on how to handle outputs in the ini
file for the future...
|Nov 12, 2015|
|4 years 6 months ||A potential arithmetic exception has been found and fixed in the
computation of view factors. Small improvements in the output
messages for the MPI version.
|4 years 6 months ||Code cleanup: getting rid of the ebalance/old directory, renaming
the NORA model into HELBIG, fixing a documentation warning. The
version numbers have been pushed to 3.0.0. This is quite out of the
blue, but the goal is to look more synchronized with
|4 years 6 months ||A method to return the sky view factors has been implemented in
order to ease models comparisons
|Nov 16, 2015|
|4 years 6 months ||For added safety, setting corr_dir to 0 when no correction could be
computed (there was already day=false). This should not be necessary
but could prevent problems in case of quick changes in the code.
|4 years 6 months ||In order to limit the damage when things go wrong, the radiation
correction factor is limited to 1.2 (instead of 2 previously). A
clean solution still has to be found though... (it seems that for
the mountains, the computed atmospheric attenuation is
|Nov 17, 2015|
|4 years 6 months ||Hoping to bring even more clarity into the run script: all the user
editable settings are grouped in the begining of the script,
including the output redirection.
|4 years 6 months ||Documentation touch up to clarify the initialization of canopy data