|Nov 12, 2015|
|3 years 11 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
|3 years 11 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.
|Nov 11, 2015|
|4 years 11 hours ||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 10, 2015|
|4 years 1 day ||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 9, 2015|
|4 years 2 days ||Improvements for the getGrid() method. It now works properly with
OpenMP and properly handles meteo forcing grids.
|4 years 2 days ||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.
|Nov 6, 2015|
|4 years 5 days ||And there was still another bug in the sky view factor... Now this
should really match the paper.
|Nov 5, 2015|
|4 years 6 days ||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 4, 2015|
|4 years 7 days ||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 3, 2015|
|4 years 8 days ||Woops... I swapped two fields
|4 years 8 days ||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).
|Nov 2, 2015|
|4 years 9 days ||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
|Oct 28, 2015|
|4 years 14 days ||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.
|4 years 14 days ||When enabling a memory use optimization for skipped pixels, a memory
management bug was introduced. In the runoff, a minor optimization
|Oct 14, 2015|
|4 years 28 days ||small fixes for the MPI version
|Oct 13, 2015|
|4 years 29 days ||Small fixes for warnings...
|4 years 29 days ||micro optimizations: no memory allocation for pixel that are skiped,
add an MPI barrier in the destructor for cleaner termination,
pre-reserving memory in some vectors. More options in run.sh for SGE
|Oct 8, 2015|
|4 years 1 month ||Fixed bugs in class Runoff:
(1) nodata values in the input runoff grids are now properly handled
(2) subwatersheds are properly identified when using a catchment
grid whose resolution is not the same as the one of the DEM
|Oct 2, 2015|
|4 years 1 month ||Some unecessary MPI barriers have been removed. A few unused typedef
have been removed. The copyright notices have been uniformized and
edited when necessary.
|Oct 1, 2015|
|4 years 1 month ||A bug has been found in the MPI version: when reading the meteo
data, only the master was able to pre-read the data but it was never
broadcasted... thus forcing all workers to re-read the data
themselves. The MPI outputs have also been cleaned up in