LWFB90 has a constant head lower boundary now

The depth of a water table can be specified (param_b90$water_table_depth), which will influence soil moisture.

Maintenance

• fixed new R CMD check warnings in R-devel due to "non-portable use of kind" in Fortran
• moved input data checks to Fortran code for speed reasons

Maintenance

Bug fixes

• Major bug fixed concerning the calculation of global radiation from sunshine duration hours (see issue #65)
• small bug concerning aggregation of output variables in BudgMon and BudgAnn using miscellaneous function process_output_LWFB90()

Changes

• default root-growth-period parameter (rgroper) was set to 0, to prevent root growth. The parameter defines the period during which the final root length density distribution is reached. The user now needs to actively switch on root growth by choosing a value > 0. It as advised to carefully investigate layer-wise root water uptake (tran) when using the root growth module for young stands.

Output for precipitation interval

Basic output is now at the precipitation interval level:

• With options_b90$prec_interval > 1, the output of run_LWFB90() now contains one row for each day
  and precipitation interval of the simulation.
• Nothing changes with daily precipitation input (options_b90$prec_interval = 1).
• Irrespective of the precipitation interval, the unit of water and vapour fluxes is mm/d.
• With respect to precipitation interval output, the list item daily_output of the return was renamed to output.
• Removed output-argument and exported the function process_outputs_LWFB90() that does the job.
• New variable added to output: sthr.
• Removed redundant variable slvp from layer_output (soil evaporation).
• Removed variable psiti from layer_output (total potential of soil layer): can be calculated by summing the layers' matrix potententials (psimi) and the gravity potentials at the soil layers' midpoints.

Maintenance & bug fixes

Changes:

• new daily output variables: daily solar (slrad), net solar (solnet), net longwave (lngnet) and net radiation above (aa) and below canopy (asubs).

Bug fixes:

• make_rootden() with method = 'betamodel' led to negative
  root density of the top layer when specified maxrootdepth was lower than soil depth.
• no more errors when using sub-daily precipitation input
• typo in hydpar_wessolek_tab() caused NAs in return for texture = 'fSms'
• soil_to_param(): some unique soil materials were omitted in the return, if they had multiple occurences in the input soil-data.frame.

Ready for CRAN submission

Most important changes:

• no more writing to file
• switched from superseded package 'doSNOW' to 'doFuture'
• rename functions and arguments for consistency

Bug fixes

Bugs fixed:

• zero division error causing infinite relawat-values in MISCDAY.ASC-output.
• 'betamodel'-method of MakeRelRootDens returned the increment of the cumulative root proportion at the soil nodes, instead of the relative root density

Release for AFM manuscript

• added new function msiterunLWFB90
• runLWFB90-function: model input (param.b90, option.b90, standprop_daily) is appended
  to the return value BEFORE evaluating the output_fun-argument. In this way,
  on-the-fly post-processing of model results including model-input is possible.
• minor bugs fixed

First public release

• 'output_fun' argument replaces 'gof_fun'-argument in runLWFB90() for more flexible output.
• minor bugs fixed
• most R CMD CHECK issues resolved, however, MacOS 10.11 El Capitan compilation error (see Issues) persists but can be resolved by updating to the latest gfortran compiler.

Pre-release version for LWF

Small patch for mrunLWFB90