Hollyhan/slm/add option to write zero ice thickness#1
Open
hollyhan wants to merge 56 commits into
Open
Conversation
New module files: init_slm_mod.f90 and user_specs_mod.f90
New files: new file: Makefile new file: sl_model_driver.f90 new file: sl_model_mod.f90 Removed file: deleted: SL_MODEL_FWTW.f90
Wrappers included are: subroutines 'read_sl' and 'write_sl' in init_slm_mod.f90
new subroutines in sl_model_mod.f90 are: 'sl_timewindow', 'set_planet', 'sl_solver_checkpoint', 'sl_solver_init', 'sl_solver'. These subroutines are called from 'sl_model_driver.f90'
…ENCHMARKED Fully benchmarked against Han et al. (GMD Time window paper)) Subroutine set_planet is called within slmodel_checkpoint in sl_model_mod.f90 script rather than in sl_model_driver.f90.
Included a subroutine called 'sl_set_unit_num()' in sl_model_mod.f90 that sets a unit number for Fortran write(*,*) statement. changed the unit numbers used by SLM to read and write data (1 and 2) to 201 since MALI uses unit numbers up to 200.
sl_model_mod.f90: 1. Subroutine 'sl_solver_init' takes in optional input arguments 'mali_iceload' and 'mali_bedrock' provided by MALI. 2. Subroutine 'sl_solver' takes in optional input argument 'mali_iceload' and output arbument 'slchange' to exchange outputs between MALI and SLM sl_model_driver.f90: Subroutine 'sl_set_unit_num' is replaced to Subroutine 'sl_solver_checkpoint' in sl_model_mod.f90.
It now correctly detects when .f90 files have been modified.
…hickness and bedTopo) 1. Take in ISM grid mask from the ice sheet model and merge with global data such that (Merged_data_GaussianGrid) = (mali_data_in_GaussianGrid) + (Global_data)(1 - mali_mask_in_GaussianGrid) 2. Clean up the code a bit
Also, renaming routines and modules and editing Makefile to reflect the name changes.
modified: Makefile
new file: namelist.sealevel => created a text namelist file based on the user_specs_mod.f90
new file: sl_init_mod.f90 => changed the name of the file 'init_slm_mod.f90', include subroutines that read in the namelist
modified: sl_model_driver.f90 => call the subroutines 'sl_drive_readnl' and 'sl_call_readnl'
modified: sl_model_mod.f90 => included a wrapper subroutine that calls the subroutine 'sl_readnl'
Remove 'config_' from the variable names
This commit creates a str_len parameter of 200 characters and uses it to define the length of all string variables used for filename or path names. This avoids filename truncation and associated read errors when long filenames are used. It also fixes the fact that the 'times' output file had been shortened to 'time'. Those truncation issues were introduced in the switch to using a namelist, such that the associated string variables could not be defined with unspecified length (*).
This allows a user/developer to look at the code to see which i/o call actually failed. Note that this will only produce useful results if the model is compiled in debug mode.
* Add debug symbols (-g) to help with debugging. (May want to remove that eventually) * Add gnu option to suppress underflow floating point exceptions, which are a nuisance but not problematic * Add build path to include path
Make filenames longer
This commit replaces fType with fType_in and fType_out, so that the input and output file types can differ (e.g. text files for input and netcdf for output). A few other details: * The .nc extension is added to NetCDF output files * unit_num is moved to user_specs_mod so it can be accessed from sl_init_mod * To get both kinds of output files, use 'both' instead of ''
Allow input and output file types to differ
Note the fix is hardcoded assuming the SLM is run with resolution where nlon is twice nlat. I think that will always be the case, but if not, a more general but complicated solution could be used.
Note I had to comment out the gnu-specific 'backtrace' command. It could be manually uncommented as needed.
This is for debugging and should eventually be reverted or pared down.
Uninitialized arrays is what was causing erratic behavior (code hanging, floating point exceptions) on large, complex problems. Initializing all arrays after allocating them made those problems go away.
This reverts commit fdc2101. Remove debugging write statements now that the code is working.
This reverts commit 1a6f63d. Remove rest of debugging statements.
Fix random occurrences of code hanging and floating point exceptions
patch zero ice thickness everywhere outside of the ice-model domain when coupled
These variables got carried over from the older version of the SLM when it was coupled to the PSU model to run over the Northern Hemispheric Ice Sheet during the Last glacial cycle. These variables are no longer used in the MALI-SLM coupling framework.
Make constants consistent with mali and others
Remove variables that are not used
Write TPW variables into a new file every timestep
One calculation where ocean is defined to be anywhere below sea level and one for where ocean is defined to be anywhere below sea level AND there is no marine-based ice.
with and without considering marine-based areas
Previously, the variables `norder` and `nglv` were parameters that were used in declaring the dimension of many variables. This has raised two issues in our recent efforts to run a suite of sensitivity coupled MALI-SLM simulations that sweeps across various values of MALI mesh resolution and the number of Gauss-Legendre nodes in latitude (represented by variable `nglv`) in the SLM. The first issue (or more like an inconvenience) was the need to code compilation needs to be done everytime `nglv` parameter was changed, and the second was the failure of compilation when the `nglv` value was set higher than 2000 because of limited memory imposed to variables that are declared with defined a dimension size. This PR addresses the two issues by 1) no longer assigning the variables `norder` and `nglv` as model parameters, and 2) making the variables that use these variables to declare their dimension as allocatable.
…list_options Change model res parameters `nglv` and `norder` as namelist options
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Previously, running the SLM coupled with an ice-sheet model always required reading in the ice thickness outside the ISM domain from external files. This caused inconvenience in cases where users wanted to consider zero-ice thickness outside the ISM domain - they still had to create files containing fields of zeros in the preprocess step. This PR resolves the issue by reading ice files from external files optional. That is when the namelist option 'patch_ice' is turned on, zere field ice thickness is assigned to the memory from the SLM directly.