This page lists a MESH configuration for the Great Lakes based on files provided from Amin Haghnegahdar in 2015, updated to work with newer versions of the MESH code.
WF_R1" parameter specified in MESH_parameters_hydrology.ini.
This setup is likely a variant of the setup presented in Haghnegahdar, 2014. Resultant of a computer crash, the exact files corresponding to the publication were lost. These files were recovered from a computer, and may correspond more closely with Amin's thesis work: https://uwspace.uwaterloo.ca/handle/10012/9108
The publication lists parameter values as well as metrics of the simulation; the information could be used to verify if the recovered setup is the same as in the publication and this note could then be redacted – at a later date.
Via verbal communication with Amin when he worked for GIWS; files were recovered by Xiaoyong Xu in Summer, 2018
These files were extracted from a local archive at the University of Waterloo by Xiaoyong Xu on July 5, 2018, and relayed by Amin Haghnegahdar on August 1, 2018. They were copied to science in June, 2019. The original files ran using a variant of Standalone MESH 1.3.006. Support pages for this version of the model no longer exist (formerly hosted via hal-depot).
Citation: Haghnegahdar, A., Tolson, B. A., Davison, B., Seglenieks, F. R., Klyszejko, E., Soulis, E. D., Fortin, V. and Matott, L.S. (2014): Calibrating Environment Canada's MESH Modelling System over the Great Lakes Basin, Atmosphere-Ocean, Atmosphere-Ocean, Volume 52, Issue 4, 281-293, DOI: 10.1080/07055900.2014.939131.
This configuration has been modified from the original files provided with updated formatting and options compatible with modern MESH code. The base options and parameter and initial values are unmodified from the files as provided. It's unclear if the original files correspond to the Haghnegahdar (2014) publication or Amin's thesis, as the original files were lost in drive failure and these were recovered from a backup, but when and how these files were generated is still unknown.
The files for this configuration can be found here: https://github.com/MESH-Model/GLakes_Haghnegahdar_2019
|MESH_drainage_database.r2c||Re-formatted, otherwise unmodified.|
|MESH_input_run_options.ini||Re-formatted, resume options deactivated. Updated to use "tb0" format input streamflow and reservoir files. Updated to write reservoir outputs and to deactivate the output of the basin-average energy balance output file.|
|MESH_input_reservoir.tb0||Reservoir options preserved, converted to "tb0" format. Original specification and inline notes preserved as comments in the header of the file. Five of six reservoirs use defined rule curves (based on WF_ROUTE reservoir routing). One reservoir uses insertion. The reservoirs have been re-ordered to put the reservoir with insertion (Lake Nipigon, "|
|MESH_input_streamflow.tb0||Converted to "tb0" format. Original specification and inline notes preserved as comments in the header of the file.|
|MESH_parameters_CLASS.ini||Re-formatted, otherwise unmodified.|
Converted to the "2.0" version of the file and defined the fields as required. Removed fields from the previous file that were not actually used. Re-formatted. Derived and added "
To use the prescribed "
|MESH_input_soil_levels.txt||Re-formatted, otherwise unmodified.|
The configuration uses CLASS, but is presumed to have been calibrated to CLASS version 3.5. The current version of CLASS in the model is 3.6.
The configuration uses WF_ROUTE, the former routing routine. The basic assumptions of the channel routing is similar to the updated RTE routine, though the configuration uses the lumped "
WF_R2" parameter for both in-channel geometry and roughness (as opposed to the derived channel geometry and explicitly Manning's n used for RTE).
The configuration uses WF_ROUTE for reservoir routing, and includes rule curves for lake representation for five of the six reservoirs (the six reservoir, first in the file, uses streamflow insertion, reading the time-series for insertion from file). WF_ROUTE uses an accumulation of inflow as the basis for the lake routing rule curve, where RTE uses an accumulation of storage in its place (makes more sense, conceptually). Thus, the existing parameters for those reservoirs are unlikely to transfer to RTE and using the reservoir routing in RTE would require re-calibration of these parameters.
Reservoir input for insertion
The start date in the reservoir input file is presumed based on the simulation start date included in the files as provided, as the "txt" format of the input reservoir file does not support specifying the start date of the record for insertion.
Overbank lumped roughness coefficient (
The files as provided are presumed to have been configured for a variant of MESH 1.3.006. This version contains the known "
WF_R1" error that carries forward from MESH 1.3.003 and onward (see the description in notes for "MESH 1.3.003 (revision 489)" on this page: https://wiki.usask.ca/display/MESH/Releases#Releases-MESH1.3).
To emulate the "
WF_R1" value that would be derived by the erroneous code. the value was explicitly calculated and provided via the hydrology (2.0) parameter file:
WF_R1" value, any version of MESH r1605 or higher must be used with these files.
- The parameters in both MESH_parameters_CLASS.ini and MESH_parameters_hydrology.ini should be compared to the Haghnegahdar (2014) publication and Amin's thesis to validate if they are the same or not (and this and the 2015 page can be updated once that has been established)
- Consider outright re-calibration, considering the version of CLASS in MESH has been updated and to use RTE routing in-place of WF_ROUTE
- Consider if an alternate rule curve or reservoir routing model should be used in place of power function
- Validate the time-series of streamflow for insertion at Lake Nipigon to the presume reference start date for the record (June 1, 2004)