notebook.community

Edit and run

FloPy

Model checker demonstration





In [1]:



    


import os
import sys
import flopy

print(sys.version)
print('flopy version: {}'.format(flopy.__version__))



    


















    






3.6.0 |Anaconda 4.3.0 (x86_64)| (default, Dec 23 2016, 13:19:00) 
[GCC 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.57)]
flopy version: 3.2.6

Set the working directory





In [2]:



    


path = os.path.join('..', 'data', 'mf2005_test')

Load example dataset and change the model work space





In [3]:



    


m = flopy.modflow.Modflow.load('test1ss.nam', model_ws=path)
m.change_model_ws('data')



    


















    






changing model workspace...
   data

By default, the checker performs a model-level check when a set of model files are loaded, unless load is called with check=False. The load check only produces screen output if load is called with verbose=True. Checks are also performed at the package level when an individual package is loaded

The check() method

Each model and each package has a check() method. The check method has three arguments:





In [4]:



    


help(m.check)



    


















    






Help on method check in module flopy.mbase:

check(f=None, verbose=True, level=1) method of flopy.modflow.mf.Modflow instance
    Check model data for common errors.
    
    Parameters
    ----------
    f : str or file handle
        String defining file name or file handle for summary file
        of check method output. If a string is passed a file handle
        is created. If f is None, check method does not write
        results to a summary file. (default is None)
    verbose : bool
        Boolean flag used to determine if check method results are
        written to the screen
    level : int
        Check method analysis level. If level=0, summary checks are
        performed. If level=1, full checks are performed.
    
    Returns
    -------
    None
    
    Examples
    --------
    
    >>> import flopy
    >>> m = flopy.modflow.Modflow.load('model.nam')
    >>> m.check()

The check class

By default, check is called at the model level without a summary output file, but with verbose=True and level=1. The check methods return an instance of the check class, which is housed with the flopy utilities.





In [5]:



    


chk = m.check()



    


















    






test1ss MODEL DATA VALIDATION SUMMARY:
  2 Warnings:
    RCH package: Mean R/T ratio < checker warning threshold of 2e-08
    RCH package: Variable NRCHOP set to value other than 3

  Checks that passed:
    Compatible solver package
    Unit number conflicts
    DIS package: zero or negative thickness
    DIS package: thin cells (less than checker threshold of 1.0)
    DIS package: nan values in top array
    DIS package: nan values in bottom array
    BAS6 package: isolated cells in ibound array
    BAS6 package: Not a number
    LPF package: zero or negative horizontal hydraulic conductivity values
    LPF package: zero or negative vertical hydraulic conductivity values
    LPF package: negative horizontal anisotropy values
    LPF package: vertical hydraulic conductivity values below checker threshold of 1e-11
    LPF package: vertical hydraulic conductivity values above checker threshold of 100000.0
    LPF package: horizontal hydraulic conductivity values below checker threshold of 1e-11
    LPF package: horizontal hydraulic conductivity values above checker threshold of 100000.0
    GHB package: BC indices valid
    GHB package: not a number (Nan) entries
    GHB package: BC in inactive cells
    GHB package: BC elevation below cell bottom

Summary array

Most of the attributes and methods in check are intended to be used by the check() methods. The central attribute of check is the summary array:





In [6]:



    


chk.summary_array



    


















    
Out[6]:








rec.array([ ('Warning', 'RCH', 0, 0, 0, 1.52404e-09, '\r    RCH package: Mean R/T ratio < checker warning threshold of 2e-08'),
 ('Warning', 'RCH', 0, 0, 0, 1.0, '\r    RCH package: Variable NRCHOP set to value other than 3')], 
          dtype=[('type', 'O'), ('package', 'O'), ('k', '<i8'), ('i', '<i8'), ('j', '<i8'), ('value', '<f8'), ('desc', 'O')])

This is a numpy record array that summarizes errors and warnings found by the checker. The package, layer-row-column location of the error, the offending value, and a description of the error are provided. In the checker, errors and warnings are loosely defined as follows:

Errors:

either input that would cause MODFLOW to crash, or inputs that almost certainly mis-represent the intended conceptual model.

Warnings:

inputs that are potentially problematic, but may be intentional.

each package-level check produces a check instance with a summary array. The model level checks combine the summary arrays from the packages into a master summary array. At the model and the package levels, the summary array is used to generate the screen output shown above. At either level, the summary array can be written to a csv file by supply a filename to the f argument. Specifying level=2 prints the summary array to the screen.





In [7]:



    


m.check(level=2)



    


















    






Errors and/or Warnings encountered.

test1ss MODEL DATA VALIDATION SUMMARY:
  2 Warnings:
    RCH package: Mean R/T ratio < checker warning threshold of 2e-08
    RCH package: Variable NRCHOP set to value other than 3

  Checks that passed:
    Compatible solver package
    Unit number conflicts
    DIS package: zero or negative thickness
    DIS package: thin cells (less than checker threshold of 1.0)
    DIS package: nan values in top array
    DIS package: nan values in bottom array
    BAS6 package: isolated cells in ibound array
    BAS6 package: Not a number
    LPF package: zero or negative horizontal hydraulic conductivity values
    LPF package: zero or negative vertical hydraulic conductivity values
    LPF package: negative horizontal anisotropy values
    LPF package: vertical hydraulic conductivity values below checker threshold of 1e-11
    LPF package: vertical hydraulic conductivity values above checker threshold of 100000.0
    LPF package: horizontal hydraulic conductivity values below checker threshold of 1e-11
    LPF package: horizontal hydraulic conductivity values above checker threshold of 100000.0
    GHB package: BC indices valid
    GHB package: not a number (Nan) entries
    GHB package: BC in inactive cells
    GHB package: BC elevation below cell bottom

DETAILED SUMMARY:
type	package	k	i	j	value	desc
Warning	RCH 	0	0	0	1.52e-09	RCH package: Mean R/T ratio < checker warning threshold of 2e-08
Warning	RCH 	0	0	0	1.00e+00	RCH package: Variable NRCHOP set to value other than 3










    
Out[7]:








<flopy.utils.check.check at 0x102aaf748>

example of package level check and summary file





In [8]:



    


m.rch.check()



    


















    






RCH PACKAGE DATA VALIDATION:
  2 Warnings:
    Mean R/T ratio < checker warning threshold of 2e-08
    Variable NRCHOP set to value other than 3











    
Out[8]:








<flopy.utils.check.check at 0x10913c6a0>

example of summary output file





In [9]:



    


m.check(f='checksummary.csv')



    


















    






test1ss MODEL DATA VALIDATION SUMMARY:
  2 Warnings:
    RCH package: Mean R/T ratio < checker warning threshold of 2e-08
    RCH package: Variable NRCHOP set to value other than 3
  see data/checksummary.csv for details.

  Checks that passed:
    Compatible solver package
    Unit number conflicts
    DIS package: zero or negative thickness
    DIS package: thin cells (less than checker threshold of 1.0)
    DIS package: nan values in top array
    DIS package: nan values in bottom array
    BAS6 package: isolated cells in ibound array
    BAS6 package: Not a number
    LPF package: zero or negative horizontal hydraulic conductivity values
    LPF package: zero or negative vertical hydraulic conductivity values
    LPF package: negative horizontal anisotropy values
    LPF package: vertical hydraulic conductivity values below checker threshold of 1e-11
    LPF package: vertical hydraulic conductivity values above checker threshold of 100000.0
    LPF package: horizontal hydraulic conductivity values below checker threshold of 1e-11
    LPF package: horizontal hydraulic conductivity values above checker threshold of 100000.0
    GHB package: BC indices valid
    GHB package: not a number (Nan) entries
    GHB package: BC in inactive cells
    GHB package: BC elevation below cell bottom











    
Out[9]:








<flopy.utils.check.check at 0x102aaf2b0>

















In [10]:



    


try:
    import pandas as pd
    df = pd.read_csv('data/checksummary.csv')
except:
    df = open('data/checksummary.csv').readlines()
df



    


















    
Out[10]:










  
    

      
      type
      package
      k
      i
      j
      value
      desc
    

  
  
    

      0
      Warning
      RCH
      0
      0
      0
      0.0
      Mean R/T ratio < checker warning threshold of ...
    

    

      1
      Warning
      RCH
      0
      0
      0
      1.0
      Variable NRCHOP set to value other than 3

checking on write_input()

checking is also performed by default when write_input() is called at the package or model level. Checking on write is performed with the same verbose setting as specified for the model. However, if errors or warnings are encountered and level=1 (default) or higher, a screen message notifies the user of the errors.

By default, the checks performed on load() and write_input() save results to a summary file, which is named after the packge or the model.





In [11]:



    


m.write_input()



    











In [ ]:

Content source: brclark-usgs/flopy

Similar notebooks:

notebook.community | gallery | about