GMS:Error Summary Plot: Difference between revisions

From XMS Wiki
Jump to navigationJump to search
 
(4 intermediate revisions by the same user not shown)
Line 4: Line 4:
==Error types==
==Error types==
Three error types are shown in the plot.
Three error types are shown in the plot.
''Mean Error'': This is the average error for the points. This value can be misleading since positive and negative errors can cancel.
*''Mean Error'': This is the average error for the points. This value can be misleading since positive and negative errors can cancel.


''Mean Absolute Average'': This is the mean of the absolute values of the errors. It is a true mean, not allowing positive and negative errors to cancel.
*''Mean Absolute Average'': This is the mean of the absolute values of the errors. It is a true mean, not allowing positive and negative errors to cancel.


''Root Mean Square'': This takes the average of the square of the errors and then takes its square root. This norm tends to give more weight to cases where a few extreme error values exist.
*''Root Mean Square'': This takes the average of the square of the errors and then takes its square root. This norm tends to give more weight to cases where a few extreme error values exist.


== Time matching ==
== Time matching ==
Line 14: Line 14:


== MODFLOW errors ==
== MODFLOW errors ==
The ''Error Summary Plot'' is most useful with models that do not make error computations from observation data. MODFLOW does compute error computations from observation data, and these errors are available in the [[GMS:MODFLOW Solution Properties Dialog|''MODFLOW properties'' dialog]]. Therefore, when using MODFLOW, you should use the MODFLOW properties dialog and not the ''Error Summary Plot''. The errors shown in the Error Summary Plot are not the same as those found in the MODFLOW properties dialog due to the time matching and interpolation issues described above. The MODFLOW properties dialog shows the errors computed by MODFLOW for the observations at the actual observation times and is therefore superior. The errors reported by MODFLOW have nothing to do with the model output times, only observation times.
The ''Error Summary Plot'' is most useful with models that do not make error computations from observation data. MODFLOW does compute error computations from observation data, and these errors are available in the [[GMS:MODFLOW Solution Properties Dialog|''MODFLOW Properties'' dialog]]. Therefore, when using MODFLOW, use the ''MODFLOW Properties'' dialog and not the ''Error Summary Plot''.  
 
The errors shown in the ''Error Summary Plot'' are not the same as those found in the ''MODFLOW Properties'' dialog due to the time matching and interpolation issues described above. The ''MODFLOW Properties'' dialog shows the errors computed by MODFLOW for the observations at the actual observation times and is therefore superior. The errors reported by MODFLOW have nothing to do with the model output times, only observation times.


== Plot creation ==
== Plot creation ==
Line 26: Line 28:
After the plot type is set in the first step of the ''Plot Wizard'', the '''Next''' button is clicked to move to the second step of the ''Plot Wizard''.
After the plot type is set in the first step of the ''Plot Wizard'', the '''Next''' button is clicked to move to the second step of the ''Plot Wizard''.


: '''''Coverage''''': Displays the name of the coverage where the current data for the plot is coming from.
* ''Coverage'': Displays the name of the coverage where the current data for the plot is coming from.
 
* ''Measurement'': This is the name of the current measurement, created in the ''Feature Objects'' | '''Attributes''' dialog, being plotted.
: '''''Measurement''''': This is the name of the current measurement, created in the ''Feature Objects'' | '''Attributes''' dialog, being plotted.
* ''Feature Objects'': Displays which feature object is utilized in the current plot, points or arcs.


: '''''Feature Objects''''': Displays which feature object is utilized in the current plot, points or arcs.


{{Navbox GMS}}
{{Navbox GMS}}

Latest revision as of 16:14, 25 August 2017

MODFLOW
Pre-processing
MODFLOW Commands
Building a MODFLOW Model
Map to MODFLOW
Calibration
Packages Supported in GMS
Saving a MODFLOW Simulation
Importing MODFLOW Files
Unsupported MODFLOW Features
Run MODFLOW
Post-processing
MODFLOW Display Options
MODFLOW Post-Processing Viewing Options
Reading a MODFLOW Simulation
Tutorials
Packages
Flow: BCF6, HUF, LPF, UPW
Solvers:

DE4, GMG, NWT, PCG,

PCGN, LMG, SIP, SOR,

SMS
Other:

BAS6, BFH, CHD1, CLN,

DRN1, DRT1, EVT1, ETS1,

GAGE, GHB1, GNC, HFB1,

HUF, LAK3, MNW1, MNW2,

OUT1, RCH1, RIV1, SFR2,

STR1, SUB1, SWI2, WEL1,

UZF1

An Error Summary plot is one of the plot types that can be created by the Plot Wizard. See the Error Summary plot section on the Plot Wizard page for a basic description. The Error Summary plot is used to display a text listing of the mean error, mean absolute error, and root mean squared error for an observation measurement in a coverage. The error values may be listed for a single time step of a dataset or for all time steps.

Error types

Three error types are shown in the plot.

  • Mean Error: This is the average error for the points. This value can be misleading since positive and negative errors can cancel.
  • Mean Absolute Average: This is the mean of the absolute values of the errors. It is a true mean, not allowing positive and negative errors to cancel.
  • Root Mean Square: This takes the average of the square of the errors and then takes its square root. This norm tends to give more weight to cases where a few extreme error values exist.

Time matching

When the measurement is transient (i.e., observed heads at various times) then it is very unlikely that the output times from the model match the times when the field measurements were taken. In this case GMS will interpolate an observed value at the model output time. If the model output time is before the first field measurement for a particular point then the interpolated value is the first measurement. Similarly if the model output time is after the last field measurement then the interpolated value is the last measurement. When the model output time is between the first and last measurement then GMS will find the 2 closest field measurments to the model output time and use linear interpolation to compute the "observed" value at that time.

MODFLOW errors

The Error Summary Plot is most useful with models that do not make error computations from observation data. MODFLOW does compute error computations from observation data, and these errors are available in the MODFLOW Properties dialog. Therefore, when using MODFLOW, use the MODFLOW Properties dialog and not the Error Summary Plot.

The errors shown in the Error Summary Plot are not the same as those found in the MODFLOW Properties dialog due to the time matching and interpolation issues described above. The MODFLOW Properties dialog shows the errors computed by MODFLOW for the observations at the actual observation times and is therefore superior. The errors reported by MODFLOW have nothing to do with the model output times, only observation times.

Plot creation

Error Summary plots are created in the Plot Wizard by setting the plot type to Error Summary. A sample plot is shown in the figure.

ErrorSummary.png


Error Summary Plot Options

After the plot type is set in the first step of the Plot Wizard, the Next button is clicked to move to the second step of the Plot Wizard.

  • Coverage: Displays the name of the coverage where the current data for the plot is coming from.
  • Measurement: This is the name of the current measurement, created in the Feature Objects | Attributes dialog, being plotted.
  • Feature Objects: Displays which feature object is utilized in the current plot, points or arcs.