Point Stat Config File Generator


In the first table below are the config file parameters that most users will need to change. Following this are more tables with parameters grouped according to function.

After each table there is a "Submit" button. Clicking any one of them should return a config file to you, unless there is an error. In that case, use your browser's "Back" button to return to this page and fix the problem.


Basic Parameters


Name Input value Explanation
Model A short text string identifying the name to be assigned to the model being verified. This text string is written out as a header column of the STAT output so that verification statistics from multiple models may be differentiated.
Begin Delta Sec Beginning and ending time offset values in seconds for observations to be used. These time offsets are defined in reference to the forecast valid time, v. Observations with a valid time falling in the window [v+beg_ds, v+end_ds] will be used.
End Delta Sec See above.
Forecast Field A comma-separated list of model variables and corresponding vertical levels to be verified. The GRIB code itself or the corresponding abbreviation may be used to specify which model fields are to be verified. A level indicator in the form ANNN, ZNNN, PNNN, or PNNN-NNN must follow each GRIB code. These indicate an accumulation interval, a single vertical level, a single pressure level, and a range of pressure levels, respectively. NNN indicates the accumulation or level value.
Obs Field Identical to the fcst_field except that it applies to the observation field. It can be left blank to use the fields identical to those specified in fcst_field.
Forecast Threshold For each grib code listed in the forecast field, one or more thresholds must be specified for use in computing discrete statistics. The thresholds may be specified using the Fortran conventions of gt, ge, eq, ne, lt, le to indicate greater than, greater than or equal to, equal to, not equal to, less than, and less than or equal to, respectively. The number of entries in thresholds must match the number of entries in vx_grib_code. However, multiple thresholds may be applied to each GRIB code by provided a space- separated list within the double or single quotes.
Obs Threshold For each observation field, one or more thresholds can be specified for use in computing discrete statistics. Specification is done exactly as for fcst_thresholds.
Forecast Wind Threshold See above.
Obs Wind Threshold When verifying winds via either the u and v components or the speed and direction, it may be desirable to eliminate winds below a certain speed. This threshold filters the winds based on speed, even when u and v winds are input. Format is the same as for fcst_thresholds.
Message Type
ADPUPA AIRCAR AIRCFT
ADPSFC ERS1DA GOESND
GPSIPW MSONET PROFLR
QKSWND RASSDA SATEMP
SATWND SFCBOG SFCSHP
SPSSMI SYNDAT VADWND
ANYAIR ANYSFC ONLYSF
Specifies a list of the message types to use for verification. At least one message_type must be provided.
Mask Station IDs A filename that contains a space-separated list of station ID's at which verification should be performed.
Mask Grid A comma-separated list of pre-defined NCEP grids over which to perform the Point-Stat verification. The predefined grids are specified as GNNN where NNN is the three digit designation for the grid. Supplying a value of FULL indicates that the verification should be performed over the entire grid on which the data resides.
Mask Polyline A comma-separated list of files that define lat/lon polygons to be used in specifying verification regions. Several masking polygons used by NCEP are predefined in the data/poly subdirectory of the MET distribution.
Output Flag Controls the type of output that the Point-Stat tool generates. Each flag corresponds to an output line type in the STAT file. Setting the flag to 0 indicates that the line type should not be generated. Setting the flag to 1 indicates that the line type should be written to the STAT file only. Setting the flag to 2 indicates that the line type should be written to the STAT file as well as a separate ASCII file where the data is grouped by line type.




Confidence Interval Parameters


Name Input value Explanation
Alpha Values A comma-separated list of alpha values to be used when computing confidence intervals. The confidence interval computed is 1 minus the ci_alpha value.
Bootstrap Interval Specifies which method should be used for computing bootstrap confidence intervals. The BCa (bias-corrected percentile) method is highly accurate but computationally intensive. The percentile method is somewhat less accurate but efficient.
Bootstrap Proportion Must be set to a value between 0 and 1 inclusive. When computing bootstrap confidence intervals over n sets of matched pairs, the size of the subsample, m, may be chosen less than or equal to the size of the sample, n. This variable defines the size of m as a proportion relative to the size of n. A value of 1 indicates that the size of the subsample, m, should be equal to the size of the sample, n.
# Resamples Defines the number of subsamples that should be taken when computing bootstrap confidence intervals. This variable should be set large enough so that when confidence intervals are computed multiple times for the same set of data, the intervals do not change much. Setting this variable to zero disables the computation of bootstrap confidence intervals that may be necessary to run in realtime or near-realtime over large domains.
Random Number Generator Defines the random number generator to be used in the computation of bootstrap confidence intervals. Subsamples are chosen at random from the full set of matched pairs. The randomness is determined by the random number generator specified.
Bootstrap Seed May be set to a specific value to make the computation of bootstrap confidence intervals fully repeatable. When left empty, the random number generator seed is chosen automatically which will lead to slightly different bootstrap confidence intervals being computed each time the data is run. Specifying a value here ensures that the bootstrap confidence intervals will be computed the same over multiple runs of the same data.




Smoothing Parameters


Name Input value Explanation
Interpolation Method A comma-separated list of interpolation methods to be used when interpolating forecast data to observation locations. The valid values which may be listed are MIN, MAX, MEDIAN, UW_MEAN, DW_MEAN, and LS_FIT for the minimum, maximum, median, unweighted mean, distance-weighted mean, and a least squares fit. Providing multiple interpolation methods indicates that statistics should be computed multiple times using a different interpolation method each time.
Interpolation Width A comma-separated list of values to be used in defining the neighborhoods over which the interpolation is performed. The neighborhood is simply a square centered on the observation point. The interp_width value specifies the width of that square.
Interpolation Threshold A number between 0 and 1. When performing interpolation over some neighborhood of points the ratio of the number of valid data points to the total number of points in the neighborhood is computed. If that ratio is greater than this threshold, the matched pair is discarded. Setting this threshold to 1 requires that the entire neighborhood must contain valid data. This variable will typically come into play only along the boundaries of the verification region chosen.




Miscellaneous Parameters


Name Input value Explanation
Rank Correlation Flag Indicates whether or not Kendall's Tau and Spearman's Rank correlation coefficients should be computed. The computation of these rank correlation coefficients is very slow when run over many matched pairs.
Grib Parameter Table Sets the GRIB table 2 parameter version, and thus indicates how to interpret GRIB codes between 128 and 255.
Temp Directory This option specifies the directory where temporary files should be written.
Output Prefix If nonempty, specifies a string to be used in the output file name. It can be useful for keeping results for different models or variables from overwriting each other.