CPT Tags

  • cpt:C

    The category number. This is a compulsory tag for the X input file in the PFV option. The tag should be set to "1" for the below-normal category, "2" for the normal category, and "3" for the above-normal category.

  • cpt:clim_prob

    The climatological probability. This is an optional tag for use in the X input file when using the PFV option. It sets the climatological probability for the corresponding category indicated by the cpt:C tag.

  • cpt:col

    In gridded files, this tag must be set to "X" or "lon", representing longitudes. In station files, this tag must be set to "station", and in index files to "index". The tag is compulsory for the first block of data, and should not be changed for subsequent blocks.

  • cpt:field

    An abbreviated name of the field variable. This tag is optional, but if present CPT will attempt to set an appropriate colour scaling (blue-to-red for temperature, and red-to-blue for precipitation), and may automatically switch on the zero-bound option for precipitation. CPT recognises the following abbreviations:

    • Precipitation
      • 'precipitation',
      • 'prate'
      • 'prcp'
      • 'prec'
      • 'pre'
      • 'pr'
      • 'rain'
      • 'rainfall'
    • Temperature
      • 'temperature'
      • 'temp'
      • 'tmax'
      • 'tmin'
      • 'tref'
      • 't2m'

    CPT will also attempt to check for consistency between the X and Y variables when running the GCM option without any skill correction. If, for example, the X-field is marked as one of the precipitation variables listed above and the Y-field contains the word "dry", CPT will assume that the Y field is some measure of dry-spells, and will issue a warning. Similarly, if the X-field is marked as one of the temperature variables listed above and the Y-field contains the word "cold", CPT will assume that the Y field is some measure of cold-spells, and will again issue a warning. It is recommended to cancel the calculation and to use variables that one would expect to be positively correlated (for example, to convert the dry-spells into a measure of wet-spells, or the cold-spells into a measure of warm spells). Alternatively, a skill correction could be applied from the GCM Options , either to the individual GCMs and/or in the model combination options ifmore than one model is being used.

  • cpt:level

    The atmospheric or oceanic level for which the data apply. This is an optional tag. It can be used to force CPT to recognise as separate fields a meteorological parameter that is measured at difference atmospheric levels for example. The units of the level do not matter.

  • cpt:M

    The ensemble member number. This is an optional tag, but when used with the GCM option, CPT will recognise individual ensemble members, and will treat these as separate fields. There is an option to sort the ensemble members, so that member 1 always contains the lowest values.

  • cpt:missing

    The missing value flag. This tag over-rides the default missing value flag. The tag is optional. Different missing values can be used for different fields, levels, members, and models (q.v.). See Replacing missing values for further details on missing value estimation. However, note that the missing vlaue options dialogue only allows you to set the missing vlaue flag for one field.

  • cpt:model

    The name of the model (typically the GCM) used to generate the data. This is an optional tag, but when used with the GCM option, CPT will recognise different models for the various forecast combination options. CPT assumes that the model is unchanged unless the value of the tag is reset.

  • cpt:ncol

    The number of columns of data in the subsequent block. In a gridded file the number of columns represents the number of longitudes, and excludes the first column of latitudes. In a station or index file, the number of columns represents the number of stations or indices, and excludes the first column containing the dates. The tag is compulsory for the first block of data, but can be excluded if it remains unchanged for subsequent blocks.

  • cpt:nrow

    The number of rows of data in the subsequent block. In a gridded file the number of rows represents the number of latitudes, and excludes the immediately preceding line of longitudes. In a station or index file, the number of rows represents the number of time steps available rather than the span from the first to last date. For example if the data spane from 2001 to 2010 then the number of years is 10, but the number of rows will be 10 only if all the years are present in the file. The number of rows excludes the lines containing the station or index names, the latitudes and longitudes and any other information lines. The tag is compulsory for the first block of data, but can be excluded if it remains unchanged for subsequent blocks.

  • cpt:row

    In gridded files, this tag must be set to "Y" or "lat", representing latitudes. In station files and index files, this tag must be set to "T", representing time. The tag is compulsory for the first block of data, and should not be changed for subsequent blocks.

  • cpt:S

    The date that the forecast data in the file were initialised (in ISO8601 format ). The cpt:S tag is only valid for GCM outputs, and should always have a corresponding cpt:T tag. It would normally be a good idea to set the cpt:S tag if the data are GCM outputs. If the cpt:S and cpt:T tags are set in the X file, CPT will assume that the Y data can be synchronous to the dates set by the cpt:T tags. If the cpt:S tag is not set, CPT will assume that the Y dates should follow the X dates set by the cpt:T tags in order to allow for a lag in the forecast. The tag is not used in station and index files. In gridded files the tag is optional; if it is present, it is assumed that the value is unchanged for subsequent blocks of data unless it is explicitly reset.

  • cpt:T

    The time period for which the data are valid (in ISO8601 format ). If the data are GCM predictions, then the tag represents the period for which the predictions apply rather than the date on which the forecast was initialised (see cpt:S tag). If the cpt:S and cpt:T tags are set in the X file, and the Y file contains monthly data, then CPT will automatically set the target season to that indicated by the cpt:T tags in the X file. Otherwise, CPT will assume that the X data represent a set of predictors that should precede the target season for the Y file. This assumption affects the default start at dates set by CPT when the files are opened. The tag is not used in station and gridded files; it is compulsory in gridded files for the first block of data, but can be excluded if it remains unchanged for subsequent blocks.

  • cpt:units

    The units in which the data are stored. This tag is optional, but if present, CPT will attempt to convert the units of the X data to those of the Y data when using the GCM option. CPT recognises the following tags and is able to convert between each:

    • Temperature
      • 'C', 'Celsius', 'Celsius_scale', 'degree_Celsius'
      • 'F', 'Fahrenheit', 'Fahrenheit_scale'
      • 'K', 'Kelvin', 'Kelvin_scale'
    • Precipitation
      • 'cm/day'
      • 'kg m-2 s-1'
      • 'm/s', 'meters/second', 'm s-1'
      • 'mm/month', 'mm/day'
      • 'mm'

The units for the Y data are used to identify whether the predictand is discrete or continuous. By default, the predictands are assumed to be continuous, but will be recognised as discrete if the Y units are set to any of:

  • 'count' or 'counts'
  • 'frequency' or 'frequencies'
  • 'number' or 'numbers'

The predictands are also assumed to be discrete if the units are 'N/A', 'undefined', or blank and if binomial or poisson regression are used. Given any other units, the predictands are assumed to be continuous regardless of the regression model. If the Y file contains multiple fields, only the first field is used to determine whether the data are discrete or continuous.

Last modified: