Skip to contents

Forecast equivalents of the catalog variables

Source: R/variables.R
forecast_variables.Rd

Where the reanalysis has a hindcast, the analysis-and-forecast products have a near-real-time equivalent running to about ten days ahead. This maps each catalog name onto the forecast product, dataset, and code that supply it.

Usage

forecast_variables()

Value

a named list keyed by catalog name, each with variable, product_id, dataset_id, and daily_dataset_id

Details

Two things make this a real mapping rather than a substitution of one identifier for another:

  • The forecast products split variables across datasets. The reanalysis serves all its physics from one dataset; the forecast has separate ones for temperature, salinity, and currents. So a forecast fetch of SST and SSS is two requests where the reanalysis is one.

  • Some codes differ. Bottom temperature is bottomT in the reanalysis and tob in the forecast. Reusing the reanalysis code would produce a failed download.

Satellite variables have no forecast: ocean colour is observation, and an observation of the future does not exist. CHL, PP, DIATO, and DINO are therefore absent here, and asking for them in forecast mode says so.

See also

accessEnvDat(), which takes mode = "forecast"

Examples

names(forecast_variables())
#>  [1] "SST"       "SSS"       "UO"        "VO"        "SSH"       "MLD"      
#>  [7] "SIC"       "BOTT"      "NO3"       "PO4"       "O2"        "NPP_MODEL"
#> [13] "CHL_MODEL" "PH"       
forecast_variables()$BOTT$variable   # "tob", not "bottomT"
#> [1] "tob"