Changes between Version 9 and Version 10 of udg/ecoms/dataserver/datasets/CFSv2


Ignore:
Timestamp:
May 16, 2015 12:34:03 PM (7 years ago)
Author:
juaco
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • udg/ecoms/dataserver/datasets/CFSv2

    v9 v10  
     1= CFSv2_seasonal dataset member setup
     2
     3
    14The ensembles members are obtained by lagged `runtimes` depending on release date. There are 12 release dates, corresponding to the 15th of each month. The members for the mid-january 1982 are the following `runtime`:
    25||= Year =||= Month =||= Day =||=  Hour                 =||
     
    811||  1982  ||   01    ||    6  || at 00, 06, 12 and 18 UTC|
    912
    10 Thus the `runtimes` for the 12 release dates, on each year, are indicated in the tables below (the default 16 initializations defining members are indicated in boldface, as taken by the [https://meteo.unican.es/trac/wiki/udg/ecoms R package] for data access. Note that due to some missing initializations in the original dataset, the configuration of the default members taken through the R interface is not directly the 16 first initializations of the corresponding lead month, but the 16 first '''complete''' initializations (i.e., available for ''all'' years in hindcast).
     13Thus the `runtimes` for the 12 release dates, on each year, are indicated in the tables below (the default 15 initializations defining members are indicated in boldface, as taken by the [https://meteo.unican.es/trac/wiki/udg/ecoms R package] for data access. Note that due to some missing initializations in the original dataset, the configuration of the default members taken through the R interface is not directly the 16 first initializations of the corresponding lead month, but the 16 first '''complete''' initializations (i.e., available for ''all'' years in hindcast).
    1114
    1215In particular, the missing initialization times (at least for precipitation) are indicated with asterisks in the table (`*` for 1989 or `**` for 1998), and the specific location of the missing runtimes indicated in the rightmost column. As a result, the total number of selectable members varies depending of the lead month chosen (this is indicated by ''n'' in the tables below). These are selected in R through the `members` argument of the function `loadECOMS` that can take any integer value from 1 to ''n'' for each particular lead month, corresponding this value to their relative order in the tables below (and excluding the members marked with asterisks).