Bug-fix CFU_load + Option to choose the number of members for the output matrix in CFU_load + Option to use only forecast with all leadtimes matching the observational period in CFU_RMS (V. Guemas)

renaming ECMWF_S4Seas* to ECMWF_S4_sea and ECMWF_S4_ann for the seasonal and annual outlook forecasts respectively

It now uses the follwing arguments (see help in header):
CFU_DTO(lon,lat,ano,yr1,yr2,yrint)

new function CUF_plot2varsvsltime, also ensuring longitude bounds are positive and CFU_load input-file frequency can be 3hourly or 6hourly

- mn and mn-hsm are not the options for data (model output generated by using Autosubmit at HPC) transferring...

- mn-dt* introduced for data transfer; where mn-dt-ecm86 and mn-dt-pr1e4400 would be dealt as ssh configurations...

Loading of monthly statistics (quantiles ...) + Loading daily means + subsampling of the loaded data at a given frequency

CFU_ratioSDRMS with confidence interval from nsdate x nmembers + unit bug fixed in CFU_plotequimap + Print changed in CFU_load

CFU_AMO() computes the Atlantic Multidecadal Oscillation (AMO) index based on the
area-averaged North Atlantic (0-60N, 80W-0) SST anomalies minus
the global (60S-60N) SST anomalies.
See the extensive description in the header of the function.

path of masks corrected + spread confidence interval over members and start dates + two leadtimes arguments

These functions detrend observed (CFU_DTO) and forecast (CFU_FTO) anomalies of, e.g., North Pacific (PDO) SSTs,
then compute the observed EOF mode (CFU_EOF) and project the anomalies onto the observed EOF pattern CFU_MODE().

Find an example below.

Detrend obs anomalies:
tmp2 <- CFU_DTO(lons,lats,XY1_ano_obs)

EOF of detrended obs anomalies::
tmp3 <- CFU_EOF(lons,lats,tmp2$ano.dt)

Detrend forecast anomalies:
tmp4 <- CFU_DTF(lons,lats,XY1_ano_exp,exp_gmt_an)

Obs PCs:
tmp11 <- CFU_MODE(lons,lats,XY1_ano_obs,tmp3,mode=1)
Forecast PCs:
tmp12 <- CFU_MODE(lons,lats,XY1_ano_exp,tmp3,mode=1)

New options : ORCA025t_v3.2 and EcEarth3; All the corresponding weigths have been computed and stored at /cfu/pub/scripts/interpolation/weigths

CFU_plotmapmostliktercile() produces a probabilistic forecast map of most likely tercile of a forecast
using the probabilistic bins computed in parallel by CFU_PB().
Similar to http://eurobrisa.cptec.inpe.br.

The data_transfer.sh is serving the purpose for transferring model output from all available HPCs...