Hola @vagudets !
"Forecast downscaling" and "downscaling through the large scale variables in analogs" features are ready in the dev-Downscaling branch. Based on the updates, if forecast data and hindcast data are provided together, the functions of the module will prefer to downscale the forecast data. For the analogs, the priority is as follows:
+if only 'data$hcst' is provided for the model part, hindcast downscaling will be applied using the same local scale variable as the predictor.
+if 'datahcst' and 'datahcstL' are provided for the model part, hindcast downscaling will be applied using the large scale variable as the predictor.
+if 'datahcst', 'datahcstL' and 'data$fcst' are provided for the model part, forecast downscaling will be applied using the same local scale variable as the predictor.
+if 'datahcst', 'datahcstL', 'datafcst', and 'datafcstL' are provided for the model part, forecast
downscaling will be applied using the large scale variable as the predictor.
Please also be aware of that data$obsL must be provided so that the function can apply downscaling using the large scale variable.
Another update, 4nn method in Intlr option has been changed to the 9nn method with principal component pre-filtering.
Salutes!
Eren
fyi @nperez