I want to downscale time-series imagery data using precipitation and evapotranspiration (temporal) and possibly even topography (static) using Google Earth Engine (GEE) Random Forest Regression. I have processed the remote sensing products to the same temporal and spatial resolution and joined them.
Typically the code would be something like:
// Create a classifier
var classifier = ee.Classifier.smileRandomForest().setOutputModel('REGRESSION').train({
features:training_data,
classProperty: 'what_I_want_to_predict',
inputProperties: ['predictor_variables']
});
var classified = predictor_variables_data.classify(classifier);
My question is:
1. How do I include temporal and static data as predictor variables (training data)? How do I sample it?
2. How does one apply the RDR model across monthly images over 2 years for example? Do you run the model 24 respective times?
Regards,
Cindy