We’ve applying naively PCA for our variables in order to reduce dimentionality. We are doing it for each (lon,lat) to reduce variables but to mantain times.
How would you say is the best way of doing it, computationaly and in regards to using the cube efficiently? is there a way of parallelizing?
Also we have questions regarding subtracting seasonality. We are working with the Colombian Cube so we would like to know the best way of doing it since we don’t have proper seasons in contrast to more northern or southern latitudes.
Thanks in advance!
Max and Daniela