I am currently using the Sentinel-1 GRD preprocessing workflow (https://sciforum.net/manuscripts/6201/manuscript.pdf). On this forum I read that number of steps included in the xml file can influence the processing speed. And that the order of nodes in the XML file does not influence the order of steps in SNAP.
So, my question is if someone has experience with applying different preprocessing steps separately. Did this increase the speed? Is it possible to simply do the preprocessing step for step? How much does the order of the different steps matter? I assume some steps need to happen at the beginning like orbit correction and that resampling is done at the end. Would like to learn more!
It is always possible to do everything step-by-step but that requires lots of IO as the intermediate products are written and read to/from disk. Try splitting the processing in two, ending at calibration, and then terrain correcting the calibrated intermediate products.
Thank you for your reply! Makes sense to start with trying this and then have a look at the difference in computation time. So, would be best to leave the order of the processing steps as it is right?