Help exporting the "SM" imagecollection.

[jvaldiviesob]

Hi, I hope you're ok. I have a question. When we export the "SM" Imagecollection (line 517), Goggle Earth Engine gets stuck, whether it is exported as Asset or to Google Drive. Has it happened to you? Have you been able to solve it? Thanks in advance.

[QianqianHan96]

Hi, I did not have this problem. I tried to run the script again, it still works. Did you run the original script or you changed to a bigger area (https://github.com/QianqianHan96/GSSM1km/blob/main/GSSM1km%20generation)?

[ccalvocm]

Greetings @QianqianHan96:

I guess @jvaldiviesob is referring to the fact that when computing anything with variable SM, it runs out of memory. This might be due to the APILand1 ImageCollection, which makes any computation and export run out of memory. I've tried to upscale APILand, reducing it, but no matter what it always makes the classification run out of memory. The fact is that you rule out that predictor, makes the code run and export, but that's undesired. Therefore, that dataset should be checked.

[QianqianHan96]

Hi Carlos,

I just checked. APILand is okay. I can display it as well. I did not rule out this predictor, it was used in regression.
Indeed if we want to print SM, seems out of memory because the computation is too much. It is not recommended to use print during big computation on GEE. I analyze the exported SM in a new script link. If you can manage to export SM to Asset or Drive, then it is okay.