Binning and Time information

Hi all

In the binning code using TIME_RANGE, where does the code take the initial date from? Do you think it would work if I try to bin mosaic level-2 images derived from another processor, that is not C2RCC or Level-2 standard from ESA?
I usually use the filter method “NONE” and it is fine, but for a 3-years dataset, I would have to create several directories for this to make 8-days composites and I would like to use the TIME_RANGE. I have tried before, but no success.
To be more clear, I am using images derived from Polymer processor.
If anyone could give me a tip on how to deal with that, I would appreciate.
Thank you a lot

Sorry, I forgot to mention that I am talking about Sentinel-3 images. =)

I thought it is worth its own thread.

In SNAP each product has a start and stop time. You can inspect this by bringing up the information window (Analysis / Information).
If this information is set depends on the input format. For most it is set. And for other it depends if it is provided in the data.
I think Polymere generates NetCDF files, right?
In principle the information can be stored in those files but I don’t know if Polymere is really doing it.
So, you can the information window. You can also check the metadata if there is any time information.
But if it is in the metadata but not in the information window it is not supported by our reader. In the case we would be interested in this use case.

Thank you Marco
Yes, I checked and Polymer Level-2 has this information. The problem is that I am working with a large region, where I mosaic 3 to 4 granules for 1 day. And the mosaiced images do not have this information, it is not passed to the mosaic output.
Do you think the solution for that would be using the Filter Method “NONE” and separate the database into 8-days folders for example?
I don’t see a way of running the binning first and then mosaic at the end, it won’t work for such a large area.
Any idea?
Thank you!

Ah, okay. Yes, separating the files into directories would ease your work.
But why do you mosaic the data first and the perform binning. Wouldn’t it be easier to just use binning?
Then you could use the time information too.

The mosaicing does not preserve the time information. This is also not easily doable for the general case. But we have thought already to improve this situation. But not implemented yet.

If you created already the mosaics and stored them as BEAM-DIMAP you could modify the dim file and add the time information. But I think you would need to modify too many files.

1 Like

Thank you
Yes, I totally get you. I was doing Binning directly, but our research group is interested in different products. For example, I am interested in bi-weekly averages for this whole area, but others want to have the daily products for other applications.
But I will definitely think about it, if it is worth to just have this two different flows.
I will talk to the computing people and see what is best in terms of programming.
Again, thank you very much

Hi Marco,
I am trying to run the binning over the level-2 products, as we discussed before. However, I am having issues to produce different outputs for each time-range defined. For example, if I loop through 2 months of data (ex: 2016/05 and 2016/06), and I want to produce 3-day averages starting on 2016-05-01, how should I set up my code? This is how I am invoking the xml:
gpt C:\Binning_scripts\binning_c2rcc.xml -Pstart=‘2016-05-01’
And the .xml is attached.
binning_c2rcc.xml (2.4 KB)

Thank you in advance,

The binning operation can only do one operation at once. So you have to call it multiple times. For each start date separately.

How you can do bulk processing is explained here.