We ingest daily images from Sentinel Hub. We have several AOI that were supposed to be ingesting new images yesterday, on the 26th.
At first we thought there was an issue in our system but after manually checking in Sentinel Hub online request builder, latest images for those AOIs are the ones from the 21st.
Is there any ingestion issue with latest images on Sentinel Hub?
we don’t see any systematic issue with ingestion. In last 24 hours there were 9437 S2 L1C, 9165 S2 L2A and 622 S1 GRD products ingested.
I suggest you check at the source (scihub.copernicus.eu) - if you find the data there but not in Sentinel Hub, do let us know. If not, I suggest you contact Copernicus Support (firstname.lastname@example.org) as they are responsible for distribution of data.
We already talked with Copernicus and they told us they had those images. We just realized that the images are in Copernicus and also in Sentinel Hub, but in a different product. We already left a message to them to understand why the images are in another product and not the one we’ve been using.
So far, all the images for those AOIs have been available when using L2A (S2A/S2B) product/data collection as shown in the image:
Sentinel-2 L1C represent “Top of the atmosphere” level products and these are typically distributed first. They are used to create “Bottom of the atmosphere” level product (Sentinel-2 L2A), which obviously come later (typically a few hours later).
L2A-level is typically better due to temporal consistency of the signal, but it depends on the use-casel.
In your case, the screenshots show that Sentinel-2 L2A product is not yet available by Copernicus for download. We don’t know the reason, why this is so, but I am sure they can provide you these information (we would be curious to get it as well).
To address your questions:
if you want to be sure you get all available data, you might indeed want to use both layers. But these kind of issues are generally rare, so it might not be wortwhile to introduce additional complexity.
We don’t know, why L2A product is not yet processed. Please, do ask Copernicus.
These issues are rare, so we imagine that you should be able to get L2A in the future. I am actually surprised that the error was not yet fixed by Copernicus/ESA.
We are already having trouble with the new images from today. The 26th were missing in L2A, today images, some are missing (at least they are not in Sentinel Hub yet) and even worse, others are cut.
This never happened before either!
Were you able to find something out from your side? We understand that it’s probably Copernicus issue, but since we are subscribed to Sentinel Hub, any help or understanding of the problem would be really helpful to us. Plus, we assume that Sentinel is better connected or knows who to talk to from Copernicus side.
We also understand that images were just out, and maybe it’s a matter of time (not sure though) but this is the first time that images ingested into our system are cut like this, so we are not sure what’s happening and we are concerned with the functioning of our systems.
As an example, from one of our AOIs, an image from S2 L2A from the 21/01/22:
This “cut” is normal - it means that some scenes are not yet processed, available.
For the today’s images - expecting to get majority of L2A within less than 6 hours after observation is too optimistic. They start coming around this time, but then it takes for some an hour or two or three longer.
For the previous case we checked and there is no issue on Sentinel Hub’s side.
Note that we don’t have any control in obsevation or processing scenarios, so that is something we cannot improve - we can just provide our general experiences. For individual requests you will have to contact the source, if you want to get information. Sentinel Hub is here to make the access to the data, which are made available, as easy as possible and as fast as possible. And we are putting a lot of effort into this.
Yes, we understand that and we really appreciate all the effort!
Please, let us understand one more thing, we ingest into our systems data from Sentinel Hub. So far this “cut” thing never happened to us. If this is something that can happen, what is the recommended way to handle this?
Because our system already ingested that image, it was already available on Sentinel Hub.
If this means that some scenes are not yet processed, is there a way we can know that? How can we understand when an image is not yet complete and we need to wait more time to ingest it?
So far if it was available we just ingested it directly.
If you are trying to fetch the data very soon after observation, this will almost certainly happen again.
The easiest option for you is probably to use Statistical API request over your area of interest, fetching only one band, just make sure you use data mask properly.
Output of this API will be a JSON containing also information on “noDataCount” - this will give you information on how much of your area is covered with the data at specific point in time, then you can make a decision, whether to use it or not.