Recently (announced on scihub on 10.2.2021) there were some problems in processing Sentinel-1 imagery in the nominal processing times. On 17.2.2021 the problem was declared as solved and the production as normal. Now there are again some delays. I see the images of yesterday over Finland are all left on SLC level, no GRD products generated yet. At least the morning images are already past the Fast20h deadline. Has the ground segment lost its capacity to process ?
This doesn’t exclude that some member from the community is able to give a reply to the question. This forum is for the SNAP and related Toolboxes not the Sentinel data generation and distribution for which the official channel above exists.
@moderators - I suggest to move this to another category also.
Thank you for your comments, Christiano Lopez. The topic of my posting may be only loosely related to s1tbx, I admit. I just did not find any better forum to address the issue.
Than you also for your advice to inform EOSupport@Copernicus.esa.int about the delays. I sent a copy of my posting to them yesterday. Today I noticed that Sentinel-1 data on scihub has changed from Fast-24h to NRT-3h between 23.2.2021 and 24.2.2021, which in my opinion is a bad mistake. I informed EOSupport@Copernicus.esa.int also about this problem (and about 4 truncated .zip packets).
I have not received any reply from EOSupport@Copernicus.esa.int to my e-mail of yesterday. Does anybody know if the e-mail address is the correct one to report anomalies in scihub ?
My experience with EOSupport is that replies can take considerable time. This goes first to some desk in the UK. Usually, you get a first reply that is typically vague or deliberately obfuscated. E.g. something about a PDGS segment that went bla, bla. Something normal people do not understand. Only if you insist, they will really sort it out. I had an issue last year when a complete ascending orbit was missing. Took me more than 2 weeks to find out that it was some programming error.
Any updates on the ongoing orbits issue would be appreciated.
No orbit products have been released (including PREORB) for more than a couple of days now.
perhaps I am missing something… I can see ORBPRE for S1A and S1B on the auxiliary data server…
you managed to fully adapt my name… it has been a while since I got it like this (ROFL)
The EO Support will seek inputs from different teams before giving an answer, it can take a bit.
Thanks for pointing to this alternate resouce - didn’t know these were available here as well. Was used to getting them from the qc site. Thanks.
@cristianoLopes - looks like only ORBPRE (long term predicts) are still streaming in. The short term predicts that could be used for processing (PREORB) seems to have stopped around same time as RESORB and POEORB.
good point, I only checked ORBPRE - random click on the list.
This seems to be a tough problem to solve. The GRD products on scihub are still NRT-3h products, no Fast-24h products so far. As EO Support adviced about DIAS systems (to access .zip files that downloaded corrupt to my computer), I checked yesterday the Sentinel-1 availability on CREODIAS. As expected (this seems to be a problem of the ground segment processor not being able to access the right orbit data, not a problem of scihub or other data access platforms) the situation was the same: only NRT-3h products, no Fast-24h products.
Actually there was a huge fire in a datacentre in Europe that has affected many services, including those offered by Copernicus. Right now things seem to be still recovering and probably some data is being recovered in downstream services…
See news on the Open Access Hub!
The news on the Open Access Hub seem to say that everything is working as specified. The issue of missing Fast-24h GRD products has not even been mentioned there. The orbit data issue due to change of server was mentioned, but even that looks like working as specified if we only look at the news on the Open Access Hub.
The issues with the S1 orbits occurred well before the OHV fire, so that is a rather lame excuse.
taken out of context. I was replying generically to YRauste’s post about missing data. Which yes could be about the missing orbits (check Marcus’ post about in on the other topic) but most probably there is still a production or dissemination recovery problem that the fire didn’t help to speed-up.
Having no direct access to the mission’s production status my guess/explanation is a good as anyone else’s.
Again this is not an official channel of the Copernicus Sentinel missions production or dissemination systems, the information available here is offered at best guess. Official communication can be found in the appropriate channels:
It seems that the official channels of the Copernicus Sentinel missions production or dissemination systems have missed this topic completely. At least I could not spot any mentioning about missing Fast-24h GRD products in the sites (in the previous post) referred to by cristianoLopes.
Beginning of this month we contacted helpdesk to ask about the issue of missing Fast 24h. This has been forwarded to technical support but unfortunately nothing since then. Our operational S1-GRD to sigma0 backscatter workflow for the Belgian collaborative ground segment is still interrupted by the absence of Fast-24h products. Also very hard to find the concrete difference between NRT-3h and Fast-24h. This is not just timeliness it seems? We did a quick comparison between both for a February acquisition where both were still available (cfr image below), and they’re not the same, so we cannot just switch to the NRT-3h stream. Does anyone know more?
If I look at Earth Engine catalogue for example, products just keep coming in, but metadata says it’s NRT-3h. Are they just mixing different timeliness products depending on what’s available? In any case, the comparison between both seems to indicate we should not do that, as they’re not consistent.
Finally, we found out about this news item: https://cophub.copernicus.eu/news/News00548
Strangely, only posted on Copernicus Services Data Hub, while also the other hubs are affected without mentioning this. Turns out that from now on, only one product will be processed and distributed, either NRT 3h or Fast 24h, not both any longer. All processing pipelines explicitly requesting Fast 24h (like ours) have to be adapted. I’m still worried a bit about the differences between the two we’ve seen like shown in my previous post, but let’s consider this official announcement as the curtain behind which we can hide in the future …
Very very odd that no more communication was given with respect to this important policy change.
This was what I was afraid of. When the budget turned out to be insufficient for producing both a draft version (read NRT-3h) and a final version (Fast-24h), someone has decided that only the draft version is produced from 24.2.2021.