OCHA-DAP/ocha-anticipy

NDVI data removed from USGS portal

Closed this issue · 3 comments

In an exciting development, NDVI data has been entirely removed from the USGS portal due to significant issues with the MODIS sensor.

Important Notice: Due to severe degradation of data quality being provided by the AQUA satellite, which carries the MODIS sensor, we have suspended access to eMODIS NDVI products until further notice. In addition, our evapotranspiration (ETa) modeling relies on MODIS derived land surface temperature products which have been degrading in quality since May of 2022. Therefore, we have chosen to suspend access to all ETa products that include data from May forward. We apologize for the inconvenience.

We will somehow need to quarantine our module lol? Maybe temporarily adding a clear warning (whatever the highest level of logging is) when the class is instantiated about the quality of data from May 2022 (if the dates of the class extend beyond that) and that no new downloading is available, an error message repeating the same thing for download, but leaving the class up for processing and loading previously downloaded data.

Uff that really sucks, worst timing ever! I just have a few questions:

  1. It sounds like this an issue with MODIS. Are there other satellites that provide NDVI to the same quality, or does this mean it's not available from any alternative source?
  2. If there are non-MODIS alternatives, do you know what they might be? Is there anything viable (i.e. public with an API) to include in a future release?
  3. Is the past MODIS NDVI data still available from USGS?

If the answer to question 3 is "yes", then do you think it's worth keeping NDVI in (with all the very clear disclaimers that you mention), i.e. for performing historical analysis?

In response.

  1. Definitely a problem with MODIS. There are other sources available!
  2. The ones I'm aware of are the Copernicus products. Not sure what the difference between them are, but one is from Sentinel (not much historical data) and another from Proba. In fact, looks like there
  3. So, I checked, and although they removed public access to download it direct from their portal, we still have historic data.

I think we definitely keep it, and I will add in a warning for users downloading after the degradation and noting data isn't available after that certain date. I will also create an issue for adding potential new data sources.

Issue with module addressed in #154, closing and will address additional sources in #153.