Data quality of fapar and LAI

Dear earth system data lab team,
I was wondering, if there is any information on data quality and potential biases regarding fapar and LAI.
When plotting time series of these variables for the Mediterranean, the years 2015 and 2016 have way higher values than previous years, it seems like offset and also variance are changed.
I attached a file, with aggregated time series of the Mediterranean (and also an example plot for a single pixel) and for comparative reasons how the data look worldwide and in the USA during that time (no obvious bias there).
Do you know any explanation for this?
Best regards,
Joscha Vogel


I noticed this issue for the FAPAR as well and I am as of yet not really sure, how exactly it is/was produced, but a growing saturation of FAPAR can be observed in the original FAPAR-TIP data (

I am particularly confused by the apperently higher relative uncertainty in the data of 2002, as compared to the data of 2016. And even though more pixels of the US and Europe are flagged as dubious in 2016 than in 2002, they don’t account for the even larger saturated areas.

Indeed, this looks worrisome. Can you please contact the data PIs of these data set, send them the images and ask for what is going on there and report this back? This would be very helpful - I could also do it, but you have more hands on experience now. Please CC me…Many thanks!

I contacted Dr. Michel van Roozendael about the issue.

1 Like