Are Landsat satellites exactly sun-synchronous?

Question

Sun-synchronous satellites are known to pass over any given point of the planet's surface at the same local mean solar time (link-to-wiki). Landsat is one of these sun-synchronous satellites.

I'm currently considering a dataset which contains a list with image-acquisition timestamps of Landsat imagery. However, I note that there is a some variance (within the order of minutes) in the moments of when the image was taken.

This is a sample of that list:

[datetime.datetime(1999, 5, 2, 1, 33, 58, tzinfo=<UTC>),
 datetime.datetime(1999, 5, 18, 1, 33, 42, tzinfo=<UTC>),
 datetime.datetime(1999, 6, 3, 1, 33, 22, tzinfo=<UTC>),
 datetime.datetime(1999, 6, 19, 1, 33, 13, tzinfo=<UTC>),
 datetime.datetime(1999, 7, 5, 1, 33, 5, tzinfo=<UTC>),
 datetime.datetime(1999, 7, 6, 1, 42, 16, tzinfo=<UTC>),
 datetime.datetime(1999, 7, 13, 1, 48, 5, tzinfo=<UTC>),
 datetime.datetime(1999, 7, 14, 1, 27, 13, tzinfo=<UTC>),
 datetime.datetime(1999, 7, 21, 1, 33, 1, tzinfo=<UTC>),
 datetime.datetime(1999, 7, 22, 1, 42, 17, tzinfo=<UTC>)

What causes this variance?

Note, shouldn't be relevant, but the images are obtained through Google Earth Engine.

Answer 1

Landsat mission statement says: "Traveling on the descending (daytime) node from north to south, the satellites cross the equator on each pass at a time that provides the maximum illumination with minimum water vapor (haze and cloud build-up). Landsat 7 crosses the equator at 10:00 a.m. +/- 15 minutes (mean local time) while Landsat 8 crosses the equator at 10:00 a.m. +/- 15 minutes (mean local time)."

The "radiation fluctuations" are part of the orbital movement of the planet/ space vessel and sketched by the Wikipedia articles for Sun-Synchronous Orbit and the different "Solar Time" termini (Mean Solar Time = UTC/ Apparent Solar Time) and the corresponding differences between the both solar times (diagram fluctuations of the sundial).

Cite: "When one says that a Sun-synchronous orbit goes over a spot on the Earth at the same local time each time, this refers to mean solar time, not to apparent solar time. The Sun will not be in exactly the same position in the sky during the course of the year."

On the other hand are the Landsat missions scanner based, with a down-link procedure for the continuous sampled data lines. These data are post-processed in different so called "Landsat Level Tier Products" (Level 1, 2 etc. and corresponding images). The procedure and dependencies of the Level Tier Products are documented and have different optimization and aggregation quality criteria. I guess based on the process level/dependencies, the provided data timestamp is in some kind aggregated.