###   NOTES with https://reentry.langbroek.org   ###

Accurate reentry forecasting is a challenge. Various factors influence the accuracy in
a negative way. There are often 'unknowns' for which guestimates have to be made, and
the actual state of the atmosphere varies over very short timescales and spatial scales
beyond the generalistation of the model atmosphere used.

Good values for mass and size for objects are often lacking: here guestimates have to
be made.

The drag area is often uncertain: apart from often uncertain object dimensions, the drag 
surface can be (significantly) less than the maximum surface, depending on object
attitude. The latter can actually vary over time, e.g. because of tumbling or because
the object self-orients along a preferred axis at some point. At low orbital altitudes, 
appendages can sometimes separate from the object, changing the drag characteristics.
Our model uses a default drag coefficient Cd of 2.2, while in reality, the true drag
coefficient of an object can be slightly different depending on shape, size and orbital
altitude.

Our Tudat model uses the NLRMSISE00 model atmosphere, with observed past, current, and
estimated future Space Weather. The latter (estimated future Space Weather) is an 
important factor introducing uncertainty. Space Weather is notoriously difficult to 
predict, and one solar outburst causing a geomagnetic storm can notably alter the
orbital evolution of an object. As the NRLMSISE00 atmosphere strictly speaking needs 
40 days of Space Weather *after* the date of interest in order to model the *current*
state of the atmosphere, there *always* will be model uncertainty when forecasting, even
over short timespans of a few hours. 

As a result, even when running the model with the very same drag area and mass parameters,
and the same starting orbit, but with an update of the Space Weather and future Space Weather
forecast inbetween the two runs *on the same day* as the previous model run (!), differences
of up to 10-15 minutes in the modelled reentry time can be introduced solely by changes in the
Space Weather forecast.

In our model, often a guestimate of the mass of the object has to be used. Even when
accurate dry masses are known, fuel remnants left in rocket stages, added masses from payload
adapters, or parts that have come off earlier can mean that the real mass is nevertheless
different.

We try to tune the model by adjusting the drag surface to the observed evolution of the
orbital altitude for the object over the previous few days to weeks. This drag surface estimate
is daily revised, to account for differences between forecast and real Space Weather, as well as
for true variations in drag surface due to attitude variations.

Some explanation of some of the notes with the tables:

* = latest forecast
X = erroneous forecast, ignore
R = reprise of previous forecast
A = aftercast (usually performed 1-2 days after the reentry to account for real Space Weather)

-------------------------------------
Dr Marco Langbroek  
Delft University of Technology
Faculty of Aerospace Engineering
-------------------------------------

document written 10 November 2025