Important usage notes
The flux files provided through this webpage contain raw inversion output.
Only certain aspects of these fluxes are sufficiently robust for interpretation purposes.
Therefore,
-
only fluxes within the period of validity should be used;
-
fluxes should be filtered to remove
variability on small spatial and temporal scales,
or other unrobust features;
-
the long-term spatial flux pattern should be looked at with great care.
Period of validity
The flux files contain the full period of the inversion calculation.
This includes the spin-up and spin-down periods,
which are influenced by edge effects.
Further, the records of some of the atmospheric sites may not span the entire inversion period.
The period of validity selects the years considered homogeneous in time.
Begin and end years (inclusive) of the period of validity are given in the download table
and in the files themselves.
Only those years should be used for any interpretative purpose.
Remark 1:
Temporal filtering should be applied to a time period that also includes
one year before and after the period of validity,
in order to reduce edge effects from the filter.
Remark 2:
The full period is relevant if the atmospheric response (change in atmospheric APO concentration
relative to the initial condition) is considered (e.g., for a forward run of an atmospheric transport model).
Filter
The numerical flux resolution provided in the files is higher than the actual resolution
in terms of information content. The latter is related to the spatial distance between the atmospheric sites,
and the temporal density of sampling (including the effect of gaps).
The appropriate type of filtering (aggregation, smoothing, compositing) depends on the application.
You are welcome to discuss this with us.