where P is the atmospheric pressure, R the constant for ideal gases, T the air
temperature, c
½ the concentration of CO 2 , h the height of the surface elements (e.g.,
trees) and c
½ t the CO 2 concentration at the top of the surface elements. In stable
night periods, the storage and advection terms, in the equations for budgets relative
to vertical flows scalar and vectorial quantities, are not negligible and eventually
cannot be measured with enough accuracy (Foken 2017).
Calculation of fluxes on calm nights, with turbulence attenuated in the surface
boundary layer, is done by applying the friction velocity rule, mentioned above,
with the exclusion of fluxes for the half-hour periods in which the friction velocity
is lower than the 0.2 ms
−1 threshold. These fluxes are replaced by gap filling, based
on values of fluxes corresponding to similar meteorological conditions. On calm
nights, with thermal stability, there is static accumulation of carbon dioxide (e.g.,
Eq. 3.226) which is removed when morning turbulence resumes. In this case, the
stored amount of CO 2 should be added to the morning vertical flux. If in this
situation the correction by the described friction velocity is applied, then the
night-time flux accumulation would be accounted twice. Thus, friction velocity
correction should be applied only if the stored CO 2 overnight is removed by lateral
advection, and thus not detected at sunrise due to the removal.
(xiv) A general process for calculating the vertical carbon fluxes in a typical
biosphere, such as a forested site, thus begins with calculating the mean
covariance over 30 min intervals., followed by rotation of coordinates so that
the component u coincides with the mean local velocity vector, thus nullifying the v and w components, and removal of linear trends. Next, the WPL
correction is applied to account for changes in air density and the Schotanus
correction for the influence of air humidity on sonic temperature and the
correction for overnight storage of CO 2 .
Data filtering should then be applied for vertical fluxes to eliminate those corresponding to vertical velocities with deviations from zero mean higher than 0.35 m;
the occurrence of high-frequency peaks in the data sets above 1%, occasional
occurrence of undulating/wave phenomena, using the absolute median deviation
around the median described above, and frictional velocities below a threshold of
0.2 ms
−1 . Data with these characteristics is of bad quality. The data selected by this
filtering is submitted to the stationarity test (item x) assigning ratings of 0, 1, and 2.
If the difference between the mean covariance for 30 min intervals as compared
with 5 min is less than 30% then a 0 rating is applied. In a similar way, if the
difference is between 30% and 50% and more than 50% ratings of 1 and 2,
respectively are applied.
Data submitted to the dynamic similarity test to the vertical wind component
(Table 3.1) are considered of good quality if the differences between the measured
and calculated values for the integral characteristics are not higher than 30%. The
final data classification will be 0, 1, or 2 if stationary test and dynamic similarity test
give a zero result, if one or both give a result of 1 and if at least one result is 2,
respectively.
100
3 Characterization of Turbulent Flow in the Surface Boundary Layer
temperature, c
½ the concentration of CO 2 , h the height of the surface elements (e.g.,
trees) and c
½ t the CO 2 concentration at the top of the surface elements. In stable
night periods, the storage and advection terms, in the equations for budgets relative
to vertical flows scalar and vectorial quantities, are not negligible and eventually
cannot be measured with enough accuracy (Foken 2017).
Calculation of fluxes on calm nights, with turbulence attenuated in the surface
boundary layer, is done by applying the friction velocity rule, mentioned above,
with the exclusion of fluxes for the half-hour periods in which the friction velocity
is lower than the 0.2 ms
−1 threshold. These fluxes are replaced by gap filling, based
on values of fluxes corresponding to similar meteorological conditions. On calm
nights, with thermal stability, there is static accumulation of carbon dioxide (e.g.,
Eq. 3.226) which is removed when morning turbulence resumes. In this case, the
stored amount of CO 2 should be added to the morning vertical flux. If in this
situation the correction by the described friction velocity is applied, then the
night-time flux accumulation would be accounted twice. Thus, friction velocity
correction should be applied only if the stored CO 2 overnight is removed by lateral
advection, and thus not detected at sunrise due to the removal.
(xiv) A general process for calculating the vertical carbon fluxes in a typical
biosphere, such as a forested site, thus begins with calculating the mean
covariance over 30 min intervals., followed by rotation of coordinates so that
the component u coincides with the mean local velocity vector, thus nullifying the v and w components, and removal of linear trends. Next, the WPL
correction is applied to account for changes in air density and the Schotanus
correction for the influence of air humidity on sonic temperature and the
correction for overnight storage of CO 2 .
Data filtering should then be applied for vertical fluxes to eliminate those corresponding to vertical velocities with deviations from zero mean higher than 0.35 m;
the occurrence of high-frequency peaks in the data sets above 1%, occasional
occurrence of undulating/wave phenomena, using the absolute median deviation
around the median described above, and frictional velocities below a threshold of
0.2 ms
−1 . Data with these characteristics is of bad quality. The data selected by this
filtering is submitted to the stationarity test (item x) assigning ratings of 0, 1, and 2.
If the difference between the mean covariance for 30 min intervals as compared
with 5 min is less than 30% then a 0 rating is applied. In a similar way, if the
difference is between 30% and 50% and more than 50% ratings of 1 and 2,
respectively are applied.
Data submitted to the dynamic similarity test to the vertical wind component
(Table 3.1) are considered of good quality if the differences between the measured
and calculated values for the integral characteristics are not higher than 30%. The
final data classification will be 0, 1, or 2 if stationary test and dynamic similarity test
give a zero result, if one or both give a result of 1 and if at least one result is 2,
respectively.
100
3 Characterization of Turbulent Flow in the Surface Boundary Layer