Level-0: Preliminary flux calculations

Info

Results from preliminary flux calculations with relaxed processing settings assist in finding the most appropriate settings for final flux calculations (Level-1).

Level-0 flux results help to find the best settings for the final flux calculations. Generally, these preliminary results are used:

• to check whether the flux processing works,

• to check the wind direction across years,

• to determine appropriate time windows for lag search for final flux calculations.

More details can be found in the documentation of the Flux Processing Chain

OPENLAG runs to determine final lag ranges

Notebook

02_L0_timelags_check

The lag between turbulent departures of wind and the gas of interest was first determined in a relatively wide time window (called OPENLAG).

IRGA75 (2004-2017, 2019)

Table 1 IRGA nominal (default) time lags and size of the lag search windows for different time periods in seconds. Used for CO2 and H2O (LE) in final flux calculations.

time period	CO2 IRGA72	H2O IRGA72	Notes
2004	0.20, 0.05-0.75	same as CO2
2005_1+4	0.25, 0.05-0.75	same as CO2
2005_2	0.20, 0.05-0.75	same as CO2
2005_3	0.20, 0.05-0.75	same as CO2
2006	0.20, 0.05-0.75	same as CO2
2007	0.20, 0.05-0.75	same as CO2
2008_1+4	0.20, 0.05-0.75	same as CO2
2008_2+3	0.25, 0.05-0.75	0.30, 0.05-0.75	different lag for H2O
2009	0.25, 0.05-0.75	same as CO2
2010	0.25, 0.05-0.75	same as CO2
2011	0.30, 0.05-0.75	same as CO2
2012	0.25, 0.05-0.75	same as CO2
2013	0.25, 0.05-0.75	same as CO2
2014	0.25, 0.05-0.75	0.30, 0.05-0.75	different lag for H2O
2015	0.25, 0.05-0.75	same as CO2
2016_1	0.25, 0.05-0.75	same as CO2
2016_2+3	0.30, 0.05-0.75	0.25, 0.05-0.75	different lag for H2O
2017_1+2	0.30, 0.05-0.75	same as CO2
2019_3	12.20, 11.90-12.70	12.15, 11.90-12.70	artificial time lag added; different lag for H2O

IRGA72 (2016-2024)

• IRGA72 OPENLAG time window size for initial lag search was from -1s to +10s between 2016 and 2018, and from -1s to +20s from 2019 onwards. The later years required a wider time window because of the introduction of an artificial time lag of approx. 8s. The goal was to find an appropriate nominal time lag, and to determine a time window for lag search as narrow as possible.

• The value -1s was used as the starting value for the time window because EddyPro had issues with a start value of 0s, in that case the lag search started e.g. at +1s for some reason (maybe bug)

• The results from the OPENLAG runs were used to define a time window as narrow as possible for the final flux calculations (Level-1)

• The histograms of found OPENLAG time lags were inspected to determine whether there was a the histogram bin with peak distribution

• The peak of the distribution was used as nominal (default) lag, time ranges around this peak were used as time window (table)

Table 2 IRGA nominal (default) time lags and size of the lag search windows for different time periods in seconds. Used for CO2 and H2O (LE) in final flux calculations.

time period	CO2 IRGA72	H2O IRGA72	Notes
2016_2	0.35, 0.05-1.00	0.55, 0.05-4.00	often faulty data
2016_3	0.30, 0.05-1.00	0.40, 0.05-4.00
2017_1+2	0.30, 0.05-1.00	0.45, 0.05-4.00
2018_1	0.35, 0.05-1.00	0.45, 0.05-4.00
2018_2	0.35, 0.05-1.00	0.40, 0.05-4.00
2018_3	only sonic	only sonic	low flow rate, no IRGA fluxes
2018_4+2019_1	0.30, 0.05-1.00	0.45, 0.05-4.00
2019_2	n.a.	n.a.	IRGA data empty
2019_4	8.40, 7.90-8.80	8.45, 8.00-13.00
2020_1	8.40, 7.90-8.80	8.45, 7.75-13.05
2021_1	8.40, 7.90-8.80	8.55, 8.00-13.00
(2021_2)	n.a.	n.a.	unknown data format
(2022_1)	n.a.	n.a.	unknown data format
2022_2	1.25, 0.95-1.40	1.40, 1.00-4.00
2022_3	8.95, 8.45-9.55	9.00, 8.20-13.05
2022_4	8.45, 8.00-9.30	8.55, 8.00-13.00
2023_1	8.45, 7.90-8.80	8.50, 7.75-13.05
2023_2	8.45, 7.90-8.80	8.50, 7.75-13.05
2024_1	8.45, 7.90-8.80	8.55, 7.75-13.05

Check wind direction across years

Notebook

03_L0_winddir_check

IRGA75 (2004-2017, 2019)

I compared histograms of wind directions between 2004 and 2017 and incl. the short time period in 2019 using Level-0 fluxes and found that a sonic orientation of 209° offset to north yields very similar results across years. It is therefore possible the the sonic orientation across all years was always close to 209°.

Here are results from a comparison of annual wind direction histograms (with bin width of 1°) to a reference period (2016-2017), all wind directions were calculated with a north offset of 209°, then a histogram was calculated for each year. The OFFSET describes how many degrees have to be added (or subtracted) to the half-hourly wind direction to yield a histogram that is most similar to the reference. All OFFSETS are small, which indicates that the wind directions are in good agreement (table).

Table 3 Wind direction offsets for the IRGA75 years 2004-2017, 2019 (in degrees) compared to a reference period (2024) from Level-0 OPENLAG runs.

YEAR	OFFSET
2004	1
2005	1
2006	1
2007	1
2008	1
2009	-1
2010	2
2011	1
2012	2
2013	3
2014	1
2015	0
2016	0
2017	0
2019	1

IRGA72 (2016-2024)

I compared histograms of wind directions between 2016 and 2024 using Level-0 fluxes and found that a sonic orientation of 209° offset to north yields very similar results across years. It is therefore possible the the sonic orientation across all years was always close to 209°.

Here are results from a comparison of annual wind direction histograms (with bin width of 1°) to a reference period (2024), all wind directions were calculated with a north offset of 209°, then a histogram was calculated for each year. The OFFSET describes how many degrees have to be added (or subtracted) to the half-hourly wind direction to yield a histogram that is most similar to the reference. All OFFSETS are small, which indicates that the wind directions are in good agreement (table).

Table 4 Wind direction offsets for the IRGA72 years 2016-2024 (in degrees) compared to a reference period (2024) from Level-0 OPENLAG runs.

YEAR	OFFSET
2016	-2
2017	-3
2018	1
2019	0
2020	1
2021	1
2022	1
2023	2
2024	0