Level-1: EddyPro settings

The Level-1 fluxes of FP2026.1 come from 18 EddyPro runs, one per setup period of the enclosed-path LI-7200 (IRGA72). All belong to the Level-1 version HS50_IRGA72_FF-202610. This page lists the settings of these runs, taken from their EddyPro project (.eddypro) and metadata (.metadata) files and checked against the EddyPro output. The runs and the merged record are described on Level-1; the setup periods on Yearly Notes.

The processing settings are the same in all 18 runs. What differs between runs are the gas analyser, its position, the time-lag windows, the source of the air temperature and whether a run computes its own spectral assessment.

Settings common to all runs

Table 1: EddyPro settings of all IRGA72 runs of FP2026.1.

Setting	Value
Software	EddyPro 7.0.9, run through fluxrun
Raw data	20 Hz, files of 6 hours, file name gives the start time
Averaging period	30 minutes; a period with more than 10 % missing raw records is not calculated
Coordinate rotation	double rotation
Detrending	block averaging
Time lag	covariance maximisation within a window per gas; the nominal lag is used when the maximum lies at the edge of the window
CO₂ and H₂O	dry mole fractions (mixing ratios) from the LI-7200, in µmol mol^-1 and mmol mol^-1; therefore no density (WPL) correction
Angle-of-attack correction	not applied
Self-heating correction	not applied (it concerns the open-path LI-7500 only)
High-frequency spectral correction	Fratini et al. (2012), with an in-situ spectral assessment
Low-frequency spectral correction	Moncrieff et al. (2004)
Quality flags	steady-state and integral turbulence tests, 0/1/2 scheme of Mauder and Foken (2004)
Random uncertainty	Finkelstein and Sims (2001)
Footprint	Kljun et al. (2004)
Raw-data tests	the nine tests of Vickers and Mahrt (1997), all enabled; spikes and values outside the absolute limits are removed before the fluxes are calculated
Spike detection	5 standard deviations for `W`, 3.5 for `U`, CO₂ and H₂O; at most 3 consecutive values count as a spike; hard flag above 1 % spikes
Absolute limits	horizontal wind 30 m s^-1, vertical wind 5 m s^-1, sonic temperature -40 to 50 °C, CO₂ 200 to 900 µmol mol^-1, H₂O 0 to 40 mmol mol^-1
Site	47.478333 N, 8.364389 E, 689 m a.s.l.; canopy height 37 m, displacement height 24.79 m, roughness length 5.55 m
Sonic anemometer	Gill, at 47 m, north offset 209°; path lengths 11 cm (horizontal) and 12.5 cm (vertical)
LI-7200	intake tube 100 cm long, 5.3 mm inner diameter, flow 12 L min^-1; path lengths 1.27 cm (horizontal) and 12.7 cm (vertical)
Output	FLUXNET format, full output, biomet and metadata files

Settings that differ between runs

Table 2: Settings that change between the IRGA72 runs. Separations are from the sonic anemometer, in cm (N northward, E eastward, V vertical). Time-lag windows and nominal lags are in seconds.

Setup periods	Gas analyser	Separation	CO₂ lag window, nominal	H₂O lag window, nominal	Air temperature from
`2016_2` to `2018_1`	IRGA72-B; the LI-7500 is in the same raw files	N +6, E -1, V -15	0.05 to 1.00, 0.30 to 0.35	0.05 to 4.00, 0.40 to 0.55	LI-7500 temperature in the raw files
`2018_2`, `2018_4+2019_1`	IRGA72-B-GN1	N +6, E -1, V -15	0.05 to 1.00, 0.30 to 0.35	0.05 to 4.00, 0.40 to 0.45	biomet file
`2018_3`	none: sonic only, `H`, `TAU` and `USTAR`				biomet file
`2019_4` to `2021_1`	IRGA72-A	N +6, E -1, V -15	7.90 to 8.80, 8.40	about 7.75 to 13.05, 8.45 to 8.55	biomet file
`2022_2`	IRGA72-A	N +6, E -1, V -15	0.95 to 1.40, 1.25	1.00 to 4.00, 1.40	biomet file
`2022_3`	IRGA72-A	N +6, E -1, V -15	8.45 to 9.55, 8.95	8.20 to 13.05, 9.00	biomet file
`2022_4`	IRGA72-A	N +6, E -1, V -15	8.00 to 9.30, 8.45	8.00 to 13.00, 8.55	biomet file
`2023_1`	IRGA72-A	N +6, E -1, V -15	7.90 to 8.80, 8.45	7.75 to 13.05, 8.50	biomet file
`2023_2` to `2026_1`	IRGA72-A	N +11, E +1, V -1	7.90 to 8.80, 8.45	7.75 to 13.05, 8.50 to 8.55	biomet file

Time lags. The lags of about 8.4 s from 2019_4 on are not the travel time through the intake tube. According to the setup notes 12 and 13 (see Yearly Notes), they include an artificial delay added by the data acquisition. 2022_2 has no such delay, and its CO₂ lag of 1.25 s lies in the expected range. The lag shifts again in 2022_3 and 2022_4, which are therefore separate runs (note 7). Lags are not comparable across setup periods.
Analyser position. The inlet was moved in March 2023 (note 23), which changes the separation from 2023_2 on. Tube length and flow are unchanged.
Air temperature. Up to 2018_1 the air temperature comes from the LI-7500, which is in the same raw files. From 2018_2 on the LI-7500 is no longer in the raw files, and the air temperature comes from the biomet file.
Biomet file. The runs up to 2025_1 read EDDYPRO_BIOMET_CH-LAE_2004-2025.csv, the 2026_1 run reads EDDYPRO_BIOMET_CH-LAE_2004-2026.09.csv. The file holds TA, RH, SW_IN, LW_IN, PA and PPFD_IN at the 47 m tower level, with the timestamp at the end of the half-hour. Over 2004 to 2025 the two files differ only in model fills, reconstructions and one half-hour on 16 May 2025.

Spectral assessment

Each run computes the in-situ spectral assessment for the high-frequency correction from its own data, except four short setup periods, which use the assessment of a neighbouring period:

Table 3: Runs that use the spectral assessment of another run.

Run	Assessment taken from
`2016_2`	`2016_3`
`2018_1`	`2017_1+2`
`2022_2`	`2022_3`
`2022_4`	`2022_3`

Runs

Table 4: The IRGA72 runs of Level-1 version HS50_IRGA72_FF-202610.

Run	Level-1 ID
`2016_2`	`Level-1_FR-20261021-175449`
`2016_3`	`Level-1_FR-20261019-195550`
`2017_1+2`	`Level-1_FR-20261019-195156`
`2018_1`	`Level-1_FR-20261021-175903`
`2018_2`	`Level-1_FR-20261019-194911`
`2018_3`	`Level-1_FR-20261021-180146`
`2018_4+2019_1`	`Level-1_FR-20261019-193809`
`2019_4`	`Level-1_FR-20261019-193534`
`2020`	`Level-1_FR-20261019-193404`
`2021_1`	`Level-1_FR-20260902-174442`
`2022_2`	`Level-1_FR-20261021-180457`
`2022_3`	`Level-1_FR-20260902-173931`
`2022_4`	`Level-1_FR-20261021-180903`
`2023_1`	`Level-1_FR-20260902-173731`
`2023_2`	`Level-1_FR-20260902-173626`
`2024`	`Level-1_FR-20260902-173438`
`2025`	`Level-1_FR-20260902-173353`
`2026_1`	`Level-1_1_FR-20261023-141128`

Each run folder holds its project and metadata file, named after the setup period and the instruments, for example CH-LAE_2025_1_[IRGA72]_HS50-B_IRGA72-A.eddypro.

Known limitations

Sonic model. All runs declare a Gill HS-50. From 2016 to 19 December 2019 the sonic was a Gill HS-100 (fieldbook). EddyPro uses the model for the path lengths of the sonic, so the HS-50 entry is nominal for that period.
Gaps between runs. The setup periods 2019_2, 2019_3, 2021_2 and 2022_1 have no IRGA72 run; see Level-1.