Modified

27 October 2026

Level-1: Final flux calculations

Level 1 is the final flux calculation with EddyPro, run with fluxrun, one run per setup period. Its FLUXNET output is the input of the post-processing chain. FP2026.1 contains it in full: 329 columns, listed in the section “L1 fluxes IRGA72” of the _INFO.md file.

The Level-1 fluxes are not quality-filtered, and FC carries no storage term. For analyses, use the columns of the post-processing chain (see QCF and Level 4.1).

Notebook: 99_IRGA72_L1_MERGED_2016-2026.09 merges the runs and checks the merged record.

NoteIRGA75 in progress

The Level-1 fluxes of the open-path LI-7500 (IRGA75) are in progress for FP2026.2 and are not part of FP2026.1. They are merged, before the self-heating correction of Level 1.1, in 99_IRGA75_L1_MERGED_noSHC_2004-2017_2019.

The IRGA72 record

The merged record runs from 11 January 2016 15:45 to 30 September 2026 23:45 (TIMESTAMP_MIDDLE), 187,937 half-hours. In FP2026.1 the columns are empty before 11 January 2016.

Table 1: The 18 Level-1 runs of IRGA72. A run covers one or more setup periods: 2017_1+2 is one run over setup periods 1 and 2 of 2017.
Run Period Note
2016_2 11 Jan - 26 May 2016 start-up of the LI-7200
2016_3 26 May - 31 Dec 2016
2017_1+2 2017
2018_1 1 - 31 Jan 2018
2018_2 31 Jan - 19 May 2018
2018_3 19 May - 7 Jun 2018 sonic only: H, no FC or LE (failed flow module)
2018_4+2019_1 7 Jun 2018 - 10 Jan 2019
2019_4 20 Mar - 31 Dec 2019 second LI-7200 unit from here on
2020 2020
2021_1 1 Jan - 13 Dec 2021
2022_2 7 Feb - 9 Mar 2022
2022_3 9 Mar - 18 Nov 2022
2022_4 18 Nov - 31 Dec 2022
2023_1 1 Jan - 13 Mar 2023
2023_2 13 Mar - 31 Dec 2023 analyser inlet repositioned
2024 2024
2025 2025
2026_1 1 Jan - 30 Sep 2026

Four setup periods have no IRGA72 run:

  • 2019_2 (11 to 17 January 2019): the LI-7200 was replaced on 11 January 2019 and both units were defective, so there are no usable gas fluxes. No sonic-only run for H has been made.
  • 2019_3 (17 January to 20 March 2019): the LI-7500 stood in for the LI-7200. Its fluxes belong to the IRGA75 record (FP2026.2).
  • 2021_2 and 2022_1 (13 December 2021 to 7 February 2022): the raw data cannot be converted (see Raw Data).

They form the two long gaps of the record: 10 January 15:15 to 20 March 18:45 2019 (69 days) and 13 December 2021 19:15 to 7 February 2022 10:45 (56 days).

Table 2: Half-hours with a Level-1 value of FC, LE and H per year, before quality control. A full year has 17,520 half-hours (17,568 in a leap year).
Year FC LE H
2016 16,378 16,558 16,876
2017 17,066 16,736 17,285
2018 15,899 15,846 17,194
2019 13,602 13,822 13,978
2020 16,125 16,232 17,433
2021 16,303 16,307 16,507
2022 15,225 15,226 15,572
2023 16,879 16,881 16,916
2024 16,943 16,944 16,993
2025 17,404 17,397 17,415
2026 (to Sep) 13,039 13,039 13,082

Settings and inputs

The EddyPro settings of all runs, and how they differ between runs, are on Level-1: EddyPro settings.

  • Time lags. The nominal lags and lag windows were chosen from the Level-0 runs; see the IRGA72 lag table on Level 0.
  • Biomet input. Six meteorological variables (TA, RH, PA, SW_IN, LW_IN, PPFD_IN) are passed to EddyPro, exported from the meteo products. The runs up to 2025 read the 2004-2025 export, the 2026_1 run the export by EDDYPRO_BIOMET_CH-LAE_2004-2026.09. EddyPro copies them into its output; the merge removes these copies. The meteorological columns of the dataset come from the meteo products.
  • Calibration gas. The CO2 concentrations of 14 December 2017 to 15 March 2019 were corrected before the runs (see Raw Data).

Merging the runs

Runs that meet at a year boundary overlap by 12 half-hours (31 December 19:15 to 1 January 00:45), and their values differ there, by up to 31.6 W m-2 in LE at the 2019/2020 boundary. Each record is taken whole from one run: before midnight from the earlier run, from midnight on from the later one. Columns without any value in the merged record are removed.

Back to top