Instrumentation
The instruments that produced the measurements in this dataset, and the periods in which each was in place. Sensor identities come from three sources, in order of authority:
- the datalogger programs, which name the channel, its sensor and its conversion constants, and are the only source for how a raw signal became a value;
- the GIN fieldbook device columns (
Device Name,Device Model), which name the physical unit and its serial number at a location; - Yearly Notes, whose
Setup ECtable is the authority for the eddy covariance setup periods, sensor separations and flux-run configuration.
The GIN export records when an entry was written, not when a device was installed. Entries dated 1 January 2016 (and 2003, 2006) are backdated inventory imports made when GIN was introduced: a device whose earliest entry carries one of these dates was already at the site, and its installation date is not in the export. Only entries that describe an event (a replacement, a repair, a device marked defective) date anything.
Eddy covariance system, 47 m
Two gas analysers were operated at this level, and the fluxes are processed per analyser rather than per year. The fluxes in FP2026.1 come from the LI-7200 (IRGA72) only. The LI-7500 (IRGA75) fluxes are in progress for FP2026.2.
| instrument | role | period |
|---|---|---|
| Gill sonic anemometer | wind components, sonic temperature | 2004-present, orientation 209°; an HS-100 (serial 000027) until 19 December 2019 and an HS-50 (serial 000063) from then on, after the GIN fieldbook, which has no sonic record before 2016 |
LI-COR LI-7500 (open path, IRGA75) |
CO2 and H2O densities | 2004 until removal on 12 December 2017; again from 17 January to 20 March 2019 (setup period 2019_3), as a stand-in for the defective LI-7200 |
LI-COR LI-7200 (enclosed path, IRGA72) |
CO2 and H2O mole fractions | from 11 January 2016 |
The two analysers ran in parallel from January 2016. The LI-7500 was removed from the site on 12 December 2017 but continued to write empty values into the data stream until it was taken out of it on 31 January 2018, which is why the setup table lists it for two further periods with no flux run. IRGA72 spans two physical LI-7200 units, exchanged inside the 2019 gap; the second has a different column set in the raw stream. The analyser was moved on the boom in March 2023, which changed its separation from the sonic. The fieldbook also records a freshly calibrated LI-7200 installed on 2 December 2025. The dates per setup period are in the setup table of the Yearly Notes.
The fieldbook records the enclosed-path system at this level as an LI-7200RS head with an LI-7550 analyser unit and an LI-7200-101 flow module.
Meteorological sensors, 47 m
These produce the tower series documented under Meteorological data.
| variable | instrument | period | source |
|---|---|---|---|
TA, RH |
Rotronic MP101A, read as a single-ended analog voltage (multiplier 0.1, offset 0) | until 21 January 2016 | CR10X logger program |
TA, RH |
Campbell Scientific CS215, on SDI-12 (multiplier 1) | 21 January 2016 to 5 June 2026 | CR1000 logger program |
TA, RH |
Campbell Scientific HygroVUE10, on SDI-12, on the channel of the CS215 | from 5 June 2026 | fieldbook |
SW_IN, SW_OUT, LW_IN, LW_OUT |
Kipp & Zonen CNR1, SN 020484, sensitivity 10.03 µV W-1 m2 (multiplier 99.7009, offset 0) applied to all four channels | 14 September 2005 until December 2021 | CR10X and CR1000 logger programs |
SW_IN, SW_OUT, LW_IN, LW_OUT |
Kipp & Zonen CNR4, SN 212965, sensitivities 13.89 (SW in), 14.38 (SW out), 10.85 (LW in), 11.33 (LW out) µV W-1 m2, one per channel | installed 14 December 2021 | fieldbook |
PPFD_IN |
Kipp & Zonen PAR LITE, SN 050590, sensitivity 5.65 µV µmol-1 m2 s (primary, incoming) | installed 8 January 2016 | fieldbook |
PPFD_OUT |
Kipp & Zonen PAR LITE, SN 050603, sensitivity 5.26 µV µmol-1 m2 s (secondary, outgoing) | installed 8 January 2016 | fieldbook |
| PAR, total and diffuse | Delta-T BF2 sunshine sensor, SN BF2116 | entries 2016-2024, active | fieldbook |
| PAR | Kipp & Zonen PQS-1, SN 110145 | at this level from July 2011, inactive in every later entry | fieldbook |
| wind speed | Vector Instruments A100LK cup anemometer, SN 16697, replacing a broken A100L | installed 8 January 2016 | fieldbook |
| - | Campbell Scientific CR1000 datalogger | from January 2016 (CR10X before) | fieldbook, logger programs |
The 8 January 2016 date covers several of these rows because they were one visit: the boom was retracted that morning and the LI-7200 with its analyser and flow-module box, the replacement cup anemometer, both Kipp & Zonen PAR sensors and a new tower box were installed together. The LI-7200’s first raw binary is three days later, on 11 January 2016. A new reflected-PAR sensor had been fitted once before, at the end of June or beginning of July 2008; that entry is a remark added after the fact and dates the work only to within about a fortnight.
The temperature and humidity sensor is identified from the logger programs alone. The fieldbook records no temperature or humidity device at the 47 m level before June 2026; the first such entry is a Campbell HygroVUE10 added on 5 June 2026, which replaced the CS215 on the same channel (see Air temperature and Relative humidity). The MP101A and CS215 identifications, and the January 2016 date of the change, rest on the CR10X and CR1000 programs and on the step in the data. The step is quantified and corrected on the Air temperature page.
The radiometer change is dated by the fieldbook, the calibration by the logger programs. The CNR1 predates GIN and has no device entry of its own, but a 2008 entry recording a relay replacement for it on the tower’s CR10X meteo logger confirms that it was in service then. The CNR4 was installed on 14 December 2021. Neither instrument change moved the shortwave series measurably; the product notebook tests this (see Incoming shortwave radiation).
One constant for four detectors. A CNR1 head holds four sensors, each calibrated separately by the factory. The logger programs convert all four with a single factor, that of the pyranometer, so the longwave and outgoing channels were converted with a factor that is not their own. The CNR4 is converted with four separate sensitivities, so the two instruments cannot be compared channel by channel without allowing for that difference. Both outgoing products flag the instrument era for this reason.
Two dates are easily confused. The CNR1 instruction first appears in the logger program of 18 August 2005, which measures all four channels. The program of 14 September 2005 changed the relay, moving the outgoing channels onto different differential inputs; the archive’s record begins on the same day. The period column above gives the record, not the program.
The serial is not fully reconciled. The logger program comments and this table give the tower CNR1 as SN 020484. The fieldbook row recording the 14 December 2021 exchange names CNR1_020522, at a different location. This is most likely a bookkeeping artefact in GIN rather than a second instrument, but it has not been resolved.
The CNR4 was installed on 14 December 2021, and the installing entry notes that the logger script still needed to be adapted to the new calibration constants. The logger program was updated on 7 January 2022, adding the four CNR4 sensitivities and new raw-voltage variables.
The two dates are 24 days apart, and the conversion the logger applied in between is not recorded. If the CNR4 was read with the CNR1’s multiplier, the values of that period are wrong by the ratio of the two sensitivities.
Two products have tested the window. Together they bound the error but do not identify the constant. Outgoing shortwave radiation compares the window against the MeteoSwiss Lägern station and finds the levels not consistent with the old constant still being in force, which argues against the worst case. Outgoing longwave radiation bounds any error there at about 1.6 W m⁻², roughly 0.5 % of the reported value, because a conversion error scales only the small net signal a pyrgeometer measures and not the large Stefan-Boltzmann term added to it.
The window is flagged, not corrected, in all four radiation products. A correction, if one is derived, has to be applied to all four channels together, since they share the head and the constant. The bound above limits the error; it does not show that there is none.
Soil profile, forest floor (FF1)
The soil sensors were replaced as a set on 19 March 2020. This is the main date for the soil variables: a change of sensor generation with no overlap, so the step across it cannot be calibrated away.
| variable | instrument | depths | period |
|---|---|---|---|
| soil water content | Decagon ECH2O EC-20 | 0.05, 0.1, 0.2, 0.3 m | until 19 March 2020 |
| soil water content | METER TEROS 12 | 0.05, 0.1, 0.2, 0.3, 0.5 m | from 19 March 2020 |
| water potential | Decagon MPS-2 | 0.05, 0.1, 0.6 m | until 19 March 2020 for the shallow depths |
| water potential | METER TEROS 21 | 0.2, 0.3, 0.5 m | from 19 March 2020 |
| soil temperature | Campbell 107 thermistor | 0.05 m | until March 2021 |
| soil temperature | Campbell 109 thermistor | 0.05 m | from 24 March 2021 |
| soil heat flux | three Hukseflux HFP01 plates | 0.05 m | three until 24 March 2021, two after |
The soil heat flux row describes three sensors rather than one. The GIN device records list three HFP01 plates at this location. One of them was discarded at the logger-box rebuild of 24 March 2021 and the two survivors were kept, which is why the product exports three plate columns of which the third ends in 2021. The plates sit metres apart under a deciduous canopy, so they are not replicates: under leaf-off sunflecks one plate can read several times another, because the light reaches them differently.
The depth is the one declared in the archive, not a surveyed figure. The field names carry 0.05 m until the end of 2011, 0.025 m from 2012 to March 2021, and 0.05 m again afterwards. The product notebook tests from the size and timing of the daily cycle whether the plates were reburied shallower. The result matches neither a move nor an unchanged depth, so the question is open. See Soil heat flux.
The 0.5 m depth exists only in the TEROS 12 generation, which is why the soil-water-content product carries a homogenised second column at four depths and deliberately none at 0.5 m. A second profile, FF2, carries MPS-2 water potential sensors at 0.05, 0.1, 0.15, 0.2 and 0.3 m. Coverage, flags and the sensor-generation step are documented under Soil water content and Soil temperature.
The subcanopy station (M1_2) carries a CS215 temperature and humidity sensor, installed on 26 March 2021 to replace a Rotronic MP100, alongside an Apogee SQ-521 and a LI-COR LI-190SA quantum sensor, an Apogee SN-500-SS net radiometer and a W200P wind vane. Forest-floor acquisition is a CR1000 with an AM16/32B multiplexer until March 2021 and an ARK-1124C from October 2021.
Profile masts
A temperature, humidity and wind profile is operated on the tower and on satellite masts at the site (T1_17.5, T1_35, M2, M3, M4). Its temperature and humidity sensors are Rotronic units of several generations: MP101, MP400H and MP402H transmitters with HygroClip S2, S3 and HC2-S3 probes, with individual heads replaced repeatedly between 2015 and 2021. Its wind sensors are Gill WindSonics. None of these profile sensors contributes to this dataset, which takes its meteorology from the 47 m level and the forest floor. They are listed here only because published descriptions of the site sometimes name them.
Published descriptions that disagree
Three sensor identities given in the literature do not match the fieldbook or the logger programs. They are listed here for readers who compare this dataset with those papers.
Shekhar et al. (2024) give the site’s temperature and humidity sensor as a Rotronic HygroClip HC2-S3. The fieldbook does hold HygroClip HC2-S3 units at CH-LAE, but on the profile masts (M3, M4, T1_35), never at the 47 m level that supplies this dataset, and the logger programs give the MP101A and CS215 above.
The same study gives soil moisture as a Decagon ECH2O EC-5. The string EC-5 does not occur anywhere in the fieldbook export; the CH-LAE profile is EC-20 and then TEROS 12, as above. The study assigns EC-20 to the subalpine site it compares against, so the two sites’ sensors appear to have been transposed.
The same study assigns the Delta-T BF2 radiation sensor, serial BF2116, to the subalpine site. That serial is registered in the fieldbook at CH-LAE at the 47 m level.
These differences do not affect the values in either dataset, only the attribution of the instruments.