Modified

27 October 2026

Vapour pressure deficit at 47 m

Half-hourly vapour pressure deficit at 47 m on the CH-LAE tower, 2004 to September 2026, complete. The file is 07_METEO_VPD_2004-2026.09 (parquet and CSV): 398,832 records on a continuous 30-minute index TIMESTAMP_MIDDLE (middle of the half-hour, UTC+01:00, no daylight saving), from 2004-01-01 00:15 to 2026-09-30 23:45. 2026 is a partial year (January to September).

VPD is not measured. It is computed by formula from air temperature and relative humidity, so it inherits the properties of those two products.

The interactive dashboard summarises the product on one page.

Method, evidence and checks:

ImportantWhich column to use

TA and RH at 47 m come from one probe, which changed on 21 January 2016 (Rotronic MP101A to Campbell CS215) and on 5 June 2026 (to a Campbell HygroVUE10). Both inputs change level at these dates, so VPD does as well.

  • VPD_T1_47_1 is computed from the measured TA and RH. It is the VPD the probe reported and matches the measured TA and RH used as biomet input to EddyPro and as drivers in the flux processing chain. Use it together with those columns, and inside one probe state.
  • VPD_T1_47_1_HOMOGENIZED is computed from the homogenised TA and RH, which put every probe state on the level of the normal CS215. Use it for anything that crosses January 2016 or June 2026 (period means, trends, year rankings, anomalies).

Changed from the 2004-2025 version of this product. There, VPD_T1_47_1 was computed from the homogenised TA and the measured RH, and no homogenised VPD existed. That mixed basis is no longer exported. The column name VPD_T1_47_1 now stands for the measured pair.

Columns

Both value columns are complete, and both flags are defined at every record and describe both value columns.

Table 1: Columns of 07_METEO_VPD_2004-2026.09.
column unit description
VPD_T1_47_1 kPa Vapour pressure deficit from the measured TA and RH. Not homogeneous across 2016-01-21 or 2026-06-05.
FLAG_VPD_T1_47_1_ISFILLED - What the record was computed from.
VPD_T1_47_1_HOMOGENIZED kPa Vapour pressure deficit from the homogenised TA and RH.
FLAG_VPD_T1_47_1_SOURCE - Which probe, and which state of it, stands behind the record.

VPD_T1_47_1 ranges from 0 to 4.104 kPa (median 0.176 kPa), VPD_T1_47_1_HOMOGENIZED from 0 to 4.183 kPa (median 0.157 kPa). VPD is zero at saturation and cannot be negative.

Table 2: The input pairs.
value column air temperature relative humidity
VPD_T1_47_1 TA_T1_47_1_gfXG (notebook 02) RH_T1_47_1 (notebook 04)
VPD_T1_47_1_HOMOGENIZED TA_T1_47_1_HOMOGENIZED_gfXG (notebook 02) RH_T1_47_1_HOMOGENIZED (notebook 04)

VPD is the saturation vapour pressure at the air temperature minus the actual vapour pressure. The Magnus coefficients are those used by ReddyProc. In the CS215 states (source codes 0, 5, 6, 7) the homogenised RH equals the measured one and the homogenised TA is not warmer than the measured one, so the homogenised VPD is never higher than the measured VPD there. On average the homogenised column is lower by 0.006 kPa in the normal CS215 state, by 0.014 kPa in the MP101A era and by 0.071 kPa in the HygroVUE10 era.

FLAG_VPD_T1_47_1_ISFILLED

Filter on == 0 for records computed from two measurements at 47 m. Every record is computed by the formula and none is modelled, but a computed record does not always rest on two measurements.

Table 3: Provenance codes and their record counts.
code meaning records share
0 measured TA and measured RH 382,158 95.82 %
5 measured RH, gap-filled TA 23 0.01 %
6 RH reconstructed from NABEL at 49 m 15,729 3.94 %
7 RH reconstructed from MeteoSwiss Lägern 922 0.23 %

Codes 6 and 7 take precedence over 5 because a reconstructed RH is the larger error: it carries about 3 % RH of uncertainty and enters the formula linearly.

FLAG_VPD_T1_47_1_SOURCE

Carried over unchanged from notebooks 02 and 04, whose source flags are identical record for record. The codes are described on Air temperature at 47 m.

Table 4: Probe and probe-state codes, their record counts and the period each covers (TIMESTAMP_MIDDLE).
code probe and state period records
0 Campbell CS215 on SDI-12, normal state 2016-01-21 14:15 to 2024-10-21 23:45 153,428
1 Rotronic MP101A, single-ended analog 2004-09-20 10:45 to 2015-12-31 23:45 197,739
2 acquisition changeover, probe undetermined 2016-01-01 00:15 to 2016-01-21 13:45 988
3 before the tower record begins 2004-01-01 00:15 to 2004-09-20 10:15 12,645
4 Campbell HygroVUE10 on SDI-12 2026-06-05 11:45 to 2026-09-30 23:45 5,641
5 probe handling or probe swap half-hour 2025-05-16 09:45 and 2026-06-05 11:15 2
6 Campbell CS215 after the handling of 16 May 2025 2025-05-16 10:15 to 2026-06-05 10:45 18,482
7 Campbell CS215 with excess daytime heating 2024-10-22 00:15 to 2025-05-16 09:15 9,907

Coverage

Both columns are complete. What varies is how much of each year rests on two measurements.

Table 5: Years in which fewer than 99 % of records are computed from two measurements. Every other year of 2005-2025 is at least 99 %, and 2026 (January to September) is 99.99 %.
year computed from two measurements why
2004 28.0 % the tower record begins on 20 September
2009 97.2 % scattered outages
2012 92.8 % logger clock error, power-supply failure and storm damage
2016 94.3 % the January outage during which the probe and logger were replaced
2019 95.7 % the November outage after the NABEL reference ended

The 2019 gap is the longest after 2018: the 47 m humidity sensor was down from 4 November 2019 04:45 to 19 November 2019 23:45, 759 records. RH there is reconstructed from MeteoSwiss Lägern, and VPD carries code 7.

The probe changes in VPD

Notebook 07 compares both columns with a VPD computed by the same formula from the MeteoSwiss Lägern TA and RH, using only records computed from two measurements (code 0), split into day and night. Only the change in the tower-minus-Lägern difference across a probe change is evidence, not the difference itself.

Table 6: The step in the tower-minus-Lägern VPD difference from the MP101A (2005-2015) to the normal CS215 (from 2017), against the median change of the same difference between two ordinary consecutive years.
column step, day ordinary year, day step, night ordinary year, night
VPD_T1_47_1 -0.0023 kPa 0.0101 kPa -0.0287 kPa 0.0071 kPa
VPD_T1_47_1_HOMOGENIZED -0.0005 kPa 0.0129 kPa -0.0140 kPa 0.0094 kPa

January 2016 is an independent test, because the 2016 transfers of 02 and 04 were fitted against NABEL, not against Lägern. The homogenised column steps less than the measured one by day and by night. The measured column’s small daytime step is not evidence of homogeneity: the two errors of the MP101A act on VPD in opposite directions (its TA reads about 1.3 °C too cold, which lowers VPD, and its RH reads too dry, which raises it), so they partly cancel. The earlier mixed basis, homogenised TA with measured RH, stepped by about -0.051 kPa by day and by night, the most of the three.

June 2026 is not an independent test. The HygroVUE10 transfers of 02 and 04 were fitted against Lägern on the same samples (the HygroVUE10 from 12 June to 30 September 2026 against the normal CS215 on the same dates of 2022-2024). Across this change the measured column steps by +0.055 kPa by day and +0.117 kPa at night, against a mean VPD of the HygroVUE10 sample of 1.39 kPa by day and 1.11 kPa at night. The homogenised column steps by +0.021 and +0.018 kPa, which is expected by construction.

Known limitations

  • The measured column is not homogeneous across 21 January 2016 or 5 June 2026. Use VPD_T1_47_1_HOMOGENIZED for anything that crosses either date, or stay inside one probe state with FLAG_VPD_T1_47_1_SOURCE.
  • A small night step remains in the homogenised column at 2016. It is -0.014 kPa against an ordinary-year change of 0.009 kPa, half the step of the measured column.
  • The June 2026 homogenisation rests on one summer and one reference. The TA and RH transfers behind it are fitted on June to September 2026 against Lägern, are untested outside summer, and are to be refitted once the HygroVUE10 record includes a winter. See Air temperature at 47 m and Relative humidity at 47 m.
  • Both humidity eras drift. Against MeteoSwiss Lägern the MP101A era drifts by -0.34 and the CS215 era by +0.44 percentage points of RH per year, more over each era than the 2016 step. The homogenised RH removes the steps, not this drift, so a VPD trend in either column, even inside one era, contains sensor movement.
  • The homogenised TA is self-consistent, not absolutely accurate. All eras are put on the radiation-shield behaviour of the MP101A, whose warm bias in sun (about +0.4 °C at strong radiation) remains. A warm bias in TA raises VPD.
  • Reconstructed humidity propagates. 4.17 % of records rest on a humidity transferred from another instrument (codes 6 and 7). Filter on FLAG_VPD_T1_47_1_ISFILLED == 0 where that matters.
  • 2026 is a partial year. It holds January to September and a dry, clear summer, so its annual statistics are not comparable with full years.
Back to top