Dataset content

Continuous soil measurements (10-min resolution) from the Pfynwald long-term irrigation experiment, covering five treatments: control, irrigation, irrigation+VPD, roof, roof+VPD. Measurements were taken at multiple plots and at three depths (-0.1, -0.8, -1.2 m below soil surface). Sensors include TEROS-21 (matric potential), TEROS-12 (volumetric water content), TEROS-21 (soil temperature), and TEROS-12 (electrical conductivity).

Processing steps

1. Raw retrieval from the LWF database (ada.v_messvar, project Pfynwald irrigation), 10-min cadence. 2. Range filtering to remove physically implausible values: - SWP: drop values ≥ 0 kPa or ≤ −8000 kPa - VWC: drop values ≤ 0 or > 0.8 m³/m³ - TEM: drop values > 40 °C or < −20 °C

3. Duplicate aggregation within the same 10-min timestamp by mean (median for SWP). 4. Temperature correction (SWP_TC): matric potential is temperature-corrected using the TEROS-21 manufacturer equation referenced to T_ref = 20 °C: - Convert SWP (kPa) to head h = |SWP|·10.197, then pF = log10(h) - ΔpF = (919.890694·ΔT − 6.945352·ΔT²)·exp(−150.754657 / (pF + 9.153084)) - SWP_TC = −10^(pF + ΔpF) / 10.197

5. No gap filling applied. Missing values remain absent from the delivered CSV.

Methods

• Database query and aggregation: R, data.table, rLWFpg
• Temperature correction: in-house implementation of TEROS-21 manufacturer formula
• See repository / script reference below

Limitations

• Values flagged out of range during step 2 are absent (not interpolated)
• Winter periods may include frozen-soil readings — these pass the range filter as physical sensor outputs but should be interpreted with caution by downstream users
• SWP_TC rows carry the messvar_id of the underlying SWP sensor; the corresponding TEM sensor used for correction is paired via site/plot/depth/datetime but not encoded in the messvar_id
• Sensor calibration uses manufacturer defaults; site-specific calibration was not performed
• Roofs were installed in spring 2024. Readings under "roof" / "roof_vpd" treatments before that reflect ambient conditions, not the experimental treatment.
• In 2025, three additional TEROS-21 sensors were added per roofed treatment at 10 cm depth to obtain a balanced sample relative to roof position (under panels, at panel edges, and in inter-panel gaps). This expansion means roofed treatments have six positions at 10 cm from 2025 onward, compared to three before.