AI disclosure: This article was drafted with AI assistance (Claude) under the guidance of a human paraglider pilot. The analysis, code, and editorial decisions were reviewed and validated by the author. Statistical findings are based on real flight data from XContest and ERA5-Land reanalysis.
The hypothesis
Every paraglider pilot knows the obvious thermal predictors: sunshine, boundary layer height, lapse rate, wind. But there’s an invisible factor literally beneath our feet that standard forecasts never show: how wet is the ground?
The physics is straightforward. Solar radiation hitting the ground can do two things: heat the air (sensible heat) or evaporate water (latent heat). Dry soil has no water to evaporate, so nearly all energy goes into heating the air — producing stronger thermals. Wet soil diverts energy into evaporation, cooling the surface and weakening thermals. This ratio between sensible and latent heat flux is called the Bowen ratio [1].
The question is: does this textbook effect actually show up in real flight data from a small Estonian airfield?
The data
We matched 114 unique flight dates (438 individual flights, 2014–2025) from Rapla/Kuusiku airfield against ERA5-Land reanalysis soil moisture (0–7 cm layer) from Open-Meteo [2]. Flight data comes from XContest [3], covering everything from 5 km beginner flights to 115 km cross-country epics.
ERA5-Land provides hourly soil moisture at ~9 km resolution. We used the daytime mean (06–18 UTC) of the top 7 cm layer — the zone that most directly influences surface energy partitioning.
The finding: 2.6× longer flights on dry soil
Splitting the 114 dates into terciles by soil moisture reveals a striking pattern:
| Tercile | SM range (m³/m³) | Median best flight | Median total km |
|---|---|---|---|
| DRY (n=38) | 0.10 – 0.22 | 41.7 km | 80.6 km |
| MEDIUM (n=38) | 0.22 – 0.32 | 19.1 km | 32.2 km |
| WET (n=38) | 0.32 – 0.44 | 15.8 km | 19.5 km |
The driest third of days produced flights 2.6× longer than the wettest third (median 41.7 km vs 15.8 km). Total kilometers flown per day showed an even larger gap (4.1×).
The overall Pearson correlation between soil moisture and best flight distance is r = -0.22 (n=114). Negative means wetter soil → shorter flights.
Removing seasonality
An obvious objection: dry soil correlates with summer, and summer has better thermals for many reasons (longer days, higher sun angle, warmer temperatures). Is soil moisture just a proxy for season?
Two controls suggest the effect is real beyond seasonality:
Within-month correlations remove the seasonal cycle entirely:
- July (n=29): r = -0.30 — SM still predicts flight quality within the best month
- May (n=17): r = -0.28 — same pattern in spring
Temperature-controlled split: When we split dates by soil temperature at the median (13.8°C) and look at the cool half only:
- Cool days (soil T ≤ 13.8°C, n=55): SM vs best_km r = -0.32
This is the key insight: soil moisture matters most on cool/marginal days. When it’s 25°C and sunny, thermals fire regardless of ground moisture. But on a cool spring day with 12°C and borderline BLH, the difference between dry and wet soil can be the difference between a 40 km cross-country and staying on the ground.
Top vs bottom flights
The 15 best flights (57–115 km) had mean soil moisture of 0.21 m³/m³. The 15 weakest flights (5–8 km) had mean soil moisture of 0.29 m³/m³.
The epic days weren’t just dry by chance — 13 of the top 15 occurred in June–August when soils are naturally driest.
Practical implications
We’ve integrated soil moisture into the eemeteo forecasting system:
- Rapla XC assessor (
rapla_xc.py): +1 bonus point when soil is dry (<0.22 m³/m³) AND temperature is below 23°C. This helps marginal cool days score higher when soil conditions compensate for lower temperatures.
- Opportunities dashboard: Visual indicators — brown circle for dry soil (good for thermals), blue circle for wet (thermals weakened).
- Morning briefing: New “Ground conditions” card showing soil moisture at Kuusiku (flying site) and Tallinn (reference point the pilot can verify against local conditions).
When to check soil moisture
- Spring (April–May): This is when SM varies most and matters most. A dry week in May can produce surprisingly good thermals even at moderate temperatures.
- After rain: 2–3 dry days are needed for topsoil to dry enough to shift the Bowen ratio.
- Marginal days: If BLH and temperature are borderline, dry soil tips the balance toward flyable conditions.
When to ignore it
- Hot summer days (>23°C): Thermals fire regardless. The bonus is deliberately suppressed above 23°C.
- Obviously wet days: If it’s raining, soil moisture is the least of your problems.
Limitations
- Sample size: 114 dates is enough to detect a signal but too few for robust threshold calibration. The tercile boundaries (0.22 / 0.32) should be treated as preliminary.
- Layer depth: ERA5-Land uses 0–7 cm; the Open-Meteo forecast API provides 0–1 cm. Absolute values may differ between these layers.
- Correlation ≠ causation: While the Bowen ratio mechanism is well-established in atmospheric science, our dataset cannot fully separate soil moisture from other correlated factors (antecedent precipitation, synoptic patterns).
- Single site: These findings are specific to Rapla/Kuusiku — a flat grass airfield at 80 m MSL in a maritime climate. Mountain sites or arid regions would have different soil moisture dynamics.
References
[1] Bowen, I. S. “The Ratio of Heat Losses by Conduction and by Evaporation from Any Water Surface.” Physical Review, vol. 27, 1926, pp. 779–787.
[2] Zippenfenig, P. “Open-Meteo — Free Weather API.” open-meteo.com, 2023. https://open-meteo.com/
[3] XContest. “World Online Paragliding Competition.” xcontest.org. https://www.xcontest.org/
[4] Muñoz Sabater, J. et al. “ERA5-Land: A State-of-the-Art Global Reanalysis Dataset for Land Applications.” Earth System Science Data, vol. 13, 2021, pp. 4349–4383.
[5] WMO. “Manual on Codes — International Codes, Volume I.1, Part A: Alphanumeric Codes.” Code Table 0901 (State of the ground). World Meteorological Organization, 2019.