A soil sample can gain water without rain or visible dew
A precision balance shows that a dry soil sample gained mass overnight, yet no rain was recorded and no droplets were visible. Has the soil captured atmospheric water? Possibly. But the balance identifies a mass change, not its source or physical form. Dew, fog deposition, vapour adsorption, handling and instrument effects need to be separated before the gain receives a hydrological name.
Soil vapour adsorption is the association of water molecules with soil surfaces. Under suitable dry conditions, it can contribute to non-rainfall water inputs without a visible liquid film. The process depends on the soil’s state and its thermal and vapour environment. A metal dew plate nearby is not a substitute measurement of the soil.
This guide follows the evidence from mass gain to attribution, using original fictional examples. It is about interpreting supplied research records, not collecting soils or manipulating desiccants. A desiccant’s capacity to attract water does not show how an undisturbed soil surface behaves or how much usable water an ecosystem receives.
Hydrology subject guide · Learning that transfers into life · How X Works hub
Choose your reading route
Route 1: Name the processes · Chapters 1–3
Route 2: Read the mass account · Chapters 4–5
Route 3: Test thermal and source evidence · Chapters 6–8
Evaluate balance resolution, drift and environmental loading
Route 4: Scale the result · Chapters 9–10
Follow retention, release and possible ecological use separately
Route 5: Transfer and report · Chapters 11–12
Full chapter index · Independent worked case · Questions and practical checks
Chapter index
Name the processes · Chapters 1–3
Separate adsorption, absorption and condensation
Understand the water-potential difference without using humidity alone
Distinguish soil behaviour from a desiccant demonstration
Read the mass account · Chapters 4–5
Write the complete lysimeter mass balance
Convert small mass gains to equivalent water depths
Test thermal and source evidence · Chapters 6–8
Evaluate balance resolution, drift and environmental loading
Use surface temperature and dew point as evidence, not a single verdict
Separate atmospheric input from local recycling
Scale the result · Chapters 9–10
Follow retention, release and possible ecological use separately
Scale from a soil sample to a heterogeneous landscape
Transfer and report · Chapters 11–12
Transfer the reasoning to wet soil, fog and saline surfaces
Report net mass, proposed mechanism and remaining ambiguity
CHAPTER 1 OF 12
1. Separate adsorption, absorption and condensation
Adsorption concerns molecules associated with a surface. Absorption commonly describes entry into the bulk of a material. Condensation involves a vapour-to-liquid phase change. Everyday speech sometimes blends these terms, but a scientific explanation should identify the process being proposed and the observation that supports it.
Soil contains a large and heterogeneous internal surface. Water can be strongly associated with mineral or organic surfaces at low water contents. This is different from assuming every new water molecule forms a freely draining droplet. A mass gain establishes that the weighed system became heavier; it does not establish the water’s mobility or availability to roots.
Dew is liquid condensation on a sufficiently cool surface. Fog consists of suspended liquid droplets that can be intercepted. Atmospheric vapour adsorption can occur without visible dew. These processes can overlap in time or space, so a record may support a combined non-rainfall input more strongly than a unique partition among mechanisms.
Start with the narrow statement: the supplied soil system gained a stated mass during a stated interval. Add “water” when non-water contributions are adequately excluded. Add “atmospheric” and “adsorption” only when the boundary and independent observations support those additional claims. Each adjective should earn its place through evidence.
CHAPTER 2 OF 12
2. Understand the water-potential difference without using humidity alone
A dry soil can have a strong affinity for water. Whether vapour moves towards it depends on the relationship between atmospheric vapour conditions and the soil’s equilibrium state at its temperature. Relative humidity describes the air relative to saturation at the air temperature; it is not by itself a complete description of that relationship.
Temperature matters because saturation vapour pressure changes with temperature and because soil water retention and equilibrium humidity are connected. A surface that warms or cools can change its tendency to exchange vapour even when the overlying air’s water-vapour amount changes little. Air temperature at a weather mast need not equal the soil-surface temperature.
Primary lysimeter and modelling work by Saaltink and colleagues examines daily thermal cycles and soil vapour adsorption. The useful general lesson is that coupled heat and water processes can produce mass oscillations. It is not a universal daily adsorption amount that should be copied from one soil or climate to another.
A practical reading question is therefore: which humidity was measured, at what temperature and height, and how was the soil’s near-surface state characterised? “Humidity increased at night” is not enough to identify adsorption. The physical comparison requires compatible atmospheric and soil variables, not a single familiar percentage.
CHAPTER 3 OF 12
3. Distinguish soil behaviour from a desiccant demonstration
A jar of silica gel can gain mass from moist air. That demonstration helps explain surface attraction but deliberately uses a highly sorptive material and a controlled boundary. It does not quantify the water gain of field soil with different minerals, pore structure, initial water content and thermal conditions.
Likewise, drying a soil strongly before an experiment changes its initial state. The resulting uptake may answer a useful material question, but it is not automatically the uptake of soil that experienced the natural previous day. Distinguish a standardised laboratory treatment from an attempt to preserve an undisturbed field state.
For a fictional comparison, material A gains 0.9 g and field-state soil B gains 0.2 g over the same nominal interval. The difference cannot be attributed solely to one environmental driver if composition, initial water content and geometry also differ. A controlled comparison changes one interpretable factor or explicitly models the differences.
The transferable idea is not that one material is “better” at producing water. It is that water exchange depends on both the environment and the material’s state. A material-capacity result, a local field flux and a sustainable water-supply claim are different questions. Do not promote one into another without the missing evidence.
CHAPTER 4 OF 12
4. Write the complete lysimeter mass balance
Take the weighed system as the complete soil sample and container, with a known exposed area. Its mass change can include rain, irrigation, fog deposition, dew, vapour exchange, drainage, leakage, removal or addition of solids, and instrument effects. The possible terms depend on the actual design. Define the system before assigning a sign to the result.
For a rain-free, root-free, closed-bottom teaching system, suppose non-water changes and drainage are excluded. A positive mass change then gives net water gain across the system’s external boundaries. It does not necessarily equal gross atmospheric adsorption because evaporation and other water inputs may occur simultaneously.
If 1.8 g enters and 0.5 g leaves during an interval, the observed net gain is 1.3 g. An endpoint balance cannot identify both gross exchanges independently. Calling 1.3 g the total adsorption would require evidence that the outward flux was negligible or had been separately accounted for.
Keep internal redistribution distinct. Water moving from the lower soil to the upper soil inside the same weighed system does not change total mass. It can change a surface-moisture sensor reading, however. A surface becoming wetter therefore supports a different inference from a complete lysimeter gaining mass. The two observations complement rather than replace one another.
CHAPTER 5 OF 12
5. Convert small mass gains to equivalent water depths
Suppose a fictional lysimeter has exposed area 0.020 m² and gains 2.4 g of water over six hours. This is 0.0024 kg/0.020 m² = 0.12 kg m⁻², equivalent to approximately 0.12 mm at freshwater density 1,000 kg m⁻³. The average net gain rate is 0.020 mm h⁻¹ over that interval.
The small depth does not make the arithmetic optional. Use kilograms rather than grams in the kg m⁻²-to-mm relationship, and use exposed area rather than total container-wall area. A depth equivalent spreads the gained water over the stated horizontal area. It does not say that a continuous liquid layer of that thickness formed on the soil.
Now double the exposed area while retaining the same total mass gain. The equivalent depth halves. Alternatively retain the area but double the exposure duration with the same net gain: total depth stays the same while the interval-average rate halves. State whether a comparison concerns total uptake, equivalent depth or average rate.
Do not multiply the overnight rate by an entire day without evidence. The same soil may lose much of the gained water later. A useful diel record preserves gains and losses separately before calculating a net daily contribution. Reporting only the positive increments can overstate the amount retained by the system.
CHAPTER 6 OF 12
6. Evaluate balance resolution, drift and environmental loading
Small gains can be of the same order as weighing uncertainty. Suppose each endpoint mass has independent standard uncertainty of 0.20 g. The difference has standard uncertainty √(0.20² + 0.20²) = 0.283 g. Over 0.020 m², this corresponds to about 0.014 mm. State the assumptions behind that calculation.
Repeated endpoint errors are not always independent. Temperature-related drift, shared calibration and persistent mechanical loading can correlate the measurements. Wind or contact with adjacent vegetation can alter balance loading. A finely printed output value is not evidence that these influences are smaller than the hydrological signal.
A blank or reference weight can reveal some drift, but its thermal and aerodynamic behaviour may differ from the soil system. If a correction is derived from a reference, state how the reference responds and what the correction assumes. Do not regard subtraction as independent validation merely because it reduces the apparent noise.
Keep negative and near-zero changes in the record. Selecting only gains above a threshold can bias a cumulative input estimate. A quality rule should be based on measurement capability and stated process criteria, not on the desired number of adsorption nights. Preserve unclassified intervals rather than forcing every small fluctuation into a hydrological category.
CHAPTER 7 OF 12
7. Use surface temperature and dew point as evidence, not a single verdict
A surface below the surrounding air’s dew-point temperature can support condensation under appropriate conditions. If the soil surface stays above dew point, visible dew is less plausible as an explanation, but adsorption can still be considered. The relevant temperature is that of the exchanging surface, not necessarily air measured two metres above it.
Temperature varies within a soil surface. Roughness, stones, vegetation and radiative exposure create small-scale differences. One sensor may not represent every part of a lysimeter. A “no dew” classification based on one temperature value needs an evaluation of spatial variability and sensor uncertainty.
For a supplied fictional interval, soil-surface temperature is 18.0 °C with uncertainty 0.3 °C and air dew point is 17.9 °C with uncertainty 0.4 °C. The difference is small compared with the combined uncertainty. Claiming condensation is impossible because 18.0 exceeds 17.9 overstates the evidence.
A stronger interpretation combines thermal conditions, visual or wetness evidence where available, fog observations, precipitation measurements and the mass record. No single proxy should be asked to provide more identification than it contains. A nearby metal plate can cool differently from soil, so its wetness is not a direct measurement of the soil’s phase or flux.
CHAPTER 8 OF 12
8. Separate atmospheric input from local recycling
Water condensed near the soil surface can originate from atmospheric vapour or from evaporation and upward transport within the soil profile. If the complete weighed system loses and regains its own water internally, that cycle does not create an external input to the total system. Source attribution requires the boundary to remain visible.
Isotope information, vapour profiles or carefully constrained modelling can provide additional evidence about sources, but each has its own mixing and fractionation assumptions. A source label is not obtained simply by observing that the soil surface became wetter. The available data may support a combined category rather than a unique atmospheric fraction.
The HESS study by Paulus and colleagues treats non-rainfall inputs as a partitioning problem using precise lysimeters and additional environmental observations. That framework illustrates why the mass series is useful but not the whole attribution argument. Its site findings are not a universal percentage for other ecosystems.
A good classroom question changes the weighed boundary. First weigh the whole column; then imagine weighing only a shallow upper sample. Internal transfer across the shallow sample’s lower boundary now changes its mass, even though it would not change the whole column’s mass. The same observed gain can therefore support different source statements depending on the system being weighed.
CHAPTER 9 OF 12
9. Follow retention, release and possible ecological use separately
Water associated with dry soil surfaces is not necessarily freely available to plants. Availability depends on how strongly it is held, where it is located and whether organisms can access it before it is released. A measured gain is a useful water-account observation, not proof of an equivalent irrigation benefit.
Consider a fictional night with 0.12 mm net gain followed by 0.09 mm net loss before noon, under an otherwise closed simplified account. The retained difference is 0.03 mm. The positive night gain and the net retained amount answer different questions. Neither alone establishes how much water passed through roots or supported microbial activity.
A soil could repeatedly adsorb and desorb small quantities while showing little net accumulation over a week. Such cycling may still matter physically or ecologically, but cumulative positive increments should not be presented as a new standing water reserve. Keep gross cycling, net retention and inferred biological use in separate categories.
If a study claims ecological importance, look for evidence matched to that claim: timing relative to drought, soil depth, organism access or independent physiological response. It is reasonable to say that non-rainfall inputs can matter under some conditions. It is not reasonable to infer a universal drought remedy from a small precision-balance gain.
CHAPTER 10 OF 12
10. Scale from a soil sample to a heterogeneous landscape
Soil texture, surface cover, initial wetness, exposure and thermal regime can vary over short distances. A small sample can represent a defined surface class, but it does not automatically represent a whole landscape. Sampling needs to address the variability that controls the exchange being inferred.
Suppose a fictional area contains 25% bare dry patches with 0.16 mm net night gain and 75% covered patches with 0.04 mm. The area-weighted gain is 0.25 × 0.16 + 0.75 × 0.04 = 0.07 mm. An equal average of the two class means would be 0.10 mm and would overrepresent the bare patches.
Replication within a class estimates some local variability. It does not substitute for measuring missing classes. Ten samples beside one tree can still miss exposed ground, shaded ground or different soils elsewhere. State how classes were defined and how areas were estimated before treating their weighted mean as a site result.
Temporal scaling has the same requirement. Observations from a selected dry spell do not establish an annual input. Rain alters initial soil wetness, while seasons alter temperature and humidity patterns. A model may help extend coverage, but the extension should remain labelled as modelled and evaluated over the conditions for which it is used.
CHAPTER 11 OF 12
11. Transfer the reasoning to wet soil, fog and saline surfaces
Change the dry-soil scenario to recently wetted soil. The affinity and vapour exchange can change, and evaporation may dominate. The old net uptake cannot be carried across merely because the sample has the same texture. Initial water state is part of the experiment, not background information that can be omitted.
Next introduce fog. Intercepted liquid droplets can add mass without the same adsorption mechanism. If the rain gauge remains dry, the event is still not automatically vapour adsorption. Use fog evidence and collection geometry, and retain a combined non-rainfall category where the data cannot distinguish the contributions.
Now add soluble salts. Hygroscopic behaviour and solution formation may affect exchange, but salt composition and concentration matter. Do not infer a field flux from the fact that a material attracts water in a package-drying demonstration. The surface’s state, atmospheric conditions and energy balance remain part of the account.
Across these changes, conservation and equivalent-depth conversion remain useful. The mechanism and source attribution do not remain fixed. A successful transfer check asks what new observation would distinguish the competing explanations. It does not merely reuse the same label for every rain-free increase in soil mass.
CHAPTER 12 OF 12
12. Report net mass, proposed mechanism and remaining ambiguity
A clear statement names the weighed boundary, exposed area, interval, initial state, mass correction, uncertainty and possible external pathways. It reports the observed net change before assigning a mechanism. If adsorption is the preferred explanation, explain which alternatives were excluded and which remain plausible.
For example: the root-free, closed-bottom soil system gained 2.4 g over six rain-free hours, equivalent to 0.12 mm. Thermal, precipitation and fog evidence supported atmospheric vapour uptake under the stated conditions, but the record did not independently separate every simultaneous inward and outward exchange. This is more informative than an unqualified claim that the soil “made” water.
Keep modelled attribution separate from measured mass change. A numerical model can connect heat, vapour and liquid processes and test whether the proposed mechanism is physically consistent. Agreement with the mass cycle is useful, but alternative parameter combinations may also fit. Independent environmental observations strengthen the inference.
The gold standard is not the most confident process label. It is a transparent explanation of what the record demonstrates, what it suggests and what additional evidence could change the interpretation. That approach makes very small water exchanges intelligible without either dismissing them as trivial or inflating them into an unsupported water-supply promise.
Independent worked case and evidence checks
An additional source test can be framed without assuming a unique answer. Two fictional soil samples have the same net mass gain, but one experienced a short fog episode and the other did not. The equality of totals does not imply equality of mechanism. In the fog-exposed sample, the balance may combine liquid deposition and vapour exchange. In the other sample, dew, adsorption and measurement effects still need consideration. A matching number is not a matching process history.
Now suppose the fog-exposed sample’s mass rose smoothly, without a clear event step. That shape does not exclude fog deposition, because deposition and simultaneous losses can vary continuously. Likewise, a rapid step does not prove rain; handling or an instrument disturbance could also create one. Use time-series shape to formulate tests, not to replace independent environmental evidence. Preserve the original time resolution before averaging, since an hourly mean can hide a short event that matters to attribution.
Finally distinguish a detection statement from a flux estimate. A study may provide strong evidence that a non-rainfall gain occurred, while its exact magnitude remains sensitive to drift correction. Another interval may have precise net mass change but weak source identification. These are different strengths and limitations. A useful results table can show measurement quality and mechanism confidence in separate columns rather than forcing both into one good-or-bad flag.
This separation travels to other hydrological instruments: confidence that something changed, confidence in how much it changed, and confidence in why it changed are not the same judgment.
A fictional soil lysimeter exposes 0.015 m². Its corrected mass rises by 1.8 g during a rain-free five-hour interval. A reference system indicates a remaining possible positive drift of 0.3 g, but the appropriate correction is not certain. Report the raw corrected-series equivalent and the drift sensitivity separately.
The 1.8 g rise is 0.0018/0.015 = 0.12 kg m⁻², or 0.12 mm. If the full 0.3 g were an artefact, the water gain would be 1.5 g, or 0.10 mm. The difference of 0.02 mm is not negligible merely because the mass values are small.
The supplied environmental record rules out recorded rain but contains a fog interval. Therefore neither 0.12 nor 0.10 mm can be identified uniquely as vapour adsorption from these observations alone. A reasonable report gives net non-rainfall gain with the unresolved mechanism noted. More decimal places would not resolve that source ambiguity.
During the following morning, corrected mass falls by 1.2 g with no other supplied water terms. The full cycle’s net gain is 0.6 g, or 0.04 mm before the uncertain drift treatment. The night input, morning release and net retained amount should remain visible rather than being collapsed into an apparently permanent 0.12 mm addition.
Changed boundary: a sensor near the upper surface indicates wetter soil while the whole lysimeter mass remains unchanged during a later interval. Internal redistribution is a possible explanation. The sensor result alone does not establish an atmospheric input. Distinguishing a state change from a boundary flux is the central transfer lesson.
Independent check: each endpoint has standard uncertainty 0.15 g, treated independently here. The mass difference uncertainty is 0.212 g, equivalent to about 0.014 mm. This is separate from the possible 0.02 mm systematic drift and the unresolved fog attribution. Keep statistical precision, correction sensitivity and source uncertainty distinct.
Worked comparison
| Observation | Supported statement | Not established alone |
|---|---|---|
| Whole lysimeter gains mass | Net mass entered the weighed system | Unique water source or mechanism |
| Surface moisture increases | The sensor’s sampled zone became wetter | An external input to the whole column |
| Soil surface is near dew point | Condensation is thermally plausible | Its amount or exclusive contribution |
| Dew plate is wet | The plate collected or condensed water | Identical exchange on soil |
| Night gain is followed by day loss | A diel water cycle occurred | Equivalent permanent retention or root uptake |
Questions and practical checks
Can soil gain water without visible dew?
Yes, vapour adsorption can contribute under suitable conditions. A mass record still needs thermal, boundary and alternative-source checks.
Does every rain-free gain equal adsorption?
No. Fog, dew, other water pathways and instrument effects can contribute.
Is adsorption the same as a layer of liquid water?
No. Equivalent depth is a mass-per-area expression, not proof that a freely mobile film formed.
Can a dew plate measure soil adsorption?
Not directly. Plate material and cooling differ from soil. It provides a different observation.
Can all positive increments be added as stored water?
Not without accounting for later release and simultaneous exchange. Gross cycling and net retained water are different quantities.
Continue with the connected subjects
Follow the neighbouring non-rainfall pathways
Dew collectors: cooling, retained water and attribution
Fog collectors: intercepted droplets and catchment transfer
Check the soil and water account
Microlysimeters: isolated-core boundaries and exposure age
Water-budget closure and what a residual can identify
Water activity and why saline surfaces require their own state description
Compare a connected measurement problem
Surface-renewal temperature ramps: examine latent heat inferred from an energy residual. Follow the changed boundary and test the relevant storage, mixing or attribution assumptions rather than transferring a correction by name alone.
Sources and scope
Saaltink and colleagues (2020): lysimeter and coupled modelling of soil vapour adsorption
Paulus and colleagues (2022): partitioning seasonal and diel non-rainfall inputs with lysimeters
Agam and Berliner (2004): coastal-desert soil-water changes and the limits of dew plates
The numerical cases in this guide are original fictional teaching examples, not measurements from a real site. This is an educational explanation, not a field safety instruction, engineering approval or an official certification. For a real project, use qualified practitioners, current product documentation and site-specific evidence. Sources were checked on 30 October 2026.