The Gold Standard of Hydrology | Fog Collectors, Wind-Normal Flux, Mesh Storage and Catchment Transfer

Chapter index

CHAPTER 1 OF 12

1. Separate fog droplets from water vapour

Fog contains small suspended liquid droplets. A passive mesh collector intercepts some of those droplets as air moves past it. Captured droplets can join and eventually drain. This is not the same process as describing every collected drop as vapour condensed directly on the mesh, and it does not mean the collector creates water from nothing.

Start with a sequence: droplets approach, some encounter and remain on the collecting material, water accumulates or moves across that material, and some reaches the measurement point. Each step can have a different limitation. Seeing the first does not guarantee completion of the last.

The Climate Technology Centre and Network overview introduces the collection-and-conveyance chain. Its broad description is useful context, not a current site assessment or universal daily yield. For a specific comparison, name the collector, location, observation interval and measured output. The word foggy alone does not quantify the liquid water available to the instrument.

Contents · Next chapter

CHAPTER 3 OF 12

3. Build a wind-normal incident-flux calculation

For a uniform teaching case, incident liquid-water mass is LWC × Uₙ × A × Δt. Here Uₙ is wind speed normal to the collecting plane, A is the defined projected collecting area, and Δt is duration. The units cancel to mass: g/m³ × m/s × m² × s = g.

Take LWC = 0.2 g/m³, Uₙ = 3 m/s, A = 1 m² and Δt = 3,600 s. Incident mass is 2,160 g. Treating liquid-water density as approximately 1 kg/L for this conversion gives 2.16 litres. That is the water carried through the chosen undisturbed reference plane under the stated uniform conditions.

The equation does not assert that all approaching air passes through the mesh unchanged or that every droplet is captured. Those effects belong to collection performance relative to the reference flux. Keep denominator and numerator definitions consistent. Comparing output against an undefined available water quantity makes an efficiency number difficult to interpret.

Contents · Previous chapter · Next chapter

CHAPTER 5 OF 12

5. Name the area in every yield

Litres per square metre is incomplete unless the square metre is identified. It may refer to projected mesh area, collecting-material area, land footprint or catchment area. These are different denominators. A result per vertical mesh area is a collector performance statistic, not automatically a depth of water delivered across the landscape.

Suppose a fictional 2 m² mesh collects 6 litres. Its collector-area yield is 3 L/m² for that period. If those 6 litres are assigned to a 1,000 m² land area for a separate accounting question, the equivalent depth is 0.006 mm, since 1 L over 1 m² corresponds to 1 mm. The change is a declared allocation, not a measured natural deposition depth.

A catchment transfer requires a model of exposure, vegetation or collector geometry, spatial variation and the area actually represented. Multiplying a ridge collector’s yield by the whole catchment area silently assumes representativeness. A small device can be a useful local index while remaining insufficient to establish catchment-wide fog-water input.

Contents · Previous chapter · Next chapter

CHAPTER 7 OF 12

7. Define efficiency at the point actually measured

A gutter collection ratio can be defined as measured gutter water divided by reference incident fog water over the same interval. In the earlier 2.16-litre incident case, 0.54 litres at the gutter gives 0.25, or 25%. The numerator and denominator must describe the same event and correctly handled other sources.

This ratio differs from the fraction of droplets reaching a fibre or the fraction initially retained. Captured water can remain stored, escape collection or drain later. A collector can have substantial initial capture and a smaller gutter yield. Conversely, drainage from water stored before the interval can make short-window gutter output exceed the current incoming reference flux.

A ratio above one is therefore a reason to inspect the accounting, not evidence that the device creates water. Check storage carryover, rain contributions, area definition, LWC representation and wind measurement. If the denominator is effectively zero or unreliable, an efficiency ratio may be undefined or uninformative. Report the observed volume separately rather than forcing a dramatic percentage.

Contents · Previous chapter · Next chapter

CHAPTER 9 OF 12

9. Distinguish fog, rain and other water sources

A wet collector can receive rain as well as fog, and water can arrive or remain through other processes. If the question concerns fog-only collection, the attribution needs evidence. Total gutter water is a valid measured quantity, but assigning all of it to fog is a separate claim.

Consider 1.2 litres measured during an interval that includes rain. If independent evidence supports 0.4 litres of rain contribution under the stated collector response, the remaining 0.8 litres is not automatically a perfect fog estimate: storage and other pathways may still matter. Show the attribution steps and uncertainty instead of subtracting a rain depth using an unrelated area.

Appearance also does not establish quality. Clear collected water may require assessment for its intended use, including the collection and storage system. This guide supplies no drinking-water approval or treatment prescription. Keep volume, source attribution, suitability and delivered service as separate questions, each with evidence appropriate to the claim.

Contents · Previous chapter · Next chapter

CHAPTER 11 OF 12

11. Aggregate by water opportunity, not by averaging percentages blindly

For several events, the combined gutter ratio is total output divided by total reference incident water. It is an incident-water-weighted summary of event ratios, not generally their arithmetic mean. The distinction is the same one encountered whenever fractions have different denominators.

Take event one with 1 litre incident water and 0.5 litres output, and event two with 9 litres incident water and 0.9 litres output. Event ratios are 50% and 10%. Their arithmetic mean is 30%, but the combined ratio is 1.4/10 = 14%. The larger event contributes most of the reference water.

A combined ratio can still hide conditions. Preserve event-level records if performance changes with wind, droplet population, wetting state or maintenance. A single seasonal average is not evidence that each dry spell is bridged. When transfer matters, test whether the new location has conditions represented in the comparison rather than importing the average unchanged.

Contents · Previous chapter · Next chapter

Independent worked case and evidence checks

Work with an idealised one-square-metre projected mesh for a two-hour event. During the first hour, LWC is 0.2 g/m³ and normal wind speed is 3 m/s. During the second, LWC is 0.1 g/m³ and normal wind speed is 2 m/s. Assume the observations represent the reference plane and remain constant within each hour.

Hour one carries 0.2 × 3 × 1 × 3,600 = 2,160 g, approximately 2.16 litres. Hour two carries 0.1 × 2 × 1 × 3,600 = 720 g, approximately 0.72 litres. The event total is 2.88 litres of incident liquid water under the model.

Gutter output is 0.30 litres in hour one and 0.42 litres in hour two. Total output is 0.72 litres. The event gutter ratio is 0.72/2.88 = 0.25, or 25%. Hour two has less incoming water but greater output, which is possible when drainage timing differs from arrival.

Suppose an independent storage accounting indicates that captured input was 1.00 litre, initial stored water was 0.10 litres and final stored water was 0.28 litres. With gutter output 0.72 litres, other losses needed for closure are 0.10 + 1.00 − 0.28 − 0.72 = 0.10 litres. This is a closed fictional balance, not a method for measuring all terms from gutter volume alone.

The initial-capture ratio in that case is 1.00/2.88, about 34.7%. The gutter ratio remains 25%. They answer different questions. Replacing one with the other would erase retained water and other losses from the chain.

Now consider only a ten-minute period after fog has ended. Incident reference water is effectively zero, but 0.08 litres drains from the wet mesh. Do not divide by zero to calculate efficiency. Record delayed drainage and connect it to an appropriate event window or unresolved starting storage.

Change the starting condition: the mesh begins completely dry rather than already holding 0.10 litres. The same incoming weather need not produce the same timing of gutter flow. A fair material comparison should recognise wetting history rather than attributing every early difference to fibre design.

Next change orientation. Keep the first hour’s wind speed at 3 m/s but turn its approach to 60° from the normal. Under the simple projection, the first-hour incident volume becomes 1.08 litres. If the second hour remains unchanged, the two-hour incident total becomes 1.80 litres. A fixed 25% illustrative gutter ratio would predict 0.45 litres, not 0.72.

That prediction assumes the ratio remains unchanged under the new direction. It is a conditional calculation, not a measured aerodynamic law. The changed direction could also alter the collection response. Label what was changed in the geometry and what was merely held constant for the exercise.

Now double projected area to 2 m² while explicitly assuming identical exposure and no interaction. Incident water doubles to 5.76 litres in the original two-hour weather. At the same assumed 25% ratio, output would be 1.44 litres. This scales the teaching model; it does not prove a real array of adjacent collectors scales linearly.

A colleague reports the original 0.72 litres as 0.72 mm of catchment input. That is valid only if the assigned horizontal accounting area is one square metre and the interpretation is explicitly an equivalent depth. Over 100 m², the same volume is 0.0072 mm; over 1,000 m² it is 0.00072 mm. Vertical mesh area has not automatically measured either land input.

For another check, consider a 10 m² collector yielding 30 litres in a day. Its mesh-area yield is 3 L/m²/day. If a separate reporting exercise allocates that water over a hectare, its equivalent depth is 30/10,000 = 0.003 mm. Both calculations are arithmetically correct, but their denominators and meanings differ.

Now test the average-weather shortcut. Two equal hours have LWC and normal wind pairs of (0.3, 1) and (0.1, 5). Mean LWC is 0.2 g/m³ and mean wind is 3 m/s. Their product is 0.6 g/m²/s. But the mean of the actual products is (0.3 + 0.5)/2 = 0.4 g/m²/s.

For one square metre over two hours, the correctly integrated incident mass is 0.4 × 7,200 = 2,880 g. The product-of-means shortcut gives 4,320 g, a 50% overestimate. The error arises because liquid water and wind did not vary independently in the supplied record. Integrate compatible observations instead of mixing separate averages casually.

Move from event performance to availability. A fictional week produces daily outputs of 10, 0, 0, 15, 5, 0 and 5 litres. Total output is 35 litres and calendar-day mean is 5 litres. Averaging only the four productive days gives 8.75 litres, which does not describe water arriving on each of the seven days.

Add a simple storage question without making a design recommendation. Start with 5 litres stored and a fixed demand of 5 litres per day. Process collection and demand as daily totals, with no losses or capacity limit. End-of-day stores become 10, 5, 0, 10, 10, 5 and 5 litres. The exercise shows how chronology matters even when weekly supply equals weekly demand.

Change the daily order to 0, 0, 10, 15, 5, 0 and 5 litres while retaining the same weekly total and starting storage. The first day empties the store, and the second day has a 5-litre shortfall. Later collection cannot retroactively supply that missing day. A weekly total alone does not establish dependable service.

Add a measurement-resolution check. A fictional volume logger records in 0.1-litre increments. An interval with no registered increment does not necessarily establish exactly zero drainage; less than one increment may remain unregistered or be assigned to a later interval. Preserve the recording convention before drawing conclusions about the shortest fog events.

Suppose the first interval physically drains 0.06 litres and the second drains 0.06 litres. A cumulative increment-based logger may register no complete increment first and one increment during the second, leaving a remainder. Reading the two rows as an exact transition from no physical drainage to drainage would misinterpret the quantised record. A longer cumulative comparison can answer a different question more reliably.

Now consider an approximate uncertainty sensitivity: measured output is 0.50 litres and reference incident volume is 2.0 litres. The nominal ratio is 25%. Holding output fixed but changing the plausible denominator from 1.6 to 2.4 litres gives ratios from about 31.3% down to 20.8%. The denominator can dominate the interpretation even when the bottle volume is easy to read.

These endpoints do not supply a confidence interval, and they omit uncertainty in output. They identify a priority for better evidence. If the choice between materials depends on a small ratio difference, checking the representative incoming flux may be more valuable than adding precision to an already adequate volume measurement. Match the repair to the source of ambiguity.

Do not extend that toy store into real tank sizing. Real systems require capacity, losses, water quality, operating constraints, demand variability and appropriate design evidence. The independent lesson is narrower and powerful: event efficiency, annual yield, spatial deposition and service reliability are connected but distinct questions.

Questions and practical checks