Why a duct can sweat while the wall beside it stays dry.
Why cold surfaces in cooled buildings get wet: an interactive model of humidity, surface temperature and room pressure, built on the Dubai hourly record.
One interior, six faces.
An engineer reads a room as a control volume: a bounded space with mass and energy crossing its faces. At whole-building scale, the enclosure has six principal faces. An individual room may border outdoors, ground, roof or other conditioned spaces, and which is which decides where heat and moisture can enter. Under typical hot-season daytime conditions in the UAE, heat generally flows inward through exposed elements. The air-conditioning works against it, and in doing so air-conditioning components, adjacent colder spaces and thermal bridges can create locally cold surfaces: supply ducts, diffusers, chilled-water pipes, coils, slabs shared with a colder neighbor.
Condensation can form on any surface whose temperature falls below the local air dew point. The coldest surface usually reaches that condition first. That inverts the intuition most people bring from temperate climates, where the cold surface is the outside wall. Here the outside wall is usually warm. The duct is cold.
The interactive model.
Move the three controls and watch the room respond. The landing state is a compliant room: 55% relative humidity, an insulated duct whose jacket is assumed at 21 °C, held at a slight positive pressure. Raise the humidity and the surface humidity at the duct climbs through ELEVATED into HIGH long before anything is wet. Change the surface to a duct with failed insulation, or a bare chilled pipe, and it drips. Take the room negative and the model indicates outdoor air, at a 24 °C dew point, moving inward through whatever leakage paths exist.
Calculated dew point, surface RH, condensation margin · Assumed outdoor condition and preset surface temperatures · Input the three controls · Practice criterion the interpretation bands · Measured only the preset room's recorded values
See what the air in this room is actually doing.
Fig. 1An engineer reads a room as a control volume, a bounded space with mass and energy crossing its faces; this one is a storage room in a private residence at 24 °C and 72.1% RH. The walls were dry. The duct sat below the 18.7 °C dew point and dripped. Duct surface shown at 15.8 °C, a representative value for a duct with failed insulation; the observed condensation requires only that it was below the 18.7 °C dew point.
What the eye seesA cool, clean storage room in a fine residence. The walls are fine. Nothing to report.
What the instrument seesRoom air at 24 °C and 72.1% RH, dew point 18.7 °C. Any surface in that room colder than 18.7 °C condenses the room's moisture onto itself, every hour the air-conditioning runs. The duct was.
For one hour in four, the air outside a building in the coastal cities of the UAE, Dubai and Abu Dhabi among them, carries a dew point above 22 °C. The dew point is the temperature at which the moisture in air turns back into water, and 22 °C is about the temperature we cool our rooms to. Wherever that air reaches a cooled surface, it condenses, and where it condenses long enough, mold follows. This is why most people who contact Indoor Sciences do so about mold, humidity or condensation, and why we answer by investigating the whole building: air, moisture, pressure and heat read as one system, the causal chain identified, corrective action prescribed and the result verified. In private residences and complex properties across Dubai and the UAE.
A hidden problem, made visible.
The room shown live in the model above, and recorded below by the practice's own instruments, is a real storage room in a private residence. The walls were dry. The duct was not. Its record is set out the way every engagement begins: what was reported, what was measured, and what the causal chain was. The prescription and the verification that follow belong to the engagement, and to the report.
The model reconstructs the room from its recorded air temperature and humidity and the observed duct condensation; because duct-surface temperature and room pressure were not recorded on that visit, those two values are explicit modelling assumptions. The photograph is the practice's own evidence from the room the model reproduces, and the thermogram shows where, in the same residence, the wall itself carried the anomaly.
Read the full investigation recordClose the investigation record
- Reported symptom
- Persistent humidity across the residence; condensation, blistering wall paint and moisture on the floor in a storage room.
- Measurements
- Storage room air 24 °C at 72.1% RH (benchmark 60% RH), dew point 18.7 °C, computed. Condensation on exposed supply ductwork; moisture on the floor beneath. Wall moisture 3.3% (handheld moisture meter, concrete scale; practice threshold 5%): no elevated moisture in the walls. Relative humidity above 60% in most rooms of the residence. Surface temperature and room pressure were not recorded on this visit; the duct surface in the model is modelled, not measured.
- Causal chain
- Widespread elevated indoor humidity, 60–74% RH across the residence, indicating latent-load removal by the air-conditioning insufficient for the moisture entering. In the storage room, 72.1% RH at 24 °C sets the dew point at 18.7 °C; the exposed supply duct, its insulation compromised, ran below that temperature and condensed the room's moisture onto itself, dripping to the floor. The wall paint blistered from the humidity; the walls themselves recorded no elevated moisture. Mechanism: humid air on a cold surface, not water in the wall.
What this changes in practice.
Look at the cold surfaces first.
In a cooled interior the condensation risk sits on ducts, diffusers, pipes and shared slabs, not on the sun-facing wall. An investigation that starts by measuring room humidity and the temperature of the coldest accessible surface reaches the mechanism faster than one that starts with the wall.
Surface humidity, not room humidity, is the criterion.
A room at 65% can carry a duct jacket at 78% surface humidity. The principal indicator is the humidity at the material, sustained over time; whether growth follows depends on the material, its temperature and the duration of exposure. The room reading is only the starting point of the calculation.
Pressure direction is a transport path.
A depressurized room admits outdoor moisture through whatever leakage exists. Direction identifies the transport path; the quantity of air and moisture moved requires envelope leakage data that this model does not hold, so the practice measures it where the symptom pattern calls for it.
The case this model reproduces is recorded on the homepage investigation record: reported symptom, measurements, and the parts of the record still to be completed.
Four interacting domains shape the indoor conditions we investigate.
Air quality
What the air carries, in a bedroom or a guest floor: microbial load, particulates, volatile compounds, gases. Measured with precision against recognized benchmarks, never inferred from how a room smells.
Moisture
Where water enters a building, travels and remains. On this coast it arrives mostly as vapor in the air rather than as a leak, and it is the origin of nearly every mold problem we are asked to see: the element most often treated at the wrong end.
Airflow and pressure
How air moves through a building, and why. Pressure decides which way air moves, and moisture and contaminants travel with it; a room under negative pressure draws in whatever surrounds it. In the storage room, humid air met a duct surface the air-conditioning had made colder than the air's dew point.
Heat flow and surface temperature
Under hot-season daytime conditions heat flows inward, and the air-conditioning fights it. The surfaces it makes cold, ducts, diffusers, chilled-water pipes, a slab beside a colder neighbor, are where the room's moisture condenses. Where humid air meets a surface below its dew point, condensation can sustain mold growth in places that remain out of sight.
The frameworkRead moreRead less
The four are read together, through the building's skin and its circulation, because they act together, and because in a hot, humid climate a failure in any one of them ends in the same place: moisture on a surface that is too cold for the air around it. The condition we work toward is control: an interior that holds its standard. Readings shown are from the illustrated investigation, not universal targets; where a value is modelled rather than measured, it says so.
Method.
Three inputs. Every displayed number is calculated from them and from the assumed conditions below; the evidence type of every value is labelled in the model.
| Input | Range | Default | Evidence type | Why it is the lever |
|---|---|---|---|---|
| Indoor relative humidity | 30 to 90% | 55% | Input | Sets the dew point of the room air, the temperature below which any surface condenses. |
| Coldest surface temperature | 10 to 25 °C | insulated duct jacket, 21 °C | Assumed | Representative assumed surface temperatures: insulated duct jacket 21 °C · duct with failed insulation 15 °C · diffuser 17 °C · bare chilled pipe 10 °C · shared slab 22 °C · exposed wall 25 °C. A picker rather than a slider, because the point is to know where to look. These are not measured case values. |
| Room pressure relative to outdoors | −10 to +10 Pa | +2 Pa | Input · direction only | Sign identifies the transport path. Airflow quantity requires envelope leakage data, which the model does not hold. Pressure does not enter the condensation verdict. |
Outdoor temperature is held constant in this model to isolate humidity, surface temperature and air-transport direction. It still influences envelope surface temperature, heat flow and drying potential in a real building.
Interpretation bands Practice criterion
| Readout | Good | Elevated | High | Basis |
|---|---|---|---|---|
| Indoor RH | < 60% | 60 to 70% | > 70% | Indoor Sciences report benchmark of 60%, consistent with DM TG 142's mold-control recommendation of below 60%. |
| Surface RH | < 70% | 70 to 80% | ≥ 80%; 100% is condensation | Material-specific critical moisture levels of roughly 75 to 85% from field validation (Johansson, Svensson and Ekstrand-Tobin, 2013). The band signals conditions associated with mold risk; growth also depends on duration, material and temperature. |
| Condensation margin | > +3 K | 0 to +3 K | < 0 K, condensation predicted | Surface temperature minus dew point. Exactly 0 K is the condensation threshold. |
Standards and guidelines referenced
| Source and edition | Scope | Published criterion | How Indoor Sciences applies it |
|---|---|---|---|
| Dubai Municipality TG 142, Mold Remediation and Control, v1 | Mold control guidance | RH below 60%, ideally 30 to 50% | The directly relevant mold-control criterion; basis of the indoor RH band. |
| Dubai Municipality TG 141, Environmental Indoor Air Quality Index, v1 | Thermal-comfort index within EIAQI | 40 to 60% RH, "comfortable" band | A comfort band, not a mold threshold; cited for context only. |
| Dubai Municipality TG 119, Indoor Air Quality for Healthy Life, DM-HSD-GU119-IAQ v4 (11 Dec 2024) | General indoor air quality | 20 to 60% RH | Broad acceptability range; the practice reports against the stricter TG 142 criterion. |
| ASHRAE Standard 55-2023 | Thermal environmental conditions for human occupancy | humidity ratio ≤ 12 g/kg | Referenced for the humidity limit of the comfort zone. |
| ASHRAE Standard 62.1-2025 | Ventilation and acceptable indoor air quality, non-residential | 65% RH at the dehumidification design condition | Applied to hotels, campuses and workplaces. |
| ASHRAE Standard 62.2-2025 | Ventilation and acceptable indoor air quality, residential | residential ventilation and moisture provisions | Applied to villas and apartments in place of 62.1. |
Worked examples.
Eight states of the same room at 24 °C, calculated with the equations above. The surface temperatures are representative assumed values except in the fifth row, which is a real storage room from one of the practice's reports: 24 °C and 72.1% relative humidity measured, condensation observed on the duct, duct surface temperature not recorded and modelled at 15.8 °C.
| Coldest surface | Indoor RH | Dew point | Surface | Surface RH | Margin | Band |
|---|---|---|---|---|---|---|
| Insulated duct jacket, 21 °C | 55% | 14.4 °C | 21 °C assumed | 66% | +6.6 K | GOOD |
| Insulated duct jacket, 21 °C | 65% | 17.0 °C | 21 °C assumed | 78% | +4.0 K | ELEVATED |
| Insulated duct jacket, 21 °C | 72% | 18.6 °C | 21 °C assumed | 86% | +2.4 K | HIGH |
| Duct with failed insulation, 15 °C | 55% | 14.4 °C | 15 °C assumed | 96% | +0.6 K | HIGH |
| The storage room: duct with failed insulation | 72.1% | 18.7 °C | 15.8 °C modelled | 100%, at saturation | -2.9 K | HIGH |
| Bare chilled-water pipe, 10 °C | 45% | 11.3 °C | 10 °C assumed | 100%, at saturation | -1.3 K | HIGH |
| Shared slab, 22 °C | 72.1% | 18.7 °C | 22 °C assumed | 81% | +3.3 K | HIGH |
| Exposed wall, 25 °C | 72.1% | 18.7 °C | 25 °C assumed | 68% | +6.3 K | GOOD |
Read the first three rows together: the same insulated duct goes from GOOD to HIGH on humidity alone, with the surface never wet. Read the last two together: at the storage room's humidity, a shared slab at 22 °C already reads 81%, while the sun-facing wall at 25 °C reads 68%.
Questions the model answers.
Why does a supply duct sweat in the UAE?
Because it is often the coldest surface in the room. Supply air leaves the cooling coil at about 12 to 15 °C and reaches the diffuser a few degrees warmer, so a bare or poorly insulated duct can sit well below the dew point of ordinary room air: at 24 °C and 55% relative humidity the dew point is 14.4 °C, and at 72.1% it is 18.7 °C. Any surface below that temperature condenses moisture onto itself. An insulated jacket keeps the touchable surface near 21 °C in this model's assumption, which stays dry at 55% and reaches mold-risk humidity above about 65%.
At what indoor humidity do conditions associated with mold arise on a cold surface?
When the relative humidity at the surface itself, not in the room air, stays above roughly 75 to 85% for long enough. Field validation puts critical surface humidity at about 75 to 79% on pine, 79 to 85% on chipboard at 22 °C. The practice bands surface RH as GOOD below 70%, ELEVATED from 70 to 80%, and HIGH at 80% and above; 100% is condensation. Growth depends on duration, material and temperature as well as humidity.
Why is the exposed wall often not the problem in this climate?
Because under hot-season daytime conditions heat generally flows inward, so the inner face of a sun-facing wall is usually warmer than the room, not colder. With a representative surface of 25 °C against room air at 24 °C and 72.1% humidity, that wall reads 68% surface RH, GOOD. Condensation and mold conditions arise instead on what the air-conditioning has made cold: ducts, diffusers, chilled-water pipes, and slabs shared with a colder neighbor. Thermal bridges can create exceptions.
What does negative room pressure admit in summer?
Outdoor air at a dew point near 24 °C or more, through whatever leakage paths exist. Dubai's June to September outdoor dew point has a median near 23 °C and a 95th percentile near 28 °C, so a cooled interior at 24 °C sits below the outdoor dew point for roughly four hours in ten across the season. A room depressurized relative to outdoors draws that air onto surfaces at 24 °C or below. Direction identifies the transport path; the quantity moved depends on envelope leakage, which must be measured.
Limitations.
The model is a simplified psychrometric illustration, and its boundaries matter more than its polish.
- Time and material are not modelled. Mold growth depends on the duration a surface spends above its critical moisture level, on the material, and on temperature. The surface RH readout signals conditions associated with mold risk at an instant; it does not predict growth.
- Surface temperatures are assumed, not calculated. The six preset surfaces carry representative temperatures. Interior surface temperatures of exposed elements depend on construction, orientation and outdoor conditions; envelope U-values and a surface-resistance calculation (e.g. ISO 6946) are outside the scope of this simplified model, and would be needed to calculate them.
- The surface-RH calculation assumes the water-vapour partial pressure at the surface equals that of the room air. Local vapour gradients and boundary-layer effects are not modelled; at a surface with poor air movement the real value can differ.
- Pressure indicates direction only. No airflow or moisture load is calculated, because no verified UAE envelope-leakage coefficient is available to the practice. The amber stream is explanatory.
- Outdoor conditions are held constant. Outdoor temperature and dew point are fixed at assumed hot-season values to isolate the three inputs.
- Ground moisture is not quantified. It is a source in a real building and is not represented in the calculation.
- Air-conditioning behavior is not modelled. An oversized unit that short-cycles leaves humidity high; a specialist looks for it first. It would be the fourth control if one were added.
Climate data.
The assumed outdoor dew point of 24 °C is drawn from the Dubai International station record (WMO 41194) for June to September, 1995–2024, n = 84,959 hours: median 23.1 °C, 95th percentile 28.0 °C. 24 °C lies between the median and the 95th percentile and is exceeded in 39.6% of June–September hours. The record is Meteostat hourly bulk data, compiled from the NOAA NCEI Integrated Surface Database (CC BY 4.0). Years covered, observation frequency, missing-data treatment and the percentile method are set out in the technical note.
Download the technical note (PDF) for consultants and project teams.
Sources.
- Alduchov, O. A. and Eskridge, R. E. (1996). Improved Magnus form approximation of saturation vapor pressure. Journal of Applied Meteorology, 35(4), 601–609. doi:10.1175/1520-0450(1996)035<0601:IMFAOS>2.0.CO;2
- Meteostat. Hourly bulk data (hourly temperature and dew point), stations 41194 Dubai International (OMDB), 41217 Abu Dhabi International (OMAA), 41196 Sharjah International (OMSJ), 41184 Ras Al Khaimah International (OMRK), 41198 Fujairah International (OMFJ) and 41218 Al Ain International (OMAL), 1 January 1995 to 31 December 2024. Retrieved 31 Aug 2026 (41194) and 4 Sep 2026 (the other stations), one file per station from Meteostat's bulk endpoint
https://bulk.meteostat.net/v2/hourly/{WMO}.csv.gz, where {WMO} stands for the station number. Compiled by Meteostat from the NOAA NCEI Integrated Surface Database (Smith, Lott and Vose, below). Licence: Creative Commons Attribution 4.0 International (CC BY 4.0). The dew-point statistics on this site are computed by Indoor Sciences from these records; years, observations, missing data and the percentile method are set out in the technical note. - Smith, A., Lott, N. and Vose, R. (2011). The Integrated Surface Database: recent developments and partnerships. Bulletin of the American Meteorological Society, 92(6), 704–708.
- Johansson, P., Svensson, T. and Ekstrand-Tobin, A. (2013). Validation of critical moisture conditions for mould growth on building materials. Building and Environment, 62, 201–209. Full text
- Dubai Municipality. Technical Guideline for Mold Remediation and Control, TG 142, v1. Document
- Dubai Municipality. Environmental Indoor Air Quality Index, TG 141, v1. Document
- Dubai Municipality. Technical Guidelines for Indoor Air Quality for Healthy Life, TG 119, DM-HSD-GU119-IAQ, Version 4, 11 December 2024. Document
- ASHRAE Standards 55-2023, 62.1-2025 and 62.2-2025. Read-only versions, ASHRAE
- Indoor Sciences. Diagnostic report, private residence, UAE, 2026. Client reference withheld.
Related
The model tells you where to look. The practice tells you what it found.
An investigation measures the actual humidity, surface temperatures and, where the symptom pattern calls for it, the pressures in the interior you are responsible for, and returns a report written to be checked, not believed.