UAE searches for "mold" do not rise and fall with the humid season; a borderline link with more humid than normal months (r = +0.30) does not hold in every check
Monthly Google Trends interest in "mold" in the UAE set against the dew point of the Dubai ERA5-Land grid cell, which passed Datalab's station check. Two tests: through the seasons, and month to month with the seasons removed. The results are reported as they came out.
Over 50 months from November 2021 to December 2025, Google search interest in "mold" in the UAE and the monthly mean dew point of the Dubai model cell correlate at +0.14 (95% interval −0.30 to +0.49): no clear link through the seasons. Searches are highest in November and lowest in June; dew point is highest in July and lowest in January. With the seasons removed, months more humid than their 1991-2020 normal go with higher search interest, r = +0.30 (95% interval +0.01 to +0.54): a clear positive link, but the checks with the mean of 4 station-checked UAE cells (+0.25, −0.04 to +0.52) and with the spelling "mould" (+0.30, −0.04 to +0.57) include zero. Figures for journalists
Google Trends values are relative (0-100) to the peak of the query in the time range and place; they are not search volumes. The series counts every search for "mold", whatever the reason, and cannot be split by reason. A correlation does not show that humidity causes searches. Only the UAE has enough weekly data; Qatar "mold" (61.2%), Kuwait "mold" (46.9%), Bahrain "mold" (16.4%) and UAE عفن (Arabic) (5%) were tested and left out.
The test results
Pearson correlation (r) between monthly series, with 95% intervals. A link is clear only when the interval excludes zero.
| Test | Months | r | 95% interval | Reading |
|---|---|---|---|---|
| "mold" vs dew point, monthly levels (seasons included) | 50 | +0.14 | −0.30 to +0.49 | no clear link |
| "mold" vs dew point, seasons removed | 50 | +0.30 | +0.01 to +0.54 | a clear positive link |
| Seasons removed, searches one month later | 49 | +0.27 | −0.13 to +0.56 | no clear link |
| Check: mean of 4 station-checked UAE cells, levels | 50 | +0.14 | −0.33 to +0.48 | no clear link |
| Check: mean of 4 station-checked UAE cells, seasons removed | 50 | +0.25 | −0.04 to +0.52 | no clear link |
| Check: spelling "mould", levels | 50 | +0.19 | −0.15 to +0.46 | no clear link |
| Check: spelling "mould", seasons removed | 50 | +0.30 | −0.04 to +0.57 | no clear link |
"mold" in the UAE, month by month
Monthly mean of the weekly Google Trends index (0-100, relative to the query's own peak week).
- mold, UAE Dec 2025 55.2
Dew point in the Dubai model cell, month by month
Monthly mean of daily mean 2 m dew point (C), ERA5-Land.
- Dew point, Dubai Dec 2025 14.1°C
The shape of a year
Searches: each month as % of the series' linear trend, averaged by calendar month (average month = 100). Dew point: 1991-2020 mean of the Dubai cell. Correlation of the two 12-month profiles: −0.06 (12 points, no interval).
| Month | "mold" searches (average month = 100) | Dew point, 1991-2020 (°C) |
|---|---|---|
| January | 100.3 | 12 |
| February | 107.7 | 12.2 |
| March | 102.1 | 13.2 |
| April | 93.6 | 14.6 |
| May | 91.4 | 17.2 |
| June | 86.3 | 20.8 |
| July | 94.5 | 22.6 |
| August | 107.1 | 22.5 |
| September | 106.1 | 21.6 |
| October | 107.5 | 19.3 |
| November | 107.8 | 16.1 |
| December | 95 | 13.8 |
Which search series carry enough data
Weeks with a non-zero value, November 2021 to October 2026. A Trends series is analysed only when at least 90% of its weeks from November 2021 to October 2026 are non-zero (batch-1 rule).
| Market and search term | Non-zero weeks | Coverage | Status |
|---|---|---|---|
| UAE: "mold" | 260 of 260 | 100% | analysed |
| UAE: "mould" | 260 of 260 | 100% | analysed |
| Qatar: "mold" | 159 of 260 | 61.2% | too sparse, not analysed |
| Kuwait: "mold" | 123 of 262 | 46.9% | too sparse, not analysed |
| Bahrain: "mold" | 43 of 262 | 16.4% | too sparse, not analysed |
| UAE: عفن (Arabic) | 13 of 262 | 5% | too sparse, not analysed |
Data provenance
- Station check
- Dubai International Airport (WMO 41194), 10.9 km: August mean dew point 22.5°C in the cell against 23.4°C at the station (−0.9°C). Rule: |August mean dew point difference| <= 1.0C (station dew point from its 1991-2020 mean vapour pressure, Magnus inverse). Cells that pass in the UAE: Dubai, Sharjah, Ajman and Ras Al Khaimah.
- Months
- Weekly Trends values averaged to calendar months (a week belongs to the month of its midpoint). Dew point: mean of daily means, months with at least 25 days. The window ends in December 2025 because the ERA5-Land series on disk ends on 2025-12-31.
- Seasons removed
- Month-to-month test: dew point as the difference from the 1991-2020 mean of the same calendar month in the same grid cell; searches as % of the series' own linear trend over the window, minus the mean of the same calendar month. Both remove the seasonal cycle, so a month-to-month link cannot come from the seasons alone.
- Uncertainty
- 95% intervals by moving-block bootstrap (5000 resamples, blocks of 6 consecutive months, fixed seed 20261114). A link is called clear only when the interval excludes zero. The 12-month profile correlation has 12 points that are not independent; no interval is given for it.
- Attribution
- Contains modified Copernicus Climate Change Service information 2026. Neither the European Commission nor ECMWF is responsible for any use that may be made of the Copernicus information or data it contains.
- Licence
- CC BY 4.0
- Generated by
- batch2/scripts/gen-datalab-b7-insights.py
- Dataset
- mould-searches-humidity.csv
For journalists
- Through the seasons, UAE search interest in "mold" and the Dubai dew point correlate at +0.14 (95% interval −0.30 to +0.49, 50 months, November 2021 to December 2025): no clear link.
- "mold" searches in the UAE are highest in November and lowest in June; the Dubai dew point is highest in July (1991-2020 mean 22.6°C).
- With the seasons removed, the correlation is +0.30 (95% interval +0.01 to +0.54); with the mean of 4 station-checked UAE cells it is +0.25 (−0.04 to +0.52).
- Interest in "mold" in the UAE rose by about 11% a year over the window (linear trend of the relative index).
- Google Trends figures are relative indices (0 to 100), not search counts.
Datalab by Laundrykhalas (2026). Searches for "mold" in the UAE against dew point, 2021-2025. https://laundrykhalas.com/datalab/insights/mould-searches-humidity/
Mould itself, or why people search: the series counts every search for the word and cannot be split by reason. Whether humidity affects searches: correlation cannot show cause, and a result of no clear link does not rule out a weak one. The dew point is outdoor air at 2 m; indoor conditions are not measured here.
Method
Weekly Google Trends values for "mold" in the UAE ("past 5 years", one query, captured 13 November 2026) were averaged to months and set against the monthly mean dew point of the ERA5-Land grid cell for Dubai, the UAE cell that passed Datalab's station dew-point check. Weekly Trends values averaged to calendar months (a week belongs to the month of its midpoint). Dew point: mean of daily means, months with at least 25 days. The window ends in December 2025 because the ERA5-Land series on disk ends on 2025-12-31. Month-to-month test: dew point as the difference from the 1991-2020 mean of the same calendar month in the same grid cell; searches as % of the series' own linear trend over the window, minus the mean of the same calendar month. Both remove the seasonal cycle, so a month-to-month link cannot come from the seasons alone. 95% intervals by moving-block bootstrap (5000 resamples, blocks of 6 consecutive months, fixed seed 20261114). A link is called clear only when the interval excludes zero. The 12-month profile correlation has 12 points that are not independent; no interval is given for it. The tests were rerun with the mean of the 4 UAE cells that passed the check (Dubai, Sharjah, Ajman and Ras Al Khaimah) and with the British spelling "mould".
Related Datalab data: dew point in Gulf cities since 1951 and UAE search demand.