Abstract
This study analyzed 36 months of meteorological data alongside mould incidence records from 200 Singapore properties to identify key weather parameters affecting indoor mould growth. Humidity showed strongest correlation (r=0.84), followed by consecutive rainy days (r=0.72) and temperature differential (r=0.68). A predictive model achieving 78% accuracy was developed for mould risk forecasting.
Weather Parameter Correlations
| Parameter | Correlation (r) | P-value |
|---|---|---|
| Ambient humidity (%) | 0.84 | <0.001 |
| Consecutive rainy days | 0.72 | <0.001 |
| Indoor-outdoor temp differential | 0.68 | <0.001 |
| Monthly rainfall (mm) | 0.61 | <0.001 |
| Average temperature | 0.34 | <0.01 |
| Wind speed | -0.28 | <0.05 |
Mould Risk Thresholds
- Low Risk: Humidity <75%, no rain for 3+ days
- Moderate Risk: Humidity 75-85%, 1-2 consecutive rainy days
- High Risk: Humidity >85%, 3+ consecutive rainy days
- Critical: Humidity >90%, 5+ rainy days, indoor temp >8°C below ambient
Predictive Model
A machine learning model was trained on the collected data achieving:
- 78% accuracy in predicting mould onset within 7 days
- 85% accuracy for 14-day predictions
- Highest performance during monsoon transitions
This model powers our Mould Alert system for proactive customer notifications.