Monitoring Survey Design, Sampling, and Data Quality
Domain 3: Monitoring Survey Design, Sampling, and Data Quality
This domain tests whether you can design a survey that answers a regulatory or impact question rather than merely collecting convenient readings. Expect items on site selection, duration, representativeness, meteorology, and validation.
Core rules
- Start with a clear objective and data quality objectives: what limit, metric, averaging time, location, and decision threshold apply.
- Select receptors and reference positions that match the permit or standard, such as a property line or a specified facade. Include background positions and, where relevant, source-off measurements.
- Sample across representative operating conditions: weekday and weekend, day, evening, and night, and the source's normal duty cycle. A short sample during an atypical lull is not representative.
- Log continuously at a defined interval, commonly 1 s or 1 min, and retain raw data. For a period Leq, energy-average the valid samples; do not average dBA values arithmetically.
- Record wind speed and direction, precipitation, temperature, and humidity. Avoid measurement when wind noise or rain affects the microphone.
- Validate data for clipping, extraneous events, wind noise, and calibration drift; document every exclusion.
Worked example
A one-hour survey yields 60 one-minute Leq values. Six minutes are invalid because of wind noise, leaving 54 valid samples. The hourly Leq is 10 log10[(1/54) sum 10^(Li/10)] over the valid samples. Excluding invalid minutes without documenting why, or simply averaging the 54 dBA numbers, changes the result and weakens the evidence.
Traps
Measuring only during a quiet period; ignoring background; using too few samples; failing to record source operating state; and treating a logged data file as self-validating.