7SSGN223 Chap.4 Environmental Monitoring and Data Quality
Environmental Monitoring and Data Quality
Begin with the decision
A monitoring programme should be designed backwards from the decision it needs to inform. Define the claim, choose a variable and unit, set spatial and temporal coverage, specify quality controls and state the result that will trigger action.
Collecting frequent data without a decision link can create detail without relevance.
Match scale to inference
One sensor can reveal timing at one location, while a wide network describes spatial pattern with different resolution and maintenance demands. Personal exposure, river-reach condition and a national indicator require different sampling frames.
State the population of places, times or people that the design can represent.
Separate precision and accuracy
A precise instrument can repeat the same biased value. Calibration against a reference, co-location, duplicates, blanks, maintenance records and metadata help distinguish environmental change from instrument change.
A method break should be marked and tested before a trend is interpreted.
Audit the denominator
Ambient water-quality indicator 6.3.2 reports the proportion of monitored water bodies with good quality. Its denominator is monitored water bodies, not every water body. Coverage by region and type therefore matters.
An expanded network can lower the percentage by adding difficult sites even if continuing sites do not deteriorate.
Compare panels and networks
A continuing panel holds sites relatively constant and can show change at established locations. A full network describes all monitored sites in each period and may shift as coverage expands. Report both when possible.
If the full-network percentage falls while the continuing panel is stable, new sites may reveal previously unobserved problems rather than deterioration everywhere. That distinction changes which investigation or intervention is justified.
Treat gaps as evidence
Missing observations may cluster during floods, equipment failures or at remote locations.
Document why values are absent and whether conclusions survive plausible values for those gaps. Preserve site, time, method, calibration and processing metadata so a method break remains visible. Quality assurance should state what each check detects, how failure is handled and who acts when the monitored result crosses the decision threshold.
What this chapter covers
- 01
Decision Question
- 02
Variable and Unit
- 03
Sampling Frame
- 04
Calibration and Quality
- 05
Indicator Denominator
- 06
Action Trigger
Worked example · free
Compute a monitored water share
- 2Calculate the original monitored percentage.
- 3Update numerator and denominator.
- 3Separate coverage change from deterioration.
Key terms
- Sampling Frame
- A Sampling Frame is the defined set of places, times or cases from which observations can enter an estimate.
- Calibration
- Calibration compares an instrument response with a reference so systematic measurement error can be detected or corrected.
- Data Coverage
- Data Coverage describes how completely observations represent the locations, periods or groups named in the claim.
Environmental Monitoring and Data Quality FAQ
Why does monitoring design start with a decision?
The decision determines which variable, scale, frequency and accuracy matter. Starting with available equipment can produce a detailed series that does not answer the policy question. A decision-first design also identifies the threshold or pattern that should trigger action.
What can cause a false environmental trend?
Sensor replacement, relocation, calibration drift, changed sampling season, missing sites and revised classification rules can all create a break. Review metadata and a continuing comparison series before attributing the pattern to environmental change or policy impact.
What does a monitored water percentage represent?
It represents the monitored water bodies that enter the indicator under its stated method. Its relevance to all water bodies depends on coverage and selection. Report water-body type, geography and missing areas, especially when the network changes between periods.
Why compare a continuing panel with the full network?
The continuing panel shows change at sites measured compatibly in both periods, while the full network describes current monitored coverage. If they diverge, additions, removals or method changes may explain the aggregate shift. Reporting both prevents broader observation from being mistaken automatically for environmental deterioration.
Assessment move
Take one policy question and design the complete loop: response variable, unit, sites, timing, calibration, missing-data rule and action trigger. Compute a percentage with the numerator and denominator printed. Then add new sites and explain why the aggregate changes.
Practise distinguishing higher precision from better validity, and write the exact metadata check that would separate a sensor break from a real environmental trend. Build a continuing-panel result and a full-network result from the same fictional data, then explain why they answer different questions. Mark observations missing during an extreme event and test two plausible values rather than deleting them silently.
Create a quality-assurance table in which every calibration, blank or duplicate has a stated purpose and response. Close by assigning responsibility for the policy trigger and identifying which method change would require a break in the time series.