Interpreting Laboratory Results & Trends

Laboratory data become most useful when operators look beyond individual numbers and evaluate patterns over time. One result may describe a moment, but a trend can reveal whether treatment is improving, deteriorating, or responding to a process change.

Good interpretation combines laboratory results with operating conditions, equipment status, flow, loading, weather, chemical feed, and instrument data.

Start with the Process Question

Before reviewing data, ask what operational question needs to be answered.

Examples include:

• Is nitrification getting worse?
• Is filter performance changing?
• Is solids separation deteriorating?
• Is chlorine demand increasing?
• Is phosphorus removal improving?

Know the Normal Baseline

A baseline is the normal range or pattern expected when the process is operating properly.

A useful baseline may include:

• average value;
• normal minimum and maximum;
• typical daily variation;
• seasonal pattern;
• normal relationship with other variables.

One Number Is Not a Trend

A single result can be affected by:

• normal process variation;
• sampling variability;
• analytical error;
• temporary loading change.

Repeated results showing the same direction provide stronger evidence of a real process change.

Trend Direction

A trend may be:

• increasing;
• decreasing;
• stable;
• cyclic;
• highly variable.

The direction should be interpreted according to the parameter.

For example, increasing effluent ammonia may be unfavorable, while increasing chlorine residual after a feed correction may be expected.

Rate of Change

Operators should consider not only the direction of change but also how quickly it is occurring.

A simple rate-of-change calculation is:

Rate of Change = New Value - Previous Value

Percent Change

A useful comparison is:

Percent Change = (New Value - Old Value) ÷ Old Value × 100

Percent Change Example

Effluent ammonia increases from 2 mg/L to 5 mg/L.

Percent Change = (5 - 2) ÷ 2 × 100

Percent Change = 150%

The absolute increase is 3 mg/L, but the relative increase is large and may deserve prompt investigation.

Average

An average can summarize a group of measurements.

A simple arithmetic average is:

Average = Sum of Values ÷ Number of Values

Average Example

Suppose daily effluent TSS values are:

• 8 mg/L;
• 10 mg/L;
• 12 mg/L.

Average = (8 + 10 + 12) ÷ 3 = 10 mg/L

Average Can Hide Important Events

An acceptable average does not prove that every individual result was acceptable or that the process was stable.

Always review:

• individual values;
• maximum values;
• minimum values;
• variability.

Minimum and Maximum

Minimum and maximum values can reveal process extremes that are hidden by an average.

For example, an average DO of 2.0 mg/L could result from stable operation near 2.0 mg/L or from repeated swings between very low and very high DO.

Rolling Average

A rolling average uses a moving group of recent results.

For example, a seven-day rolling average uses the most recent seven days and updates as each new result becomes available.

Rolling averages can smooth short-term noise and reveal longer-term direction.

Do Not Smooth Away Important Problems

Rolling averages are useful for trends but can hide short-duration spikes.

Operators should review both:

• raw data;
• smoothed trend.

Variability

A process can become unstable before its average changes significantly.

Increasing variability may indicate:

• poor control;
• intermittent loading;
• equipment cycling;
• instrument problems;
• process instability.

Compare Similar Time Periods

Wastewater and drinking-water processes may show daily or seasonal patterns.

Useful comparisons may include:

• same hour on different days;
• weekday versus weekend;
• summer versus winter;
• wet weather versus dry weather.

Seasonal Trends

Seasonal changes can affect:

• temperature;
• source-water quality;
• biological reaction rates;
• oxygen solubility;
• chemical demand;
• algae.

A result that is unusual in winter may be normal in summer.

Flow Matters

Concentration data should often be reviewed together with flow.

A lower concentration can still represent a higher mass load when flow increases.

Mass Loading

A common wastewater calculation is:

Load, lb/day = Flow, MGD × Concentration, mg/L × 8.34

Loading Example

Day 1:

• Flow = 2 MGD;
• BOD = 200 mg/L.

Load = 2 × 200 × 8.34 = 3,336 lb/day

Day 2:

• Flow = 4 MGD;
• BOD = 150 mg/L.

Load = 4 × 150 × 8.34 = 5,004 lb/day

Although BOD concentration decreased, total BOD loading increased.

Normalize Data When Appropriate

Normalization allows more meaningful comparison when operating conditions change.

Examples include:

• mass per day;
• chemical use per million gallons;
• energy use per unit flow;
• solids produced per unit load.

Correlations

A correlation means two variables change in a related way.

Examples may include:

• increasing ammonia with decreasing DO;
• increasing flow with increasing effluent TSS;
• increasing turbidity with filter head loss;
• increasing temperature with changing biological activity.

Correlation Does Not Prove Cause

Two variables can move together without one directly causing the other.

Operators should use:

• process knowledge;
• additional measurements;
• equipment information;
• controlled observations

before deciding on cause and effect.

Process Lag

A process response may occur after a delay.

This is called process lag.

Examples include:

• influent load reaching an aeration basin later;
• chemical adjustment affecting downstream water later;
• source-water change appearing at the plant after travel time;
• distribution-system change appearing hours or days after treatment.

Detention Time and Data Interpretation

When comparing upstream and downstream data, account for hydraulic detention time.

The effluent measured at 10:00 a.m. may not correspond to influent collected at 10:00 a.m.

Example of Process Lag

If a basin has approximately six hours of detention time, an influent upset occurring at noon may not appear at the downstream sampling point until several hours later.

Use Related Parameters Together

One parameter often becomes more meaningful when interpreted with another.

Useful combinations include:

• DO and ammonia;
• pH and alkalinity;
• BOD and TSS;
• MLSS and settleability;
• chlorine dose and residual;
• filter turbidity and head loss;
• flow and concentration.

DO and Ammonia Example

Rising effluent ammonia together with falling DO can support an oxygen-limitation hypothesis.

If ammonia rises while DO remains adequate, operators should look for other causes such as:

• low temperature;
• low pH;
• low alkalinity;
• toxicity;
• insufficient solids retention time.

pH and Alkalinity Example

Falling pH combined with falling alkalinity can indicate loss of buffering capacity.

Stable pH with declining alkalinity may still be important because the remaining buffer is being consumed.

BOD and TSS Example

If effluent BOD and TSS rise together, solids carryover may be contributing to the higher BOD.

If BOD rises while TSS remains normal, incomplete soluble organic removal may deserve greater attention.

Chlorine Dose and Residual

If chlorine dose increases but residual decreases, chlorine demand may have increased.

Possible reasons include:

• higher organic matter;
• ammonia;
• reduced compounds;
• source-water change.

Outliers

An outlier is a value that differs substantially from surrounding data.

An outlier may represent:

• a real process event;
• sampling error;
• analytical error;
• instrument failure;
• data-entry error.

Do Not Automatically Delete Outliers

An unusual result should be investigated rather than removed simply because it does not fit the trend.

Investigating an Outlier

Review:

• sample ID;
• sample location;
• collection time;
• laboratory QA/QC;
• instrument calibration;
• process conditions;
• equipment alarms;
• other related measurements.

Bad Data Can Look Like a Process Problem

Examples include:

• fouled DO probe suggesting low oxygen;
• dirty turbidity cell suggesting poor filtration;
• incorrect dilution factor producing a false high laboratory result;
• wrong sample label making two locations appear reversed.

Real Process Problems Can Look Like Bad Data

Do not dismiss unusual results merely because they are surprising.

A real shock load, equipment failure, chemical spill, or hydraulic event may produce values outside the historical range.

Use Independent Evidence

A result is more credible when independent evidence supports it.

Examples include:

• high effluent ammonia plus declining nitrate production;
• high TSS plus visible clarifier carryover;
• high conductivity plus known industrial discharge;
• high turbidity plus poor filter performance.

Control Limits

Facilities may establish operating ranges or internal control limits for important process parameters.

These limits can help identify changes before a regulatory or process failure occurs.

Operating Limit Versus Regulatory Limit

An internal operating target is not the same as a regulatory limit.

An operator may take action well before a compliance limit is approached.

Warning Limits

A practical control system may use:

• normal range;
• warning range;
• action range.

This helps operators respond before the process becomes unstable.

Control Charts

A control chart plots measurements over time against expected limits or statistical boundaries.

It can help identify:

• gradual drift;
• sudden shift;
• increasing variability;
• repeated unusual patterns.

Trend Before Limit Exceedance

A process may show deterioration before a formal limit is exceeded.

Examples include:

• ammonia slowly rising;
• filter turbidity gradually increasing;
• residual chlorine declining;
• SVI increasing over several days.

Data Frequency Matters

A parameter measured once per week cannot reveal short-term variation as well as a continuously monitored parameter.

Interpret trends according to:

• sampling frequency;
• process response time;
• importance of the parameter.

Do Not Compare Unequal Data Carelessly

Before comparing results, confirm:

• same units;
• same analytical method;
• same sample type;
• same location;
• similar timing.

Method Changes

A change in analytical method can create a shift in results even when the process itself has not changed.

Document method changes so trend interpretation remains meaningful.

Instrument Replacement

Replacing or recalibrating an instrument can also create an apparent trend shift.

Operators should document:

• instrument changes;
• calibration changes;
• sensor replacement.

Process Changes Must Be Marked on Trends

Useful trend charts should identify important events such as:

• chemical-feed change;
• new equipment startup;
• maintenance outage;
• storm event;
• process setpoint change.

Before-and-After Comparison

When evaluating an operational change:

1. establish the baseline;
2. make one controlled change where practical;
3. allow sufficient process response time;
4. compare post-change data with the baseline;
5. confirm that other major conditions did not change simultaneously.

Avoid Changing Too Many Variables at Once

If several operating settings change simultaneously, it becomes difficult to determine which change caused the result.

Short-Term Versus Long-Term Trends

Short-term trends are useful for:

• rapid process control;
• equipment problems;
• shock loads.

Long-term trends are useful for:

• seasonal patterns;
• asset deterioration;
• chemical consumption;
• process optimization.

Visual Inspection Still Matters

Laboratory data should be combined with field observations such as:

• clarifier appearance;
• filter condition;
• foam;
• odor;
• sludge blanket;
• color;
• equipment sound.

Data Entry Errors

Trend charts can be distorted by incorrect data entry.

Common errors include:

• wrong decimal point;
• wrong unit;
• wrong date;
• wrong sampling location;
• transposed digits.

Example of a Decimal Error

A result of 2.5 mg/L entered as 25 mg/L can appear to be a major process upset.

Unexpected values should be checked against original laboratory records.

Missing Data

Missing data should not automatically be replaced with estimated values unless the applicable procedure specifically allows it.

Missing results should remain clearly identifiable.

Non-Detect Results

A non-detect should not automatically be entered as zero unless the applicable data procedure specifically requires that treatment.

The reporting threshold should be considered during trend analysis.

Data Resolution

Do not interpret differences smaller than the practical resolution or precision of the method as meaningful process changes.

For example, a change from 2.01 to 2.02 may not be operationally significant if the method uncertainty is much larger.

Operational Significance

A statistically visible change is not always operationally important.

Ask:

• Does the change affect treatment?
• Does it approach an operating limit?
• Is it persistent?
• Does another parameter confirm it?

Prioritize Trends by Risk

Give greater attention to trends that can affect:

• public health;
• permit compliance;
• treatment stability;
• critical equipment;
• chemical safety.

Example: Nitrification Trend

Suppose over several days:

• effluent ammonia rises;
• alkalinity falls;
• pH falls slightly;
• DO remains adequate.

This pattern may point toward insufficient alkalinity rather than oxygen limitation.

Example: Clarifier Trend

Suppose:

• SVI increases;
• sludge blanket rises;
• effluent TSS begins increasing.

Together, these results provide stronger evidence of worsening solids separation than any one result alone.

Example: Filter Trend

Suppose:

• filter turbidity slowly increases;
• head loss rises;
• filter run time approaches its usual endpoint.

The combined trend may indicate the filter is approaching the need for backwash.

Example: Increasing Chlorine Demand

Suppose:

• chlorine dose remains constant;
• residual gradually decreases;
• source-water organic matter increases.

This pattern is consistent with increasing chlorine demand.

Example: Wet-Weather Wastewater Trend

During a storm:

• flow increases sharply;
• influent concentration may decrease because of dilution;
• total hydraulic and mass loading can still change significantly;
• clarifier performance may deteriorate because of hydraulic loading.

Concentration alone would not describe the entire event.

Keep Trend Displays Simple

Useful trend charts should:

• use clear units;
• show time consistently;
• avoid unnecessary variables on one chart;
• identify important events;
• use appropriate scales.

Do Not Manipulate Chart Scale to Hide Problems

Changing the vertical scale can make a large change look small or a small change look dramatic.

Trend displays should support objective interpretation.

Document Operator Response

When data lead to an operational change, record:

• what trend was observed;
• what action was taken;
• when the action occurred;
• what result followed.

Build Process Knowledge Over Time

Historical data help operators learn how their specific facility responds to:

• weather;
• loading;
• chemical changes;
• equipment changes;
• seasonal conditions.

Common Data-Interpretation Mistakes

• Reacting to one result without checking the trend.
• Looking only at averages and ignoring extremes.
• Comparing concentration without considering flow and loading.
• Ignoring process detention time.
• Assuming correlation proves causation.
• Deleting outliers without investigation.
• Ignoring laboratory QA/QC.
• Comparing results with different units or methods.
• Making several process changes at once.
• Ignoring instrument or sampling changes that affect the trend.
• Treating a non-detect as an exact zero without considering the reporting basis.
• Using chart scales that distort the significance of changes.

A Practical Trend-Review Sequence

1. Define the process question.
2. Confirm the parameter, units, sample location, and method.
3. Establish the normal baseline.
4. Review raw values, average, minimum, maximum, and variability.
5. Review the direction and rate of change.
6. Compare concentration with flow and mass loading where relevant.
7. Review related parameters.
8. Account for detention time and process lag.
9. Review equipment, chemical-feed, weather, and operating changes.
10. Investigate outliers and laboratory QA/QC.
11. Make a controlled operational response when justified.
12. Continue trending to verify the response.

A Practical Unexpected-Result Review

1. Check whether the value is physically reasonable.
2. Confirm the sample ID, location, and collection time.
3. Confirm units and reporting basis.
4. Review laboratory QA/QC and instrument calibration.
5. Compare with previous results.
6. Compare with related process measurements.
7. Review alarms, equipment status, and recent operating changes.
8. Repeat or verify the result when appropriate.
9. Document the final interpretation.

What to Remember for the Exam

• A trend is more informative than one isolated result.
• A baseline describes the normal operating range or pattern.
• Rate of change can be as important as the absolute value.
• Averages can hide high, low, and unstable individual results.
• Rolling averages smooth short-term variation but can hide spikes.
• Increasing variability can indicate process instability even when the average remains unchanged.
• Concentration and mass loading are different, and flow must be considered when evaluating load.
• Normalized data can improve comparison when flow or operating conditions change.
• Correlation does not prove that one variable caused another.
• Process lag and detention time must be considered when comparing upstream and downstream results.
• Related measurements should be interpreted together.
• Outliers can represent real process events or bad data and should be investigated.
• Laboratory QA/QC, calibration, sample information, and process data help determine whether an unusual result is credible.
• Internal operating limits can provide warning before regulatory or process limits are reached.
• Control charts and trends can reveal drift, shifts, and increasing variability.
• Changes in analytical method, instrument, sampling location, or process operation should be documented on trend records.
• Do not make many process changes at once when trying to determine cause and effect.
• Trend interpretation should combine laboratory results with field observations and equipment status.
• Data-entry errors, wrong units, and decimal mistakes can create false process trends.
• Good operators use data to detect change early, make controlled adjustments, and verify the process response.