11.3 Data Quality Objectives (DQOs), Analytical Methods & PARCC Parameters

Data Quality Objectives (DQOs), Analytical Methods & PARCC Parameters

Environmental sampling and analytical data form the legal, technical, and financial foundation for multi-million-dollar remediation projects, Superfund delistings, RCRA corrective actions, and toxic tort litigation. To ensure environmental data are scientifically valid, defensible, and of known quality, the U.S. Environmental Protection Agency (EPA) established structured frameworks: the 7-Step Data Quality Objectives (DQO) Process, Quality Assurance Project Plans (QAPPs), the PARCC Data Quality Indicators, and EPA SW-846 Standard Test Methods.

A Certified Hazardous Materials Manager (CHMM) must know how to establish tolerable decision error limits, calculate statistical precision and accuracy, interpret detection limits (MDL vs. PQL), and select appropriate SW-846 laboratory analytical methods.

---

1. The EPA 7-Step Data Quality Objectives (DQO) Process (EPA QA/G-4)

The DQO process is a systematic, iterative planning protocol developed by the EPA (guidance document EPA QA/G-4) that ensures the type, quantity, and quality of environmental data collected are appropriate for their intended decision-making purpose.

+-----------------------------------------------------------------------------------------+
|                        THE EPA 7-STEP DQO PLANNING PROCESS                              |
|                                                                                         |
|   [STEP 1] STATE THE PROBLEM                                                            |
|   - Assemble scoping team, define conceptual site model (CSM), identify budget/deadlines|
|                                     |                                                   |
|                                     v                                                   |
|   [STEP 2] IDENTIFY THE GOAL OF THE STUDY                                               |
|   - Specify principal study question, identify alternative management actions           |
|                                     |                                                   |
|                                     v                                                   |
|   [STEP 3] IDENTIFY INFORMATION INPUTS                                                  |
|   - Determine data sources, required basis for action levels, sampling/analytical methods|
|                                     |                                                   |
|                                     v                                                   |
|   [STEP 4] DEFINE THE BOUNDARIES OF THE STUDY                                           |
|   - Specify spatial boundaries (horizontal/vertical DUs) and temporal boundaries/seasons|
|                                     |                                                   |
|                                     v                                                   |
|   [STEP 5] DEVELOP THE ANALYTIC APPROACH (DECISION RULE)                                |
|   - Define parameter of interest (mean, 95% UCL) & "If... Then..." decision logic      |
|                                     |                                                   |
|                                     v                                                   |
|   [STEP 6] SPECIFY PERFORMANCE OR ACCEPTANCE CRITERIA                                   |
|   - Set tolerable limits on Type I (alpha) and Type II (beta) decision errors; Gray Reg |
|                                     |                                                   |
|                                     v                                                   |
|   [STEP 7] DEVELOP THE PLAN FOR OBTAINING DATA                                          |
|   - Optimize sampling design (simple random, stratified, systematic grid, ISM)          |
+-----------------------------------------------------------------------------------------+

Step-by-Step Breakdown:

Step 1: State the Problem

Establish a multi-disciplinary scoping team, define the environmental problem, summarize existing site data, construct an initial Conceptual Site Model (CSM), and identify regulatory programs and financial/schedule constraints.

Step 2: Identify the Goal of the Study

Formulate the primary study question (e.g., "Does lead concentration in surficial soil exceed the residential direct-contact screening level?") and identify the resulting alternative actions (e.g., "If yes, excavate and dispose off-site; if no, leave in place without institutional controls").

Step 3: Identify Information Inputs

Identify required data types, field measurement parameters, laboratory analytical methods (SW-846), appropriate sample matrices, and regulatory benchmark values.

Step 4: Define the Boundaries of the Study

Define the geographic boundaries of the site, vertical depth intervals (e.g., $0\text{--}6,\text{inches}$ for direct contact, $0\text{--}15,\text{feet}$ for construction worker exposure), temporal boundaries (e.g., seasonal high water table), and discrete Decision Units (DUs).

Step 5: Develop the Analytic Approach (The Decision Rule)

Combine the statistical parameter of interest with the action level into an unambiguous "If... Then..." decision rule:

"If the true 95% Upper Confidence Limit (95% UCL) of the mean concentration of trichloroethene in groundwater exceeds the Safe Drinking Water Act Maximum Contaminant Level (MCL) of $5,\mu\text{g/L}$, then initiate pump-and-treat remedial action; otherwise, continue semi-annual monitoring."

Step 6: Specify Performance or Acceptance Criteria (Decision Error Analysis)

Because environmental data contain inherent sampling and measurement variability, decisions are subject to statistical uncertainty. DQO Step 6 establishes tolerable limits for two distinct decision errors:

• Type I Error (False Positive / $\alpha$ Error): The decision maker concludes that contamination exceeds the action level when, in reality, it does not. (Consequence: Unnecessary, expensive remediation of clean soil/water).
• Type II Error (False Negative / $\beta$ Error): The decision maker concludes that contamination is below the action level when, in reality, it exceeds it. (Consequence: Failure to remediate hazardous contamination, resulting in unacceptable human health or ecological exposure).
• The Gray Region: A range of concentrations immediately adjacent to the action level where the consequences of a decision error are relatively minor. The boundary where $\beta$ error is specified (often set at $80%\text{ to }90%$ of the action level).

+-----------------------------------------------------------------------------+
|                        DECISION ERROR MATRIX (STEP 6)                       |
|                                                                             |
|   TRUE SITE CONDITION   | DECISION: SITE IS CLEAN  | DECISION: SITE EXCEEDS |
|   ----------------------+--------------------------+----------------------- |
|   Site is CLEAN         | Correct Decision         | TYPE I ERROR (alpha)   |
|   (Below Action Level)  | (1 - alpha)              | (False Rejection)      |
|                         |                          | [Unnecessary Cleanup]  |
|   ----------------------+--------------------------+----------------------- |
|   Site is CONTAMINATED  | TYPE II ERROR (beta)     | Correct Decision       |
|   (Above Action Level)  | (False Acceptance)       | (1 - beta = Power)     |
|                         | [Unmitigated Risk/Harm]  |                        |
+-----------------------------------------------------------------------------+

Step 7: Develop the Plan for Obtaining Data

Select the most cost-effective sampling design (simple random, stratified, systematic grid, or incremental sampling methodology) and calculate the required sample size ($N$) to achieve the error tolerances established in Step 6.

---

2. Quality Assurance Project Plans (Current EPA QAPP Standard)

For environmental information operations conducted by or for EPA, the QAPP documents the project objectives, data generation, assessment, and data review needed to produce information of known quality. The current governing EPA standard is CIO 2105-S-02.0, Quality Assurance Project Plan (QAPP) Standard (July 2023), supported by EPA guidance updated in 2025. It replaced the former QA/R-5 requirements. Many agencies and legacy templates still organize content into the four useful groups below, but the project must follow the current EPA standard and the applicable organization-specific requirements rather than treating the old four-group outline as universally mandatory:

+-----------------------------------------------------------------------------+
|               COMMON QAPP CONTENT ORGANIZATION (LEGACY A-D GROUPS)          |
|                                                                             |
|   [GROUP A] PROJECT MANAGEMENT                                              |
|   - Project organization, QA manager role, problem definition, project      |
|     description, quality objectives, training certifications, records.      |
|                                                                             |
|   [GROUP B] DATA GENERATION & ACQUISITION                                   |
|   - Sampling process design, sampling handling/custody, SW-846 analytical   |
|     methods, field/lab QC samples, instrument calibration, supplies.        |
|                                                                             |
|   [GROUP C] ASSESSMENT & OVERSIGHT                                          |
|   - Field audits, laboratory performance evaluations, management reports.   |
|                                                                             |
|   [GROUP D] DATA VALIDATION & USABILITY                                     |
|   - Data review, verification, validation protocols, user reconciliation.   |
|     (Evaluation against PARCC parameters).                                  |
+-----------------------------------------------------------------------------+

---

3. PARCC Parameters (Data Quality Indicators)

During data review and validation, environmental data may be evaluated using the PARCC data quality indicators. The QAPP must define project-specific acceptance criteria; PARCC is a useful framework, not a source of universal numeric limits:

+-----------------------------------------------------------------------------+
|                        THE FIVE PARCC PARAMETERS                            |
|                                                                             |
|   P - PRECISION        : Agreement among repeated independent measurements. |
|   A - ACCURACY / BIAS  : Closeness of a measured value to the true value.   |
|   R - REPRESENTATIVENESS: Degree to which data represent true conditions.  |
|   C - COMPLETENESS     : Percentage of usable data versus planned data.      |
|   C - COMPARABILITY    : Confidence that data can be compared across time/lab|
+-----------------------------------------------------------------------------+

1. Precision

Quantifies the random error and reproducibility of repeated measurements under identical conditions. Measured via Field Duplicates and Laboratory Matrix Spike Duplicates (MSD) using Relative Percent Difference (RPD):

$\text{RPD} = \frac{|C_1 - C_2|}{\frac{C_1 + C_2}{2}} \times 100 = \frac{2 \cdot |C_1 - C_2|}{C_1 + C_2} \times 100$

• Example benchmarks: Some project QAPPs use RPD limits in the 20%–30% range for aqueous samples and 30%–50% for solids. The analytical method, concentration relative to reporting limits, and matrix can require different criteria.

2. Accuracy / Bias

Quantifies the systematic error (closeness to the known true value). Measured via Matrix Spikes (MS), Laboratory Control Samples (LCS), and Surrogate Spikes using Percent Recovery ($%R$):

$\%R = \frac{C_{\text{spiked}} - C_{\text{unspiked}}}{C_{\text{added}}} \times 100$

• Example benchmarks: A project might use 70%–130%, but method- and analyte-specific control limits or laboratory statistically derived limits govern when specified.

3. Representativeness

The degree to which sample data accurately and precisely represent environmental conditions at the sampling location and time. Ensured by proper sampling design (e.g., ISM), proper well screen placement, zero-headspace VOC sampling, and avoiding stagnant water collection.

4. Completeness

The percentage of valid, legally defensible, and usable data obtained compared to the total number of planned measurements:

$\%\text{Completeness} = \frac{\text{Number of Valid / Usable Measurements}}{\text{Total Number of Planned Measurements}} \times 100$

• Example benchmark: A QAPP might require at least 90% completeness. The actual decision-quality target is project-specific.

5. Comparability

The confidence with which one data set can be compared to another across different sampling events, seasons, laboratories, or analytical techniques. Achieved by standardizing units (e.g., $\mu\text{g/L}$, $\text{mg/kg}$), using standardized EPA SW-846 test methods, consistent sampling SOPs, and standard reporting limits.

+-----------------------------------------------------------------------------+
|                     WORKED EXAMPLE: RPD & % RECOVERY                        |
|                                                                             |
|   Problem 1: Precision (RPD)                                                |
|   - Parent soil sample Lead concentration (C1): 180 mg/kg                   |
|   - Co-located field duplicate Lead concentration (C2): 220 mg/kg           |
|                                                                             |
|   RPD = [ |180 - 220| / ((180 + 220) / 2) ] * 100                           |
|   RPD = [ 40 / (400 / 2) ] * 100 = (40 / 200) * 100 = 20.0%                 |
|   Result: 20.0% <= the example project 30.0% limit -> ACCEPTABLE FOR THIS QAPP      |
|                                                                             |
|   Problem 2: Accuracy (% Recovery)                                          |
|   - Unspiked groundwater Benzene concentration: 10 ug/L                     |
|   - Spiked groundwater sample result: 55 ug/L                               |
|   - Known mass of Benzene added: 50 ug/L                                    |
|                                                                             |
|   %R = [ (55 ug/L - 10 ug/L) / 50 ug/L ] * 100 = (45 / 50) * 100 = 90.0%   |
|   Result: 90.0% is within the example project 70-130% limits -> ACCEPTABLE    |
+-----------------------------------------------------------------------------+

---

4. Sensitivity & Detection Limits (MDL vs. PQL/LOQ)

Analytical sensitivity is defined by a strict hierarchy of detection and quantitation limits:

+-----------------------------------------------------------------------------------------+
|                         DETECTION LIMIT HIERARCHY & DATA QUALIFIERS                     |
|                                                                                         |
|    0.0 ug/L               MDL (e.g., 0.8 ug/L)             PQL / LOQ (e.g., 2.5 ug/L)   |
|   +-----------------------+--------------------------------+--------------------------> |
|   |     "U" FLAG          |            "J" FLAG            |      QUANTITATIVE DATA |
|   | Undetected            | Estimated Detection            | Quantified with Known  |
|   | (< MDL)               | (MDL <= Value < PQL)           | Precision & Accuracy   |
|   +-----------------------+--------------------------------+--------------------------> |
+-----------------------------------------------------------------------------------------+

• Method Detection Limit (MDL): A statistically derived method detection threshold under 40 CFR Part 136 Appendix B. The current procedure estimates an MDL using method blanks and low-level spikes and controls the probability of falsely claiming detection; laboratory reporting and U/J flags must follow the method and project QAPP.
• Practical Quantitation Limit (PQL) / Limit of Quantitation (LOQ) / Reporting Limit (RL): The lowest concentration that can be reliably achieved and quantitatively measured with specified precision and accuracy during routine laboratory operating conditions (typically $3\times\text{ to }10\times\text{ MDL}$).
• "J" Flag (Estimated Value): Commonly used for a detected result below the quantitation/reporting limit or for another validation qualification. It communicates estimated magnitude; it is not a universal statement that one result has exactly 99% certainty. Follow the laboratory qualifier definition and QAPP.

---

5. Standard EPA SW-846 Analytical Test Methods

EPA publication SW-846 (Test Methods for Evaluating Solid Waste: Physical/Chemical Methods) specifies the official test procedures for RCRA compliance and Superfund site cleanups: