Ecotoxicology

    Choosing Background Concentrations for Pollution Indices: Why Your Baseline Matters

    How to choose and report soil or sediment background concentrations for CF, PLI and Igeo, with a worked baseline-sensitivity example in Labreador.

    LabreadorNovember 21, 20267 min read

    Two researchers can analyse exactly the same sediment sample and report different Pollution Load Indices. Neither calculation has to be wrong: they may simply have used different background concentrations.

    That is why choosing a baseline is part of the scientific method, not a final setting to adjust until the results look convincing. This guide explains how the choice affects CF, PLI and Igeo, and how to make it explicit in Labreador's Pollution Indices module.

    Background is not a safety limit

    A background concentration describes the reference condition against which enrichment is assessed. It may come from a suitably characterised local reference site, a pre-industrial sediment layer, or a published geochemical dataset appropriate to the study.

    A regulatory limit answers a different question: whether a concentration meets a particular standard or screening criterion. Substituting a drinking-water limit for a sediment background does not turn CF into a health-risk calculation.

    Keep these questions separate:

    • CF, PLI and Igeo: how does the measured material compare with a specified geochemical reference?
    • Water HPI: how do water concentrations compare with the standards selected for that calculation?
    • Dietary THQ and HI: how does an estimated intake compare with toxicological reference doses?

    In Labreador, soil/sediment and water have separate calculation paths. A reference in mg/kg cannot be used as though it were a water standard in mg/L.

    Match the reference to the sample

    Before accepting background values, check four things.

    1. Matrix and setting. Soil, sediment, water and fish tissue are not interchangeable. For sediment, geological setting and grain-size distribution can also affect natural metal concentrations.
    2. Analytical comparability. A partial extraction and a near-total digestion need not measure the same operationally defined fraction. Compare the preparation and analytical methods, not just the element names.
    3. Units and weight basis. Samples and references must be expressed in compatible units and on the same wet- or dry-weight basis. mg/kg and µg/g are numerically equivalent; dry weight and wet weight are not automatically equivalent.
    4. Reference quality. A nearby location is not necessarily uncontaminated. Explain how the reference was selected and what evidence supports its use.

    A published average can provide a transparent reference when a defensible local baseline is unavailable. It is not automatically representative of every catchment. Report that limitation instead of presenting a global average as a measured local condition.

    Worked example: unchanged sample, different PLI

    The following values are hypothetical, not recommended background concentrations. All refer to sediment in mg/kg dry weight.

    ElementSampleReference AReference B
    Cu301020
    Zn903060
    Pb18612

    The equations are:

    CF(element) = C_sample / C_background
    PLI = (CF_Cu × CF_Zn × CF_Pb)^(1/3)
    Igeo(element) = log2(C_sample / (1.5 × C_background))
    

    Using Reference A:

    CF_Cu = 30 / 10 = 3
    CF_Zn = 90 / 30 = 3
    CF_Pb = 18 / 6 = 3
    PLI = (3 × 3 × 3)^(1/3) = 3
    Igeo_Cu = log2(30 / (1.5 × 10)) = 1
    

    Using Reference B:

    CF_Cu = 30 / 20 = 1.5
    CF_Zn = 90 / 60 = 1.5
    CF_Pb = 18 / 12 = 1.5
    PLI = (1.5 × 1.5 × 1.5)^(1/3) = 1.5
    Igeo_Cu = log2(30 / (1.5 × 20)) = 0
    

    The sample has not changed. Doubling every background concentration halves every CF and the PLI; it lowers each Igeo by one unit. Both PLIs remain above the baseline level of 1, but the apparent magnitude of enrichment differs substantially.

    Under the Tomlinson interpretation used in Labreador, PLI < 1 indicates no pollution relative to the chosen reference, PLI = 1 is the baseline level, and PLI > 1 indicates progressive deterioration. These are reference-dependent index labels, not proof that a site is biologically safe or unsafe.

    Use replicate references without hiding missing values

    If you have multiple suitable reference samples, Labreador lets you mark imported rows as baseline replicates. Their arithmetic mean is calculated separately for each mapped element.

    Hypothetical Cu reference replicates: 8, 10, 12 mg/kg
    Mean Cu background = (8 + 10 + 12) / 3 = 10 mg/kg
    

    Measured zeros remain observations in that mean. Missing, invalid and unresolved censored entries are not zeros and do not become numerical reference observations. Because eligibility is element-specific, the number of usable reference measurements can differ between elements; report those counts and the original reference data.

    A mean background of zero cannot serve as the positive denominator required for CF. No usable reference observations means an unavailable baseline, not permission to reuse a value from an earlier analysis. Replication helps describe a reference but does not establish that the site is suitable or automatically propagate reference uncertainty into the indices.

    Why PLI and MPI can tell the same statistical story

    MPI is the geometric mean of the measured concentrations, without a background denominator. For a complete, strictly positive element panel and one common background vector:

    MPI(sample) = (C_1 × C_2 × ... × C_n)^(1/n)
    MPI(background) = (B_1 × B_2 × ... × B_n)^(1/n)
    PLI(sample) = MPI(sample) / MPI(background)
    

    Every sample's MPI is therefore divided by the same positive constant. The group tests offered for PLI and MPI can give identical test statistics and p-values under these conditions. That is expected scaling behaviour, not two independent confirmations of pollution. Labreador displays a note explaining this relationship.

    This identity does not justify comparing different element panels. Dropping a missing metal changes what the geometric mean summarises. Labreador reports incomplete or zero-containing PLI/MPI panels as unavailable rather than silently calculating an index from fewer mapped elements. MPI also retains concentration units and has no universal safety threshold.

    A practical Labreador workflow

    1. Open Pollution Indices and select the soil/sediment matrix before importing data.
    2. Paste or upload your measurements, assign sample identifiers and element columns, and confirm the units. Check that samples and background share a weight basis.
    3. Mark suitable reference rows as baseline replicates, or enter your documented background values for each element.
    4. Review per-element CF and Igeo alongside the complete-panel PLI. Investigate unavailable results instead of treating them as zero.
    5. Keep the same justified reference and element panel when comparing sites. If you examine an alternative reference, label it as a separate sensitivity analysis.
    6. Export the results and report the baseline source, values, units, basis, analytical method, reference sample counts and missing-data handling with them.

    Choose the primary baseline before interpreting the indices. A sensitivity analysis can show how conclusions depend on that assumption; it should not become a search for whichever baseline produces the preferred classification.

    What to write in your methods section

    Use a statement like this, replacing every bracketed item with your actual study information:

    CF, PLI and Igeo were calculated for [element panel] using
    [reference source] as background. Sample and reference concentrations
    were expressed in mg/kg dry weight and determined using [method].
    Background values were [values by element]; where replicate reference
    samples were used, arithmetic means and usable counts were reported.
    Incomplete or nonpositive PLI panels were reported as unavailable.
    Calculations were performed in Labreador, cited via its deposited release.
    

    For the equations and broader interpretation, see the pollution index calculator guide. For concentration-basis errors, see dry weight versus wet weight.

    References

    • Håkanson, L. (1980). An ecological risk index for aquatic pollution control: a sedimentological approach. Water Research, 14(8), 975–1001. doi:10.1016/0043-1354(80)90143-8.
    • Tomlinson, D. L., Wilson, J. G., Harris, C. R., and Jeffrey, D. W. (1980). Problems in the assessment of heavy-metal levels in estuaries and the formation of a pollution index. Helgoländer Meeresuntersuchungen, 33, 566–575. doi:10.1007/BF02414780.
    • Müller, G. (1969). Index of geoaccumulation in sediments of the Rhine River. GeoJournal, 2, 108–118.