A role (or roles) for passive sampling in regulatory monitoring Ian Allan, Branislav Vrana, Richard Greenwood, Graham Mills, Alfhild Kringstad, Erling Bratsberg, Katherine Langford & Chris Harman
Why would anyone use PS? Complex calculation of C TWA and understanding of principle of passive sampling Need for ingenuity during sampler deployment Never sure samplers will be there upon return! Costs due to the need for blanks/control samplers (including PRC data) Cannot directly compare passive sampling data with EQS for nonpolar organics
Why not? Cost efficiency when representative data is needed (time integrative nature of the sampling) wanted Dissolved concentrations A better measure of surface water quality Limits of detection Variability of passive sampling data
Monitoring tasks Testing for compliance (with EQS) Monitoring long-term trends Measurement of riverine fluxes Source tracking and assessment of spatial distribution Linking exposure and effects Contaminant speciation
This presentation 1. Improving representativeness Combining bottle sampling and passive sampling 2. Measurement of riverine fluxes (e.g. OSPAR) Total/dissolved fluxes of trace nonpolar organics 3. Long-term trends Surface water concentrations of polycyclic aromatic hydrocarbons (PAHs) 4. Testing for compliance (e.g. WFD) Contaminant concentrations at low yet relevant concentrations
1. Sampling representativeness Characteristic: Time-integrative sampling Reduce monitoring burden Support spot sampling with passive sampling Use of the Chemcatcher and DGT samplers for monitoring trace metals Fieldwork in the River Meuse (NL)
1. Sampling representativeness Passive sampling for metals Reduce monitoring burden Support spot sampling with passive sampling Develop knowledge of water body specific contaminant speciation and partitioning data Use of the Chemcatcher and DGT samplers for metals Measurement of a labile fraction of metals and total/filtered fractions by two procedures
1. Sampling representativeness Time-weighted average concentrations The DGT: - Polyacrylamide diffusive gel layer and a chelex resin for accumulation of metals - Elution with 1M HNO 3 and analysed by ICP-MS The Chemcatcher: - Cellulose diffusion-limiting membrane and chelating disk as a receiving phase -Elution with HNO 3 and extract is analysed by ICP-MS DGT Uptake rates R S (ml/hr) 2.5 2.0 1.5 1.0 0.5 Ag Al Cd Co Cr Cu Fe Mn Ni Pb Zn C DGT = M DGTΔg DtA Uptake rate (ml/day) 6 5 4 3 2 Copper C Chemcatche r = M Chemcatcher R S t 0.0 0 10 20 30 40 Temperature (degc) 1 0 20 40 60 80 Turbulence (rpm)
1. Sampling representativeness Passive sampling for metals River Meuse Eijsden (NL) In 2005 Concentration (μg L -1 ) 2.5 2.0 1.5 1.0 0.5 Cd Total C Team A Total C Team B Filtered C Team A Filtered C Team B 0.0 0 5 10 15 20 25 Time (Days)
1. Sampling representativeness Passive sampling for metals Deviation from the spot sampling mean (%) 430 200 150 100 50 0-50 -100 Cd Cu Ni Pb Zn Metal Comparison of passive sampling data (7,14,21 and 28 day exposures) with mean filtered concentrations by spot sampling (reference value) Spot sampling - Team A Spot sampling - Team B DGT RP DGT OP Chemcatcher Spot sampling variability
2. Riverine fluxes Characteristic: LODs & time-integrative sampling Monitoring of direct inputs and discharges from rivers (e.g. for OSPAR region) Contaminant present: Dissolved (and complexed for metals) Sorbed to SPM, dissolved organic matter & colloids Can we estimate all of these from one single measurement? And with what confidence?
2. Riverine fluxes PCBs in the Drammenselva River (Norway) Measurement of dissolved and particulate matterassociated contaminants Passive sampling with SPMDs Monitoring of SPM with continuous-flow centrifugation/large volume water sampling
2. Riverine fluxes Limits of detection for PCBs 1000 Limit of detection (pg L -1 ) 100 10 CB209 CB180 CB156 CB138 CB153 CB105 CB118 CB101 CB52 CB28 1 X Data 0.1 Passive sampling LVWS - PUF LVWS - Filter SPM sampler Centrifuge
Contaminant concentrations 2. Riverine fluxes 10 Estimation of distribution between phases (e.g K POC ) CB180 CB156 1 X Data CB138 CB153 CB105 CB118 CB101 CB52 CB28 Passive sampling LVWS - PUF LVWS - Filter SPM sampler Centrifuge Concentration (pg L -1 )
2. Riverine fluxes Contaminant fluxes Drammenselva River How does C TWA vary with C total? LODs for bottle sampling for PCBs 1 ng L -1 (0.1 for HR-GC) Riverine flux: 10 kg/year to 10 g/year estimates Put into perspective: Uncertainties on flow & SPM/POC measurements Variability of SPM levels with depth and along river sections or transects
3. Long term trends Characteristics: LODs, method with low variability As part of many legislative texts Do they need to be based on C total? Or are dissolved phase concentrations sufficient? Knowledge of contaminant partitioning
3. Long term trends Sampling with LDPE/silicone Reproducibility of sampler production Variability of masses absorbed (24 & 51 days) Further long-term data
3. Long term trends PAH Masses accumulated 140 120 Silicone strip SPMD 1.4 1.2 100 1.0 ng sampler -1 80 60 40 (m 51d /51)/(m 24d /24) 0.8 0.6 0.4 20 0 PYR FLUO CHRY BbjF BeP X Data 51 day exposure Silicone strip 0.2 SPMD 0.0 3.5 4.0 4.5 5.0 5.5 6.0 Log K OW
3. Long term trends Trend monitoring with passive samplers Reproducibility of sampler production: PRC spike RSD < 10 % Variability in masses absorbed RSD 5 % with identical sampler surface area and deployment Similar data quality for 24 and 51 d exposures Do we need to know about changes in contaminant distribution between phases with changes in total concentrations?
4. Testing for compliance Issues with LODs for standard bottles sampling for non-polar organics (e.g. WFD) Are EQS for bottom sediments an acceptable replacement? What are the challenges? Setting dissolved phase EQS? What about measuring dissolved concentrations and estimated C total empirically?
Conclusions Passive samplers offer possibilities for improvements in data quality for regulatory monitoring tasks In most cases, a knowledge and understanding of contaminant speciation and partitioning (specific to each water body) is a prerequisite Apparent equilibrium partitioning and/or safety factors?
Acknowledgements SFT Norwegian State pollution control authority Nathalie Guigues - BRGM Jesper Knutsson Chalmers University of Technology - Goteborg