· Research Centre for Toxic Compounds in the Environment, Masaryk University, Brno, Czech...

Research Centre for Toxic Compounds in the Environment, Masaryk University, Brno, Czech RepublicInstitute of Biostatistics and Analyses, Masaryk University, Brno, Czech Republic

Team of authors:J. Klánová, L. Dušek

J. Borůvková, R. Hůlek, K. Šebková, J. Gregor,J. Jarkovský, L. Kohút, J. Hřebíček, I. Holoubek

Brno, Czech RepublicJanuary 2012

The initial analysis of the Global Monitoring Plan(GMP) reports and a detailed proposal to developan interactive on-line data storage, handling,and presentation module for the GMPin the framework of the GENASIS databaseand risk assessment tool

The initial analysis of the Global Monitoring Plan (GMP) reports and a detailed proposal to develop an interactive on-line data storage, handling, and presentation module

for the GMP in the framework of the GENASIS database and risk assessment tool

Research Centre for Toxic Compounds in the Environment Masaryk University Brno, Czech Republic

Institute of Biostatistics and Analyses Masaryk University Brno, Czech Republic

Team of authors:

J. Klánová, L. Dušek J. Borůvková, R. Hůlek, K. Šebková, J. Gregor, J. Jarkovský, L. Kohút, J. Hřebíček, I. Holoubek

Brno, Czech Republic January 2012

The initial analysis of the Global Monitoring Plan (GMP) reports and a detailed proposal to develop an interactive on-line data storage, handling, and presentation module for the GMP in the framework of the GENASIS database and risk assessment tool

i

List of contents:

List of abbreviations......................................................................................................................... ii Executive summary.......................................................................................................................... iii Review of the GMP reports content................................................................................................ 1

Chapter 1. Background......................................................................................................... 3 Chapter 2. Goals of the Global Monitoring Plan data review.............................................. 4 Chapter 3. Chemicals in GMP reports – overview............................................................... 5

Analytical review of the GMP reports content.............................................................................. 11 Chapter 4. Analytical outcomes I. Ambient air data reported in Global Monitoring Plan regional reports..................................................................................................................... 13 Chapter 5. Analytical outcomes II. Human tissues data reported in Global Monitoring Plan regional reports............................................................................................................. 20

Pilot methodological proposals........................................................................................................ 25 Chapter 6. Methodology I. GMP regional reports – data standardization and proposal for database structure................................................................................................................. 27 Chapter 7. Methodology II. Information and communication technologies........................ 34 Chapter 8. Methodology III. Statistical data processing...................................................... 38

Pilot proposal for the data collection system.................................................................................. 55 Chapter 9. Proposal for data structure in future data collection campaigns – electronic data capture system for GMP............................................................................................... 57 Chapter 10. Pilot on-line reporting over GMP regional reports........................................... 66

Conclusions........................................................................................................................................ 75 Annexes.............................................................................................................................................. 77

Annex 3.1. Chemicals in the Global Monitoring Plan regional reports - overview Annex 4.1. Ambient air data in GMP reports – overview Annex 4.2. Ambient air data in GMP reports – data structure Annex 5.1. Human tissues data in GMP reports – overview Annex 5.2. Human tissues data in GMP reports – data structure Annex 6.1. Overview of variables reported in the GMP reports and their content analysis Annex 10.1. On-line data visualization: user manual


ii

List of abbreviations

AMF Action message format ARIMA Autoregressive integrated moving average DDD Dichlorodiphenyldichloroethane DDE Dichlorodiphenyldichloroethylene DDT Dichlorodiphenyltrichloroethane dl-PCBs Dioxin-like polychlorinated biphenyls GENASIS Global Environmental Assessment Information System GIS Geographic Information System GMP Global Monitoring Plan HCB Hexachlorobenzene HCH Hexachlorocyclohexane ICT Information and communication technologies LOQ Limit of quantification lw Lipid weight OLAP Online analytical processing PAHs Polycyclic aromatic hydrocarbons PBDEs Polybrominated diphenylethers PCBs Polychlorinated biphenyls PCDDs Polychlorinated dibenzo-p-dioxines PCDFs polychlorinated dibenzofurans PeCB Pentachlorobenzene POPs Persistent organic pollutants RDMS Relational database management system SC Stockholm Convention SQL Structured query language TEF Toxic Equivalency Factor TEQ Toxic equivalent UNEP United Nations Environment Programme


iii

Executive summary

This report is prepared as a review of available Global Monitoring Plan (GMP) data on persistent organic pollutant (POPs) concentrations in three environmental matrices: ambient air, breast milk and human blood that serve as core media to monitor and examine effectiveness of the measures adopted by the Stockholm Convention on POPs.

A wide team of authors performed a critical review of the GMP reports on qualitative attributes, i.e. reported chemicals and sampling frequency of reported concentration levels. The main part of the review summarizes data available for 12 chemicals (including their recommended congeners, isomers and degradation products) that were mandatory to report in the GMP in 2008 and identifies other reported data on POPs that could be used in the future data collection campaigns.

The outcome of the review consists of three parts: review of available GMP data, methodical proposal how to improve statistical processing of UNEP-GMP data in the future and finally, the proposal of a new e-data capture system suitable for the next GMP campaigns. Each of these chapters respected principal characteristics of examined environmental matrices. Key conclusions are highlighted in this executive summary.

Review of available GMP data

Compounds and parameters found in the GMP reports were sorted by their relation to the Stockholm Convention, by reported data content and its appropriateness for further trend analyses. The classification strictly followed scope of the Stockholm Convention and the GMP Guidance document and their amendments in time. Reported compounds were therefore classified into four groups: 1) original 12 POPs included in the Stockholm Convention in 2001, their congeners, isomers and degradation products determined in the GMP Guidance (2007); 2) additional 10 POPs included into the SC in 2009 and 2011 and specified in the updated GMP Guidance; 3) all other compounds, their sums and toxic equivalents related to the Stockholm Convention but not specified in any of the GMP Guidance documents; 4) compounds found in the GMP reports but not related to the SC.

All regional GMP reports contain 171 variables (including concentration data on congeners, isomers, transformation products, various summations and toxic equivalents – TEQs). Analysing the primary pool of reported parameters, 149 of them were related to 12 original Stockholm Convention POPs and 10 additional POPs. Out of this number, 58 +7 were listed as compounds highly recommended for monitoring and evaluation in the GMP Guidelines documents for 12 original and 10 additional POPs, respectively. Remaining parameters (84) consisted of non-recommended congeners, various sums and toxic equivalents, often not correctly identified. There were also 22 other reported parameters with no relation to the SC.


iv

The qualitative part of the review concluded that most of the original 12 POPs were reported in all matrices and existing data would thus enable a time-related comparison with data from future sampling campaigns.

Although the GMP Guidelines from 2007 would not cover the current Stockholm Convention scope as amended, some of the additional 10 POPs were also identified in the reviewed reports (8 in ambient air and 4 in both breast milk and human blood, respectively).

In addition, reported sampling frequency was evaluated to determine content of the GMP data usable in time trend analyses. Chemicals with concentration reported over time span of at least 4 consecutive years were marked as suitable for baseline time series analyses („historical time series available”), while other records belong to the family of „point estimates” or „occasional reports“. Moreover, historical time series may be very useful in investigation of background time changes and trends in POPs concentrations before adoption of the Stockholm Convention.

Availability of time series differed in the examined matrices, both within the group of original 12 POPs and additional 10 POPs. Altogether, a sufficient number of 12+10 POPs were reported for ambient air monitoring data in the category “historical time series available” for at least one country from three UN regions with “point estimates“ being available in all UN regions. Records spread over 1990 – 2008, however majority of the time series originate from recent measurements (after 2004). Ambient air monitoring data thus represent a relevant basis for pair-wise comparison with the future GMP data collection campaigns.

POPs detected in human tissues were mostly reported as lipid-adjusted concentrations and mostly as annual point estimates. Available data are suitable for future statistical processing. The spectrum of chemicals and their spatial and temporal coverage reported for breast milk and human blood was a significantly lower than that for ambient air. Nevertheless, all of the original 12 POPs in the Stockholm Convention were reported.

Although reported data in analyzed reports were, in general, rated suitable for planned statistical processing, the audit of the content revealed serious challenges related to data standardization in the GMP report. The reports suffer from the lack of standardized taxonomy for POPs, their isomers, transformation products and summations which were frequently used. Heterogeneity of the data is further enhanced by reporting various toxic equivalents (TEQ) (based on WHO TEF values from various years) rather than concentrations of the individual PCDDs, PCDFs and PCBs congeners. Unclear identification of units, time and spatial scales of the reported concentrations as well as insufficient specification of aggregated data belong to other frequently identified drawbacks.

Another source of variation is associated with different structure of reports provided by different UN regions; some of the national records included detailed primary data including even rarely measured compounds while others were based only on sums of key groups of POPs. Low standardization of the regional GMP reports also contributed to a remarkable variation in reported concentration of examined chemicals.


v

Nevertheless, a significant part of available data allowed development and testing of a pilot database. We conclude that data processing can continue (after expert review and approval) in further analyses including assessment of the contamination baselines.

Data from all GMP reports were re-coded into an electronic database template, standardized in structure and in content of data fields and validated with respect to their spatial and temporal variability. This analysis definitely confirmed a reliable coverage for all original 12 POPs included in the Stockholm Convention, in both ambient air and human tissues monitoring.

The review and pilot processing of data resulted in conclusion that available GMP data can be used for baseline statistical processing. Data can be analyzed as annually aggregated time series or at least as relevant point estimates, prepared for comparison with the next data collection campaign. However, detailed standardization and relevant quality scoring of individual records are needed prior any future processing. The methodical solution of this problem is proposed in a section dealing with the structure of the future GMP database in chapter 6. Although available data were inserted into e-data capture forms, some uncertainties could not be solved at this stage (i.e. missing values). The information relevant to that particular parameter is stored in the database but such record could not be used in the statistical processing at the moment. Fully standardized re-coding of available data prepared for the direct linking with the next GMP data collection campaigns in the same database appears to be effective way how to solve all limits in data standards.

Methodical proposal how to improve statistical processing of UNEP-GMP data

The content of the GMP reports was analyzed and summary overview of the data structure and content suitable to run statistical summaries were prepared (see Annex 6.1). Based on this analysis, several methodical concepts were proposed to improve future GMP data collection campaigns, associated data management and processing:

• proposal on how to solve main problems with statistical processing of unclear inputs, i.e. reconstruction of the sample distribution variability in a heterogeneous field of used variability measures, estimation of minimum effect size for future comparison of GMP campaigns and a robust methodology of trend quantification and assessment;

• proposal on how to solve data heterogeneity and how to implement data collection standards.

The main added value of this part is the successful pilot quantification of effect size, relevant for statistical comparison of POPs concentrations. Using the model data from the Košetice station (Czech Republic), statistically significant quantified difference could be determined for POPs concentration data, provided that they were detected in the same matrix and by same methodology. This approach might be helpful in comparison of data between the former and future GMP records. It will be further verified using a larger international data set


vi

Based on the test data, it has been shown that although annually aggregated concentration data sets definitely loose the seasonal variability, they are still suitable for comparative statistical processing. The overall trends can be assessed on annually-adjusted data provided that the statistically significant and detectable difference (effect size) is known/determined. Therefore, annually aggregated data were recommended as a minimum for the relevant statistical comparison of former and future GMP reports.

Finally, a robust statistical methodology was applied for pair-wise comparison of consecutive time measurements or for trend estimation in short time series and tools allowing for processing of heterogeneous data were built in order to keep maximum information for future comparison.

Proposal for a new electronic data capture system, suitable for the next GMP reports

The required IT support necessarily includes centralized data management, database customized for POPs concentration data and its standardized superstructure allowing comparison of former and future GMP data collection campaigns, and thereby quantification of time trends. These tasks represent a challenge for several disciplines including environmental informatics, statistics and software development.

A comprehensive ICT background for planned processing of the UNEP-GMP data was proposed and preliminary designed to fulfil required functionality: standardization of already obtained GMP data, database suitable for archiving and processing of POPs data and on-line data visualization system. The development followed criteria and standards adopted for the information system GENASIS (www.genasis.cz), which was also used for the pilot visualization of available GMP data (www.genasis.cz/unep). The pilot version of the on-line data visualization was prepared and tested on the re-coded GMP data. Prepared graphical and classificatory functions are password protected and allow authorized users to sort data, study their structure and completeness. However, we recognize that further validation procedures are strongly needed. That is why the on-line visualization is opened only to authorized data administrators at the moment. Finally validated POPs data should be further re-coded to a robust GMP database that would potentially serve as a centralized storage and visualisation system for the next GMP campaign.

In addition to its functionality, security management of the proposed solution is highly emphasized. Only password protected data administrators are approved to use the pilot on-line data visualization tools. The system was developed using common and advanced standards adopted for information systems of high quality, e.g. ISO 9001:2001/2009, ISO/IEC 20000-1:2006, ISO/IEC 27001:2006.

Proposal for the standardized GMP database based on fully parametric data sheets that could be used in the next GMP data collection campaign is one of the outputs of our work. The whole system is proposed to improve quality of collected global data set on POPs occurrence


vii

in the environment and to strengthen the ownership and responsibility of the local data administrators.

In the pilot database proposal, the set of parameters was arranged hierarchically, corresponding to the ontology of the planned information system. Three main levels are prioritized in the ontology: a) identification of the sampling site, b) spatial data description, c) measurement – value items. Besides primary data (measurement level), the data collection system will be equipped with classifiers facilitating data validation and processing: coded information on spatial aggregation (if used/applicable), coded information on time aggregation (if used/applicable). Only annually aggregated values will be accepted and additional classifier linking new GMP reports with formerly reported variables is proposed. We are of the view that next GMP data collection campaign could contribute to the retrospect identification of missing records such as incomplete time span of reported data, location, missing records and to the verification of uncertain records and thereby increase the number of validated and standardized parameters collected in the first set of GMP reports.

Chapter 1. Background

3

Chapter 1.

Background

The aim of the Stockholm Convention on Persistent Organic Pollutants (POPs) is to protect human health and the environment from adverse effects of POPs that are released to the environment from use, production, unintentional production of POPs and from stockpiles and wastes containing POPs.

The Article 16 of the Convention requires that effectiveness of measures adopted by the Convention to eliminate or significantly reduce POPs releases must be regularly evaluated. The first overview of the geographical distribution of POPs was established by the first Global Monitoring Plan data collection campaign in 2008. The reported global data on POPs occurrence are gathered in five regional reports that were adopted by the 4th meeting of the Conference of the Parties in Geneva in May 2009 and next global data collection campaign is to occur in 2014.

In addition, number of chemicals that are currently subject to the Stockholm Convention has significantly increased from 12 original POPs that were mandatory to report in 2008. Next collection campaign would have to collect information on 22 POPs and a significant number of their congeners, isomers, precursors, transformation or degradation products as outlined in the recently updated GMP Guidance document.

This task could represent a challenge for many Parties, but it also offers a great opportunity to significantly increase our knowledge on POPs occurrence, compare collected data of the former collection campaign and thereby determine trends. Use of existing analytic and visualisation tools combined with a robust expert knowledge could lead to a rapid determination of the global priority actions that would significantly improve global POPs and chemicals management.

Chapter 2. Goals of the Global Monitoring Plan data overview

4

Chapter 2.

Goals of the Global Monitoring Plan data review

This report represents the outcome of the pilot inspection over the content of the Global Monitoring Plan (GMP) regional reports. The GMP reports gathered data on occurrence of persistent organic pollutants in the world in environmental matrices selected as core matrices for the purpose of the Stockholm Convention:

• breast milk • human blood • ambient air, (active and passive samples)

pertaining to the decision SC-2/13 on the effectiveness evaluation and relevant guidance to the Global Monitoring Plan (GMP).

The goals of the analysis consisted of the following:

1. Perform a critical review of the Global Monitoring Plan regional reports content 2. Propose steps for standardization of the monitoring data 3. Design the GENASIS-GMP database module to visualize the content of the GMP reports 4. Propose data model and electronic format templates for the next GMP phases

The resulting report describes the outcome of our work in sections:

• Pilot statistical summary of reported data and identification of records suitable for concentration comparison, spatial and/or time-related comparisons (chapters 3, 4 and 5)

• Critical summary of statistical measures used in the GMP records and analysis of their mutual compatibility; pilot statistical processing of collected variability data (chapter 6)

• IT infrastructure needs for future data collection, archiving, processing and reporting (chapter 7)

• Statistical methodology proposal (chapter 8) • Proposal of a new database system for GMP electronic data capture system (chapter 9) • Development of pilot online reporting system running over already accessible data and

facilitating further data validation (chapter 10)

Chapter 3. Chemicals in the GMP reports – overview

5

Chapter 3.

Chemicals in the GMP reports – overview

3.1. Introduction

This chapter explains the approach used to analyze the data content of the regional GMP reports prepared in year 2009 regarding its scope and time scale. Reported variables were at first divided into four groups according to their occurrence in the Stockholm Convention and the Global Monitoring Plan Guidelines.

The second part of this chapter describes spatial and temporal availability of data on chemicals of interest in each core matrix, and identifies the data sets with a potential for future analysis of the time trends.

3.2. Chemicals in the Stockholm Convention and in the Guidance Document for the Global Monitoring Plan under the Stockholm Convention

Global Monitoring Plan reports represent a comprehensive source of information carrying concentration data on selected chemicals and environmental matrices – active and passive samples of ambient air, samples of human milk and blood.

GMP, as an important tool for effectiveness evaluation of the Stockholm Convention, was proposed according to the Article 16 of the Convention. Chemicals banned or restricted by the Stockholm Convention in all SC Parties are defined in its Annexes A, B and C. At the time of the first GMP data collection, twelve chemicals were included in the Stockholm Convention (aldrin, dieldrin, DDT, endrin, HCB, heptachlor, chlordane, mirex, PCBs, PCDDs, PCDFs and toxaphene).

Annexes A, B and C of the Stockholm Convention list chemicals under their general name (e.g. DDT) with no additional chemical specification. Additional guidance for the 12 POPs has been therefore provided in the chapter 2 of the GMP Guidance document (2007) summarizing the isomers/congeners of the parent POPs as well as their degradation products and metabolites recommended for monitoring (see table 3.1).


6

Table 3.1: Compounds recommended for reporting in the GMP Guidance 20071

Chemical Parent POPs Transformation products

Aldrin Aldrin

Chlordane Cis- and trans-chlordane Cis- and trans-nonachlor, oxy-chlordane

DDT 4,4’-DDT, 2,4'-DDT 4,4'-DDE, 2,4'-DDE, 4,4'-DDD, 2,4'-DDD

Dieldrin Dieldrin

Endrin Endrin

HCB HCB

Heptachlor Heptachlor Heptachlorepoxide (A and B)

Mirex Mirex

Polychlorinated biphenyls (PCBs) Sum of 7 PCBs* (28, 52, 101, 118, 138, 153 and 180)

PCB WHO-TEQ (12 congeners)

Polychlorinated dibenzo-p-dioxines (PCDDs) and polychlorinated dibenzofurans (PCDFs)

2,3,7,8-substituted PCDDs/Fs (17 congeners)

Toxaphene Congeners Parlar 26, Parlar 50 and Parlar 62

3.3. Proposed grouping of reported chemicals

All reported variables were at first divided into groups related to their occurrence in the Stockholm Convention and the Global Monitoring Plan Guidelines and relevance of existing records for further data collection was also examined.

Together, all regional GMP reports contain 171 variables (including concentration data on congeners, isomers, transformation products, various summations and toxic equivalents – TEQs). Analysing the primary pool of reported parameters, 58 of them were related to 12 original Stockholm Convention POPs) as specified in the chapter 2 of the Guidance on the GMP for Persistent Organic Pollutants, 20071 (compounds highly recommended for monitoring and evaluation).

Additional seven of all reported variables (alpha-HCH, beta-HCH, gamma-HCH, PeCB, endosulfan I, endosulfan II and endosulfan SO4) were not obligatory at the time of the first GMP report, but they are very important for the future as they are related to the additional 10 compounds that were listed in the Stockholm Convention in 2009 and in 2011, respectively.

1 Guidance on the Global Monitoring plan for Persistent Organic Pollutants, February 2007, amended in May 2007 (http://chm.pops.int/Implementation/GlobalMonitoringPlan/Overview/tabid/83/Default.aspx)


7

Their specification can be found in the revised Guidance on the GMP for Persistent Organic Pollutants, 20092.

The third group (84 variables) comprises parameters related to the Stockholm Convention POPs, but not specified in the Guidelines. Among them, there are various PCB congeners (not recommended in the Guidance) and often not very well defined summations of the individual compounds (56 together) as well as variety of 28 toxic equivalents (TEQs for dioxins, furans and PCBs based on various TEFs, often not cited correctly).

In the fourth group there are 22 chemicals or parameters with no relation to the Stockholm Convention (mostly PAHs (16 PAHs + sum of PAHs). We believe that only the first two groups of parameters should be reported in the next GMP data collection campaigns as parameters from the third group can be calculated in the (global) database and parameters currently reported under the fourth group have no relation to the effectiveness evaluation of the Stockholm Convention.

3.4. Variability of the GMP reported set of chemicals among matrices and UN regions

According to the decision SC-2/13 (Annex I, paragraph 3) the GMP requires monitoring of two core matrices for the purpose of the effectiveness evaluation: ambient air and human tissues (breast milk or blood serum).

Each core matrix reported in the GMP reports is characterized by a different profile of assessed compounds. It is due to different sampling techniques applied, as well as objectively determined relevance of chemicals in given matrices.

Only twenty parameters in total were reported in all matrices (see table 3.2), eleven of those parameters are the individual compounds included in the GMP Guidance as of 2007 and three are among the additional 10 POPs.

Table 3.2: List of parameters tracked in all matrices

Aldrin DDTs Heptachlor p,p-DDE PCB 153 Alpha-HCH Dieldrin Indicator 7 PCBs p,p-DDT PCBs Beta-HCH Gamma-HCH Mirex Parlar 26 Trans-chlordane Cis-chlordane HCB p,p-DDD Parlar 50 Trans-nonachlor

2 Draft revised guidance on the global monitoring plan for persistent organic pollutants; UNEP/POPS/COP.5/INF/27; http://www.chem.unep.ch/pops/GMP/New_POPS_monitoring_guidance-COP.5-INF-27.English.pdf


8

Pilot analysis also revealed a significant variability in the range of chemicals reported from the individual UN regions. Out of 42 parameters reported in all reports, 38 are recommended in the Guidelines (most of them are PCDDs/Fs congeners).

3.5. Parameters reported in the individual matrices

Principal aim of the first GMP report was to establish a baseline concentration for the future establishment of the long-term trends and effectiveness evaluation of the Stockholm Convention. It inevitably means that future analyses must focus on chemicals measured representatively in all parts of the world, and to the extent that allows for analysis of their variability and time trends. Therefore, we propose further sorting of the chemicals included in the GMP reports according to accessibility of their long-term data.

For a purpose of the pilot data analysis, the parameters reported in the first GMP report but not related to the Stockholm Convention (group four from chapter 3.3) were omitted as they are not relevant for the effectiveness of the Stockholm Convention.

Parameters related to original 12 POPs and additional 10 POPs (chemicals adopted in 2009 and endosulfan) of the Stockholm Convention were kept separately (Figures 3.1 – 3.3), both groups were further divided according to the accessibility of long-term data in the existing reports.

Time series include parameters with a long-term monitoring record sufficient for at least basic trend evaluation. The threshold for including chemical into this group was availability of annually aggregated data for four years of monitoring at a minimum. All other parameters – i.e. parameters with less than four years of accessible data or with a large amount of missing values in consecutive years - allow for point estimates only and are thus included under that heading. Nevertheless these data can still be used for establishment of the baseline concentrations for the chemicals of interest.

All parameters1

Ntot = 74, NSC = 38

Additional 10 POPsOriginal 12 POPs

Air

Time series Point estimates Point estimates

Ntot = 8, NSC = 7Ntot = 66, NSC = 31

Ntot = 39, NSC = 27 Ntot = 27, NSC = 4

Time series

Ntot = 4, NSC = 4 Ntot = 4, NSC = 3

Figure 3.1: Parameters reported in ambient air, Ntot stands for overall count of parameters reported in the GMP reports and related to the SC, while Nsc shows a subgroup of parameters monitoring of which was recommended in the Guidances on GMP.


9

Breast milkAll parameters1


Ntot = 118, NSC = 61

Ntot = 4, NSC = 3Ntot = 114, NSC = 58

Ntot 0, NSC = 0 Ntot = 4, NSC = 3Ntot = 17, NSC = 5 Ntot = 97, NSC = 53

Time series Point estimates Point estimatesTime series

Figure 3.2: Parameters reported in breast milk , Ntot stands for overall count of parameters reported in the GMP reports and related to the SC, while Nsc shows a subgroup of parameters monitoring of which was recommended in the Guidances on GMP.

All parameters1


Human blood

Time series Point estimates Point estimatesTime series

Ntot = 81*, NSC = 53

Ntot = 4, NSC = 3Ntot = 77, NSC = 50

Ntot = 0, NSC = 0 Ntot = 4, NSC = 3Ntot = 6, NSC = 2 Ntot = 71, NSC = 48

Figure 3.3: Parameters reported in human blood, Ntot stands for overall count of parameters reported in the GMP reports and related to the SC, while Nsc shows a subgroup of parameters monitoring of which was recommended in the Guidances on GMP.

For more detailed information on chemicals reported for each matrix see Annex 3.1.

3.6. Compounds reported in the GMP reports – conclusions

The first step in the analysis of the regional GMP reports was to determine the relation of all reported parameters to the Stockholm Convention. Reported variables were divided into four groups according to their occurrence in the Stockholm Convention and the Global Monitoring Plan Guidelines.

The first group are the original 12 POPs (“Dirty Dozen”) comprised in the Stockholm Convention in 2001, their isomers and degradation products defined in the GMP Guidance. The second group comprises parameters related to additional 10 POPs included in the Stockholm Convention in 2009 and 2011 (“Nasty Nine” in 2009 and endosulfan in 2011) and specified in the updated GMP Guidance. The third group consists of all the other compounds related to the Stockholm Convention, their sums and toxic equivalents; however, these parameters are not specified in any of the GMP Guidance documents. The final group are all other compounds found in the GMP reports with no relation to the Stockholm Convention.


10

Data collected under the first round of the GMP reports (2009) provided baseline information on the pollution levels. Statistically based, time-related comparison of this baseline data with data coming from future reports will be possible for majority of recommended compounds as shown in the figures 3.1-3.3.

We note that such comparison will be complicated by a fact that the first GMP reports were not standardized. They contain two types of information: either concentration data on the individual congeners/isomers / transformation products or concentration summation of their congeners/isomers/transformation products. In some cases, chemicals are even reported under their general name with no specification of congeners/isomers/transformation products included. This fact has significantly increased heterogeneity and uncertainty of the GMP primary data, as it is unclear how and whether the summation was performed and which individual compounds were determined.

Frequent reporting of the temporally-aggregated values without clear instruction on such aggregation made a problem even more complex. Only a few monitoring stations worldwide are capable of providing statistically relevant regular time series with primary concentration measures; most of the other data are aggregated.

Therefore, it is highly desirable to significantly improve reporting requirements and standardization in the future. Fully standardized data fields, linked spatially to already collected GMP data, would enable time-related trend comparison for all POPs included in the Stockholm Convention.

Chapter 4. Analytical outcomes I. Ambient air data reported in Global Monitoring Plan regional reports

13

Chapter 4.

Analytical outcomes I. Ambient air data reported in Global Monitoring Plan regional reports

This chapter summarizes data on POPs in ambient air available in the first GMP report data collection in a pilot summary statistics. The main aims of such analysis were to provide content overview and identify statistically valuable data content. The analysis is called „pilot“, because it has been performed over re-coded GMP data without relevant expert approval and review.

4.1. Available GMP data for ambient air and aims of the pilot analysis

Two approaches are nowadays commonly used for the ambient air sampling. Cheaper and technically easier is a passive air sampling. Passive samplers are installed in the environment and ambient air is sampled by natural airflow of air masses through the sampling media inside the sampler. Major disadvantage is a semi-quantitative nature of data as the air flow cannot be measured directly. Second technique used is an active sampling, where active sampler is set on a site and surrounding air is actively pumped by electronically operated pump through the filters. Advantage of this technique is higher precision of reported concentration values; on the other hand, active samplers are rather expensive and demand stable power supply.

The GMP reports present records from both active and passive air sampling, with all consequences for data quality and content. Passive sampling provides concentration estimates that are directly aggregated in time and, therefore, the variance behind point estimates is significantly different from outcomes of active sampling. That is why both sampling techniques were reported separately and are also presented separately in following statistical summaries. In addition, passive sampling data were provided in two different reporting formats: as the air concentrations or the amounts of chemicals captured in sampling media.

For a purpose of future data evaluation, it is, however, necessary to have all air data in the same reporting format. Therefore, passive air data reported in the amounts of chemicals captured in sampling media were re-calculated to the air concentrations using theoretical sampling volumes of passive samplers. The value of theoretical sampling volume was estimated based on the long-term intercalibration of passive and active samplers. Standardization of the reporting format allowed for joining passive and active air data in the same data set.


14

4.2. Ambient air: active sampling

Concentration data on 71 parameters were published in the GMP reports for active air sampling. The list of measured parameters covers 47% of all parameters specified in chapter 2 of the Guidance on Global Monitoring Plan for POPs (2007).3,4 All of the original 12 POPs in the Stockholm convention (SC) are fully covered by active air monitoring data. Moreover, some chemicals of the additional 10 POPs were already reported. These chemicals were not obligatory for reporting in 2009, but they are important for future evaluation and comparisons. The reported chemicals from the latter group are summarized in the table 4.1.

Table 4.1: Compounds from the group of additional 10 POPs already reported in GMP reports for ambient air: active sampling (2009)

Alpha-HCH Beta-HCH Gamma-HCH PeCB Endosulfan I Endosulfan II Endosulfan SO4

In total, active air sampling data were reported for 4 UN regions (active air monitoring was not reported in the African regional report) in 31 countries and subregions.

Records covered a period of 20 years (1990 – 2009); however, a majority of data was collected recently (after 2004), see table 4.2.

DDT isomers, dieldrin, cis-chlordane, heptachlor and HCB were the most frequently reported chemicals, however the set of chemicals is strongly region-specific; therefore, a spatial comparison of parameters among the regions is often impossible.

Large fraction of data on active air sampling presented in table 4.2 were reported as the long-term series (four years or more) and are suitable for establishment of the long-term trends (table 4.3).

4.3. Ambient air: passive sampling

Data from the passive air monitoring programmes are available in all GMP reports, for 58 countries and geographical subregions. The overall number of reported parameters is 37; 19 out of those parameters (table 4.4) were from the group of congeners, isomers and degradation products recommended for monitoring in the Guidance document on GMP (2007). In addition, eight chemicals (alpha-HCH, beta-HCH, gamma-HCH, PeCB, sum of 3

3 Guidance on the Global Monitoring Plan for Persistent Organic Pollutants, Preliminary version, February 2007, Amended in May 2007, United Nations, Geneva - GE.07.00630 - April 2007, UNEP/CHEMICALS/2007/2

4 http://chm.pops.int/Implementation/GlobalMonitoringPlan/Overview/tabid/83/Default.aspx


15

PBDEs, endosulfan I, endosulfan II, endosulfan SO4) out of the 10 new POPs group were also reported.

Table 4.2: Overview of available ambient air: active sampling data

Region Country Years N of parameters N of entries

Asia and Pacific

Cambodia

2006

17

17

China 2004 28 546 Hong Kong SAR 1998 - 2007 11 38 Indonesia 2005 - 2006 20 40 Japan 1997 - 2007 24 831 Mongolia 2006 22 22 Philippines 2006 21 21 Republic of Korea 2005 - 2007 21 84 Thailand 2006 - 2007 21 42 Vietnam 2005 - 2006 22 43

CEEC

Czech Republic

1989 - 2009

19

261

Slovakia 1997 7 8 GRULAC

Chile

2002 - 2003

1

6

Mesoamerican subregion 2002 - 2004 10 650 Southern Cone subregion 2000 - 2001 2 24

WEOG

Arctic Canada

1993 - 2006

9

106

Australia 2002 - 2003 2 2 Austria 2005 - 2006 12 12 Canada 1990 - 2005 24 121 Finland 1996 - 2005 2 33 Germany 2005 - 2006 12 12 Greenland 2004 - 2005 5 6 Iceland 1995 - 2005 5 72 New Zealand 1996 - 1997 7 7 Norway 1993 - 2006 7 101 Russian Federation 1993 - 2002 8 40 Spain 1994 - 2007 2 5 Sweden 1993 - 2005 4 35 Switzerland 2005 - 2006 12 12 UK 1991 - 2006 2 6 USA 1990 - 2005 26 148


16

Table 4.3: Parameters reported in the GMP reports in ambient air: active sampling in a sufficiently long time series

Country Japan Hong Kong SAR Czech Republic Canada Canada Finland Iceland

Site Hateruma Island Kosetice Alert Kinngait Pallas Storhofdi

Aldrin X

Beta-HCH X X

cis-Chlordane (= alpha) X

trans-Chlordane (= gamma) X

cis-Nonachlor X

trans-Nonachlor X

Chlordane + Nonachlor X X X

o,p-DDD X

o,p-DDE X

o,p-DDT X

p,p-DDD X X

p,p-DDE X X

p,p-DDT X X X

5 DDTs X

DDT group X X

Dieldrin X X

dl-PCB I-TEQ

Endrin X

alpha-HCH X X

delta-HCH X

gamma-HCH X X

HCB X X X X X

Heptachlor X

Heptachlorepoxide cis- (= exo, B) X

Heptachlorepoxide trans- (= endo, A) X

Mirex X

Oxychlordane X

Parlar 26 X

Parlar 50 X

Parlar 62 X

PCB 101 X

PCB 118 X

PCB 138 X

PCB 153 X

PCB 180 X

PCB 28 X

PCB 52 X

Indicator 7 PCBs X

Sum 10 PCBs X X X

PCBs X

PCDDs/Fs (17) X

PCDDs/Fs I-TEQ X

PCDDs/Fs TEQ

PeCB X


17

Table 4.3: Parameters reported in the GMP reports in ambient air: active sampling in a sufficiently long time series (continue)

Country Norway Norway Norway Spain Sweden Sweden Sweden USA USA

Site Lista Svalbard Zeppelin Catalonia Aspvreten Raö Rorvik remote rural

Aldrin

Beta-HCH

cis-Chlordane (= alpha)

trans-Chlordane (= gamma)

cis-Nonachlor

trans-Nonachlor Chlordane + Nonachlor X X

o,p-DDD

o,p-DDE

o,p-DDT

p,p-DDD

p,p-DDE

p,p-DDT X

5 DDTs DDT group X X

Dieldrin

dl-PCB I-TEQ X X

Endrin

alpha-HCH delta-HCH gamma-HCH

HCB X X X

Heptachlor

Heptachlorepoxide cis- (= exo, B)

Heptachlorepoxide trans- (= endo, A)

Mirex

Oxychlordane

Parlar 26

Parlar 50

Parlar 62

PCB 101

PCB 118

PCB 138

PCB 153

PCB 180

PCB 28

PCB 52

Indicator 7 PCBs X X X

Sum 10 PCBs X X

PCBs

PCDDs/Fs (17) X

PCDDs/Fs I-TEQ X X

PCDDs/Fs TEQ X X

PeCB


18

Table 4.4: GMP Guidance (2007) recommended chemicals reported in ambient air: passive sampling

Cis-chlordane o,p-DDD p,p-DDE PCB 138 PCB 52 Dieldrin o,p-DDE p,p-DDT PCB 153 Trans-chlordane HCB o,p-DDT PCB 101 PCB 180 Trans-nonachlor Heptachlor p,p-DDD PCB 118 PCB 28

Passive air monitoring data were reported from all UN regions in 56 countries (see table 4.5). PCBs, DDTs, HCB and chlordane were the most frequently reported chemicals.

Table 4.5: Overview of available ambient air: passive sampling data

Region Country Years N of parameters N of entries Africa Democratic Republic of Congo 2008 19 152

Egypt 2008 25 122 Ethiopia 2008 19 114 Ghana 2005 - 2008 26 269 Kenya 2008 19 513 Malawi 2005 8 8 Mali 2008 19 570 Mauritius 2008 19 114 Nigeria 2008 19 152 Republic of Congo 2008 19 114 Senegal 2008 19 95 South Africa 2005 - 2008 26 481 Sudan 2008 19 114 Togo 2008 19 114 Tunisia 2008 19 114 Zambia 2008 19 114

Asia and Pacific Fiji 2006 - 2007 19 684 CEEC Armenia 2008 19 779

Belarus 2007 - 2008 19 665 Bosnia and Herzegovina 2006 19 171 Bulgaria 2007 19 570 Croatia 2007 19 475 Czech Republic 2003 - 2008 32 23778 Estonia 2006 - 2007 19 475 Hungary 2007 19 475 Kazakhstan 2008 19 779 Kyrgyzstan 2008 19 475 Latvia 2006 19 475


19

Lithuania 2006 19 418 Macedonia 2007 19 456 Moldova 2007 19 665 Montenegro 2007 19 665 Poland 2007 32 663 Romania 2006 - 2007 19 1501 Russia 2007 32 453 Serbia and Montenegro 2006 19 608 Slovakia 2006 19 1045 Slovenia 2007 19 646 Ukraine 2008 19 570

GRULAC Andean subregion 2005 - 2008 9 118

Caribbean subregion 2005 - 2008 8 40 Mesoamerican subregion 2005 - 2006 9 79 Southern Cone subregion 2002 - 2008 9 131

WEOG Australia 2005 - 2006 18 178

Bermuda 2005 - 2006 18 89 Canada 2005 - 2006 18 379 Finland 2005 - 2006 18 74 France 2005 - 2006 18 74 GAPS 2005 - 2006 6 10 Iceland 2005 - 2006 18 88 Ireland 2005 - 2006 18 89 Italy 2005 18 104 Norway 2005 18 72 Spain 2005 - 2006 18 150 Turkey 2005 - 2006 18 88 USA 2005 - 2006 18 279

First data reported on ambient air-passive sampling are from 2003. The only site with at least 4 years of data available for establishment of the long-term trends is Košetice, Czech Republic.

Chapter 5. Analytical outcomes II. Human tissues data reported in Global Monitoring Plan regional reports

20

Chapter 5.

Analytical outcomes II. Human tissues data reported in Global Monitoring Plan regional reports

POPs detected in human tissues – breast milk and blood – were most often reported as lipid-adjusted concentrations in the data sets of GMP reports. The range of compounds and their spatial and temporal coverage reported for breast milk and human blood is significantly lower than that for ambient air. Nevertheless, all of the original 12 POPs in the Stockholm Convention as well as GMP Guidance 2007 recommended analytes were reported. A high heterogeneity and low degree of standardization is enhanced by reporting various toxic equivalents (TEQ) (based on WHO TEF values from various years) rather than concentrations of the individual PCDDs, PCDFs and PCBs congeners. However, a significant part of standardized data allows, after expert review and approval, for further processing and assessment of the contamination baselines.

5.1. Introduction

POPs detected in human tissues are usually reported as lipid-adjusted concentrations as shown in the data sets of GMP reports. Profile of reported chemicals in human tissues – breast milk and human blood – is expectedly partially different from that reported for ambient air (see chapter 4).

GMP1 reports cover 121 variables (congeners, isomers or transformation products) in breast milk and 82 in human blood; however, a large fraction of data is reported only as an outcome of occasional sampling campaigns performed at one given time point. Time series suitable for modelling of the temporal trends are much less frequent than in the case of ambient air (see chapter 4). Nevertheless, reported data allow for establishment of the baseline concentration levels.

5.2. Breast milk

Human milk should be sampled according to the WHO protocol (WHO/ECEH, 1996).5 The most common substances tracked in the milk all over the world are dibenzo-p-dioxins and

5 WHO/ECEH, 1996. Levels of PCBs, PCDDs and PCDFs in human milk. Second round of WHO-coordinated exposure study. Environmental Health in Europe Series 3, 121pp.


21

dibenzofurans (often reported as a sum of PCDDs/Fs) and polychlorinated biphenyls (PCBs). All these groups of compounds are usually expressed as toxic equivalents (TEQ) of 2,3,7,8-tetrachlorodibenzo-p-dioxin (in pg/g of lipid). Calculation of the TEQs can be based on different sets of toxic equivalency factors (TEF) that are updated on the regular basis (i.e. I-TEF from 1989, WHO-TEF from 1998, and 2005) which unfortunately adds to the complexity of reported data.

In the GMP Guidance, it is recommended to report two groups of PCBs: “12 dioxin-like PCB congeners” with TEFs that are reported together with PCDDs/Fs. The other group are so-called “indicator PCBs”. They are reported either as a sum 6 PCBs (PCB 28, PCB 52, PCB 101, PCB 138, PCB 153 and PCB 180) or 7 PCBs, where PCB 118 is also included together with previously mentioned six PCBs.

Large spatial coverage (breast milk data reported from 41 countries) and a sufficient time scale of collected data in some countries allow for a comparison among regions and basic evaluation of trends. Reported chemicals also adequately followed the recommendation of the Stockholm Convention; all of the original 12 POPs as well as analytes specified in the Guidance on GMP (2007) were covered in the first GMP report. There were 37 parameters reported in all UN regions (table 5.1) that enabled spatial variation analysis.

Table 5.1: Parameters reported in all UN regions – breast milk

13 PCDDs/Fs isomers Heptachlorepoxide cis- OCDF PCB 101 PCB 180

Dieldrin o,p-DDD Oxychlordane PCB 118 PCB 52 Endrin group o,p-DDE p,p-DDD PCB 138 Toxaphene HCB o,p-DDT p,p-DDE PCB 153 Trans-chlordane Heptachlor OCDD p,p-DDT PCB 156 Trans-nonachlor

Pilot inspection of variability in concentration time series suggested that at least 4-year period of monitoring is needed for an assessment of the temporal trends. Parameters reported in the GMP reports based on four and more years lasting monitoring within one country) are presented in table 5.2.

5.3. Human blood

There were four types of blood samples reported in the GMP reports: maternal blood, core blood, serum and plasma. In this chapter, we provide summary information on availability of data from any blood samples. GMP data on POPs in human blood are commonly reported as concentrations of chemicals in the lipid fraction of the fluid (lipid content-adjusted values), similar to breast milk. In blood, there were 82 parameters (congeners, isomers, transformation products and TEQs) reported covering all original 12 SC POPs. Similarly to breast milk, a relatively large fraction of compounds was reported only occasionally. Analytes recommended for reporting by the GMP Guidance are covered in the regional reports by 85%.


22

The only missing items are dioxin-like PCB congeners (PCB 105, 114, 123, 156, 157, 167 and 189) and cis-nonachlor. Several chemicals from the group of additional 10 POPs were reported (alpha-HCH, beta-HCH, gamma-HCH) in human blood as well.

The results of human blood monitoring were reported from 17 countries; spatial coverage is thus not as widespread as in the case of breast milk. The most frequently reported substances were HCB, DDE and DDT (mainly p,p- isomers), PCB 153, dieldrin and beta-HCH. However, only HCB and DDE (or p,p-DDE isomer) were reported in all UN regions.

The most compact datasets were reported from Japan (dl-PCBs and PCDDs/Fs WHO-TEQs between 2002 and 2006) and from Germany (4 cities, PCB 153 and HCB between 1995 and 2006 and between 1998 – 2006 respectively). Data reported from all the other countries can be only used for establishment of the baseline concentration levels, not the time trends. DDTs and PCBs are the most frequently reported groups of compounds (table 5.3).

Table 5.2: Parameters reported in the GMP reports for breast milk in a sufficiently long time series

Country India Japan Belgium Finland Germany Norway Sweden Czech Republic

DDE X X DDT X 4 DDTs X X DDTs X X Dieldrin X HCB X X X X Oxychlordane X Oxychordane + Nonachlor X PCDDs/Fs I-TEQ X PCDDs/Fs TEQ PCDDs/Fs WHO-TEQ X PCB 118 X PCB 153 X Indicator 6 PCBs X X X Sum 3 PCBs X Sum 10 PCBs X PCBs X X

Table 5.3: Parameters reported in the GMP reports for human blood in a sufficiently long time series

Country India (human blood)

Japan (human blood)

Germany (blood plasma)

Brazil (blood serum)

p,p-DDE X DDTs X HCB X PCDDs/Fs WHO98-TEQ X dl-PCB WHO98-TEQ X PCB 153 X


23

5.4. Conclusion

Breast milk and human blood data reported in the first GMP report provided a relatively sufficient set of chemicals recommended for monitoring under the Stockholm Convention. Most of data are reported as aggregated (pooled samples, cohorts attributed to the whole country) on one-year base. Annually aggregated data can be used for robust estimation of trends if at least four consecutive time points are available.

The time and spatial scales covered by human blood monitoring were narrower than in the case of breast milk. The results of the human blood monitoring were reported from 17 countries only (breast milk was reported from 41 countries), and a sufficiently long time series were rather scarce. Nevertheless, reported data are valuable due to a wide variety of reported parameters. The analytes recommended for the reporting by the GMP Guidance 2007 were covered by 85%.

In general, data reporting POPs in human tissues represent reliable dataset for an assessment of baseline levels and for simple time trends; the main source of such valuable data are the worldwide sampling campaigns organized by the WHO. The WHO has already run four6 2-year rounds (first in 1987, last in 2007) and has collected data about a wide range of chemicals (larger group of chemicals than POPs) in the breast milk from all parts of the world. Unfortunately, not all data (even in aggregated form) from the existing WHO campaigns were reported in the GMP reports. Additional information is available in these original reports:

• Campaign 1: Environmental Health Series No. 34 (1989), Levels of PCBs, PCDDs, and PCDFs in breast milk, WHO Regional Office for Europe, Copenhagen, Denmark

• Campaign 2: Environmental Health Series No. 3 (1996), Levels of Polychlorinated dibenzo-p-dioxines (PCDDs), polychlorinated dibenzofurans (PCDFs) and polychlorinated biphenyls (PCBs) in human milk: Second round of WHO-coordinated exposure study, WHO Regional Office for Europe, Copenhagen, Denmark

• Campaign 3: Van Leeuween, FXR, Malish R. Results of the third round of WHO-coordinated exposure study on the levels of PCBs, PCDDs and PCDFs in human milk. Organohologen Compounds (2002)56:311-316

• Campaign 4: Fourth WHO-Coordinated Survey of Human Milk for Persistent Organic Pollutants: Guidelines for Developing a National Protocol

6 http://whqlibdoc.who.int/euro/1994-97/EUR_ICP_EHP_M02_03_05.pdf http://www.who.int/foodsafety/chem/POPtechnicalnote.pdf Van Leeuween, FXR, Malish R. Results of the third round of WHO-coordinated exposure study on the levels of PCBs, PCDDs and PCDFs in human milk. Organohologen Compounds (2002)56:311-316 http://www.who.int/foodsafety/chem/POPprotocol.pdf

Chapter 6. Methodology I. GMP regional reports – data standardization and proposal for database structure

27

Chapter 6.

Methodology I. GMP regional reports – data standardization and proposal for database structure

6.1. Introduction

This chapter describes the steps of the data standardization including main identified challenges in the original data sets. The outcome of the chapter is presentation of the database structure and description of its entities including description of their relation.

6.2. Identified challenges in the standardization of GMP reports data

Several data issues have been identified in the GMP reports reviewing process. The discrepancies and uncertainties of reported data originated in a lack of intercalibration of reporting processes among various countries and regions and standardization of reporting format. This chapter summarizes main systematic errors, demonstrates adopted approaches and solutions applied during a re-coding of the GMP reports into electronic data form.

The following list contains main challenges encountered:

1. Limited standardization of the used nomenclature (i.e. trans-chlordane vs. gamma-chlordane or o,p‘-DDD vs. o,p-DDD).

• Solution: The names of all parameters were standardized.

2. Reporting parameters as “chlordanes”, “PCBs” or “DDTs” without specification of the isomers or degradation products summed up.

• Solution: Whenever possible, missing information was found in the text. When not found, parameters that were not specified properly were kept separately as new parameter.

3. Difference in labelling of PCBs sums.

• Solution: The general label used was „Sum x PCBs“, where x stands for a number of congeners. In the case of PCBs indicator, it was distinguished whether 6 or 7 congeners were used in the sum. Both cases were kept as a separate parameter.

4. Not specified whether “heptachlor” means just a parent compound or includes its metabolites as well (“heptachlor group”).


28

• Solution: Whenever possible, missing information was found in the text. When not found, all categories were kept separately.

5. Limit of Quantification (LOQ) not reported.

• Solution: This parameter is needed for future statistical evaluation of data. Currently, it is available for 10% of data only. Available values were recorded; however, methodology for future statistical evaluation has to be adopted as a next step.

6. Concentration values reported in various units.

• Solution: For each parameter/matrix (or matrix fraction) pair only one concentration unit was allowed. All concentration values were re-calculated to this unit if necessary.

7. Concentration units not clearly defined in the table captions of GMP reports.

• Solution: In most cases, units were found in the relevant parts of the report text. In cases where an appropriate unit could not be found in the GMP report it was identified based on the expert opinion.

8. Unclear reporting of units in the breast milk analysis (ng or pg/ml, ng or pg/l, ng or pg/g and ng or pg/g lipid adjusted, even ng/l lw).

• Solution: Information was found in the text or appropriate units were indirectly identified from the values. Data reported per liquid volume and per lipid content were kept separately as they are not directly transferrable. Applies to blood samples as well.

9. Results for the passive air monitoring were reported in two different formats: as air concentrations or quantity of chemicals captured in sampling media (ng/m3 or ng/sample).

• Solution: Passive air data reported in the quantity of chemicals captured in sampling media were re-calculated to the air concentrations using theoretical sampling volumes of passive samplers. The value of theoretical sampling volume was estimated based on the long-term intercalibration of passive and active samplers.

10. Results for human blood were reported in four different formats: as human blood, core blood, plasma or serum.

• Solution: These four categories were kept separately as four independent matrices since their results are not directly comparable.

11. TEQ (WHO-TEQ): missing or wrong specification of the TEF values used for calculation of the TEQ (year). In addition to commonly used WHO-TEQ98 and WHO-TEQ2005 values, values of WHO-TEQ1995, 1997, and 2001 were reported.


29

• Solution: WHO-TEQ97 specification was replaced by WHO-TEQ98. WHO-TEQ1997 and WHO-TEQ2001were kept together with other not specified WHO-TEQ parameters separately from correctly identified parameters (WHO98-TEQ and WHO2005-TEQ) because various TEQ versions are not directly comparable.

12. Subregion Northern Scandinavia without any specification of involved countries.

• Solution: Such subregions were kept in database in place of the country. For visualization, data was assigned to all involved countries (Norway, Finland and Sweden) of the Scandinavian Peninsula when creating the maps.

13. Ambient air data reported as spatially aggregated values, often without specific information on number and type of aggregated sites.

• Solution: The GMP requires evaluation of the long-term trends of POPs in ambient air at background sites. Information on classification of the sampling sites is often not provided in the first GMP reports but has to be included in the database as well as in future reporting templates or tools.

Keeping the individual parameters identified in the GMP Guidance document, and their well defined sums and TEQs apart from those that are not well defined, is a basic requirement allowing for data visualization and interpretation. However, it also means working with a very high number of parameters (almost double when compared to the set recommended by the GMP Guidance).

6.3. Transfer of the GMP reports content into a database – general approach

Transforming data from regional reports into a form that was suitable for both database development and analytical processing was a complex task. Data had to be extracted from textual form, tables, citations and charts into a common standardized data structure with shared property lists (matrices, chemicals, units…). The methodology used can be divided into a series of five consecutive steps described below. The resulting output is given in the section 6.3.

6.3.1. Transfer of the data: consecutive steps

Identification of collected data: First of all, an overall inspection of all regional reports was performed. The main goal was to identify reported matrices, measured chemicals (compounds), used values and aggregation characteristics as well as sampling frequency and time ranges. Further analytical outcomes were designed according to output of this inspection.

Datasets volume identification: Determination of volume (range and size) of datasets was equally important, because appropriate storage technology had to be chosen. Reported


30

datasets are not of high volume: they consist of several thousands of records, which is a relatively small volume for modern database systems. Nevertheless, there is a need for fast reporting and fast outcome generation over multidimensional data. A relational database engine was used as a main storage solution in order to enable fast sorting, aggregation and selection.

Design of a common data structure: For the purposes of the data extraction from particular regional reports, a common data structure was used allowing for maintenances of all collected values – both directly measured and aggregated. Used common data structure consists of:

• UN Region • Country • Site • Matrix • Parameter • Year – start • Month – start • Day – start • Year – end • Month – end • Day – end • Aggregation (in case of aggregated value)

o Mean o Median o Minimum o Maximum o SD – standard deviation o SE – standard error o GM – geometrical mean

• Measured value (in case of directly measured value) o Value o LOQ

• Unit

Extraction of data from regional reports into a common data structure: During the process of data extraction from regional reports, a number of problems had to be solved as described above in part 6.1. All missing information had to be looked up in the supporting text of the reports or was completed based on the context knowledge. Most of the problems identified in 6.1 were caused by the non-existent standardized form for data collection, which inevitably brought a level of uncertainty into further processing and results.

Transformation of extracted data into a database: For the purposes of analytical outcomes from regional reports, annual aggregation of values was used. Aggregations were calculated and prepared using specialized software (SPSS). Import of re-calculated data into the database was done by automated procedures to ensure data consistency and integrity.


31

6.4. Transfer of the GMP reports content into a database - Outcome: Aggregated data structure

Aggregated values were stored in a relational database consisting of 10 basic entities:

• UN Region • Country • Site • Matrix • Compound • Parameter • Record • Aggregation • Percentile • Unit

The database structure used and interactions among its entities are shown in figure 6.1.

Figure 6.1: Relational database developed to store the re-coded content of the GMP reports

The following text provides a detailed description of each entity and its content.

Record

The “Record” serves as a basic entity to hold all information together.

• year – indicates for which year the aggregation was performed; • note – space to store supplementary information if needed.


32

Matrix

“Matrix” is a data structure to hold a list of matrices (passive air, human blood…)

• name – full name of matrix; • description – space to add some additional description if needed.

Aggregation

The “Aggregation” entity stores basic aggregation characteristic such as:

• n – number of aggregated values; • sd – standard deviation; • se – standard error; • mean – value of mean; • mean LoQ – value of limit of quantification for mean; • geomMean – geometrical mean; • geomMeanLoQ – value of limit of quantification for geomMean; • unit – unit of aggregated values.

Percentile

“Percentile” is a data structure item to store any percentile for a given aggregation. In most cases, 0th percentile as minimum, 100th percentile as maximum and 50th percentile as median are used.

• level – level of percentile (50th, 75th…) • value • valueLoQ – value of limit of quantification for a given percentile value

Compound

“Compound” is a code list containing 22 compounds under the Stockholm Convention (original twelve and additional ten POPs) without further specification.

• name – full name of compound

Parameter

“Parameter” is a specific congener, isomer or degradation product reported in the GMP report (see column E of the Annex 6.1), which falls under one of the items specified in the “Compounds” code list.

• name – full name of parameter

Country

The “Country” item holds a list of countries / states enrolled into reports.


33

• name – full name of country • code – short name or abbreviation

Site

“Site” is a data structure which identifies specific site, from which the parameter is reported.

• name – full name of site • site classification

UN region

“Region” is a data structure item that contains code list of geopolitical regional groupings according to the United Nations.

• name – full name of regional group • description – space to add some additional description if needed

Additional information: LoQ values

Information on limit of quantification (LOQ) associated with every reported concentration value is crucial for future statistical evaluation of the GMP data. Common practice is that data reported as “below limit of quantification” (<LOQ) are replaced by a halved LOQ rather than zero for the purposes of statistical evaluation). Thus, the exact value of the limit of quantification should be attached to every measured and reported concentration value. Table 6.1 shows options of <LOQ data processing.

Table 6.1: Possible cases in processing LOQ values

Case #1 Case #2 Case #3 Value Exact number NULL NULL ValueLoQ NULL or exact number Exact number -1 The measured value is

known and is above LOQ The measured value is under LOQ and LOQ is known

The measured value is under LOQ and LOQ is unknown

NULL in database systems is equivalent to void cell (value).

In the first GMP reports, the LOQ values are available only for 10% of reported concentration values and they are not included in the first version of the database. It means that only cases #1 and #3 can be found in the existing output: the constant -1 indicates that measured value is under LOQ but the exact value of LOQ is not known.

For a purpose of the pilot statistical evaluation of data, LOQ data should be included in the next version of database. In majority of cases when the LOQ value is unknown, it can be replaced either by zero or by a half of the lowest reported value. However, it is crucial to include LOQ among required parameters in the standardized reporting format for the future monitoring reports.

Chapter 7. Methodology II. Information and communication technologies

34

Chapter 7.

Methodology II. Information and communication technologies

7.1. Introduction

Chapter 6 (data standardization) of this report proposes principles for processing of existing GMP data and suggest areas where standardization is critical. This chapter briefly describes IT requirements and methodology used for archiving of obtained data, on-line database suitable for future data collection campaigns and reporting system. We draw our experience from the development of the criteria and standards used in the GENASIS information system (www.genasis.cz)

7.2. Main database server

Main database server holds all data needed to create online interactive reports of analytical outcomes. We chose the relational database management system (RDMS) because of its stability, interoperability and ability to guard data integrity and achievement of high level. The MSSQL Server on Windows Server operating system is used as RDMS. This arrangement is considered a standard solution for small, medium and enterprise level applications.

7.3. Data access server

Data access server works as a middle ware technology enabling online reporting interface to communicate with the main database server. In addition, it can be used as a place to implement data caching strategies.

Database selection queries, data transformations and some supporting calculations are performed on the server. Data access layer runs on Apache web server. It was created in PHP with support of Zend Framework. Communication between data access server and online reporting interface over network is implemented by binary communication protocol AMF.

7.4. Reporting system description

Reporting system architecture is illustrated in figure 7.1. All reporting outcomes are designed to be fully accessible online via web forms or interactive SW tools. Graphical and SW tools


35

libraries of the GENASIS system (www.genasis.cz) were used for the pilot online reporting tool over already accessible GMP data (as described in chapters 9 and 10).

Reporting system task is to enrich and visualize analytical outcomes. It should combine two main data resources:

• Main database, which stores aggregated data, analytical outcomes calculation results and support metadata;

• Map layers, which hold spatial data.

Figure 7.1: Architecture of reporting system applied for processing of GMP data

7.5. Business intelligence to support interactive reports

Basic concepts of business intelligence processing are implemented to speed up online analytical outcomes over GMP data. Core concepts of OLAP methodology are used and basic data structures in business intelligence layer are transformed into fast-to read multi-dimensional data structures.


36

7.6. User interface technologies for interactive reports

Online reporting interface plays role of a client's application, which combines spatial data resources and tabular analytical outcomes from main database in one place. The interface is able to display both textual and multimedia information in combination with interactive charts, map windows, tabular data, etc.

User can customise reports by selecting data and their range, filtering values by categories (i.e. UN region, matrix, compounds and other parameters). User can also click on item of interest to get complementary information.

Performance boosters are used to speed up work with online reporting interface. Application screens are cached, so that there is no need to recreate them again after switching views. Asynchronous data processing is implemented as well as speed up screen loading time. User interface was created using Adobe Flex technology, which needs to run a Flash Player plug-in installed in user browser.

Current version of the online pilot reporting tool uses GMP data collected in the 1st global sampling campaigns and recorded in the reports. Standard procedures require that data are verified and approved for online processing at first. Since that step was not performed, the online reporting interface is password protected. Access is permitted only to authorized users. Disclaimer on the title page informs all users that they are using pilot version of the online reporting tool and that presented data still need to be verified and approved for on-line presentation.

7.7. Using GIS to present spatial and time-related data

Standard GIS technologies are used to visualize spatial data for better, faster and comprehensive understanding. Therefore the online reporting tool presents selected analytical outcomes as map layers presenting given features as shapes. Map layers created in ArcMap desktop software are published as map service via ArcGIS Server and enabled to load into online reporting interface. They show all spatial data together with base (background) map(s) and desired interactivity. Time-related data are animated, so that time progress can be easily shown (in a slideshow).

To present spatial data in pilot online reporting tool several ESRI software company technologies were used: ArcMap to create base map layers (background), ArcGIS Server to publish map layers as web map services and ArcGIS API for Flex to create map window in online reporting interface.


37

7.8. Data and system security

Data quality and security is a prerequisite for correct functioning of the proposed system. The following text is an overview of data protection and security system used for www.genasis.cz, which was used as model for the online pilot reporting tool.

High quality servers with hardware encryption of disc array are used.

Servers are operated in internal IBA/RECETOX data centre in the Masaryk University network with multiple level security:

• Central university network traffic monitoring • Firewall of IBA/RECETOX subnet and dedicated firewalls on each server • Servers are placed in a separated subnet dedicated only for server traffic • Server operating system is updated on regular basis • Physical access to data centre is strictly limited to authorized personnel • Data centre is under nonstop camera monitoring, access is possible only via electronic

cards and the building has its own physical security • Servers have backup power supply and there is specialized mode prepared in case of

fire or accident

All development and implementation steps are guaranteed by ISO certificates: EN ISO 9001:2001 (Quality Management Systems), ISO/IEC 20000-1:2006 (Information Security Management Systems), and ISO/IEC 27001:2006 (IT Service Management.

7.9. Conclusions

Based on the above we propose to keep the same methodical standards for the creation of the global database and information system, which should cover future GMP data gathering campaigns. If the GENASIS-like infrastructure and tools are selected for processing of data originating from future GMP sampling campaigns, some development customizing the tools for the POPs ontologies and supporting user’s skills and comfort is needed.

Chapter 8. Methodology III. Statistical data processing

38

Chapter 8.

Methodology III. Statistical data processing

8.1. Introduction

This chapter describes issues that must be solved prior to any processing of environmental data. Every chemical released into the environment is measured and evaluated through its concentration, no matter in which matrix it is observed. Temporal variations of the concentration characterize the fate of a compound in the environment, but the concentration variations can represent two different issues – either the overall trend or the seasonality component of the trend. Prior to data processing it must also be clear whether the record contains primary or aggregated values (usually yearly or multiple-year aggregated values). Though the aggregation limits the quantification of influence of within-year variance, it facilitates the overall trend assessment.

In addition, a record containing annual values (both means and medians) cannot be directly interpreted and used as point values without any variance. Nevertheless, point values can still be of use for future comparison of trends if a variance could be determined.

All of the above described issues occur readily in the analyzed GMP data records:

• Substantial part of the data reported in GMP were published as yearly (or multiple-year, aggregated) values.

• Annual values (means and medians) were often reported without any relevant variance.

This chapter explains below how to determine a variance of existing point values via model time series and describes variability inspection and extrapolation by non-parametric methods that would be of use for processing existing GMP records. All statistical tools proposed are complemented by a robust list of references in the final part of this chapter.

8.2. Model data series and goals of the analysis

The variance of already obtained values can be indirectly estimated from model time series, but main added value of the methodology is not an artificial reconstruction of the sample distribution, but estimation of the so-called minimum effect size, i.e. to quantify the difference between two pair-wise point estimates (e.g., obtained in the former and future GMP reporting), which can be treated as statistically significant regarding some variance accompanying the concentration estimates.


39

The variability analysis of air pollution data from long-term monitored model site, Košetice, Czech Republic, is shown including well founded extrapolation that could be also applied to existing data in GMP reports. The Košetice dataset was chosen because of the quality of data collection at this station. The Košetice observatory is a worldwide-known background station involved in several international monitoring networks (e.g. EMEP) such as long-term regular monthly monitoring of ambient air by passive samplers. Highly standardized sample handling and analyses provide excellent quality of data outcomes.

Analysis of the Košetice dataset can demonstrate variation hidden in values aggregated the way as in some of the GMP records. Furthermore, the analysis of species-specific data will document which sources of variability (between or within year variability) are the most important ones. The revealed variability will then be used for the reconstruction of the sample data distribution based on log-normal model to show the usability of one-year variability data for future processing. Finally, the minimum effect size will be computed for all relevant parameters, thus quantifying the differences detectable as statistically significantly different from the geometric mean of analyzed data under the prerequisite of revealed one-year variability.

8.3. Methodology

A standard descriptive statistics was used for the selected data monitoring. Mean and standard deviation were used for primary non-transformed data to show the overall variability and influence of asymmetric sample distribution (see table 8.1). Due to the skewed distribution, the median supplemented by 5th – 95th percentile range together with geometric mean were also computed. Mean and standard deviation of data transformed by natural logarithm are also provided, because it is used in subsequent evaluation of time series variability and minimum effect size.

The variability sources of time series were analysed using one way analysis of variance within year as an explaining factor using log transformed data (see table 8.2.). Estimate of the minimum effect size was based on one-sample t-test comparing time series with 12 measurements (12 months during one year sampling). For the results see table 8.3. The computation was performed using log-transformed data, annually adjusted measure of variability and grand mean of time series are expressed as lower and upper border of statistically significant differences around geometric mean of primary data.

The estimated mean and standard deviation of log transformed data were used for the reconstruction of log-normal distribution of data from one-year aggregation and compared to the distribution of sampled values (see figures in this chapter). All compounds are described both by parametric statistics (mean, standard deviation) on primary non-transformed data and due to the skewed distribution also by non-parametric statistics and results based on natural log transformed data (geometric mean; mean and standard deviation of transformed data). The variability was assessed both for the whole time series and after subtraction of annual differences (year-adjusted measure of data variability).


40

Table 8.1: Basic description of analyzed model dataset (Košetice station, N=62; 6 years of sampling)

Primary data Log transformed data

Compound Mean (SD) Geometric mean

Median (5-95th percentile) Mean (SD) SD adjusted for years

influence

alfa-HCH 3.14 (2.28) 2.48 2.79 (0.60; 5.94) 0.91 (0.75) 0.72

beta-HCH 0.82 (1.02) 0.38 0.44 (0.10; 2.81) -0.96 (1.30) 1.00

delta-HCH 0.13 (0.12) 0.11 0.10 (0.10; 0.28) -2.20 (0.41) 0.41

gamma-HCH 5.31 (3.12) 4.35 4.75 (1.26; 10.14) 1.47 (0.70) 0.52

HCB 22.11 (17.82) 14.66 13.74 (5.04; 54.30) 2.69 (1.18) 0.76

o,p-DDD 0.19 (0.17) 0.15 0.10 (0.10; 0.52) -1.91 (0.66) 0.66

o,p-DDE 0.19 (0.15) 0.16 0.10 (0.10; 0.44) -1.85 (0.61) 0.61

o,p-DDT 0.16 (0.21) 0.12 0.10 (0.10; 0.62) -2.11 (0.57) 0.57

p,p-DDD 0.57 (1.08) 0.30 0.32 (0.10; 1.77) -1.21 (1.02) 0.75

p,p-DDE 6.02 (2.64) 5.46 5.90 (2.44; 9.85) 1.70 (0.46) 0.46

p,p-DDT 0.82 (0.76) 0.52 0.53 (0.10; 2.27) -0.65 (1.03) 1.03

DDD 0.67 (1.07) 0.38 0.46 (0.10; 1.77) -0.96 (1.03) 0.73

DDE 6.13 (2.64) 5.58 5.91 (2.72; 9.94) 1.72 (0.45) 0.45

DDT 0.91 (0.80) 0.65 0.74 (0.20; 2.34) -0.43 (0.83) 0.76

DDTs 7.71 (3.14) 7.11 7.39 (3.43; 12.60) 1.96 (0.41) 0.39

PCB 101 1.00 (1.62) 0.54 0.57 (0.10; 2.53) -0.61 (1.05) 0.82

PCB 118 0.31 (0.50) 0.20 0.10 (0.10; 0.64) -1.63 (0.84) 0.81

PCB 138 0.89 (1.25) 0.52 0.58 (0.10; 2.44) -0.65 (1.03) 0.82

PCB 153 1.31 (1.41) 0.91 0.92 (0.23; 4.00) -0.10 (0.86) 0.83

PCB 180 0.56 (0.72) 0.34 0.38 (0.10; 1.51) -1.08 (0.97) 0.97

PCB 28 1.32 (0.70) 1.14 1.11 (0.57; 2.65) 0.13 (0.58) 0.53

PCB 52 2.00 (1.64) 1.50 1.26 (0.54; 5.28) 0.40 (0.75) 0.63 6 indicator PCB

sum 7.07 (5.71) 5.46 5.09 (1.78; 18.68) 1.70 (0.71) 0.59

PeCB 2.74 (1.80) 2.14 2.30 (0.82; 6.40) 0.76 (0.83) 0.83

41

Table 8.2: Sources of variability in time series – analysis for selected pollutants Year Variability source

Compound 2003 2004 2005 2006 2007 2008 Between years

Within years p

alfa-HCH 3.40 (2.75; 4.21) 2.49 (1.97; 3.14) 2.67 (1.73; 4.12) 3.63 (2.88; 4.57) 1.95 (1.26; 3.03) 1.43 (0.61; 3.33) 16.0% 84.0% 0.076

beta-HCH 1.38 (0.24; 8.00) 0.27 (0.15; 0.49) 0.18 (0.11; 0.30) 1.04 (0.66; 1.63) 0.38 (0.18; 0.79) 0.24 (0.10; 0.62) 29.4% 70.6% 0.001

delta-HCH 0.10 (0.10; 0.10) 0.11 (0.09; 0.14) 0.15 (0.10; 0.23) 0.10 (0.10; 0.10) 0.10 (0.10; 0.10) 0.10 (0.10; 0.10) 14.2% 85.8% 0.116

gamma-HCH 4.91 (2.98; 8.09) 4.36 (3.21; 5.91) 5.27 (4.15; 6.69) 6.43 (4.72; 8.76) 3.33 (2.01; 5.51) 2.38 (1.38; 4.12) 21.2% 78.8% 0.018

HCB 39.58 (29.11;53.81) 31.89 (24.64;41.28) 14.90 (4.34; 51.22) 11.41 (9.59; 13.58) 8.28 (6.98; 9.82) 9.60 (8.37; 11.01) 20.6% 79.4% 0.021

o,p-DDD 0.18 (0.12; 0.27) 0.13 (0.09; 0.18) 0.14 (0.09; 0.22) 6.8% 93.2% 0.541

o,p-DDE 0.18 (0.12; 0.27) 0.13 (0.10; 0.17) 0.17 (0.11; 0.25) 6.6% 93.4% 0.556

o,p-DDT 0.11 (0.09; 0.13) 0.13 (0.09; 0.18) 0.14 (0.07; 0.25) 2.6% 97.4% 0.846

p,p-DDD 0.10 (0.10; 0.10) 0.20 (0.09; 0.41) 0.55 (0.28; 1.09) 0.45 (0.33; 0.62) 0.25 (0.17; 0.36) 0.24 (0.16; 0.37) 23.2% 76.8% 0.010

p,p-DDE 6.37 (5.23; 7.76) 5.71 (4.43; 7.36) 5.88 (4.94; 6.99) 6.00 (4.38; 8.23) 3.98 (3.07; 5.15) 5.75 (4.12; 8.03) 11.7% 88.3% 0.209

p,p-DDT 0.78 (0.35; 1.74) 0.84 (0.53; 1.33) 1.03 (0.77; 1.38) 0.23 (0.13; 0.38) 0.27 (0.14; 0.49) 0.76 (0.50; 1.18) 38.4% 61.6% <0.001

DDD 0.10 (0.10; 0.10) 0.20 (0.09; 0.41) 0.56 (0.28; 1.10) 0.68 (0.51; 0.90) 0.40 (0.30; 0.54) 0.40 (0.28; 0.58) 29.0% 71.0% 0.001

DDE 6.37 (5.23; 7.76) 5.71 (4.43; 7.36) 5.89 (4.96; 6.99) 6.31 (4.70; 8.47) 4.13 (3.20; 5.32) 5.95 (4.29; 8.26) 11.7% 88.3% 0.211

DDT 0.78 (0.35; 1.74) 0.84 (0.53; 1.33) 1.04 (0.77; 1.39) 0.36 (0.24; 0.54) 0.44 (0.26; 0.73) 0.91 (0.58; 1.45) 25.9% 74.1% 0.004

DDTs 7.53 (7.00; 8.11) 7.45 (5.71; 9.72) 8.01 (6.95; 9.24) 7.64 (5.99; 9.74) 5.24 (4.17; 6.59) 7.43 (5.39; 10.25) 13.7% 86.3% 0.133

PCB 101 0.71 (0.58; 0.86) 1.24 (0.67; 2.31) 1.09 (0.68; 1.75) 0.37 (0.23; 0.58) 0.19 (0.13; 0.27) 0.40 (0.28; 0.56) 45.6% 54.4% <0.001

PCB 118 0.16 (0.06; 0.39) 0.24 (0.13; 0.45) 0.31 (0.20; 0.49) 0.14 (0.10; 0.19) 0.14 (0.10; 0.20) 0.23 (0.13; 0.38) 15.3% 84.7% 0.089

PCB 138 0.89 (0.60; 1.32) 1.27 (0.80; 2.03) 1.20 (0.88; 1.65) 0.29 (0.18; 0.45) 0.19 (0.13; 0.27) 0.32 (0.22; 0.48) 59.1% 40.9% <0.001

PCB 153 0.52 (0.18; 1.52) 1.50 (0.97; 2.32) 1.98 (1.40; 2.82) 0.69 (0.47; 0.99) 0.45 (0.34; 0.59) 0.72 (0.50; 1.03) 44.1% 55.9% <0.001

PCB 180 0.67 (0.47; 0.94) 0.55 (0.30; 1.00) 0.70 (0.49; 1.01) 0.21 (0.14; 0.32) 0.13 (0.10; 0.18) 0.33 (0.19; 0.59) 43.8% 56.2% <0.001

PCB 28 1.14 (0.56; 2.29) 1.36 (1.06; 1.74) 2.04 (1.68; 2.47) 0.97 (0.82; 1.14) 0.64 (0.44; 0.95) 1.04 (0.92; 1.18) 43.8% 56.2% <0.001

PCB 52 1.04 (0.81; 1.33) 2.70 (2.02; 3.60) 3.30 (2.49; 4.37) 1.13 (0.84; 1.51) 0.70 (0.56; 0.87) 1.04 (0.82; 1.30) 64.1% 35.9% <0.001

6 PCB sum 5.39 (4.36; 6.66) 9.99 (7.39; 13.51) 10.99 (8.43; 14.33) 3.92 (3.09; 4.98) 2.45 (1.99; 3.00) 4.05 (3.38; 4.85) 66.5% 33.5% <0.001

PeCB 3.12 (1.80; 5.43) 1.51 (0.65; 3.52) 4.06 (3.10; 5.32) 1.89 (1.48; 2.40) 1.62 (1.31; 1.99) 1.95 (1.53; 2.50) 20.5% 79.5% 0.022 Note: Data are described by geometric mean and its 95% confidence interval. The variability sources are analyzed using one way analysis of variance with year as analyzed factor based on data transformed by natural logarithm.


42

Table 8.3: Minimum effect size for statistical comparisons

Log transformed data Primary data

Description Mean SD adjusted for the influence of

years

Detectable difference Geometric mean Lower detectable

border Upper detectable

border

alfa-HCH 0.91 0.72 0.46 2.48 1.57 3.91

beta-HCH -0.96 1.00 0.63 0.38 0.20 0.72

delta-HCH -2.20 0.41 0.26 0.11 0.09 0.14

gamma-HCH 1.47 0.52 0.33 4.35 3.14 6.04

HCB 2.69 0.76 0.48 14.66 9.06 23.73

o,p-DDD -1.91 0.66 0.42 0.15 0.10 0.23

o,p-DDE -1.85 0.61 0.39 0.16 0.11 0.23

o,p-DDT -2.11 0.57 0.36 0.12 0.08 0.17

p,p-DDD -1.21 0.75 0.48 0.30 0.19 0.48

p,p-DDE 1.70 0.46 0.29 5.46 4.08 7.30

p,p-DDT -0.65 1.03 0.65 0.52 0.27 1.01

DDD -0.96 0.73 0.46 0.38 0.24 0.61

DDE 1.72 0.45 0.29 5.58 4.19 7.43

DDT -0.43 0.76 0.48 0.65 0.40 1.06

DDTs 1.96 0.39 0.25 7.11 5.54 9.12

PCB 101 -0.61 0.82 0.52 0.54 0.32 0.91

PCB 118 -1.63 0.81 0.52 0.20 0.12 0.33

PCB 138 -0.65 0.82 0.52 0.52 0.31 0.88

PCB 153 -0.10 0.83 0.53 0.91 0.54 1.54

PCB 180 -1.08 0.97 0.62 0.34 0.18 0.63

PCB 28 0.13 0.53 0.34 1.14 0.81 1.59

PCB 52 0.40 0.63 0.40 1.50 1.00 2.24

6 indicator PCB sum 1.70 0.59 0.37 5.46 3.76 7.92

PeCB 0.76 0.83 0.53 2.14 1.26 3.62

The computation is based on one-sample t-test comparing time series with 12 measurements (12 months during one-year sampling). The test is computed using data transformed by natural logarithm, annually adjusted measure of variability and grand mean of time series. The detectable alternative is expressed both as the detectable difference of transformed data and the lower and upper borders of statistically significant differences around the geometric mean of primary data.


43

0

0.05

0.1

0.15

0.2

0.25

0 2 4 6 8 10

Freq

uenc

y

0

0.05

0.1

0.15

0.2

0.25

0 0.5 1 1.5 2 2.5

Freq

uenc

y

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 0.05 0.1 0.15 0.2 0.25

Freq

uenc

y

0

0.05

0.1

0.15

0.2

0.25

0 2 4 6 8 10 12

Freq

uenc

yalfa-HCH (ng.filter-1)

Primary data Data reconstruction

beta-HCH (ng.filter-1)

delta-HCH (ng.filter-1) gama-HCH (ng.filter-1)

Figure 8.1: Reconstruction of sample data distribution based on one year data – HCH

These charts show the comparison of the real-data distribution (in black) and data reconstruction (in grey) based on log normal model distribution and annually adjusted measure of variability.

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 10 20 30 40 50 60

Freq

uenc

y

HCB (ng.filter-1)

0

0.05

0.1

0.15

0.2

0.25

0 2 4 6 8 10

Freq

uenc

y

PeCB (ng.filter-1)


Figure 8.2: Reconstruction of sample data distribution based on one year data – HCB, PeCB


44


0

0.05

0.1

0.15

0.2

0.25

0 5 10 15

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 0.1 0.2 0.3 0.4 0.5

Freq

uenc

y

0

0.05

0.1

0.15

0.2

0.25

0 0.1 0.2 0.3 0.4 0.5

Freq

uenc

y


o,p-DDD (ng.filter-1)

o,p-DDE (ng.filter-1)

0

0.05

0.1

0.15

0.2

0.25

0 0.5 1 1.5

Freq

uenc

y

p,p-DDD (ng.filter-1)

Freq

uenc

y

p,p-DDE (ng.filter-1)

Figure 8.3a: Reconstruction of sample data distribution based on one year data – DDTs


45

0

0.05

0.1

0.15

0.2

0.25

0 5 10 15

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.5 1 1.5 2 2.5 3

Freq

uenc

y

p,p-DDT (ng.filter-1)

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.1 0.2 0.3 0.4

Freq

uenc

y

o,p-DDT (ng.filter-1)

0

0.05

0.1

0.15

0.2

0.25

0 0.5 1 1.5 2

Freq

uenc

y

DDD (ng.filter-1)

Freq

uenc

y

DDE (ng.filter-1)

0

0.05

0.1

0.15

0.2

0.25

0 0.5 1 1.5 2 2.5

Freq

uenc

y

0

0.05

0.1

0.15

0.2

0.25

0.3

0 5 10 15

Freq

uenc

y

DDT (ng.filter-1) DDTs (ng.filter-1)


Figure 8.3b: Reconstruction of sample data distribution based on one year data – DDTs



46

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0 0.2 0.4 0.6 0.8 1

Freq

uenc

y

0

0.05

0.1

0.15

0.2

0.25

0 0.5 1 1.5 2 2.5

Freq

uenc

y

0

0.05

0.1

0.15

0.2

0.25

0.3

0 1 2 3 4

Freq

uenc

y

0

0.05

0.1

0.15

0.2

0.25

0 0.5 1 1.5

Freq

uenc

y

PCB 118 (ng.filter-1)



PCB 153 (ng.filter-1) PCB 180 (ng.filter-1)

Figure 8.4a: Reconstruction of sample data distribution based on one year data – PCBs I



47

0

0.05

0.1

0.15

0.2

0.25

0.3

0 5 10 150

0.05

0.1

0.15

0.2

0.25

0 0.5 1 1.5 2 2.5

Freq

uenc

y


0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.5 1 1.5 2 2.5 3

Freq

uenc

y

0

0.05

0.1

0.15

0.2

0.25

0.3

0 1 2 3 4 5

Freq

uenc

y

PCB 28 (ng.filter-1) PCB 52 (ng.filter-1)

Freq

uenc

y

6 indicator PCB sum (ng.filter-1)


Figure 8.4b: Reconstruction of sample data distribution based on one year data – PCBs II


8.4. Results

The figures 8.1 – 8.4.b above show that model dataset from the Košetice station allows to reconstruct data of sample distribution for most of the important POPs based on annual values.

Basic statistical description of the assessed parameters was performed both for original and transformed data (see table 8.1). Time series decomposition via analysis of variance with year as explaining factor was performed (table 8.2).The analysis of variance was performed on transformed data and figures show that a significant part of the total variation is explained by the “within-year variation”. This supports the commonly used practice when a yearly based aggregation is used as a tool to eliminate a major part of seasonality effects.


48

On the basis of these results we can conclude that annually aggregated data reported to the GMP reports could be used for future processing. But we note that, the within-year components of variance, are not determinable in these aggregated data points.

Finally, minimum effect size relevant for statistical comparisons was computed. This parameter helps to identify a statistically significant difference(s) between time-related monitoring data (see table 8.3).

8.5. Variability inspection – conclusion

The main added value of the study presented in the part 8.5 is the successful quantification of effect size relevant for statistical comparison of concentrations. It proves that for environmental data it is possible to determine quantified difference(s) in the POPs concentration(s) that could be regarded as statistically significant, provided that they were detected in the same matrix and by same methodology. Such extrapolated boundaries should nevertheless be verified using more model data sets. Then they could serve for comparison of data between the former and future GMP records.

When reviewing the data collected in the GMP report, we have noted a low degree of standardization of data model and of statistical reporting. A number of variance ranges are used, including simple range values over time-aggregated data. In some cases records contain only point estimates without any variance.

This chapter concludes that even such data can be partially used for the data processing when using the robust statistical model. This is the only way how to reconstruct a sample data distribution.

Finally, annually aggregated data can be fully used for relevant statistical comparison as shown in the figures above.

The charts show that the most important variance component is associated with within-year dynamics (seasons, weather changes, etc.) in case of POPs air concentration. It has been shown that annually aggregated data loose this within-year dimension, but the overall trend can be assessed on annually-adjusted data provided that the statistically significant difference (effect size) is known/determined.

Based on test data results we can conclude that used methodology is able to:

a) re-construct sample distribution data in empty fields of the report b) quantify minimum effect size to be used in the future for comparisons of consecutive

GMP data collection campaigns


49

and thereby well responds to the issues identified in existing GMP data set in the chapters 6 and 7.

8.6. Proposal of methodology of trend analyses

Methodology of trend analyses is another important methodical issue to be solved in relation to data currently available in the GMP reports. The principal goal is time-related comparison of concentration values and assessment of potential trends related to the POPs contamination in a given matrix. Given the future needs and for the purposes of the effectiveness evaluation a special attention is be paid to the comparison of two consecutive data series.

The main issue consists of several parts. One represents exact time series analysis (i.e., variability decomposition, model selection and quantification of components) and the other is how to realize pair-wise, time-related comparison of existing concentration data.

As shown in the chapter 3 and earlier in this chapter, a significant part of the data pool in GMP reports exists in annually aggregated values and their consecutive number is limited, or point estimates of concentration data are the only source of information available.

The proposal below consists of robust statistical methodology on how to analyze time series limited in number of time points and on how to extract reliable information on hypothesized trend. Our experience shows that use of the robust, non-parametric methodology is the only way forward for existing GMP data pool due to:

• low availability of a very time series over historical data • data available are usually not complemented with a relevant typology of the sample

distribution and with variability of the estimates

The trend analysis of time series is generally computed using parametric statistical methods like linear regression, or more sophisticated models like ARIMA.

However ARIMA usually produces satisfactory results (due to its power and flexibility), it is a very complex technique requiring a great deal of experience of the researcher (Bails & Peppers, 1982). And the same is valid for other parametric statistical methods.

Therefore parametric techniques are not the best solution for routine reporting and automated processing of environmental monitoring data as environmental time series are generally characterized by asymmetric distribution, outliers occurrence and gaps in the data.

Robust non-parametric techniques are thus methods of choice. The text below suggest to use two non-parametric methods and shows their applications and effects on a test dataset from MONET network (Košetice sampling site) as provided in the part 8.7.3.


50

8.6.1. Daniel’s test

The Daniel’s test (Daniel, 1950) is an application of Spearman rank correlation coefficient computed as correlation between the values of time series and their ranks in the time series.

The value of Daniel’s test for X (time) and Y (concentrations) is given as:

nn

dr

n

ii

s −−=∑=3

1

261

1

Where di is a difference between X and Y ranks and n is a number of X-Y pairs.

The statistical significance of this trend test is usually provided in tables, for example in Zar, 1999. This test is optimal for testing against linear trends.

8.6.2. Mann-Kendal test and Mann-Kendal test with seasonal component

The Mann-Kendal test is a non-parametric test for detection of trend in time series based on binarization of time series steps. The test was proposed by Mann (1945) and Kendall (1975) and has many extensions for seasonality (Hirsch and Slack, 1982) or incorporation of covariates (Libiseller and Grimvall, 2002).

The description of Mann-Kendall test below is drawn from Libiseller (2004). The MK statistic for time series {Zk, k = 1,2,…, n} of data is defined as

)sgn( jij

i ZZT −= �<

2

where

<−=>

=0,1

0,00,1

)sgn(xif

xifxif

x

3

If no ties between the observations are present and no trend is present in the time series, the test statistic is asymptotically normally distributed with

( ) 0=TE and ( ) ( )( ) 18/521 +−= nnnTVar 4


51

If the response variable is measured of several (ω ) seasons, the seasonal Mann-Kendall test (or Hirsch-Slack test) is computed by dividing the data into ω subseries at first, where every series represents a season..

( )∑<

−=lk

jkjlj ZZsignT ω,1=j

5

This is the Mann-Kendall statistics for season j, which is summed over all seasons to obtain the seasonal statistics.

∑=

=ω

1jjTS

6

S is asymptotically normal distributed with mean value zero and variance

[ ] [ ] ( )∑∑≠==

+=ωω

jggj

gjj

j TTCovTVarSVar1,1

7

With

[ ]( )( ) ( )( )

18

5215211∑=

+−−+−=

m

iiiijjj

j

tttnnnTVar

8

where jn is the number of nonmissing observations for season j, m is the number of tied

groups and it is the size of the i-th tied group.

8.6.3. Working example

Test dataset is HCB data from MONET Košetice station for 6 years. The analysis was performed for both primary data (see figure 8.5.) and yearly aggregated values (see figure 8.6.). As seen in the figures 8.5 and 8.6. both techniques are suitable for a very heterogeneous data.


52

0

10

20

30

40

50

60

70

80

90

2003

-09

2003

-11

2004

-1

2004

-3

2004

-5

2004

-7

2004

-9

2004

-11

2005

-1

2005

-3

2005

-5

2005

-7

2005

-9

2005

-11

2006

-1

2006

-3

2006

-5

2006

-7

2006

-9

2006

-11

2007

-1

2007

-3

2007

-5

2007

-7

2007

-9

2007

-11

2008

-1

2008

-3

2008

-5

HC

B (n

g.fil

ter-1

)

Time

Season N Mann-Kendall Mann-Kendall (SD) Z p1 5 -2 4.08 -0.49 0.6242 5 -2 4.08 -0.49 0.6243 5 -6 4.08 -1.47 0.1424 5 -6 4.08 -1.47 0.1425 5 -4 4.08 -0.98 0.3276 4 -4 2.94 -1.36 0.1747 4 -4 2.94 -1.36 0.1748 4 -4 2.94 -1.36 0.1749 4 -2 2.94 -0.68 0.497

10 5 -8 4.08 -1.96 0.05011 5 -4 4.08 -0.98 0.32712 4 -2 2.94 -0.68 0.497

Comb. 55 -48 29.61 -1.62 0.105

Mann Kendall test

N Rs t p55 -0.69 -7.47 <0.001

Daniel’s test

Figure 8.5: Daniel’s and Mann-Kendal test on primary HCB data

0

5

10

15

20

25

30

35

40

45

2003

2004

2005

2006

2007

2008

HC

B (n

g.fil

ter-1

) –ye

ar a

ggre

gatio

n

Time

N Mann-Kendall Mann-Kendall (SD) Z p6 -13 5.32 -2.44 0.015

Mann Kendall test

N Rs t p6 -0.94 -5.66 0.005

Daniel’s test

Figure 8.6: Daniel’s and Mann-Kendal test on HCB data with year aggregation


53

8.7. References

Bartlett, M.S. (1947) The use of transformations. Biometrics. 3, 39-52.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (second ed.). Lawrence Erlbaum Associates, Hillsdale, New Jersey.

Holoubek, I; Klanova, J; Jarkovsky, J; Kohoutek, J (2007) Trends in background levels of persistent organic pollutants at Kosetice observatory, Czech Republic. Part I. Ambient air and wet deposition 1996-2005. Journal Of Environmental Monitoring. 9 (6), 557-563.

Limpert, E.; Stahel, W.; Abbt, M. (2001) Log-normal Distributions across the Sciences: Keys and Clues. BioScience. 51 (5), 341–352.

Zar, J.H. (2009) Biostatistical Analysis (5th Edition). Prentice Hall, Upper Saddle River, New Jersey 07458.

Bails, D. G. and Peppers, L. C. (1982) Business fluctuations: Forecasting techniques and applications. Englewood Cliffs, NJ: Prentice-Hall.

Libiseller, C. (2004) MULTMK/PARTMK – A program for the computation of multivariate and partial Mann-Kendall tests. http://www.mai.liu.se/~cllib/welcome/PMKtest.html.

Daniels, H.E. (1950) Rank correlation and population models. Journal of Royal Statistical Society Series B, 12, 171–181.

Hirsch, R.M., Slack, J.R. and Smith, R.A. (1982) Techniques of Trend Analysis for Monthly Water Quality Data, Water Resources Research 18(1), 107-121.

Kendall, M.G. (1975) Rank Correlation Methods, Charles Griffin, London.

Libiseller, C. and Grimvall, A. (2002) Performance of Partial Mann-Kendall Test for Trend Detection in the Presence of Covariates. Environmetrics 13, 71-84.

Mann, H.B. (1945) Nonparametric Tests against Trend. Econometrica 13, 245-259.

Zar, H. (1999) Biostatistical analysis, Fourth edition. Prentice-Hall. Upper Saddle River, New Jersey.

http://www.mai.liu.se/~cllib/welcome/PMKtest.html

Chapter 9. Proposal for data structure in future data collection campaigns – electronic data capture system for GMP

57

Chapter 9.

Proposal for data structure in future data collection campaigns – electronic data capture system for GMP

Standardized version of database suitable for the next GMP data collection campaign was proposed. The set of parameters was arranged in a hierarchical way corresponding to a suitable ontology of planned information system. Three principal levels are standardized and prioritized by: a) identification, b) spatial data, c) measurement – value. In addition to primary data (measurement level), this system will be supplemented with classifiers facilitating data validation and processing: coded information on spatial aggregation, (if used/applicable), coded information on time aggregation, (if used/applicable). Moreover, only annually aggregated values will be accepted and a classifier linking new GMP reports with formerly reported parameters is introduced.

9.1. Reasoning behind the proposal

Analysis in chapter 6 identified problems encountered with transfer of data reported to GMP reports into a pilot database. Lacking standardization in key parameters caused loss of some of the reported data for interregional comparison in the coding and validation process. The losses were due to data difference in time, range of chemicals, monitoring method and sometimes also gaps in information provided.

Moreover, the identification of reported measurements and values was not standardized; therefore, unclear units, LOQs or values recalculated per different base in the GMP reports can be found. Data series allowing spatial and time comparison are rather scarce.

All these above mention problems limit the data processing and reduce the information value of the reports.

Based on experience in chapters 3-5 we propose a model for future data collection campaigns. The proposal is based on hierarchical structure of data fields, containing standardized parameters with predefined content in all dimensions necessary for future data processing (values, units, measurement method, LOQs, description of data aggregation, etc.) and it thereby limits the loss of any reported data to a minimum. Standardized way of data collection should also facilitate the retrospective control of already reported GMP records.

The proposal is prepared in a form of electronic template described in detail in this chapter and in chapter 10. We believe that new collection system could be directly implemented as comprehensive electronic data capture system, supplied with centralized data management.


58

We are aware that any such proposal must be first reviewed by the relevant Stockholm Convention Expert group(s) and also approved by the Convention`s governing body prior to its implementation.

9.2. GMP data templates proposal

The new structure of collected GMP data is summarized in figures 9.1 and 9.2. The structure of the data fields covers important parameters that must be reported in a fully standardized way: geographical identification and time of reported data, measurement – value – units chain and definition of data aggregation (if applied).

Figure 9.1 displays the key data fields in a hierarchical scheme. The scheme also represents a logical sequence of records in a database. The following list highlights the most important data fields and information, which are required as obligatory in the proposed system (see Figure 9.2 for details):

• contact identification of the data administrator responsible for data insertion; • identification of site reported and identification of any type of spatial aggregation (if

used); • predefined set of reported chemicals (POPs); • definition of method used, including corresponding LOQ; • identification of units used for reported concentration values; • description of time aggregation (if used); • definition of variability (an obligatory field for aggregated data).

Report/region

Matrix

Site spatial aggregation specific Site

Country

Compound

Method

Value primary value in time aggregated in time

Unit

Code list – user selects appropriate item from predefined code list.

…E-mail…

…Name…Name

Institution

…Phone…

…Institution…E-mail

Phone

Contact information:

Initiation of database records must be supplemented with contact identification of the

responsible data administrator (obligatory

item)

In addition to primary data (measurement level), the system will be supplemented with classifiers,which will facilitate data validation and processing:• Coded information on spatial

aggregation, if used • Coded information on time

aggregation, if used (only annually aggregated values will be accepted)

• Classifier linking new GMP reports with formerly reported items

Figure 9.1: GMP reporting – data collection structure proposal


59

• Regional group / Report• Matrix

• Country

• Site

• Compound• Method

• Value• Unit

LEVEL A. Identification: a structure defining report (region) affiliation and matrix – country identification. Obligatory additional component: contact identification of the responsible data administrator.

LEVEL B. Spatial data: standardized identification of sampling site(s) – standard format of coordinates, records on spatial aggregation (if applied) and link to former GMP reports (if given site is already reported in GMP)

LEVEL C. Measurement - value data: obligatory selection of coupled data fields: compound-method and value-unit,identification of time aggregation (if applied) and variability measures

Figure 9.2-1: GMP electronic data capture system – data sheets proposal

• Regional group / Report

• Country• Matrix

• Site• Compound

• Method• Value

• Unit

Asia and PacificAfricaCEECGRULACWEOG

» Next



60

AlbaniaArmeniaAzerbaijanBelarusBosnia and HerzegovinaBulgariaCroatiaCzech Republic

• Regional group / Report• Country

• Matrix• Site


• Value• Unit

EstoniaGeorgiaHungaryLatviaLithuaniaForm. Yugoslav Rep. of MacedoniaMontenegroMoldova

PolandRomaniaRussiaSerbiaSlovakiaSloveniaUkraine

» Next



• Matrix

• Site• Compound

• Method• Value

• Unit

Air activeAir passiveBreast milkHuman blood

» Next



61


• Matrix• Site – applies for air sampling


• Value• Unit

Spatial aggregation

Specific Site

…Site 1…

…Site 2…

…Site 3…

…Site identification…

…Site 4…

…Site 5…

Monitoring network/project identification…Project title…

Site for this matrix already reported in GMP reports?NoYes …Compound 1… …Year1…

…Compound 2…

…Compound 3…

…Compound 4…

…Year2…

…Year3…

…Year4… » Next



• Matrix• Site – applies for human tissues sampling


• Value• Unit

Pooled sample

Individual samples

…Site 1…

…Site 2…

…Site 3…

…Site 4…

…Site 5…

Monitoring network/project identification…Project title…

Site for this matrix already reported in GMP reports?NoYes …Compound 1… …Year1…

…Compound 2…

…Compound 3…

…Compound 4…

…Year2…

…Year3…

…Year4… » Next



62


• Matrix• Site

• Compound

• Method• Value

• Unit

Select compound to be reported: …Select…ChlordaneDDTsDieldrinEndrinHeptachlorHexachlorobenzenePCBsSum of 6 PCBs

» Next



• Matrix• Site


• Value• Unit

Unit …Select…ng/gpg/g

Method …Select…Capillary GC

Quadrupole MSIon Trap MS

LOQ …LOQ…

» Next



63


• Matrix• Site


• ValuePrimary values(s)

Anually aggregated value(s)

…Time… …Value…Time Value Variablity measures

Not relevant …Min… …Max… …SD… …SE…Available

…Time… …Value… Not relevant …Min… …Max… …SD… …SE…Available

…Year… …Value…Year Value Variablity measures

…Min… …Max… …SD… …SE…

…Year… …Value… …Min… …Max… …SD… …SE…

…Year… …Value… …Min… …Max… …SD… …SE…


The primary data (measurement level) records are supplemented with classifiers facilitating data validation and processing:

• Coded information on spatial aggregation (if used); • Coded information on time aggregation (if used – only annually aggregated values will

be accepted); • Classifier linking new GMP reports with formerly reported items;

The participating subjects will be allowed to report primary data directly via user friendly sheets enabling export from their data repositories. Spatial aggregation must be identified in a parametric way (site level). Time aggregation must be based on annual mean and/or median concentration of a particular chemical and supplied with an appropriate variability measure.

9.3. Principles of data validation included in the proposed data collection system

The new data templates are proposed to allow a direct and immediate data validation during the reporting process (see fig 9.2.3-9). Local data administrators will be informed about obligatory data fields through automated electronic queries. Each (completed/filled) record must therefore contain all required information such as proper unit as an obligatory attribute of concentration value.


64

Additional validation mechanism is generated from logical links among data fields. These automatically working tools control, for example, proper coding of time points in a consecutively reported data time series, concentration values etc. Different data validation procedures will be used for different matrices (air, breast milk, human blood).

Furthermore, the system is designed to allow at least partial retrospective validation of already collected GMP data. Administrators will be asked to identify whether the currently reported site/matrix/chemical had already been reported in former GMP data collection campaign. If so, a detailed identification of already reported data would be requested (see Figure 9.2-5, 9.2-6). This additional retrospective validation provides that newly reported items could be directly linked to the former reports and records and both old and new data can be processed in a correct, pair-wise mode. If needed, local data administrators may be contacted by global GMP database administrator to verify some of the formerly reported concentration data.

9.4. Building the new data collection system and its benefits

The sequence described below leads to a fully electronic on-line data collection working in a centralized data management controlling reported data in real time:

1. Validation and expert review of the proposed data structure and data fields 2. Definition of obligatory code lists 3. Recognition of obligatory (e.g. units) and voluntary (e.g. name of monitoring projects)

data records 4. Development of pilot, preliminary version of the database and data collection web

forms 5. On-line testing and review of the system; the test version of the database will be

accessible via project website, all potential users being able to test it and suggest modifications

6. Development of the database final version, including control and validation links among data fields

7. Implementation of final database in an authorized system for electronic data collection

Steps 5 and 6 would verify whether the system development was successful and the new data collection structure is able to control and link reported data in real time.

We see the principal benefits of the proposal as follows:

• Automatic linking of localities with formerly reported data will be performed prior to archiving of the records.

• Data fields will be filled and managed by local administrators and data managers via web forms generated by the central database. Therefore, local administrators can fully manage all procedures related to the data reporting and validation.

• Local storage of completed data reports will be possible.


65

• The system can automatically generate electronic reports, namely tables and predefined graphical summaries.

• And finally and most importantly - the regional data managers can also describe and summarize the reported data in real time, without any delay due to centralized data processing (in contrast to GMP1 phase where the Regional Organization Groups had to create all tables from nationally collected data and only then could write a report).

9.5. Conclusions – main added value of the proposed system

The proposed system eliminates problems and limits encountered during the review of the first Global Monitoring Plan data collection (2009). Its benefits/outputs are as follows:

1. Correct identification of primary data sources. Correct recognition of both primary and aggregated data, scoring of their weight (given by the way of aggregation and completeness).

2. Search is controlled interactively by spatial and time coordinates; way of data aggregation is clearly identified. Spatially aggregated and time aggregated data will be supplied/complemented with variability measures.

3. Standardized validation of reported data, controlled by defined code lists. 4. Correct statistical summary provided for both primary and aggregated data sources.

Reported mean/median concentration values will be supplied with proper variability measures.

5. Standardized nomenclature of key data fields, in particular chemicals and concentration units.

6. Matching of relevant reports coming from the same place as system enabling pair-wise trend analyses.

Chapter 10. Pilot on-line reporting over GMP regional reports

66

Chapter 10.

Pilot on-line reporting over GMP regional reports

This chapter describes an on-line reporting system proposal for future use for the purposes of the GMP reporting. The pilot on-line reporting tool is transparent and provides a simple overview of data collected within environmental monitoring programmes using specifically designed analytical tools for the visualization of POPs data. Quick information on presence/absence of selected compounds monitoring in a particular country and year is available, as well as aggregated concentrations of the compounds described by their mean, median, minimum, and maximum. Several analytical tools have been created for the visualization of in manner and scale. The user can choose among four analytical tools how to visualize and filter the data including a world map with colour-coded countries performing active monitoring programmes, visualization of sampling frequencies of selected compounds, and measured values after one-year aggregation.

10.1. Introduction

The conclusions of chapter 3-5 show that data collected and presented in the GMP reports provide sufficiently large pool of information that merits on-line presentation. Such presentation requires visualization tools that would enable quick and easy searching, viewing, and analysis.

The visualization tools have been developed on the database of aggregated data that was described in the chapters 6 and 7. The proposed on-line pilot reporting tool is available on-line www.genasis.cz/unep. The website provides easily accessible information on performance of monitoring programmes in individual countries and/or regions sorted according to matrices, time, and compounds. Annually aggregated concentration values are available. All data on individual compounds show whether the particular compound was part of 12 original POPs or 10 POPs that were subsequently added to the Stockholm Convention.

There are four analytical tools that present data from all UNEP-GMP reports in a different manner and scale:

• World map – monitoring overview; • Sampling frequency – compounds; • Sampling frequency – years; • Measured values (concentration levels).

http://www.genasis.cz/unep


67

Data processing and development of the visualization tools has been based on our previous experience with the GENASIS portal development and its analytical tools architecture. A similar technological platform and processing procedures were used. A detailed manual for the on-line reporting system is available in the Annex 10.1.

10.2. Aims and user group of the on-line pilot reporting

All data records accessible in the GMP reports were re-coded, standardized and digitized (see also chapter 4) and subsequently stored in a pilot version of the database system. After these procedures, a final set of data consisting of point estimates (occasional reporting), annually aggregated short-time series, and outcomes of well established monitoring programmes that are reliably reported, was obtained. Due to diversity of the input the information and their validity cannot be fully unified at present. That is one of reasons to propose an on-line descriptive browser in order to facilitate search over the data storage in the future.

Password protected access to the pilot on-line reporting tool is provided to a selected group of users. At the moment, main user group of the pilot on-line reporting tool could be data managers who could make use of the tool in a final validation and standardization of already collected data.

Subsequent desirable step would be that validated and approved data would be reported into a similar reporting system enriched by codes distinguishing records according to their quality, completeness, and validity. Such upgraded on-line reporting system would be also password protected.

www.genasis.cz/unep

Figure 10.1: Title page of the on-line pilot reporting tool (www.genasis.cz/unep).


68

10.3. On-line data visualization 1: Interactive world map – monitoring overview

An interactive world map displaying monitoring activities in the world in individual years. Differentiation according to environmental matrices and years is available. The colour-coding represents countries that performed monitoring and reported their data to the GMP reports.

10.3.1. Required data

Records that were successfully re-coded from the original GMP reports were used as data source. Aggregations were calculated on one-year basis for each country. The results were used to create separate map layers for each matrix. Data in map layers are time related allowing the use of time filters. World map in the background facilitates user`s orientation in and understanding of the output.

10.3.2. Input data

User interaction in the monitoring overview tool is to choose from predefined values of available filters and to control the map window (move, zoom in or zoom out) by mouse. To select the matrix, no direct input is needed – all available values are predefined and user chooses from the pop-up menu.

Figure 10.2: On-line data visualization 1 – world map indicating countries with monitoring for a particular compound and year.


69

10.3.3. Output

The output is an interactive map window with overlaying feature layers visualizing the progress of monitoring activities in each country (see figure 10.2). Users are thus informed on both monitoring points and sampling frequency.

10.3.4. Standardized attributes

A standardized code list of matrices was used to keep consistency with other reports (outputs).

10.3.5. Filters

The monitoring overview (figure 10.2) consists of the map window with two main controls to switch between available matrices (top) and to control current time range (years) at the bottom of the window. Time slider controls change time range or just show the time-related existence of monitoring activities in a slideshow.

10.3.6. Used software tools

This output in figure 10.2 is completely built on ArcGIS technologies provided by the ESRI software company. Particular map layers were prepared in ArcMap Desktop application and published as a web map service via ArcGIS server. Adobe Flex and ArcGIS API for Flex were used to build user interface with map window. Map layers with analytical outcomes are directly streamed from GENASIS ArcGIS server and the background-map is loaded from arcgisonline.com servers.

10.4. On-line data visualization 2: Sampling frequency – compounds

The chart in figure 10.3 shows sampling frequency of individual compounds in reporting countries. Filters for the selection of region, matrix, and year are available. Compounds are listed on x-axis, countries on y-axis. Compounds included in the GMP Guidance as recommended analytes and/or the Stockholm Convention are marked and highlighted. Summary information for each “active” (= data available) compound/country combination is accessible after clicking on a relevant mark and is supplemented with summary statistics.

10.4.1. Input data

It is required to select a matrix and year. Input values can be changed later in an interactive way, according to the user's needs.


70

10.4.2. Output

As an output, the user obtains a table chart with grid. The top row represents a list of currently selected compounds. The left column shows the list of selected countries. Marked grid cells indicate that there are some reported values for the corresponding country and within the selected time period in the on-line pilot reporting tool. Details can be obtained via mouse hover or mouse click.

Use dropdown lists to select from available values to filter the output. Filters marked

with the grey cross can be deactivated

Hover the mouse cursor over a cell to get additional information about

aggregated value

List of countries

List of compounds

Figure 10.3: On-line data visualization 2 – sampling frequency for individual countries and compounds


Aggregated data from all GMP reports were used as a source for all chart reports showing sampling frequency. In this case, several additional parameters such as compound groups, regional groups plus indicators showing whether the given compound is – or is not subject to the GMP Guidance or to the Stockholm Convention (red and blue mark sign respectively) are used.


Standardized code lists are used for each filter control. Data entering the reporting are extracted from the standardized database records.


71

10.4.5. Filters

There are several filters available:

• Matrix – choosing a matrix is required to receive the output report • Time – choosing a year is also mandatory to receive the output report • Regional group filter helps to aim at a particular group of countries • Country filter shows only sites and regions from a given country • Two checkbox filters to show only compounds subject to the Guidance of GMP /

compounds included in the Stockholm Convention


Selection of aggregated values from database is based on filter (values of conditions). The primary records in original GMP reports were computationally standardized in order to fit the generally usable and applicable aggregation level (i.e., annually aggregated point estimates).

10.5. On-line data visualization 3: Sampling frequency – years

See figure 10.4 for a chart showing the data on sampling frequency. The user selects region, matrix, and compound (or group of compounds) to be displayed. Years are listed on x-axis, countries on y-axis.

10.5.1. Input data

It is required to select a matrix and compound, for which the report is created. These values can be changed later in an interactive way, according to the user's needs.

10.5.2. Output

The user obtains a table chart with grid. The top row is a list of currently selected years. The left column shows the list of selected countries. Marked cells indicate that there are some reported values for the relevant country and time period in the on-line pilot reporting tool. Details can be obtained via mouse hover or mouse click.


Aggregated data from all GMP reports were used as a source for all these chart reports. In this case, several additional parameters are used which are joined to core aggregations.


72


Standardized code lists are used for each filter control. Data entering the reporting are extracted from the standardized database records.

10.5.5. Filters

There are several filters available:

• Matrix – choosing a matrix is required to receive the output report • Time range – choosing time range to be displayed is also mandatory • Regional group filter helps to aim at a particular group of countries Country filter

shows only sites and regions from a given country


Selection of aggregated values from database is based on filter (values of conditions). The primary records in GMP reports were computationally standardized to fit the generally usable and applicable aggregation level, i.e. annually aggregated point estimates.

Figure 10.4: On-line data visualization 3 – sampling frequency for individual countries and years


73

10.6. On-line data visualization 4: Measured values – summary statistics

This chart in figure 10.5 is designed for a direct comparison of measured values and concentrations, using a standard set of robust summary statistics. The user selects region, matrix, year and compound. Concentrations levels of that compound in different locations (countries, specific localities) are displayed in form of box-and-whisker plots with mean/median value, minimum and maximum.

10.6.1. Input

The user selects desired combination of matrix, compound a year. No other direct inputs are needed.

10.6.2. Output

The output is box-and-whisker like diagram, allowing also a various groupings of geographical entities (countries, regions, monitoring programmes etc.).

Figure 10.5: On-line data visualization 4 – concentrations of selected compounds presented in form of box-and-whisker plots


74


A standard set of filters is available to the user to optimize the processed part of the GMP database. The processed data are already prepared and coded in the database, including re-coded summary statistics from the GMP records.


Standardized and shared attributes across other reports and outcomes code lists of compounds and matrices are used.

10.6.5. Filters

There are three basic filters (matrix, compound a year) which user must always select.


All aggregated values are stored in database and are shared among other reports and outcomes. Robust summary statistics are prepared in all existing views (if the ranges or variance data were reported to GMP reports).


75

Conclusions

This report comprises analysis over the content of data on the POPs concentrations in the core environmental matrices: breast milk, human blood and ambient air collected in the Global Monitoring Plan (GMP) regional reports in 2008. All reported variables were described qualitatively in detail and, if necessary, transformed or completed using the text part of the GMP reports (i.e. nomenclature, time scale, sampling frequency).

Reported POPs were divided into four groups according to their relation to the Stockholm Convention and the Global Monitoring Plan Guidelines. The first group are the initial 12 POPs comprised in the Stockholm Convention in 2001, their congeners, isomers and degradation products defined in the GMP Guidance. The second group comprises parameters related to additional 10 POPs subsequently added to the Convention up to 2011. Third group contains all other parameters related to the Stockholm Convention, but none pertaining to any of the GMP Guidance documents. Finally, fourth group are chemicals found in the regional reports but with no relation to the Convention.

All five analyzed reports contain 171 variables (including concentration data on congeners, isomers, transformation products, various summations and toxic equivalents – TEQs). Their distribution among groups described above was 58, 7, 84 and 22 respectively. However, only twenty parameters in total were reported in all matrices (see table 3.2.), eleven of those parameters are the individual compounds included in the GMP Guidance as of 2007 and three parameters are among the additional 10 POPs. In addition, 42 parameters were reported in all reports, 38 out of them are recommended in the GMP Guidelines.

We believe that only the first two groups of parameters should be reported in the next GMP data collection campaigns as parameters from the third group can be mostly calculated in the (global) database and parameters currently reported under the fourth group have no relation to the effectiveness evaluation of the Stockholm Convention. In case that reporting of calculated /aggregated data is preferred, there must be a clear guidance provided on which sums or TEQs are acceptable.

Subsequently, statistically valuable data were identified for all matrices. Only a few monitoring stations worldwide are capable of providing statistically relevant regular time series with primary concentration measures; most of the other data reported are aggregated. Frequent reporting of the temporally-aggregated values without clear instruction on such aggregation makes potential comparison of collected GMP reports data with future campaigns a complex task. All challenges regarding the completeness and heterogeneity of primary records were recorded and taken into account in the proposal for standardized electronic template as most of the problems described in the chapter 6.1 were caused by the non-existent standardized form for data collection, which inevitably brought a level of uncertainty into further processing and results.


76

Our endeavour was to create a practical solution for the next collection campaign, namely in data collection, handling, storage and presentation so that maximum cost-effectiveness and sustainability is achieved while using existing infrastructure, capitalising and maximising the expert knowledge in UN regions. Moreover, the whole data collection system should improve quality of collected global data set on POPs occurrence in the environment and strengthen the ownership of the data management and reporting by local data administrators.

We have carefully designed a relational database consisting of 10 entities as described in detail in the part 6.3. We draw from our experience with the development of the GENASIS database as described in the chapter 7 and 9. The set of linked parameters was arranged hierarchically. Three principal levels are standardized and prioritized by: a) identification, b) spatial data, c) measurement – value. Annually aggregated values from the GMP reports were transferred into the database and successfully tested in the pilot version of the on-line data visualization tool (www.genasis.cz/unep).

Its graphical and classificatory functions allow authorised user to sort the GMP data, to study their structure and completeness. The pilot on-line reporting tool is transparent and provides a simple overview of data collected within environmental monitoring programmes using specifically designed analytical tools for the visualization of POPs data. Quick information on presence/absence of selected compounds monitoring in a particular country and year is available, as well as aggregated concentrations of the compounds described by their mean, median, minimum, and maximum. There are four analytical tools that present data from all UNEP-GMP reports in a different manner and scale:

• World map – monitoring overview; • Sampling frequency – compounds; • Sampling frequency – years; • Measured values (concentration levels).

We recognize nevertheless that further validation procedures are strongly needed to be able fully use information comprised in GMP reports and link it effectively with the future collection campaigns.

Chapter 3 – Annex 3.1 Chemicals in the Global Monitoring

Plan regional reports - overview

Compounds in GMP and Stockholm Convention

• Chemicals included in the Stockholm Convention (SC) are listed in its Annexes

A,B and C1.

• Annexes contain only general names of chemicals (e.g DDT) without exact specification of isomers, congeners or transformation products that should be monitored. The Guidance document on Global Monitoring Plan solves this discrepancy in its chapter 2, where it specifies a list of additional chemicals (recommended analytes) related to chemicals in the Stockholm Convention as of 2007.

• Available GMP regional reports (2009, 1st data collection campaign) contain chemicals reported either as congeners/isomers/transformation products or as a summation of individual compounds.

• Reported summations without specification of their content of congeners/isomers or additional information increases heterogeneity and uncertainty in the data.

1 Stockholm Convention on Persistent Organic Pollutants (POPs), as amended in 2009 (http://chm.pops.int/Convention/ConventionText/tabid/2232/Default.aspx)

Background

Compounds in GMP and Stockholm Convention

Compounds to be identified in GMP reports of 20092 Chemical Parent POPs Transformation products

Aldrin Aldrin

Chlordane Cis- and trans-chlordane Cis- and trans-nonachlor, oxychlordane

DDT 4,4’-DDT, 2,4'-DDT 4,4'-DDE, 2,4'-DDE, 4,4'-DDD, 2,4'-DDD

Dieldrin Dieldrin

Endrin Endrin

HCB HCB

Heptachlor Heptachlor Heptachlorepoxide

Mirex Mirex

Polychlorinated biphenyls (PCBs) Sum of 7 PCBs* (28, 52, 101, 118, 138, 153 and 180), PCB WHO-TEQ (from 12 congeners)

Polychlorinated dibenzo-p-dioxines (PCDDs) and polychlorinated dibenzofurans (PCDFs)

2,3,7,8-substituted PCDDs/Fs (17 congeners)

Toxaphene Congeners Parlar 26, Parlar 50 and Parlar 62

* It is recommended to analyze and report seven PCB congeners separately. 2 Guidance on the Global Monitoring Plan for Persistent Organic Pollutants, February 2007, amended in May 2007

(http://chm.pops.int/Implementation/GlobalMonitoringPlan/Overview/tabid/83/Default.aspx)

Overview of chemicals identified in GMP reports

Aldrin

Alpha hexachlorocyclohexane (a-HCH)*

Beta hexachlorocyclohexane (b-HCH)*

Lindane (g-HCH)*

Chlordane

Chlordecone

Dieldrin

Endrin

Mirex

Toxaphene

Heptachlor

Hexabrombiphenyl

Hexabromdiphenyl ether

Heptabromodiphenyl ether

Hexachlorbenzene (HCB)

Pentachlorbenzene (PeCB)

Polychlorinated biphenyls (PCB)

Tetrabromodiphenyl ether (TBDE)

Pentabromodiphenyl ether (PBDE)

DDT

Perfluorooctane sulfonic acid (PFOA), it salts and perfluoroctane sulfonyl fluoride

Polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDF)

Endosulfan

Chemicals in the Stockholm Convention

Chemicals identified in the GMP reports (isomers, sums, transformation products)**

Cis- (alpha-) chlordane, Trans- (gamma-) chlordane, Oxychlordane, Cis- and Trans-nonachlor, 8 various summations of chlordane related compounds (CC, TC, OC, CN, TN)

Dieldrin; Endrin, Endrin ketone and Endrin; Mirex

Various PCBs: indicator PCBs (PCB 28, PCB 52, PCB 101, PCB 118, PCB 138, PCB 153, PCB 180) and 27 others. Various PCBs WHO-TEQ and PCBs I-TEQ, Sum of 3, 10, 14, 21, 25, 36, 48 PCBs, dl-PCBs (also with WHO-TEQ). Mono-ortho and Non-ortho PCBs in TEQs

Parlar (Toxaphene) 26 isomers, Parlar (Toxaphene) 50, Parlar (Toxaphene) 62, Toxaphene (as a group), Parlar 40, Parlar 41, Parlar 44

DDT, DDE, DDD, o,p’-DDT, o,p’-DDE, o,p’-DDD, p,p’-DDT, p,p’-DDE, p,p’-DDD, Sum of DDTs, Sum p,p-DDTs (p,p-isomers together), Sums of various number of DDT isomers, DDT + p,p-DDE

Reported only as a Sum of 3 PBDEs

PCDDs and PCDFs (13 congeners) in sum or separately - PCDDs, PCDFs, PCDDs/Fs – often given as various TEQs: WHO-TEQ, Nordic-TEQ and I-TEQ, OCDD, OCDF

Endosulfan I, Endosulfan II, Endosulfan SO4, Endosulfans (sum of)

* In some reports, HCHs are given only as sum of HCH isomers

** Highlighted chemicals are recommended by Guidance on GMP 2007 Chemicals written in italics were not identified in any of the analyzed GMP regional reports.

Aldrin

Hexachlorbenzene (HCB), Pentachlorbenzene (PeCB)

a-HCH, b-HCH, g-HCH, d-HCH, HCHs (sum of)

Heptachlor, Heptachlorepoxide, Heptachlorepoxide cis-, Heptachlorepoxide trans-, Heptachlor group


For purposes of the pilot data analysis, all 171 reported variables were divided into the following groups:

1) Original 12 SC POPs – 12 original Stockholm Convention POPs and their congeners/isomers/degradation products specified in Guidance on the GMP for POPs (2007) as recommended analytes - group 1

2) Additional 10 SC POPs – chemicals not obligatory for reporting in 2009, but desirable to report due to their listing in the Stockholm Convention in 2009 and 2011, respectively. Individual congeners/isomers/degradation products are specified in a revised Guidance on the GMP for POPs (2009) - group 2

3) SC related parameters – parameters not specified in the Guidance on GMP but related to the Convention (e.g. summations, TEQs, etc.) - group 3

4) Other chemicals – all other reported parameters not specified in the Guidance or not related to the Convention at all (e.g. PAHs) - group 4

Classification of chemicals


• Group 1 contains 58 parameters related to the original 12 POPs.

• Group 2 are 7 parameters related to the additional 10 POPs (nine new POPs + endosulfan)

• Group 3 gathers 84 parameters (28 various TEQs and 56 individuals and summations).

• Group 4 are the remaining 22 parameters reported (mainly PAHs, but also fungicides, herbicides, insecticide and others) having no relation to the Stockholm Convention.

Data variability • 67 of all parameters were reported in one matrix only, 20 parameters were

reported in all matrices, 42 parameters in all UN regions.

• All of 12 original POPs are identified in GMP reports. All exactly specified chemicals given in the Guidance on GMP 2007 are also fully covered by at least one combination of matrix and region.

Chemicals reported in GMP - overview


• As one of the main tools for effectiveness evaluation is trend analysis, the

chemicals in each group have been further divided according to their data availability:

1. Chemicals with a relatively long-term follow-up, sufficient amount of data for baseline trend evaluation at a minimum: regularly reported chemicals with time series available. The threshold for inclusion of a chemical into this group is at least four years of consecutive monitoring. When a longer monitoring occurred, a decent amount of missing values in the time line are acceptable.

2. Chemicals without at least four-year long monitoring data or with a lot of missing values in consecutive years. such data allow for a basic assessment of concentration levels. Because of the scarce reporting they allow for point estimates only.

Time availability of reported chemicals


Chemicals and their groups identified in GMP reports

1,2,3,4,6,7,8-HpCDD Endrin PCB 118 1,2,3,4,6,7,8-HpCDF Endrin ketone PCB 138 1,2,3,4,7,8,9-HpCDF HCB PCB 153 1,2,3,4,7,8-HxCDD Heptachlor PCB 180 1,2,3,4,7,8-HxCDF Heptachlorepoxide cis- PCB 77 1,2,3,6,7,8-HxCDD Heptachlorepoxide trans- PCB 81 1,2,3,6,7,8-HxCDF Mirex PCB 105 1,2,3,7,8,9-HxCDD o,p-DDD PCB 114 1,2,3,7,8,9-HxCDF o,p-DDE PCB 123 1,2,3,7,8-PeCDD o,p-DDT PCB 126 1,2,3,7,8-PeCDF OCDD PCB 156

2,3,4,6,7,8-HxCDF OCDF PCB 157 2,3,4,7,8-PeCDF Oxychlordane PCB 167

2,3,7,8-TCDD p,p-DDD PCB 169 2,3,7,8-TCDF p,p-DDE PCB 189

Aldrin p,p-DDT Trans-chlordane Cis-chlordane PCB 28 Trans-nonachlor Cis-nonachlor PCB 52 Parlar 26, 50 and 62

Dieldrin PCB 101

Additional 10 POPs

Original 12 POPs N = 58

N = 7

Alpha-HCH Endosulfan I PeCB Beta-HCH Endosulfan II

Gamma-HCH Endosulfan SO4

All reported variables are enlisted (N = 171)

Other chemicals N = 22

Acenaften benzo(k)fluoranten naftalen

Acenaftylen dibenzo(ah)antracen pyren

Antracen fenantren Sum 16 PAHs, Sum 4 HCHs

Benzo(a)antracen fluoranten Dacthal and trifluralin

Benzo(a)pyren fluoren Delta-HCH

Benzo(b)fluoranten chrysen Chlorthalonil Benzo(ghi)perylen indeno(123cd)pyren

SC related parameters N = 84

4 DDTs DDD, DDE, DDT DDT group 5 DDTs DDT + p,p-DDE DDTs

dl-PCB I-TEQ Mono-ortho PCBs WHO 1995-TEQ PCDDs/Fs Nordic-TEQ

dl-PCB TEQ Mono-ortho PCBs WHO 1997-TEQ PCDDs/Fs TEQ

dl-PCB WHO-TEQ 1997 Mono-ortho PCBs WHO 1998-TEQ PCDDs/Fs WHO 1995-TEQ

dl-PCB WHO-TEQ 1998 Mono-ortho PCBs WHO-TEQ PCDDs/Fs WHO 1997-TEQ

dl-PCB WHO-TEQ 2001 Non-ortho PCBs WHO 1995-TEQ PCDDs/Fs WHO 1998-TEQ

dl-PCB WHO-TEQ Non-ortho PCBs WHO 1997-TEQ PCDDs/Fs WHO 2001-TEQ

dl-PCBs (12) Oxychordane + Nonachlor PCDDs/Fs WHO-TEQ

Endosulfans p,p-DDTs PCDFs (10)

Endrin group Parlar 40, 41 and 44 PCDFs WHO 1995-TEQ

Heptachlor group PCB 28 + 31, 44, 49, 70 PCDFs WHO 1997-TEQ Heptachlorepoxide PCB 74, 99, 110, 170 Sum 3, 10, 14, 21 PCBs

Chlordane PCB 183, 187, 194, 196 Sum 25, 36, 48 PCBs Chlordane + Nonachlor PCB 202, 206, 208, 209 Sum 3 PBDEs

Chlordane + Oxychlordane PCBs Total dioxins TEQ Chlordane + Oxychl. + t-nonach PCDDs (7) Total dioxins WHO 1995-TEQ Chlordane + trans-nonachlor PCDDs WHO 1995-TEQ Total dioxins WHO 1997-TEQ

Chlordane group PCDDs WHO 1997-TEQ Total dioxins WHO-TEQ Indicator 6 PCBs PCDDs/Fs (17) Toxaphene Indicator 7 PCBs PCDDs/Fs I-TEQ Trans-chlordane + cis-nonachlor

Ntot: number of all reported parameters, NSC: number of parameters specified in the Guidance on GMP as recommended chemicals

* Parameters not related to the Stockholm Convention are not evaluated.

All parameters Ntot = 74*, NSC = 38

Additional 10 POPs Original 12 POPs

Ambient air

Chemicals in GMP reports: ambient air

Aldrin, Dieldrin, Endrin

Cis-chlordane, Cis-nonachlor

Heptachlor, HCB

Heptachlorepoxide cis- and trans-

Chlordane + nonachlor

Mirex

o,p-DDD, o,p-DDE, o,p-DDT

p,p-DDD, p,p-DDE, p,p-DDT

5 DDTs, DDT group

Oxy-chlordane

Parlar 26, 50, 62

PCB 28, 52, 101,118, 138, 153, 180

PCBs, dl-PCBs I-TEQ

Indicator 7 PCBs, Sum 10 PCBs

PCDDs/Fs (17)

PCDDs/Fs I-TEQ

PCDDs/Fs TEQ

trans-Chlordane (= gamma)

trans-Nonachlor

2,3,7,8-TCDD and 2,3,7,8-TCDF

DDE and DDT + p,p-DDE

dl-PCB I-TEQ, WHO-TEQ

dl-PCB WHO-TEQ 1997, dl-PCBs (12)

Chlordane, Chlordane + Trans-nonachlor

Indicator 6 PCBs, PCBs

PCDDs (7), PCDFs (10)

PCDDs WHO-TEQ 1997

PCDDs/Fs WHO-TEQ, WHO-TEQ 1997

PCDFs WHO-TEQ 1997

Sum 21, 25 and 48 PCBs

Total dioxins WHO-TEQ 1997

Toxaphene

Alpha-HCH

Beta-HCH

Gamma-HCH

PeCB

Time series Point estimates Point estimates

Ntot = 8*, NSC = 7 Ntot = 66, NSC = 31

Ntot = 39*, NSC = 27 Ntot = 27, NSC = 4

Time series

Ntot = 4, NSC = 4 Ntot = 4, NSC = 3

Endosulfan I

Endosulfan II

Endosulfan SO4

Sum 3 PBDEs

Chemicals in GMP reports: ambient air

• In total, 43 chemicals regularly reported and related to the Stockholm Convention are suitable for time series analysis.

• Long-term time series availability (current chemicals in Stockholm Convention only):

Ambient air

Japan (Hateruma Island) 2004 - 2007 22 parameters, all of them cover original 12 POPs

Hong Kong 1998 - 2007 PCBs, PCDDs/Fs (17), PCDDs/Fs I-TEQ

Canada (Kinngait) 1994 - 2002 Alpha-HCH, Beta-HCH, DDT group, Dieldrin, HCB, Chlordane + Nonachlor, Sum 10 PCBs

Canada (Alert) 1998 - 2005 HCB, Sum 10 PCBs, DDT group, Chlordane+nonachlor

Finland (Pallas) 1998 - 2005 Indicator 7 PCBs, p,p-DDTs

Iceland (Storhofdi) 1995 - 2005 5 DDTs , HCB, Chlordane+Nonachlor, Sum 10 PCBs

Norway (Svalbard) 1998 - 2007 HCB, DDT group, Chlordane+nonachlor, Sum 10 PCBs

Norway (Lista) 1998 – 2003 HCB

Spain (Catalonia) 1994 - 2007 PCDDs/Fs TEQ

Sweden (Aspverten) 1998 - 2005 Indicator 7 PCBs

Sweden (Rao) 2002 - 2005 Indicator 7 PCBs, p,p-DDTs

Sweden (Rorvik) 1998 - 2001 Indicator 7 PCBs

Czech Republic (Kosetice) 1996 -2008 15 parameters, 11 of them cover original 12 POPs, 4 of them cover additional POPs

USA (remote and rural) 1999 -2002 dl-PCB TEQ, PCDDs/Fs TEQ, PCDDs/Fs I-TEQ



All parameters1


Breast milk

Chemicals in GMP reports: breast milk

4 DDTs

DDE and DDT

DDTs

Dieldrin

HCB

Indicator 6 PCBs

Oxychlordane

Oxychlordane + nonachlor

PCB 118

PCB 153

PCBs

PCDDs/Fs I-TEQ

PCDDs/Fs TEQ

PCDDs/Fs WHO-TEQ

Sum 3 and 10 PCBs

17 PCDDs/Fs isomers PCB 101, 105, 110, 114, 118

Aldrin PCB 123, 126, 138, 153, 156

Cis-chlordane and cis-nonachlor PCB 157, 167, 169, 170, 180

DDD, DDT group, DDTs PCB 183, 187, 189, 194, 196

dl-PCB TEQ, I-TEQ, WHO-TEQ PCB 202, 206, 208, 209, 28+31

dl-PCBs WHO-TEQ 97and 01 PCB 28, 44, 49, 52, 70, 74, 77

Endrin, Endrin group PCB 81, 99, PCBs

Endrin ketone, heptachlor PCDDs WHO-TEQ 1997

Heptachlor group, epoxide PCDDs/Fs TEQ, I-TEQ, Nordic

Heptachlorepoxide cis- (trans-) PCDDs/Fs WHO-TEQ 95, 97,01

Chlordane (group) PCDDs/Fs WHO-TEQ

Chlord.+oxychlord. (+t-nonach.) PCDFs WHO 1997-TEQ

Chlord. +oxychl. + trans-nonach Sum 10 PCBs

Mirex Sum 3 PCBs

o,p- and p,p- isomers of DDX Total dioxins TEQ

Mono-orth PCBs WHO-TEQ, 97 Total dioxins WHO-TEQ

Non-ortho PCBs WHO-TEQ, 97 Toxaphene

Indicator 7 PCBs trans-Chlordane (= gamma)

Parlar 26, 40, 41, 44, 50, 62 trans-Nonachlor

Alpha-HCH

Beta-HCH

Gamma-HCH

Endosulfans

Ntot = 118*, NSC = 61

Ntot = 4, NSC = 3 Ntot = 114, NSC = 58

Ntot = 0, NSC = 0 Ntot = 4, NSC = 3 Ntot = 17, NSC = 5 Ntot = 97, NSC = 53

Time series Point estimates Point estimates Time series

Chemicals in GMP reports: breast milk

• In total, 17 chemicals regularly reported and related to the Stockholm Convention are suitable for time series analysis.


• The rest of regularly (repeatedly) reported chemicals allow for time-related general assessment of concentration level(s), but not for exact time series analysis.

Breast milk

Japan 1972 - 2006 4 DDTs PCBs

Japan 1986 - 2006 Oxychlord+nonachl

Japan 1980 - 2006 HCB

India 1979 – 1994 DDTs

Sweden 1997 - 2006 4 DDTs Sum 10 PCBs PCDDs/Fs WHO-TEQ 1995

Sweden 1967 – 1997 DDE DDT PCBs

Sweden 1967 - 2007 Dieldrin

Sweden 1972 – 2007 HCB Oxychlordane

Sweden 1972 - 2006 PCB 153

Belgium and Finland 1988 – 2007 Indicator 6 PCBs

Germany 1993 – 2003 PCDDs/Fs I-TEQ HCB DDE

Germany 1997 – 2003 HCB DDE

Germany 2000 – 2004 Sum 3 PCBs

Czech Rep. 2000 – 2007 DDTs HCB PCB 118

Norway 1988 – 2007 Indicator 6 PCBs

All parameters1


Human blood

Chemicals in GMP reports: human blood

Time series Point estimates Point estimates Time series

Ntot = 81*, NSC = 53

Ntot = 4, NSC = 3 Ntot = 77, NSC = 50

Ntot = 0, NSC = 0 Ntot = 4, NSC = 3 Ntot = 6, NSC = 2 Ntot = 71, NSC = 48



17 PCDDs/Fs isomers PCDDs WHO-TEQ 95, 97

Aldrin, dieldrin and endrin PCDDs/Fs WHO-TEQ, 97

Cis-chlordane and cis-nonachlor PCDFs WHO-TEQ 95, 97

DDD, DDE, DDT Sum 3, 14 and 36 PCBs

dl-PCB WHO-TEQ Total dioxins WHO-TEQ 95

dl-PCBs WHO-TEQ 97and 01 Toxaphene

Heptachlor and mirex Trans-chlordane

Heptachlorepoxide cis- (trans-) Trans-nonachlor

Chlordane (group)

Indicator 6 and 7 PCBs

Mono-ortho PCBs WHO-TEQ 95

Non-ortho PCBs WHO-TEQ 95

Non-ortho PCBs WHO-TEQ 97

o,p- isomers of DDX

p,p-isomers of DDD and DDT

Oxychlordane

Parlar 26, 40, 41, 44, 50, 62

PCB 28, 52, 77, 81, 101, 118

PCB 126, 138, 169, 180, PCBs

DDTs

dl-PCBs WHO-TEQ 1998

HCB

p,p-DDE

PCB 153

PCDDs/Fs WHO-TEQ 1998

Alpha-HCH

Beta-HCH

Endosulfans

Sum 3 PBDEs

• In total, 6 chemicals regularly reported and related to the Stockholm Convention

are suitable for time series analysis.


• The rest of regularly (repeatedly) reported chemicals allow for time-related general assessment of concentration level(s), but not for exact time series analysis.

Chemicals in GMP reports: human blood

Human blood

Japan 2002 - 2006 dl-PCBs WHO-TEQ 1998 PCDDs/Fs WHO-TEQ 1998

India 1982 – 2005 DDTs

Germany 1995 - 2006 PCB 153

Germany 1998 - 2006 HCB

Brazil 1997 – 2001 p,p-DDE


• All reported parameters were divided into four principal groups: – Original 12 SC POPs: data related to original 12 SC chemicals (individual analytes specified in

GMP Guidance, 2007): 58 parameters in total – Additional 10 SC POPs: non-obligatory chemicals for reporting, but they are important due to

their presence in the updated list of target compounds of the SC (2009 and 2011), individual compounds specified in revised Guidance on the GMP (2009) in GMP reports: 7 parameters in total

– SC related chemicals – parameters not specified in the Guidance but related to the Stockholm Convention (e.g. summations, TEQs, etc.): 84 parameters in total

– Other chemicals: all other reported parameters not specified in the Guidance or not related to the SC at all: 22 parameters in total

• Trend assessment is possible only if a sufficiently long-term monitoring was performed. Reporting frequency of chemicals was performed for each matrix separately.

• Data reported in the GMP reports are relatively heterogeneous, not standardized in content and time scale. Nevertheless, the final database represents a suitable basis for future comparison of concentration levels.

• The largest data set was obtained for ambient air monitoring. A major part of the original 12 SC POPs is covered sufficiently for time series analysis. Despite the fact that not all compounds of interest are available in a sufficiently long time scale, most of them can be analyzed at least in annually measured time series.

Conclusions

Chapter 4 – Annex 4.1 Ambient air data in GMP reports -

overview

Annex 4.1. Ambient air data in GMP reports

Ambient air data in GMP reports: active sampling


Ambient air: active sampling – data description Concentration data on 71 parameters were published in the GMP reports.

2,3,7,8-TCDD Endosufan SO4 Oxychlordane PCDDs/Fs I-TEQ

2,3,7,8-TCDF Endosulfan I (alpha) p,p-DDD PCDDs/Fs TEQ

5 DDTs Endosulfan II (beta) p,p-DDE PCDDs/Fs WHO 1997-TEQ

Aldrin Endrin p,p-DDT PCDDs/Fs WHO-TEQ

Alpha-HCH Gamma-HCH p,p-DDTs PCDFs (10)

Beta-HCH HCB Parlar 26 PCDFs WHO 1997-TEQ

cis-Chlordane (= alpha) Heptachlor Parlar 50 Sum 10 PCBs

cis-Nonachlor Heptachlorepoxide Parlar 62 Sum 21 PCBs

DDE Heptachlorepoxide cis- (= exo, B) PCB 28 Sum 25 PCBs

DDT + p,p-DDE Heptachlorepoxide trans- (= endo, A) PCB 52 Total dioxins TEQ

DDT group Chlordane PCB 101 Total dioxins WHO 1997-TEQ

DDTs Chlordane + Nonachlor PCB 138 Toxaphene

Dieldrin Indicator 6 PCBs PCB 53 trans-Chlordane (= gamma)

dl-PCB I-TEQ Indicator 7 PCBs PCB 180 trans-Chlordane + cis-Nonachlor

dl-PCB TEQ Mirex PCBs trans-Nonachlor

dl-PCB WHO 1997-TEQ o,p-DDD PCDDs (7) Delta-HCH

dl-PCB WHO-TEQ o,p-DDE PCDDs WHO 1997-TEQ Sum HCHs

dl-PCBs (12) o,p-DDT PCDDs/Fs (17)


Ambient air: active sampling – data description • All of the original 12 POPs in the Stockholm Convention (SC) are fully

covered by ambient air active sampling in GMP reports.

• GMP reports data comprised 47% of recommended analytes specified in chapter 2 of the Guidance on GMP1,2 (2007).,

• Moreover, some chemicals of the group 2 (additional 10 POPs) were already reported

Compounds from the group 2 reported in GMP reports in ambient air active sampling data (2009)3



3 http://chm.pops.int/Convention/ConventionText/tabid/2232/Default.aspx, www.pops.int


Alpha-HCH Beta-HCH Gamma-HCH PeCB Endosulfan I Endosulfan II Endosulfan SO4

Ambient air: active sampling – data description • In total, active air sampling data were reported from 4 UN regions (not

reported in the African regional report).

• Spatial coverage of the ambient air active sampling data in GMP reports comprises 31 countries and subregions.

• Records covered a period of 20 years (1990 – 2009); however, a majority of data was collected recently (after 2004), see table 4.2.

• Data suitable for the long-term analyses were provided from 16 countries.

• DDT isomers, dieldrin, cis-chlordane, heptachlor and HCB were the most frequently reported chemicals, however the set of chemicals is strongly region-specific; therefore, a spatial comparison of parameters among regions is often impossible.


Map - data availability

• Countries with available data from ambient air active sampling stratified according to number of reported parameters.


Countries reporting data from active air monitoring

• Availability of ambient air: active sampling data from Asia and Pacific, CEEC and GRULAC Regions.

** Mesoamerican subregion includes Belize, Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, Mexico and Panama ** Southern Cone subregion covers Chile, Uruguay, Paraguay, Argentina and Brazil



Cambodia 2006 17 17China 2004 28 546

Hong Kong SAR 1998 - 2007 11 38Indonesia 2005 - 2006 20 40

Japan 1997 - 2007 24 831Mongolia 2006 22 22

Philippines 2006 21 21Republic of Korea 2005 - 2007 21 84

Thailand 2006 - 2007 21 42Vietnam 2005 - 2006 22 43

Czech Republic 1989 - 2009 19 261Slovakia 1997 7 8

Chile 2002 - 2003 1 6Mesoamerican subregion 2002 - 2004 10 650Southern Cone subregion 2000 - 2001 2 24

GRULAC

Asia and Pacific

CEEC

Countries reporting data from active air monitoring

• Availability of ambient air: active sampling data from WEOG* Region.

* European Alps sub-region covers 3 mountain tops: Zugspitze (Germany), Sonnblick (Austria) and Weissfluhjoch (Switzerland)



Arctic Canada 1993 - 2006 9 106Australia 2002 - 2003 2 2Austria 2005 - 2006 12 12Canada 1990 - 2005 24 121Finland 1996 - 2005 2 33

Germany 2005 - 2006 12 12Greenland 2004 - 2005 5 6

Iceland 1995 - 2005 5 72New Zealand 1996 - 1997 7 7

Norway 1993 - 2006 7 101Russian Federation 1993 - 2002 8 40

Spain 1994 - 2007 2 5Sweden 1993 - 2005 4 35

Switzerland 2005 - 2006 12 12UK 1991 - 2006 2 6

USA 1990 - 2005 26 148

WEOG

Ambient air data in GMP reports: passive sampling


Ambient air: active sampling – data description • GMP reports contained data on 37 parameters reported from ambient air:

passive sampling

• Out of these, 19 parameters were from the group of chemicals

recommended for monitoring in the Guidance document on GMP (2007)

Alpha-HCH Gamma-HCH p,p-DDD PCBs

beta-HCH HCB p,p-DDE PeCB

cis-Chlordane (= alpha) Heptachlor p,p-DDT Sum 3 PBDEs

dacthal Heptachlorepoxide PCB 101 Sum 48 PCBs

DDT group Chlordane + trans-Nonachlor PCB 118 trans-Chlordane (= gamma)

delta-HCH chlorothalonil PCB 138 trans-Nonachlor

Dieldrin Indicator 7 PCBs PCB 153 trifluralin

Endosufan SO4 o,p-DDD PCB 180

Endosulfan I (alpha) o,p-DDE PCB 28

Endosulfan II (beta) o,p-DDT PCB 52


Ambient air: passive sampling – data description • GMP reports data comprised 32% of recommended analytes specified in

chapter 2 of the Guidance on GMP1,2 (2007) GMP Guidance specified parameters identified in ambient air: passive monitoring data in GMP reports

• In addition, eight chemicals (alpha-HCH, beta-HCH, gamma-HCH, PeCB, sum of 3 PBDEs, endosulfan I, endosulfan II, endosulfan SO4) out of the 10 new POPs group were also reported.



3 http://chm.pops.int/Convention/ConventionText/tabid/2232/Default.aspx, www.pops.int

cis-Chlordane (= alpha) o,p-DDD p,p-DDE PCB 138 PCB 52

Dieldrin o,p-DDE p,p-DDT PCB 153 trans-Chlordane (= gamma)

HCB o,p-DDT PCB 101 PCB 180 trans-Nonachlor

Heptachlor p,p-DDD PCB 118 PCB 28


Ambient air: passive sampling – data description • Passive air sampling data were reported from all UN Regions, and 56

countries and subregions Overview of countries with available passive air monitoring data

Andean subregion* Croatia Ghana Macedonia Romania Togo

Armenia Czech Republic Hungary Malawi Russian Federation Tunisia

Australia Democratic Republic of Congo Iceland Mali Senegal Turkey

Belarus Egypt Ireland Mauricius Serbia Ukraine

Bermuda Estonia Italy Mesoamerican subregion* Slovakia USA

Bosnia and Hercegovina Etiopia Kazakhstan Moldavia Slovenia Zambia

Bulgaria Fiji Kenya Montenegro South Africa

Canada Finland Kyrgyzstan Nigeria Southern Cone subregion*

Caribbean subregion France Latvia Norway Spain

Congo GAPS** Lithuania Poland Sudan

* In Andean subregion is: Bolivia, Colombia, Ecuador, Peru and Venezuela In Mesoamerican subregion are included: Belize, Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, Mexico and Panama In Southern Cone subregion are included: Chile, Uruguay, Paraguay, Argentina and Brazil In Caribbean subregion are included: Antigua and Barbuda, Bahamas, Barbados, Cuba, Dominica, Dominican Republic, Grenada, Guyana,

Haiti, Jamaica, Saint Kitts and Nevis, Saint Lucia, St. Vincent and the Grenadines, Suriname, Trinidad and Tobago ** GAPS covers more than 50 sites on seven continents - http://www.ec.gc.ca/rs-mn/default.asp?lang=En&n=22D58893-1


Ambient air: active sampling – data description

• Overall time range is a relatively limited as the first reported record is from 2003.

• The only site for ambient air passive sampling with at least 4 years of data available for establishment of the long-term trends is Košetice, Czech Republic.

• PCBs, DDTs, HCB and chlordane were the most frequently reported chemicals.


Map - data availability

• Countries with available data from ambient air passive sampling stratified according to number of reported parameters


Countries reporting data from passive air monitoring • Availability of ambient air: passive sampling data from Africa, and Asia and

Pacific Regions



Democratic Republic of Congo 2008 19 152Egypt 2008 25 122

Ethiopia 2008 19 114Ghana 2005 - 2008 26 269Kenya 2008 19 513

Malawi 2005 8 8Mali 2008 19 570

Mauritius 2008 19 114Nigeria 2008 19 152

Republic of Congo 2008 19 114Senegal 2008 19 95

South Africa 2005 - 2008 26 481Sudan 2008 19 114Togo 2008 19 114

Tunisia 2008 19 114Zambia 2008 19 114

Asia and Pacific Fiji 2006 - 2007 19 684

Africa

Countries reporting data from passive air monitoring

• Availability of ambient air: passive sampling data from CEEC Region



Armenia 2008 19 779Belarus 2007 - 2008 19 665

Bosnia and Herzegovina 2006 19 171Bulgaria 2007 19 570Croatia 2007 19 475

Czech Republic 2003 - 2008 32 23778Estonia 2006 - 2007 19 475

Hungary 2007 19 475Kazakhstan 2008 19 779Kyrgyzstan 2008 19 475

Latvia 2006 19 475Lithuania 2006 19 418

Macedonia 2007 19 456Moldova 2007 19 665

Montenegro 2007 19 665Poland 2007 32 663

Romania 2006 - 2007 19 1501Russia 2007 32 453

Serbia and Montenegro 2006 19 608Slovakia 2006 19 1045Slovenia 2007 19 646Ukraine 2008 19 570

CEEC

• Availability of ambient air: passive sampling data from GRULAC Region: data were aggregated over the larger subregions described below


In Andean subregion are included: Bolivia, Colombia, Ecuador, Peru and Venezuela In Caribbean subregion are included: Antigua and Barbuda, Bahamas, Barbados, Cuba, Dominica, Dominican Republic, Grenada,

Guyana, Haiti, Jamaica, Saint Kitts and Nevis, Saint Lucia, St. Vincent and the Grenadines, Suriname, Trinidad and Tobago In Mesoamerican subregion are included: Belize, Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua, Mexico and Panama In Southern Cone subregion are included: Chile, Uruguay, Paraguay, Argentina and Brazil



Andean subregion 2005 - 2008 9 118

Caribbean subregion 2005 - 2008 8 40

Mesoamerican subregion 2005 - 2006 9 79

Southern Cone subregion 2002 - 2008 9 131

GRULAC

• Availability of ambient air: passive sampling data from WEOG Region




Australia 2005 - 2006 18 178Bermuda 2005 - 2006 18 89Canada 2005 - 2006 18 379Finland 2005 - 2006 18 74France 2005 - 2006 18 74GAPS 2005 - 2006 6 10

Iceland 2005 - 2006 18 88Ireland 2005 - 2006 18 89

Italy 2005 18 104Norway 2005 18 72

Spain 2005 - 2006 18 150Turkey 2005 - 2006 18 88

USA 2005 - 2006 18 279

WEOG

GAPS covers more than 50 sites on seven continents - http://www.ec.gc.ca/rs-mn/default.asp?lang=En&n=22D58893-1

Annex 4.2. Ambient air data in UNEP-GMP reports – data structure

Chapter 4 – Annex 4.2 Ambient air data in UNEP-GMP

reports – data structure

• Important note: following tables are for illustration purposes only! • Presented data structure serves as source for online analytical outcomes at

www.genasis.cz/unep (visualizations 2 and 3). See chapter 10 for further details.



Air-active sampling – illustration of data structure

Data available

Data not available

Air-passive sampling – illustration of data structure


Data available

Data not available

Chapter 5 – Annex 5.1 Human tissues data in GMP

reports - overview

Human tissues data in GMP reports - overview

PART I. Breast milk data in the GMP

reports

Human tissues data in GMP reports - breast milk

• Breast milk has been recommended as a matrix suitable for evaluation of a

transfer of chemicals from the environment into human body. • Monitoring of breast milk is coordinated by WHO. The sampling should be

performed according to the WHO protocol (WHO/ECEH).1 • A range of chemicals reported in the GMP reports for human milk was a

relatively wide.

• The most frequently reported compounds were PCDDs, PCDFs and PCBs – usually expressed as Toxic Equivalents – TEQs. However, a plethora of sets of Toxic Equivalency Factors (TEFs) was used for calculation of TEQs.

1 WHO/ECEH, 1996. Levels of PCBs, PCDDs and PCDFs in human milk. Second round of WHO-coordinated exposure study. Environmental Health in Europe Series 3, 121pp.

Introduction


List of parameters reported in breast milk data

1,2,3,4,6,7,8-HpCDD DDTs Mono-ortho PCBs WHO 1997-TEQ PCB 70 PCB 202

1,2,3,4,6,7,8-HpCDF Dieldrin Mono-ortho PCBs WHO-TEQ PCB 74 PCB 206

1,2,3,4,7,8,9-HpCDF dl-PCB I-TEQ Non-ortho PCBs WHO 1997-TEQ PCB 77 PCB 208

1,2,3,4,7,8-HxCDD dl-PCB TEQ Non-ortho PCBs WHO-TEQ PCB 81 PCB 209

1,2,3,4,7,8-HxCDF dl-PCB WHO 1997-TEQ o,p-DDD PCB 99 PCBs

1,2,3,6,7,8-HxCDD dl-PCB WHO 2001-TEQ o,p-DDE PCB 101 PCDDs WHO 1997-TEQ

1,2,3,6,7,8-HxCDF dl-PCB WHO-TEQ o,p-DDT PCB 105 PCDDs/Fs I-TEQ

1,2,3,7,8,9-HxCDD Endosulfans OCDD PCB 110 PCDDs/Fs Nordic-TEQ

1,2,3,7,8,9-HxCDF Endrin OCDF PCB 114 PCDDs/Fs TEQ

1,2,3,7,8-PeCDD Endrin group Oxychlordane PCB 118 PCDDs/Fs WHO 1995-TEQ

1,2,3,7,8-PeCDF Endrin ketone Oxychordane + Nonachlor PCB 123 PCDDs/Fs WHO 1997-TEQ

2,3,4,6,7,8-HxCDF Gamma-HCH p,p-DDD PCB 126 PCDDs/Fs WHO 2001-TEQ

2,3,4,7,8-PeCDF HCB p,p-DDE PCB 138 PCDDs/Fs WHO-TEQ

2,3,7,8-TCDD Heptachlor p,p-DDT PCB 153 PCDFs WHO 1997-TEQ

2,3,7,8-TCDF Heptachlor group Parlar 26 PCB 156 Sum 10 PCBs

4 DDTs Heptachlorepoxide Parlar 40 PCB 157 Sum 3 PCBs

Aldrin Heptachlorepoxide cis- (= exo, B) Parlar 41 PCB 167 Total dioxins TEQ

Alpha-HCH Heptachlorepoxide trans- (= endo, A) Parlar 44 PCB 169 Total dioxins WHO-TEQ

Beta-HCH Chlordane Parlar 50 PCB 170 Toxaphene

cis-Chlordane (= alpha) Chlordane + Oxychlordane Parlar 62 PCB 180 trans-Chlordane (= gamma)

cis-Nonachlor Chlordane + OxyChlordane + trans-Nonachlor PCB 28 PCB 183 trans-Nonachlor

DDD Chlordane group PCB 28 + 31 PCB 187

DDE Indicator 6 PCBs PCB 44 PCB 189

DDT Indicator 7 PCBs PCB 49 PCB 194

DDT group Mirex PCB 52 PCB 196

• A total of 121 parameters in breast milk were identified in GMP reports


Parameters reported in breast milk

• All of the original 12 POPs as well as analytes specified in the Guidance

document on Global Monitoring Plan (2007) were covered in the first GMP reports.

• alpha-HCH, beta-HCH, gamma-HCH and endosulfans were parameters classified

in the group 2 (additional 10 POPs) which were also identified in the GMP reports.

• Total coverage of all parameters specified in both GMP Guidelines (2007 and 2009) for breast milk is 85%.

1 Guidance on the Global Monitoring Plan for Persistent Organic Pollutants, Preliminary version, February 2007, Amended in May 2007, United Nations, Geneva - GE.07.00630 - April 2007, UNEP/CHEMICALS/2007/2 2 http://chm pops int/Implementation/GlobalMonitoringPlan/Overview/tabid/83/Default aspx


Parameters reported in breast milk

• A reliable spatial coverage has been achieved – in total, data from 41 countries have been reported.

• 37 of all published parameters (all relate to the original 12 POPs) were reported from all UN regions allowing for spatial comparison.

• Long-term data for some parameters were provided from 8 countries.

13 PCDDs/Fs isomers Heptachlorepoxide cis- OCDF PCB 52 PCB 156

Dieldrin o,p-DDD Oxychlordane PCB 101 PCB 180

Endrin group o,p-DDE p,p-DDD PCB 118 Toxaphene

HCB o,p-DDT p,p-DDE PCB 138 Trans-chlordane

Heptachlor OCDD p,p-DDT PCB 153 Trans-nonachlor


Maps • Countries reporting breast milk data in GMP reports


Maps • Countries reporting breast milk data in the long-term time series

Human tissues data in GMP reports - overview

PART II. Human blood data in the GMP

reports

Human tissues data in GMP reports - human blood

List of parameters reported in human blood data

15 PCDD/F congeners Heptachlorepoxide trans- (= endo, A) Parlar 26 PCBs

Aldrin Chlordane group Parlar 40 PCDDs WHO 1995-TEQ

Alpha-HCH Indicator 6 PCBs Parlar 41 PCDDs WHO 1997-TEQ

Beta-HCH Indicator 7 PCBs Parlar 44 PCDDs/Fs WHO 1997-TEQ

Cis-chlordane (= alpha) Mirex Parlar 50 PCDDs/Fs WHO 1998-TEQ

Cis-nonachlor Mono-ortho PCBs WHO 1995-TEQ Parlar 62 PCDDs/Fs WHO-TEQ

DDD, DDE and DDT Non-ortho PCBs WHO 1995-TEQ PCB 101 PCDFs WHO 1995-TEQ

DDTs Non-ortho PCBs WHO 1997-TEQ PCB 118 PCDFs WHO 1997-TEQ

Dieldrin o,p-DDD PCB 126 Sum 14 PCBs

dl-PCB WHO 1998-TEQ o,p-DDE PCB 138 Sum 3 PCBs

dl-PCB WHO-TEQ o,p-DDT PCB 153 Sum 36 PCBs

Endosulfans OCDD PCB 169 Total dioxins WHO 1995-TEQ

Endrin OCDF PCB 180 Toxaphene

Gamma-HCH Oxychlordane PCB 28 Trans-chlordane (= gamma)

HCB p,p-DDD PCB 52 Trans-nonachlor

Heptachlor p,p-DDE PCB 77

Heptachlorepoxide cis- (= exo, B) p,p-DDT PCB 81

• A total of 82 parameters (chemicals, their isomers, congeners, products of transformation and TEQs) for human blood were identified in GMP reports


Parameters reported in human blood

• Reported parameters covered all of the original 12 POPs.

• GMP reports data comprised 85% of congeners, isomers and degradation products recommended by the GMP Guidance (dl PCB congeners were missing).

• Only HCB and DDE were reported in all UN regions

• Some compounds from the group 2 (additional 10 POPs) were also identified in GMP reports:

Overview of compounds from the group 2 that were NOT identified in GMP reports:

Chlordecone Hexabrombiphenyl ether Pentachlorbenzene (PeCB) Pentabromodiphenyl ether (PBDE)

Hexabrombiphenyl Heptabromodiphenyl ether Tetrabromodiphenyl ether (TBDE) Perfluorooctane sulfonic acid (PFOA),

it salts and perfluoroctane sulfonyl fluoride


List of countries reporting human blood data • Concentration data were reported from 17 countries or subregions

• The most compact datasets were reported from Japan (dl-PCBs and PCDDs/Fs

WHO-TEQs between 2002 and 2006) and from Germany (4 cities, PCB 153 and HCB between 1995 and 2006 and between 1998 – 2006 respectively).

• The longest (even though incomplete) monitoring record was reported from India (DDTs, 1975-2005).

• The largest set of parameters were reported from the Russian Federation (37) and Japan (30).

Alaska (USA) Ghana Mexico Slovakia

Brazil Greenland Northern Scandinavia United States

Canada Iceland Norway

Faroe Islands India Romania

Germany Japan Russian Federation


Maps • Countries reporting human blood data into GMP reports

Annex 5.2. Human tissues data in UNEP-GMP reports – data structure

Chapter 5 – Annex 5.2 Human tissues data in UNEP-GMP

reports – data structure

• Important note: following tables are for illustration purposes only! • Presented data structure serves as source for online analytical outcomes at

www.genasis.cz/unep (visualizations 2 and 3). See chapter 10 for further details.



Breast milk – illustration of data structure

Data available

Data not available

Human blood– illustration of data structure


Data available

Data not available

Annex 6.1.

Overview of variables reported in the GMP reports and their content analysisChemical Parent POPs Transformations products Parameter reported in GMP Content analysis

1. Aldrin Aldrin Aldrin

2. Chlordane cis-Chlordane (= alpha) cis-Chlordane (= alpha)

trans-Chlordane (= gamma) trans-Chlordane (= gamma)

Oxychlordane Oxychlordane

cis-Nonachlor cis-Nonachlor

trans-Nonachlor trans-Nonachlor

Chlordane CC+TC

Chlordane + Nonachlor CC+TC+CN+TN

Chlordane + trans-Nonachlor CC+TC+TN

Chlordane + Oxychlordane CC+TC+OC

Chlordane + OxyChlordane + trans-Nonachlor CC+TC+OC+TN

Oxychordane + Nonachlor OC+CN+TN

trans-Chlordane + cis-Nonachlor TC+CN

Chlordane group CC+TC+OC+CN+TN

3. DDT o,p-DDT o,p-DDT

p,p-DDT p,p-DDT

o,p-DDE o,p-DDE

p,p-DDE p,p-DDE

o,p-DDD o,p-DDD

p,p-DDD p,p-DDD

DDTs composition not specified

DDT group = p,p DDE + p,p DDD + p,p DDT + o,p DDT + o,p-DDE + o,p-DDD

4 DDTs = p,p DDE + p,p DDD + p,p DDT + o,p DDT

5 DDTs = p,p DDE + p,p DDD + p,p DDT + o,p DDT + o,p-DDE

DDT = p,p DDT + o,p DDT

DDE = p,p DDE + o,p DDE

DDD = p,p DDD + o,p DDD

p,p-DDTs = p,p DDE + p,p DDD + p,p DDT

DDT + p,p-DDE = p,p DDE + p,p DDT + o,p DDT

4. Dieldrin Dieldrin Dieldrin

Original 12 POPs in the Stockholm Convention

5. Endrin Endrin Endrin

Endrin ketone Endrin ketone

Endrin aldehyd

Endrin group = Endrin + Endrin ketone

6. HCB HCB HCB

7. Heptachlor Heptachlor Heptachlor

Heptachlorepoxide cis- (= exo, B) Heptachlorepoxide cis- (= exo, B)

Heptachlorepoxide trans- (= endo, A) Heptachlorepoxide trans- (= endo, A)

Heptachlor group = heptachlor + heptachlorepoxide cis- + heptachlorepoxide trans-

Heptachlorepoxide = heptachlorepoxide cis- + heptachlorepoxide trans-

8. Mirex Mirex Mirex

9. PCB PCB 28 PCB 28

PCB 52 PCB 52

PCB 101 PCB 101

PCB 118 PCB 118

PCB 138 PCB 138

PCB 153 PCB 153

PCB 180 PCB 180

PCB 77 PCB 77

PCB 81 PCB 81

PCB 105 PCB 105

PCB 114 PCB 114

PCB 123 PCB 123

PCB 126 PCB 126

PCB 156 PCB 156

PCB 157 PCB 157

PCB 167 PCB 167

PCB 169 PCB 169

PCB 189 PCB 189

PCBs composition not specified

Sum 3 PCBs = 138 + 153 + 180

Sum 10 PCBs = 28 + 52 + 101 + 105 + 118 + 138 + 153 + 156 + 167 + a 180

Indicator 6 PCBs = 28 + 52 + 101 + 138 + 153 a 180

Indicator 7 PCBs = 28 + 52 + 101 + 138 + 118 + 153 a 180

dl-PCBs (12) = 77 + 81 + 105 + 114 + 118 + 123 + 126 + 156 + 157 + 167 + 169 + 189

Sum 14 PCBs

Sum 21 PCBs

Sum 25 PCBs

Sum 36 PCBs

Sum 48 PCBs

PCB 44

PCB 49

PCB 70

PCB 74

PCB 99

PCB 110

PCB 170

PCB 183

PCB 187

PCB 194

PCB 196

PCB 202

PCB 206

PCB 208

PCB 209

PCB 28 + PCB31

dl-PCB TEQ TEQ type used not specified

dl-PCB WHO-TEQ TEQ type used not specified

dl-PCB WHO98-TEQ TEQ

dl-PCB I-TEQ TEQ

Mono-ortho PCBs WHO98-TEQ TEQ

Non-ortho PCBs WHO-TEQ TEQ type used not specified

Mono-ortho PCBs WHO-TEQ TEQ type used not specified

Non-ortho PCBs WHO98-TEQ TEQ

10. PCDD 2,3,7,8-TCDD 2,3,7,8-TCDD

1,2,3,7,8-PeCDD 1,2,3,7,8-PeCDD

1,2,3,4,7,8-HxCDD 1,2,3,4,7,8-HxCDD

1,2,3,6,7,8-HxCDD 1,2,3,6,7,8-HxCDD

1,2,3,7,8,9-HxCDD 1,2,3,7,8,9-HxCDD

1,2,3,4,6,7,8-HpCDD 1,2,3,4,6,7,8-HpCDD

OCDD OCDD

PCDDs (7) Sum of 7 PCDDs

PCDDs WHO-TEQ TEQ type used not specified

PCDDs WHO98-TEQ TEQ

11. PCDF 2,3,7,8-TCDF 2,3,7,8-TCDF

1,2,3,7,8-PeCDF 1,2,3,7,8-PeCDF

2,3,4,7,8-PeCDF 2,3,4,7,8-PeCDF

1,2,3,4,7,8-HxCDF 1,2,3,4,7,8-HxCDF

1,2,3,6,7,8-HxCDF 1,2,3,6,7,8-HxCDF

2,3,4,6,7,8-HxCDF 2,3,4,6,7,8-HxCDF

1,2,3,7,8,9-HxCDF 1,2,3,7,8,9-HxCDF

1,2,3,4,6,7,8-HpCDF 1,2,3,4,6,7,8-HpCDF

1,2,3,4,7,8,9-HpCDF 1,2,3,4,7,8,9-HpCDF

OCDF OCDF

PCDFs (10) Sum of 10 PCDFs

PCDDs/Fs (17) = PCDDs + PCDFs

PCDFs WHO-TEQ TEQ type used not specified

PCDDs/Fs I-TEQ TEQ

PCDDs/Fs Nordic-TEQ TEQ

PCDDs/Fs TEQ TEQ type used not specified

PCDDs/Fs WHO98-TEQ TEQ

PCDDs/Fs WHO-TEQ TEQ type used not specified

PCDFs WHO98-TEQ TEQ

Total dioxins WHO98-TEQ TEQ (Total dioxins = PCDDs + PCDFs + dl-PCBs)

Total dioxins TEQ TEQ type used not specified

Total dioxins WHO-TEQ TEQ type used not specified

12. Toxaphene Parlar 26 Parlar 26

Parlar 50 Parlar 50

Parlar 62 Parlar 62

Toxaphene

Parlar 40

Parlar 41

Parlar 44

Chemical Parent POPs Transformations products Parameter reported in GMP Content analysis

1. Chlordecone Chlordecone

2. Alpha-HCH Alpha-HCH Alpha-HCH

3. Beta-HCH Beta-HCH Beta-HCH

4. Gamma-HCH (= Lindane) Gamma-HCH Gamma-HCH

5. Hexabrombiphlenyl PBB 153

6. PeCB PeCB PeCB

7. tetra+penta BDE BDE 47

BDE 99

BDE 100

BDE 153

BDE 154

8. hepta+octa BDE BDE 175/183

Sum 3 PBDEs 47 + 99 + 100

9. PFOS/PFOSF PFOS

PFOSF

10. Endosulfan Endosulfan I (alpha) Endosulfan I (alpha)

Endosulfan II (beta) Endosulfan II (beta)

Endosulfan SO4 Endosulfan SO4

Endosulfans not specified

Additional 10 POPs in Stockholm Convention

Chemical Parent POPs Transformations products Parameter reported in GMP Content analysis

dacthal

delta-HCH

chlorothalonil

trifluralin

Sum 4 HCHs

Sum 16 PAHs

naftalen

acenaftylen

acenaften

fluoren

fenantren

antracen

fluoranten

pyren

benzo(a)antracen

chrysen

benzo(b)fluoranten

benzo(k)fluoranten

benzo(a)pyren

indeno(123cd)pyren

dibenzo(ah)antracen

benzo(ghi)perylen

Other compounds in GMP reports that are not a subject to the Stockholm Convention

Chapter 10 – Annex 10.1 On-line data visualization

User manual

Introduction

• Web site www.genasis.cz/unep presents POPs data from environmental monitoring programmes as provided to GMP regional reports. Data are sorted by matrices, time, and compounds of interest.

• Four analytical outcomes are available: – World map – monitoring overview – Sampling frequency – compounds – Sampling frequency – years – Measured values

http://www.genasis.cz/unep

www.genasis.cz/unep

Enter password to access online data

visualisation

Main menu Click on this logo to return

to this page again

Description and links to all available reports

Use this dropdown list to switch

among matrices Drag to zoom in and out. You can use scroll

button on your mouse to zoom

as well.

You can move in the map by

dragging it

Use these buttons to move, play or pause

slideshow in the selected time scale



Hover the mouse cursor over a cell to get additional information about

aggregated value

List of countries

List of compounds



List of countries

List of years

Hover the mouse cursor over a cell to get additional information



Reported values are shown as a box-and-whisker plot. A plot

consist of minimum, maximum, mean and median.

If a value is unavailable, appropriate sign is not

displayed

User manual

• Previous slides are based on randomly generated data and are provided just for demonstration purposes of the tools available in the on-line visualisation.

• Please use the link (www.genasis.cz/unep) to access pilot online reporting tool and view POPs data reported in the first GMP reports.

Date post:	19-Oct-2019
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

· Research Centre for Toxic Compounds in the Environment, Masaryk University, Brno, Czech...

Documents