SMS supports EGCSA with rejection of proposed scrubber water Emission Factors at IMO

Following an in-depth review, Sustainable Maritime Solutions has provided a briefing document to EGCSA, which roundly rejects Swedish/German proposals for substance Emission Factors in scrubber water overboard discharges at IMO.

The briefing document, reproduced below, supports a submission to the 82nd session of the IMO Marine Environment Protection Committee by IBIA – The International Bunker Industry Association, which calls for Emission Factors to be developed from a large sample analysis programme using IMO agreed standards rather than the flawed use of historical studies by Swedish researchers.

SMS found it surprising that multiple weaknesses were not recognized and cited by the Swedish led academic team when the Emission Factors were first published. How and why is not clear but to ensure well informed and correct decision making, expert advice that is both transparent and balanced is of course vital.

Overview and recommendation

This briefing document relates to a submission to IMO by Germany (MEPC 78/9/3, 15 Apr 22)- Unified and representative emission factors for environmental risk assessment of the discharge water from exhaust gas cleaning systems, which is scheduled to be reviewed at MEPC 82.

Following an in-depth review of the source data, it is strongly recommended that the proposed Emission Factors are rejected, as the validity is questionable for multiple reasons.

  • The source data studies are historical and largely involve early adopters of marine EGCS technology, which may not be representative of current systems;
  • It is not possible to validate the calculation methodology for the emission factors as the source data does not have complete engine power and overboard discharge flow rate data;
  • Questions around source data quality and how data has been used in calculations;
  • Only a very few sample data points have been used to calculate some emission factors, which significantly reduces reliability;
  • There is a lack of assurance that source data studies have followed proper sampling procedures on board ship and that representative samples have been taken;
  • There is evidence that standard test methods used for laboratory analysis of samples have not been used, so result data cannot be confirmed as comparable and fully representative;
  • Several substances are not detected in a large proportion of samples. However, because various analysis methods have been used with very wide ranging limits of substance detection, statistical treatment of the sample data has resulted in the average concentrations being artificially high;
  • Application of an overly high discharge flow rate and sampling below recommended engine loads has increased the proposed emission factors.

Instead it is recommended that representative emission factors be developed from a large data set of samples, gathered globally under an accepted standard protocol, with engine load range levels above 50%, and analysed at laboratories with equipment providing common low detection levels. All relevant operational data used in the calculation of the Emission Factor should be included, e.g. ship, location, time/date, engine load and discharge water flow rates.

Detailed comments on this matter have been submitted to IMO by IBIA (MEPC 82/5/1, 26 July 2024). Explanatory non-technical notes are contained below.

Why is this important?

Exhaust gas cleaning systems are installed on over 5000 ships consuming approximately 30% of the total annual fuel consumption by the global fleet. Scrubbing technology enables cost effective transportation of goods across the world using residual fuel with a well-to-wake carbon footprint that is lower than distillate fuel. Without EGCS the cost of goods would increase for all. The technology will also perform a key role in on board carbon capture systems and in the reduction of particulate matter emissions regardless of the fuel in use.

Protecting the environment and preserving the economic and technological benefits of EGCS are vital as shipping transitions to a lower carbon future. It is therefore vital that the use of EGCS is not unnecessarily restricted and decisions are based on fully validated and representative data.

Where are Emission Factors used?

Emission factors are used in environmental risk assessments and IMO has adopted Risk Assessment Guidelines for EGCS overboard discharges [1]. Broadly speaking a risk assessment predicts the environmental concentration of a substance in a body of water e.g. a sea area, based on inputs of the substance to the water and distribution and losses of the substance over time. This is compared to the predicted concentration at which there will be no effect, based on toxicity testing with live organisms together with application of a conservative safety factor, or an existing Environment Quality Standard.

[1] MEPC.1/Circ.899, 10 June 2022

What are emission factors?

Emission factors are an integral component of an environmental risk assessment as enable the input of substances to a body of water to be modelled. The emission factors proposed in the submission by Germany are for metals, polyaromatic hydrocarbons (PAH) and nutrients in open loop EGCS discharges. If not representative, the quantity of substances entering the water may be under or over-stated and the risk assessment  compromised.

The concentration of a substance in a EGCS discharge, the overboard flow rate of the EGCS discharge and engine power at the time of sampling are used for calculation of an emission factor; each must be accurately measured. Following correct on board sampling procedures and standardized laboratory analysis methods are vital as substance concentrations in EGCS overboard discharges are at parts per billion level.

How have the proposed emission factors been developed?

The emission factors proposed by Germany are the same as those included in a paper submitted to IMO by Sweden in 2022 [2] and are based on a Chalmers University of Technology literature review of studies of EGCS discharges between 1993, when marine scrubbing technology was at a prototype stage, and 2018.

Most studies were undertaken on ships of early EGCS adopters (e.g. ro-pax and ro-ro ship operators) in North-West Europe and the Baltic.

Using laboratory analysis results reported in the source data studies, the Chalmers group calculated an average concentration for each substance of interest which was then multiplied by an average ‘specific’ discharge water flow rate to give the proposed emission factor. The specific discharge water flow rate used was twice the nominal rate typically associated with open loop EGCS.

To assist with understanding, four straightforward calculation formulae are shown in an Appendix to this document. The source data is freely available online to download and examine. It has been used in a number of papers by the same group of authors and the EU EMERGE project [3].

[2] PPR 9/INF.16, 28 Jan 2022

[3] https://research.chalmers.se/en/publication/527206; and https://research.chalmers.se/en/publication/539968

What average data are the proposed emission factors based on?

Average substance concentrations of metals, PAHs and nutrients have been developed from sets of laboratory sample analysis results in source data studies. Typically each data set includes concentrations of several substances from analysis of an EGCS discharge water sample. However, a few are average or median values determined from analysis of a number of samples in a single study. Most data sets do not include all substances of interest. This is further discussed below.

The average specific discharge flow rate is developed from some of the sets of data using the engine power and overboard discharge flow rate where given.

It is not possible to calculate individual emission factors for the substances in all data sets and then calculate an overall average of those because not all overboard discharge flow rates and engine powers are reported. It is therefore not possible to confirm if the methodology used to calculate the proposed emission factors is truly valid.

There are several source data studies, so why are the proposed emission factors of questionable validity?

Source data quality

Of the 15 studies providing source data, 2 are unpublished, one includes unpublished analysis data sets from samples not specifically taken for the study and one includes data from 2 samples not taken at the overboard discharge, but at the scrubber outlet before cleaning.

Unpublished studies lack data visibility which means assessment and confirmation of data quality is difficult and may not be possible. Sampling procedures and analysis methods cannot be readily confirmed. Including samples from locations other than the overboard discharge also suggests a lack of quality control when compiling the source data used to calculate the proposed emission factors.

Small number of sample data points used in the emission factor calculation

While there are several sets of source data, most do not include all substances of interest. This is particularly true for a number of the metals, for which average substance concentrations and thus emission factors are based on 10 or less sample data points. One metal is based on just 2 data points!

The authors of the original literature review quote a ‘95% confidence interval’, however for robust statistical analysis the substance concentrations in a small number of samples (especially less than 30) may not give a true ‘real-world’ average and outliers can have a disproportionately large effect on the average, causing it to be skewed and unreliable.

In summary, the reliability of an average based on a small data set of 10 or fewer is much reduced and there is an increased risk of bias.

Lack of assurance that proper sampling procedures have been followed on board ship and that representative samples have been taken

Concentrations of substances in EGCS discharges are at parts per billion level, it is therefore vital recommended on board sampling procedures are followed, including strict adherence to cleanliness, use of correct containers and transportation requirements so that sample integrity is not compromised.

Based on hands-on experience the 2021 EGCS Guidelines [4] contain highly pertinent information on sampling and the supporting data to be collected. Only a small number of the source data studies have descriptions of the sampling process and all were undertaken before the 2021 EGCS Guidelines were adopted.

The lack of information on sampling in the source data studies means it cannot be confirmed that correct procedures have been followed and samples fully representative of EGCS overboard discharges have been taken.

 [4] MEPC.340(77) 26 November 2021

Evidence that standard test methods used for analysis of samples at the laboratory have not been used

It is well known that different sample preparation and laboratory analysis methods can result in different levels of target substance recovery from a sample of liquid and any particles within it. IMO’s EGCS Risk Assessment Guidelines therefore recommend that specific ISO and EPA standard test methods in the 2021 EGCS Guidelines be used to ensure comparability of laboratory results.

Of the published source data studies, just one confirms use of a standard test method recommended in the 2021 EGCS Guidelines. Of the others the test methods are either unreported, withdrawn or a mix of modified ‘in-house’ and standard methods.

 As a simple illustration of the potential for unrepresentative results, it is not known if individual metal concentration values used to determine averages are from total metals analysis, i.e. dissolved plus particle-bound metals; or only metals dissolved in the liquid sample (i.e. particle bound metals are not included).

 The lack of information and use of non-standard test methods in the source data studies means it cannot be confirmed that analysis data are comparable and fully representative of EGCS overboard discharges.

The effect of different test methods and statistical treatment of substance concentrations below the limit of detectability in a large proportion of samples

Average substance concentrations have been used to calculate the proposed emission factors, however, for several metals and PAHs the concentrations are below the limits of detectability for the various test methods used in over half of the source data samples.

Rather than considering  these concentrations to be zero, 50% of the detection limit is applied when calculating the average. The origins of this practice appear unclear but seem to have begun in the 1970s in response to the growing need for standardized methods to handle non-detects in environmental data. It became more formalized in the 1980s and achieved widespread adoption by the 1990s, driven largely by regulatory requirements and the need for consistent, unbiased data analysis (i.e. neither an under-estimation nor over-estimation) in environmental monitoring.

While the above is now standard practice, in the source data the limits of detectability for each substance differ by a factor of up to 200 for metals and over a factor of 600 for PAHs depending on the analytical test method used.

To illustrate the effect, if the same sample were to be analysed using 2 analysis methods, the first with a detection limit of 4 and the second with a detection limit 200 times higher i.e. 800 and the substance of interest was not detected in either sample, the values for the average calculation would be 2 and 400 using the 50% rule, giving an average of 201 rather than the more accurate average of 2.

Various test methods have been used in the source data studies and limits of detectability range significantly. When non-detectable concentrations are replaced with 50% of the detection limit, the test methods with high limits of detectability can cause an artificial increase in the average concentration for the substance.  It is therefore important that sample analysis data be determined consistently using recommended standard methods with similar laboratory detection limits for each substance tested and that the limits be as close as possible to current laboratory equipment sensitivities.

The effect of using of an overly high specific discharge flow rate

Each proposed emission factor is the product of the average concentration of the substance multiplied by an average specific discharge flow rate. The emission factor will therefore change in direct proportion to the average specific discharge flow rate used.

90m3/MWh is used as the average specific flow rate, so 45m3/MWh (the nominal rate associated with open loop EGCS) would half each emission factor, 180m3/MWh would double each emission factor.

The source data set does not contain complete discharge rate and engine power data for each sample, so 90m3/MWh is applied to substance concentrations in samples that may have been taken at higher or lower flow rates, although a review of the available specific discharge flow rate data shows the median value to be 63 m3/MWh and over half of the samples to have been taken at less than 75m3/MWh. Emission factors at these values would be up to 30% lower than those proposed.

In the source data high specific discharge flow rates were typically found when ships were operating and manoeuvring at low engine powers near ports and anchorages rather than when underway at sea where ships spent the majority of the time.  A small number of these high specific discharge flow rates have heavily skewed the overall average value in a relatively small data set. However in the last ten years, there have been improvements in EGCS scrubbing efficiency and seawater flow control to reduce energy consumption, e.g. by use of variable frequency drive pumps. The proposed emission factors are based on historical data and so do not take into account advances in technology that can reduce specific discharge flow rates.

Furthermore the 2021 EGCS Guidelines recommend sampling at greater than 50% engine load as do EPA and ISO sampling protocols.  If specific discharge flow rates recorded below 50% engine load were removed when calculating the average from the source data, all proposed emission factors would be reduced by 30%.

The use of overly high specific discharge flow rates results in a directly proportional increase of emission factors. The historical source data does not have complete specific discharge flow rate data and the average value is heavily skewed by a small number of data points not reflective of the majority of ship operations. Advances in EGCS technology and sampling at recommended engine loads, both of which can reduce specific discharge flow rates and thus emission factors, have not been taken into account.

APPENDIX

Formulae applied for each substance (metal, PAH, nutrient):

Emission Factor substance (mg/MWh) = Average substance concentration in open loop discharge water (µg/l)  x  Average specific discharge flow rate (m3/MWh)

Where

Average substance concentration in open loop discharge water (µg/l)  =Sum of substance concentrations in the source data studies / Number of substance concentrations in the source data studies

And

Average specific discharge flow rate (m3/MWh) = Sum of specific discharge flow rates in the source data studies / Number of specific discharge flow rates in source data studies

Where

Specific discharge flow rate (m3/MWh) = Open loop EGCS overboard discharge flow rate (m3/h) / Engine power at time of sampling (MW)

Links:

MEPC.340(77) 2021 Guidelines

MEPC.1-Circ.899 – 2022 Guidelines For Risk And Impact Assessments Of The Discharge Water From Exhaust Gas Cl… (Secretariat)

MEPC 82-5-1 – Comments on document MEPC 7893 on emission factors for environmental riskassessment of the… (IBIA)

MEPC 78-9-3 – Unified and representative emission factors for environmental risk assessmentof the discha… (Germany)-1

PPR 9-INF.16 – Current knowledge on the impact on the marine environment of large-scale use of exhaust ga… (Sweden)