This website uses non-intrusive cookies to improve your user experience. You can visit our cookie privacy page for more information.

Higher tier drainflow from MACRO

Background

Higher tier drainflow exposure assessments using the MACRO model have become the standard calculation method submitted by Applicants in order to refine aquatic risk assessments in the UK. These assessments are needed when the aquatic risk assessment fails the UK first tier drainflow calculation method (see first tier drainflow PECsw).

To harmonise this area of the exposure assessment an automated Excel tool has been developed by HSE, supported by key stakeholders. The Excel tool runs MACRO v4.3b simulations, based on the original approach of Brown et al (20041)). Simulations are generated for four vulnerable drained soil classes with dry, medium and wet weather climate scenarios for parent and metabolite compounds. The tool includes a step-by-step user guide along with the necessary soil and climate files. The higher tier MACRO drainflow tool and associated files (including the MACRO v 4.3b model) can be downloaded (Compressed (zipped) folder containing necessary files and 17.5MB in size).

There is also the option for higher tier drainflow assessments to be performed using the Webfram model (see higher tier drainflow from Webfram).  Applicants should note that either approach to refining first tier exposure assessments is acceptable and it is not necessary to pass both methods. Applicants are encouraged to use either of the harmonised tools as soon as possible in regulatory submissions to the UK. However HSE will continue to accept individual Applicant approaches to running higher tier MACRO simulations for a period of 6 months (ie for submissions submitted to HSE up to June 2018).

Further guidance on the selection of appropriate parameters for use with the MACRO model can be found in the research report for defra project 'Development of guidance on parameter estimation for MACRO.

The general approach to refining first tier drainflow exposure assessments using either higher tier approaches implemented into MACRO or Webfram is considered valid by HSE. However after a number of years of experience with submissions, HSE released the following guidance on the presentation of higher tier drainflow modelling results for UK submissions.

Typically the maximum daily concentrations on an annual basis from each of up to 30 years’ worth of a simulation have been generated. These values have to be compared with a Regulatory Acceptable Concentration (RAC) derived from the Ecotoxicology database to estimate the numbers of years where the Toxicity Exposure Ratio (TER) would be unacceptable (ie where the maximum daily concentration exceeds the Regulatory Acceptable Concentration (RAC) on at least one day per year). This is normally presented for each soil and weather scenario separately. An overall level of exceedance can be estimated if the individual percentages failing at each scenario are weighted to reflect the extent to which each scenario represents drained soils for the relevant crop of interest in England and Wales. An example of this form of presentation is provided at the end of this guide in Appendix A.

An additional refinement that has been used for some substances has been to calculate the 90th percentile value of daily concentrations for each scenario. This value has been calculated by ranking all daily drainflow concentrations greater than 0.001µg/l (to weight the assessment in favour of the periods of peak drainflow events) before calculating the 90th percentile. The 90th percentile value for each scenario has then been compared directly with the regulatory concentration. The rationale behind this approach has been that the 90th percentile level has often been used in environmental assessments as being an appropriate realistic worst case value (eg typically a 90th percentile spray drift value is derived from the larger database of experimental drift values for use in aquatic risk assessments; additionally the FOCUS groundwater scenarios aim at describing an overall vulnerability approximating to the 90th percentile).

However there were concerns about the appropriateness of the different methods of presenting the results, particularly the use of a 90th percentile drainflow value. Although there is a general understanding that the 90th percentile level may be an appropriate realistic worst case value in other areas of environmental risk assessment, HSE considers that the frequency of drainflow events predicted by models such as MACRO should also be considered when assessing the overall protectiveness of such an approach. For example, when a 90th percentile spray drift value is used for a product that may be applied once per year, the resulting Predicted Environmental Concentration (PEC) value implies that for 9 out of 10 spray events for the receiving water body (ie a static 30 cm deep ditch) the PEC would be lower than predicted. Conversely, 1 out of 10 would be greater.

When a 90th percentile drainflow concentration is selected, it would mean that for 9 out of 10 drainflow events for that water body, the PEC would also be lower than predicted. However drainflow events are generally predicted to occur much more frequently than spray drift events. Therefore, for this type of drainflow modelling, the 90th percentile concentration value is not considered comparable to a concentration derived from a 90th percentile spray drift value, nor is it comparable to the protection goals of the FOCUS groundwater scenarios, in either the temporal frequency of exceedances or the spatial level of exceedance. Therefore it is not possible to say that the 90th percentile drainflow concentration protects either 90 percent of soil scenarios, or that it protects 90 percent of the simulated years. For this reason the use of the 90th percentile drainflow concentration approach was no longer be considered a valid approach to refine such exposure assessments by HSE.

Therefore, since March 2009, all such risk assessments have been based initially on a consideration of the maximum daily concentrations versus the RAC on an individual scenario basis to determine the number of years where the risk assessment would be acceptable or not. Further details in the form of a worked example are included in Appendix A.

As regards a suitable precedent in terms of the numbers of combined scenario years failing the relevant TER trigger, HSE proposes that a pass rate of 90 %& should be set as a guide to an appropriate value (ie the total scenario years failing plus any land area that remains unquantified2 by the modelling should be no greater than 10 %). Such a trigger is likely to be appropriate for use against effects endpoints derived from tests on species with the potential for rapid recovery (eg first tier data on algae and aquatic plants). Such a trigger may not be appropriate when the critical risk is identified to be against other species eg fish, or when the RAC has been derived from higher tier data from any species.

Historically the SEISMIC tool has been used to aid interpretation of the results of higher tier drainflow assessments, particularly for determining the relative extent of soil, crop and climate scenarios in the UK. Since the SEISMIC tool is no longer available to all Applicants, and there is currently no other agreed method for determining the extent of soil, crop and climate scenarios in the UK, Applicants may be unable to undertake the full scenario years type calculations outlined in Appendix A. Note this is only a problem when performing higher tier assessments using the MACRO model, since the full scenario years type calculations are included as standard in the Webfram tool. This means that some Applicants will be unable to calculate the combined scenario year’s value for comparison against the 90 % pass rate (ie the total scenario years failing plus any land area that remains unquantified by the modelling should be no greater than 10 %). In the absence of a universally available tool, HSE proposes that regulatory decisions should principally be based on a detailed consideration of the significance of the individual levels of RAC exceedance within each scenario. Such a consideration could take into account the ecology of the specific group of organisms of concern in the aquatic risk assessment, as well as a consideration of the relative distribution of each scenario within the agricultural landscape where this information is available. It is therefore important that the detail of all scenarios be presented in the final regulatory submission in order that each assessment can be fully evaluated by HSE (see the worked example in Appendix A for further details). This approach is considered justifiable because based on experience it is highly unlikely that the combined scenario years trigger of 10 % would be breached without the individual scenario trigger also being breached.

It is recommended that any higher tier drainflow approaches to support a UK approval should be discussed with HSE in the first instance if they are likely to deviate significantly from the guidance outlined above. For example, if you intend to include additional refinements to the assessment such as calculating exposure concentrations that exceed the regulatory concentration in a time window appropriate to the critical effect endpoint this will need to be fully justified from both an exposure and effects point of view. In addition if you intend to take into account either the duration of exposure relative to the time scale of effects, recovery effects and/or the effects of repeated exposures these issues will again need to be fully supported from both the exposure and effects side.

Footnotes

  1. Brown, Dubus, Fogg, Spirlet and Gustin (2004) Exposure to sulfosulfuron in agricultural drainage ditches: field monitoring and scenario based modelling. Pest . Manag. Sci. 60 pp765-776.
  2. Note that in the automated Excel tool very wet climate scenarios (>850 mm rainfall) are not modelled.  However, the results for the wet scenarios can be used as a surrogate for results from these very wet areas to minimise the area not quantified by the model. This is consistent with how this area is handled in the Webfram model.

Appendix A

Example of presentation of higher tier MACRO drainflow modelling results based on full reporting of scenario exceedances

Risk assessment based on a RAC of 5.0µg/l for 'Substance X' for a simulated use on wheat.

Number of years with ditch concentrations of 'Substance X' >5.0 µg/l on at least one day for each of the modelling scenarios for actual daily concentrations (total years simulated per scenario = 30 ; values in parentheses are percentages of years exceeding 5.0µg/l on at least one day per year):

Number of years of 30 year simulation with daily ditch concentration >5.0µg/l on at least one day.

Soil Dry (<625 mm per annum) Medium (625-750 mm per annum) wet (750-850 mm per annum)
Denchworth 5/30 (16.7) 10/30 (33.3) 8/30 (26.7)
Hanslope 4/30 (13.3) 6/30 (20.0) 5/30 (16.7)
Brockhurst 1/30 (3.3) 2/30 (6.7) 1/30 (3.3)
Clifton 0/30 (-) 1/30 (3.3) 0/30 (-)
Quorndon 0/30 (-) 0/30 (-) 0/30 (-)

Proportion of total wheat growing land in England and Wales accounted for by each scenario (figures taken from Brown et al.,2004).

Soil type Extent of soil within each climate scenario ( %) Total extent
Dry Medium Wet >850mm p.a. ( %)
Undrained  -  -  -  - 45.9
Peaty soils  -  -  -  - 3.5
Denchworth 2.7 3.0 0.7 0.5 7.0
Hanslope 9.0 5.6 0.5 0.4 15.5
Brockhurst 4.8 7.6 1.8 0.9 15.1
Clifton 1.5 5.2 1.6 0.9 9.2
Quorndon 2.5 0.9 0.3 0.2 3.8
Total 20.4 22.3 4.9 2.9 100

Results for each scenario can be weighted according to their prevalence across England and Wales.  For example the 16.7 % exceedance predicted at the 'Denchworth dry' scenario above can be multiplied by the value of 2.7 % which represents its relative distribution within the total wheat growing land of England and Wales: 16.7 % * 2.7 % = 0.45 %.  The cumulative level of scenario year exceedances can be obtained by summing the relative contributions for all 15 scenario combinations:

Soil type Extent of soil within each climate scenario ( %) Total extent
Dry Medium Wet ( %)
Denchworth 16.7 * 2.7 %
= 0.45
33.3 * 3.0 % = 1.00 26.7 * 0.7 %
= 0.19

= 1.64
Hanslope 13.3 * 9.0 %
= 1.20
20.0 * 5.6 %
= 1.12
16.7 * 0.5 %
= 0.08

= 2.4
Brockhurst 3.3 * 4.8%
= 0.16
6.7 * 7.6 %
= 0.51
3.3 * 1.8 %
= 0.06

= 0.73
Clifton 0 * 1.5 %
= 0
3.3 * 5.2 %
= 0.17
0 * 1.6 %
= 0

= 0.17
Quorndon 0* 2.5 %
= 0
0 * 0.9 %
= 0
0 * 0.3 %
= 0

= 0

Combining all years and scenarios and considering the number of scenario years with actual concentrations of 'Substance X' where concentrations were greater than 5.0µg/l on at least one day, this occurred in a combined total of 5.0 % of scenario-years.  The concentration was <5.0µg/l in 92.1 % of scenario years and the risk was not quantified in 2.9 % of scenario years (the unquantified area represents the proportion of wheat growing land where annual rainfall was >850mm).

In addition, a consideration of the significance of the individual levels of exceedance within each scenario should be provided by Applicants. For example in the worked example above the maximum number of exceedances is seen at the 'Denchworth medium' scenario and here an assessment of the acceptability of the level of exceedance of 10 out of 30 years should be provided in the regulatory submission. As previously stated such a consideration could take into account the ecology of the specific group of organisms of concern in the aquatic risk assessment (eg algae, aquatic plants, invertebrates etc), as well as a consideration of the relative distribution of each scenario within the agricultural landscape. As highlighted above, in the absence of availability of the SEISMIC tool, regulatory decisions based on MACRO modelling should principally be based on a detailed consideration of the significance of the individual levels of exceedance within each scenario only, and not need to consider the combined scenario years. Based on experience, HSE now differentiates between the following groups when making regulatory decisions:

Algae and aquatic plants

When the risk to aquatic plants and algae are being assessed via the use of MACRO, we use the criteria of no more than 10 % overall failure rate and no more than 60 % in any one scenario. These regulatory precedents were set during discussions with the predecessor of the Expert Committee on Pesticides, the ACP, and mean that no more than 18 years out of 30 should exceed the RAC based on first tier data.  Note that if an applicant has carried out Webfram, then the same criteria should be used.

Fish and aquatic invertebrates

There are no clear criteria (for either outputs from MACRO or Webfram) for the acceptance of exceedances for aquatic invertebrates and fish. This is due to our more limited experience of higher tier drainflow assessments for these groups.

In light of this, it is therefore necessary to consider the issue on a case-by-case basis considering such issues as the exposure profile in regards to the magnitude, frequency and the duration of the exceedances (only available from MACRO), the time to effect (requires raw data and suitable assessment intervals) and the recovery or potential species that may be affected. Please consider these exceedances further against the following metrics:-

However in order that the Webfram model can be used in situations where the critical effect endpoint is derived from a fish or aquatic invertebrate study, and to simplify the assessments using information from the MACRO generated exposure profiles when the level of RAC exceedance is low, HSE have proposed the following additional criteria:

Updated 2017-12-12