Sittig et al.

Environ Sci Eur (2020) 32:86

https://doi.org/10.1186/s12302-020-00362-1

RESEARCH Open Access

Consideration of risk management practices

in regulatory risk assessments: evaluation
of field trials with micro‑dams to reduce
pesticide transport via surface runoff and soil
erosion
Stephan Sittig1* , Robin Sur2, Dirk Baets3 and Klaus Hammel2

Abstract
Background: On sloped agricultural fields, water and sediment can be transported downhill as runoff and erosion.
This process can cause losses of valuable top soil material, water resources for plant availability, and nutrients as well
as transport of plant protection products (PPP) into adjacent surface water bodies. In the European and the US risk
assessment for the registration of PPP, runoff and erosion are numerically calculated with the simulation model PRZM
which uses the USDA runoff curve number (CN) concept for the water movement. Results from runoff field trials were
used to estimate the effect of dedicated management practices in terms of mitigating runoff and erosion, i.e. creating
micro-dams between the ridges of potato fields or in maize cultivation on model input parameters.
Results: Application of different cultivation and tillage techniques (micro-dams/bunds) showed a consistent
decrease of the measured quantities of runoff, erosion, and PPP transport as well as of the calculated CN and pre-
dicted environmental concentrations in surface water. The results presented here support the approach to quan-
titatively consider in-field risk mitigation measures (if applied) in the context of regulatory surface water exposure
calculations, as proposed by the SETAC MAgPIE workshop.
Conclusion: Based on these data, a robust case can be made to consider innovative runoff mitigation for risk assess-
ment purposes by, e.g. lowering the CN in the exposure scenarios. In the assessment presented herein, an average
decrease in the mean of the derived CN of 86 of 21 points (± 11, 10th percentile: 12) for potatoes could be derived.
For maize, the mean calculated CN of 73 was lowered on average by 3 points.
Keywords: Pesticides, Mitigation, Runoff, Soil erosion, Micro-dams, Risk assessment, Predicted environmental
concentrations

Background plant protection products (PPP) into adjacent surface

On sloped agricultural fields, water and sediment can water bodies. The erosion of soil material from agricul-
be transported downhill as runoff and erosion. These tural fields has several short-term and long-term conse-
processes cause loss of valuable soil, nutrients and quences: from the removal of seeds and a covering of the
emerging plants to the loss of fertility and organic sub-
stance. Pesticides are generally transported in solution
*Correspondence: sittig.stephan@t‑online.de
1
knoell Germany GmbH, Konrad‑Zuse‑Ring 25, 68163 Mannheim,
via runoff or sorbed to soil particles via erosive trans-
Germany port. The relation between both transportation path-
Full list of author information is available at the end of the article ways depends amongst others on the physio-chemical

© The Author(s) 2020. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and
the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material
in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material
is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the
permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat​iveco​
mmons​.org/licen​ses/by/4.0/.
Sittig et al. Environ Sci Eur (2020) 32:86 Page 2 of 10

and environmental properties of the individual substance prevent soil erosion [40]. Generally, several mitigation
under investigation [33]. strategies can be applied to reduce the amount of super-
In the European and US risk assessment framework, ficial runoff [35]: vegetated buffer/filter strips, grassed
surface runoff is one building block for the estimation depressions, waterways, ditches or wetland. Furthermore,
of potential risks for the aquatic environment, i.e. for measures of Good Agricultural Practice can be applied,
surface water bodies adjacent to agricultural fields. In e.g. conservation tillage [38] or the usage of specific
Europe and the US, runoff of PPP is calculated with the machinery to improve infiltration. All these practices aim
Pesticide Root Zone Model [36], which uses the USDA to reduce the runoff velocity, increase the capability of
runoff curve number (CN) concept [43] to quantify the the soil to retain water on the surface and subsequently
amount of runoff water. In this concept, assuming a high to facilitate infiltration.
CN indicates a relatively large runoff susceptibility of a The technique to construct micro-dams between the
field compared to a lower CN. Hawkins [23] has dem- ridges of ridge–furrow tillage systems is a largely known
onstrated the importance of an accurate determination agricultural technique [22, 26, 37, 39]. Other terms than
of the CN, preferably via direct measurements of runoff micro-dams are furrow-diking, tied ridges, furrow dam-
following precipitation. The European registration proce- ming, basin tillage, basin listing, microbasin tillage, diked
dure is based on Regulation (EC) No 1107/2009. A cor- furrows, basin tillage [26, 37]. Corresponding devices are
responding modelling framework for environmental fate commercially available, e.g. the Barbutte from Cottard
was established by the FOrum for Co-ordination of pesti- (France), the Dycker from Grimme (Germany), or the
cide fate models and their Use [18, 19]. To calculate pre- partition inserting machine from Netagco-Rumpstad/
dicted environmental concentrations for different surface AgriMaas (The Netherlands). Distinct strategies in terms
water bodies (PECsw), predefined standard scenarios (for of tillage and plough management to preserve the fields
runoff, drainage and drift) are applied. These are based and increase infiltration are also applied in maize cultiva-
on European conditions in different geographical regions tion, e.g. disc or drum ploughs to create small dams or
assuming certain percentiles of representativeness. To holes between the rows.
derive PECsw following from runoff, the output of daily Elhakeem and Papanicolaou [15] and Oliveira et al. [28]
runoff fluxes, erosion masses and pesticide mass out- presented general procedures to derive CN from field tri-
puts from agricultural fields given by PRZM simulations als. Hence, such data can be used to assess the effect of
is applied as input for the surface water model FOCUS micro-dams or other management practices between the
TOXSWA [5]. With the latter, concentration dynamics in ridges of potato or maize fields and to quantify the miti-
surface water and corresponding sediment are calculated gation effect by deriving the CN reduction. The SETAC
for model water bodies representing streams, ditches and MAgPIE workshop [35] proposed a lowering of the CN
ponds. The procedure to derive PECsw is currently under by 3 points after the application of in-field bunds in row
revision to take a 20-year simulation period into account crops.
with the aim of more robust calculations of PECsw [14]. The objective of the present work is to enlarge the small
In the US, the legal basis for pesticide regulation is the database for the effects in reducing runoff, soil erosion
Federal Insecticide, Fungicide and Rodenticide Act [42] and PPP transport following the application of certain
and the Endangered Species Act [17]. Runoff/erosion and mitigation measures in potato (i.e. micro-dams) and in
drift are the only entry routes for regulatory exposure maize (similar tillage techniques) cultivation. To this end,
assessments. The corresponding aquatic exposure end- runoff CN were calculated based on rainfall–runoff rela-
points are EEC (estimated environmental concentration tions reported in field studies. Furthermore, reductions
for ecological risk assessments) and EDWC (estimated in soil erosion and in pesticide loadings in these field
drinking water concentration for dietary risk assess- studies were derived quantitatively. The herein calculated
ments). These values are calculated within the PWC shell CN reductions can be used in environmental risk assess-
(pesticide in water calculator) with PRZM coupled with ment to quantify the effects of the optional mitigation
the variable volume water model (VVWM) with a pond measures in terms of runoff. A potential representation
as the receiving water body [16]. in the concept of calculation of predicted environmental
Due to the planting in rows and the resulting bare soil concentrations for surface water (PECsw) is presented as
in between (together with low plant density), it is reason- well.
able to assume that the cultivation of row crops such as
potatoes or maize leads to a higher vulnerability to sur- Methods
face runoff compared to wheat or root crop [2]. The Ger- Trials under consideration in this study
man Environmental Protection Agency recommends In this paper, seven studies (newly designed and from lit-
the “earthing up of transverse ridges for potato crops” to erature) were evaluated—Table 1 lists the experimental
Sittig et al. Environ Sci Eur (2020) 32:86 Page 3 of 10

Table 1 Details of the studies under investigation; n.a.: not available; n.ap.: not applicable
Potatoes Maize
Olivier et al. [29] Goffart et al. Aurbacher Areas [3] Areas [4] CIPF [7] UCL [41]
[21]a et al. [6]b

2009 2010

Device Barbutte Barbutte Self-developed Prototypes (Grimme) Disc and drum Disc and drum
prototypes plough plough
Soil type Sandy loam Various, up to Silty “loess Silt loam Loam Sandy loam n.a.
28% sand soils”
Irrigation No No Yes Yes Yes No No
Micro-dam height (cm) 10–17 10–17 20–27 ≈5 13 Several cm
Distance between the 1.5 1.6 1.5–8 1.6 1.5 Several cm
micro-dams (m)
Plot length/area 30 m 30 m Unknown ≈5m Unknown 72 m2 75 m2
Slope (%) >3 < 3 and > 3 2–10 <3 <4 9, 16 10–12
Plant protection product(s) Yes, n = 4 Yes, n = 5 No No No Yes, n = 4 Yes, n = 1
applied?
The studies of Olivier et al. [29] and CIPF [7] were newly designed and more detailed results were available for evaluation by the authors
a
All information given as mean of 5 trials over 2 years, consisting of trials both with slope < 3% and > 3%
b
Mean values over 35 trials reported only

details. Runoff CN were calculated based on the reported applying 30 mm of rain [4]. In the first trials, the Barbutte
measurements of precipitation and runoff (see Chap- from Cottard (France, see Fig. 1) was applied to construct
ter 2.2 for the calculation procedure). Reductions in run- micro-dams. Similar devices were used in the remaining
off volumes, erosion masses and plant protection product three trials.
loads were derived quantitatively. In Additional file 1, all The two maize studies were conducted with the appli-
experimental results, together with the calculated runoff cation of two types of micro-dam-creating techniques:
CN can be found (Additional file 1: Tables S1–S13). disc and drum plough. Observations were reported over
The five potato studies were conducted either on one a complete season [7] or from two storm events [41],
site for a complete season [29], reported as annual sums respectively. In Fig. 2, the two devices and the resulting
from three sites over 2 years [21], in a research project patterns on the agricultural fields are shown.
compiling 35 trials on silty loess soil trials in south-west- Generally, in 4 of the 7 studies, plant protection prod-
ern Germany [6], in a single-event study design using ucts were applied (cf. Tables S14, S15 in Additional
both pre-wetted and dry soil [3], or an irrigation trial, file 1 for information on mobility). In the potato trials

Fig. 1 Cottard Barbutte equipment and in-furrow micro-dams (Fotos: Dirk Baets)
Sittig et al. Environ Sci Eur (2020) 32:86 Page 4 of 10

Fig. 2 Devices and resulting structures after usage in the cultivation of maize: upper row, disc plough; lower row, drum plough (Fotos: Dirk Baets)

in Olivier et al. [29], aclonifen [10], linuron [20], flufen- (P − 0.2 ∗ S)2
acet [30], and metribuzin [9] and in Goffart et al. [21], Q= . (1)
P + 0.8 ∗ S
fluazinam [11], mancozeb [31], aclonifen, flufenacet, and
metribuzin were used. In the maize trials, flufenacet, Equation (1) can be solved for S (also called: daily
isoxadifen-ethyl [32], tembotrione [13], terbutylazin [12] watershed parameter) yielding
were applied in the CIPF [7] trial, and flufenacet in the   1/2 
UCL [41] trial. 2
S = 5 P + 2Q − 4Q + 5PQ . ( 2)

Concept of the runoff curve number and derivation

The relation between S [mm] and CN is given as [24]:
from the trials
In the USDA CN approach (as used in the European
  
S = 25.4 mm ∗ 1000 CN − 10 , (3)
regulatory context), direct runoff is regarded, i.e. a com-
bination of channel runoff, surface runoff, and subsur- which can be solved for CN as:
face flow [43]. In this study, measured relations between
1000
precipitation and runoff were used for the calculation of CN =  . (4)
runoff CN, representing the transformation of total pre- 10 + S 25.4 mm
cipitation amounts into total runoff amounts. For given P and Q we solve Eq. ( 2) for S and insert the
In the representation of runoff after Mockus [27], a solution into Eq. (4) to finally obtain CN.
potential maximum retention S (mm) after beginning The theoretical maximum CN of 100 leads to S = 0 and
of the runoff is defined. Assuming an initial abstraction finally to Q = P. The CN is the driving parameter of sur-
(consisting of interception, infiltration during early parts face runoff in the risk assessment models. Figure 3 shows
of the storm, and surface depression storage) of 0.2 * S, the relation between precipitation and runoff for differ-
runoff Q [mm] is calculated based on precipitation P ent CN for the example of one of the trials under investi-
(mm): gation, i.e. example results from Areas [3].
Sittig et al. Environ Sci Eur (2020) 32:86 Page 5 of 10

Fig. 3 Relationship between rain and runoff expressed by the runoff curve number (CN). The maximum of 100 describes a 1:1 line; CN 92 and 73
are examples from one of the trials under investigation (Areas, 2005)

The curve number concept used in the regulatory [43]. In the trials considering a complete season of
framework rain–runoff relations (i.e. [7, 29]), the CN were aver-
In the context of risk assessment, CN are defined sce- aged rainfall-weighted due to the assumption that the
nario-specific and crop-specific, being a function of soil reliability of the CN is proportional to the height of the
type, soil drainage properties, and management prac- rainfall event. In general, when averaging over vari-
tices. Furthermore, crop stages are taken into account ous experimental conditions, and not taking detailed
as defined in [19, Appendix D]. For potatoes and maize, information on the specific moisture conditions into
the stages of emergence and maturation are defined account, the assumption of intermediate moisture con-
in scenario R1 with the average CN of 82, being 87 at ditions was judged to be adequate.
harvest time, and 91 during fallow. Average CN refer to
intermediate moisture conditions, named antecedent
moisture condition (AMC) II. Example higher‑tier calculations of maximum predicted
Curve numbers show variability depending on sev- environmental concentrations for surface water
eral boundary conditions, such as the AMC, based on (PECsw,max)
a 5-day precipitation history before each event. This To investigate the effects of lowering CN in the risk
index, denoting AMC II as average conditions, AMC I assessment, example calculations for FOCUS surface
as dry and AMC III as wet conditions, is calculated by water scenarios were conducted. To this end, all stand-
FOCUS PRZM (i.e. PRZM 3.22) on a daily basis. This ard values for the CN for AMC II for the early stage, crop
depends on the AMC in the upper soil layers [44]. The maturation, harvest and fallow were exemplarily lowered
CN are adjusted using the following rules [36]: by 10% (and the CN for AMC I and III were calculated by
the model, as usually, described above). For the three sub-
4.2 ∗ CNAMC II stances aclonifen, fluazinam and metribuzin, which show
CNAMC I = , (5)
10 − 0.058CNAMC II a relatively immobile, intermediate and mobile behaviour,
respectively (see Additional file 1: Table S16 for basic
23 ∗ CNAMC II substance properties) maximum PECsw (PECsw,max)
CNAMC III = . (6) were calculated exemplarily. All other parameters were
10 + 0.13CNAMC II
applied unchanged from the original FOCUS surface
The actual applied CN in the PRZM model are cal- water scenarios for potatoes.
culated in a linear interpolation while being restricted The FOCUS surface water scenarios considered here
in variability and never reaching the endpoints of AMC are stream scenarios with a plant protection product
I or AMC III. For further information please refer to treated field of 1 ha adjacent to the water body. This
Young and Fry [44]. water body is connected to a 100-ha upstream catch-
Here, the aim was to arrive at CNs for AMC II by ment, of which 20% are treated with plant protection
deriving the mean of all CN from the unique events product. This fraction of treated area can be also con-
sidered as the aeric fraction of the target crop in the
Sittig et al. Environ Sci Eur (2020) 32:86 Page 6 of 10

catchment. Such a setting would occur for example maximum predicted environmental concentrations
if four different agricultural crops were grown tak- (PECsw,max), for example FOCUS assessments for
ing 80% of the area and 20% would be non-arable land potatoes and three example substances.
such as pasture or forest (see [19] for details on the
scenario definition). The resulting concentration in the Observed effects of the management practices
stream is essentially the sum of all mass fluxes (from in the measurable quantities
20 + 1 ha) divided by all water fluxes (from 100 + 1 ha) All mitigation measures lead to a decrease in runoff from
to the water body. The micro-dams are assumed to the agricultural fields (Table 2)—by 86% (± 7%) on aver-
be installed on the treated area only so that the gen- age in the potato trials and by 51% (± 9%) in the maize
eration of runoff water and mass fluxes are reduced on trials, respectively. Similarly, the eroded sediment quanti-
21 ha. On the remaining 80 ha the runoff water entry ties were lowered—by 90% (± 6%) and 71% (± 6%), PPP
is unchanged, runoff mass entries are zero. This setup loss was reduced (where applicable) by 88% (± 4%) in
can be described by the formula: potato trials and by 46% (± 12%) in maize trials.

mitigation no mitigation 1−f

PECsw = PECsw ∗ (7) Effects in derived curve numbers
1 − fa ∗ f
The runoff reductions are reflected in lower CN, which
with the runoff reduction factor: were subsequently derived herein. Table 2 lists averages
and percentiles; in Additional file 1, all single data are
runoff fluxno miti − runoff fluxmiti listed.
f = , (8)
runoff fluxno miti
where fa = fraction area treated (= 21 ha/101 ha = 0.208, Potatoes
FOCUS definition for stream scenarios). The evaluation of the field studies suggests that the
If for example the mitigation reduces runoff water potential application of micro-dams justifies a reduc-
and mass fluxes from the treated area by one half tion of the average runoff CN for surface water exposure
(f = 0.5), the PECsw is reduced by about a fraction of modelling. Here, the calculated reduction is on average
0.45. This procedure is in accordance with the estab- 21 points from the mean CN of the untreated trials of
lished practice to consider mitigation of runoff entries 86. Keeping in mind, runoff being a non-linear function
in European environmental risk assessment [25], e.g. of CN, an estimation for an analogous average reduction
using the software SWAN [8]. of the default value of 82 (e.g. during growing phases) in
The runoff reduction factor f was calculated using the FOCUS standard scenarios R1 and R3 can be made
the runoff mass fluxes as evaluated by PRZM (and to be approximately 20 points, the 10th percentile being
used as input for TOXSWA) on the day of the corre- approximately 10 points (here: 12 points). For R2 with
sponding PECsw,max causing event in the simulations the default value of 78, the values of the proposed reduc-
with micro-dams. These dates for with or without tions can be assumed to be identical.
micro-dam installation were different in the R1 and R3 In Olivier et al. [29], a rainfall-weighted average (each
scenarios and identical in R2. In R1 and R3, the micro- individual event is weighted by the corresponding
dams were causing a delay, i.e. a later event delivered amount of precipitation) of CN = 83 for the untreated
PECsw,max under the mitigated condition. setup was derived, being the result of 13 single events.
A corresponding
 PECsw,max  for a situation without Hence, this result was close to the standard FOCUS value
no mitigation
micro-dams PECsw , that was required for of CN = 82. In 4 of the 17 events, the basin at the end of
the calculation of the mitigated PECsw,max the untreated trial site was overflown. Consequently, the
mitigation
( PECsw ), using Eqs. (7) and (8) was derived with “true” extent of the mitigation could not be inferred in
the TOXSWA metamodel [1], since it is not an output these cases, and the overall average constitutes an under-
of TOXSWA. estimation of the overall actual mitigation.
The two trials from Goffart et al. [21] consisted of three
locations, each observed for 1 year. The absolute values
Results and discussion of the calculated CN were much lower than those for
Table 2 lists the various effects of the application of the other studies (Additional file 1: Table S3). Due to
micro-dams in potatoes and the distinct measures in the conceptual difference, the CN derived here were not
maize cultivation as derived herein, both in measur- included in the calculation of the averages.
able quantities and in subsequently derived runoff In Aurbacher et al. [6], the reported overview table of
CN. Table 3 shows the effects of a CN reduction on 35 studies allowed for a calculation of one single average
Sittig et al. Environ Sci Eur (2020) 32:86 Page 7 of 10

Table 2 Effects of micro-dams in potato cropping and distinct strategies in maize cultivation on runoff quantities,
erosion, plant protection product (PPP) loads and curve numbers (CN; means), as means with standard deviations, 10th
percentiles, and ranges; n.a.: not applicable
Potatoes Maize
c
Olivier et al. 29] Goffart et al. [21] a,g
Aurbacher et al. Areas [3] Areas [4] CIPF [7]d,e UCL [41]d,e
[6]b
2009 2010

Runoff reduction 86% (± 14%) 79% (± 19%), 98% 81% (69%–93%) 84% 53% (± 24%) 48% (± 6%)
10th pt.: 82% 10th pt.: 61% (53%–98%) 36%, 70% 52%, 44%
(47%–100%)
Runoff change − 86% (± 7%), 10th percentile: − 80% − 51% (± 9%), 10th percentile:
− 45%
Erosion reduction 85% (± 35%) 88% (± 17%) 97% n.a. n.a. 67% (± 15%) 76% (± 2%)
10th pt.: 67% 10th pt.: 71% (58%– 57%, 78% 77%, 74%
(0%–100%) 100%)
Erosion change − 90% (± 6%), 10th percentile: − 86% − 71% (± 6%), 10th percen-
tile: − 68%
PPP ­reductionf n = 4, 91% (± 5%) n = 5, 84% (± 20%) n.a. n.a. n.a. n = 4.50% n = 1.41%
10th pt.: 86% 10th pt.: 62% (56%– (± 17%) (± 7%),36%,
(87%–96%) 100%) 38%, 62% 46%
PPP change − 88% (± 4%), 10th percentile: − 85% − 46% (± 12%), 10th percen-
tile: − 39%
CN untreated 83 – – 75 92 95 68 78
CN treated 73 – – 39 73 78 67/64d 75/73d
Mean: 65 Mean: 74
→CN reduction 10 (12%) – – 36 (48%) 19 (21%) 17 (18%) 3 (4%) 4 (5%)
CN ­changeg Mean CN untreated/treated: 86/66 (reduction: 21 ± 11; 10th percentile: 12) − 5% (± 0.7%)
(− 25% (± 16%), 10th percentile: − 14%)
a
%-tal reductions in runoff, erosion and PPP reduction given as mean values of 5 trials over the 2 years
b
Mean values over the 35 reported trials available only
c
Runoff was generated “delayed” in this trial, both CN and %-tal runoff reduction based on the 2 measurements after 40 mm of artificial rain
d
Two distinct techniques were applied: disc and drum plough
e
Runoff, erosion, and PPP reduction percentages: all evaluations are given as mean values over the distinct treatments
f
Average reductions of the sums of PPP solved in runoff water and sorbed to the sediment; over all PPP applied
g
Due to the conceptual difference of one reported value for a complete season, CN from Goffart et al. [21] were not included here (see Additional file 1: Table S3 for
details)

Table 3 Results from lowering the standard values of the runoff curve number (CN; intermediate antecedent moisture
conditions, AMC II) by 10%: maximum predicted environmental concentrations for surface water (PECsw,max) (µg/l),
calculated using FOCUS PRZM and TOXSWA and the standard scenarios for potatoes; The recalculated PECsw,max are
based on runoff mass fluxes from FOCUS PRZM, considering a dilution from the upper catchment in the stream scenarios
Aclonifen (immobile) Fluazinam (mod. mobile) Metribuzin [9] (highly mobile)
(1200 g ha−1) [10] (750 g ha−1) [11] (350 g ha−1)
Regular CN CN − 10% PECsw,max Regular CN CN − 10% PECsw,max Regular CN CN − 10% PECsw,max
red. (%) red. (%) red. (%)

R1 ­streama 1.43 0.35 76 0.85 0.20 76 0.61 6.7E−03 99

R2 ­streamb 0.50 0.20 60 0.21 0.08 60 1.21 0.42 66
R3 ­streamc 1.83 0.50 73 1.32 0.33 75 21.1 0.22 99
The substances under investigation are aclonifen, fluazinam and metribuzin, applied once ± 14 days relative to emergence
a
ACL and FLZ: event causing PECsw,max occurs on May 20th and on May 30th 1984 without and with the application of micro-dams, respectively; MTB: on May 7th
and on May 20th 1984, respectively
b
All three substances: both with and without the application of micro-dams, PECsw,max causing event occurs on March 11th 1977
c
All three substances: event causing PECsw,max occurs on April 2nd and on April 20th 1980 without and with the application of micro-dams
Sittig et al. Environ Sci Eur (2020) 32:86 Page 8 of 10

value of reduction. A CN reduction of 36 points (from 75 PECsw,max taking the effect of dilution from the upper-
to 39) was derived from the annually reported data. stream catchment into account. Table S16 in Additional
After the application of artificial rain of 40 mm in the file 1 shows the detailed results after the recalculation, i.e.
trial of Areas [3], an average reduction of 19 points (CN PRZM runoff fluxes with and without the application of
92 to 73) was observed over the two setups with dry and micro-dams, the corresponding reduction factors f and
humid initial conditions, respectively. The CN in the later PECsw,max.
study of Areas [4] was reduced to a very similar extent: In the stream scenarios R1 and R3, the specific runoff
after 30 mm of artificial rain, the CN was decreased by 17 events that lead to the PECsw,max are later ones due to
points (CN 95 to 78). the application of micro-dams (R1: 10 days for ACL and
FLZ, 13 days for MTB; R3 for all three: 18 days). There,
the PECsw,max-decreasing effect of globally lowering the
Maize CN is higher than for R2.
The outcome of the maize trials investigating the applica- Due to the interplay of runoff and erosion driven by
tions of disc plough or drum plough can be summarized runoff, a general statement whether highly mobile (less
as a reduction of the average CN of 73 to 70 points. This sorptive) or less mobile (higher sorptive) substances are
also implies a reduction of 3 points in the FOCUS stand- more affected by a reduction of CN cannot be given.
ard scenarios R1–R4.
For the trial of CIPF [7], conducted over a complete
Consequences for risk assessment and risk management
season and averaged over two repetitions of the treat-
The trials under investigation in this study show a great
ments on different slopes, an overall precipitation-
extent of heterogeneity due to the differences in the exe-
weighted reduction of 3 points (from 68 to 65) was
cution of the field trials (Table 1) and the type of report-
derived (Additional file 1: Table S14). The differences
ing (i.e. either reported for single events or as multiple
between the two applied techniques and slopes were con-
events over a complete season). Nevertheless, the aver-
siderably small.
aged results are considered suitable to quantify the
In UCL [41], the average over the two reported events
effects of micro-dams in potato and maize cultivation.
was CN = 78 for the untreated setup and CN = 74 for the
In risk assessment in case micro-dams or similar tech-
setups after treatment with disc or drum plough.
niques are applied, either a conservative default mitiga-
tion effectiveness can be used or a higher-tier assessment
Example calculations of predicted environmental can be conducted, as, e.g. deriving PRZM CN for specific
concentrations measures as presented here for micro-dams. The SETAC
In the numerical simulations for risk assessment using MAgPIE workshop [35] proposed to generally lower the
FOCUS PRZM, edge-of-field concentrations in runoff mean CN by 3 points (or alternatively reduce the surface
are not necessarily reduced when lowering the CN, since water concentration by 50%) to account for micro-dams
both water and mass fluxes are reduced. However, PECs or other in-field bunds. The findings in this report for
in the receiving water body will be reduced by dilution potato cultivation suggest reductions of CN far beyond
due to the contribution of unmodified streamflow from the recommendation of the MAgPIE workshop of only
the upstream catchment. a 3-point reduction, which was however based on one
A considerable reduction of the concentrations reach- study only, which is not generally accessible.
ing the surface water body is achieved when the instal- Generally, the choice of CN has considerable conse-
lation of micro-dams causes a time-delay of the mass quences: for small-to-moderate rain events the predicted
entries by eliminating events closer to application and, amount of runoff is proven to be highly sensitive to
consequently, more time for leaching and degradation wrong assumptions of CN [23]. For example, an overes-
becomes available. timation of CN by 10%, for a precipitation of 5 mm and
To consider the effects of micro-dam application ade- CN = 80 (which is in the range of the standard values for
quately, the fact that only the treated agricultural field is potatoes and maize) leads to an overestimation of runoff
assumed to be equipped with micro-dams can be taken by 100%.
into account. Then other areas of the upper-stream
catchment contribute to a dilution with the inflow of the General considerations
not-mitigated amount of water in the surface water body. For the effective prevention of surface runoff and erosion
Table 3 lists the PECsw,max after globally decreas- it is advised to minimize runoff generation on-site (via
ing the value for the CN in the input files for the stand- increasing infiltration) and to maximize runoff retention
ard potato scenario by 10% (and otherwise conducting off-site (e.g. via buffer strips). On-site runoff prevention
a regular assessment) and additionally recalculating the should always be the preferred option because it reduces
Sittig et al. Environ Sci Eur (2020) 32:86 Page 9 of 10

the risk of high runoff flow velocities and the generation Abbreviations
AMC: Antecedent moisture condition; CN: Curve number; EFSA: European
of rill-flow which makes off-site runoff retention much Food Safety Agency; FOCUS: FOrum for Co-ordination of pesticide fate models
more difficult and expensive. It has been reported, that and their Use; MAgPIE: Mitigating the risks of plant protection products in the
an increase in the quantity of runoff leads to an exponen- environment; PEC: Predicted environmental concentration; PECsw: Predicted
environmental concentration for the surface water compartment; PPP: Plant
tial increase in pesticide transfer from the field, whereas protection product; PRZM: Pesticide root zone model; SETAC​: Society of Envi-
subsequent measures, i.e. buffer strips lead to a linear ronmental Toxicology and Chemistry; TOXSWA: TOXic substances in Surface
decrease in pesticide transfer when capturing runoff and Waters; USDA: United States Department of Agriculture.
erosion [35]. Acknowledgements
Besides the reduction of runoff, the practice to create The authors thank the Centre Wallon de Recherches Agronomiques (CRA-W)
micro-dams is reported to lead to an improved water and and Ulg-Gembloux Agro-Bio Tech for collaboration in the potato trials and
Centre Indépendant de Promotion Fourragère (CIPF) in the maize trials.
nutrient supply in the agricultural field and consequently
to a yield increase [22, 34, 37]. Hence, the application of Authors’ contributions
these mitigation practices shows many benefits and can SS analyzed and interpreted the field data (newly designed and from litera-
ture), calculated the curve numbers and predicted environmental concentra-
adequately be included in the risk assessment, if applied tions. RS guided the evaluation and was major contributor to the manuscript.
by the farmer. DB was responsible for the conduction and evaluation of the field trials by
the contractor companies and provided very valuable practical expertise. KH
provided the strategy to derive the predicted environmental concentrations
Conclusion according to the European risk assessment framework and was a major con-
Following the application of micro-dams, the runoff tributor to the manuscript. All authors read and approved the final manuscript.

water volume in the field trials was substantially reduced, Funding

on average by 86% for potatoes and 51% for maize. The The work on this manuscript and the evaluations shown therein were funded
reduction percentages for the plant protection products by Bayer AG. The trials of Olivier et al. [29] and CIPF [7] were conducted in
behalf of the Bayer AG.
were very similar to the reduction of the water runoff,
on average by 88% for potatoes and 46% for maize. Even Availability of data and material
more reduced was the eroded soil mass, on average by All data generated or analyzed during this study are included in this published
article, i.e. in additional file.
90% for potatoes and 71% for maize.
The reductions of the runoff water volume were trans- Ethics approval and consent to participate
lated into reduction of the runoff curve number (CN). In Not applicable.

the regulatory exposure assessment context, generally Consent for publication

realistic worst-case assumptions are made for distrib- Not applicable.
uted driving variables such as CN to safeguard a suffi-
Competing interests
ciently conservative assessment. Based on the findings in The authors declare that they have no competing interests.
this report, we propose a 10-point reduction of the CN
for the FOCUS R1–R3 potato scenarios which conserva- Author details
1
knoell Germany GmbH, Konrad‑Zuse‑Ring 25, 68163 Mannheim, Germany.
tively represents the 1­ 0th percentile of measured reduc- 2
Bayer AG Division Crop Science, Alfred‑Nobel‑Str. 50, 40789 Monheim,
tions and a 3-point reduction (mean from two studies) of Germany. 3 Bayer AG Division Crop Science, Jan‑Emiel Mommaertslaan 14,
the CN for maize to account for the mitigation of runoff 1831 Machelen, Belgium.

and erosion losses from treated fields by micro-dams. Received: 18 February 2020 Accepted: 25 May 2020
As conclusion from this study, we recommend the
adoption of CN reductions due to micro-dams in pota-
toes and other comparable measures in maize into the
regulatory exposure assessment. We propose an absolute References
reduction of the CN in risk assessment in accordance 1. Adriaanse PI, Van Leerdam RC, Boesten JJTI (2017) The effect of the runoff
size on the pesticide concentration in runoff water and in FOCUS streams
with the findings in this study. simulated by PRZM and TOXSWA. Sci Total Environ 584–585:268–281
In risk assessment, the installation of micro-dams can 2. Areas (2012) Association Régionale pour l’Etude et l’Amélioration des
be considered in a higher-tier calculation which will lead Sols: Expérimentations sur les pratiques culturales 2001–2010: synthèse
des résultats de ruissellement et d’érosion. http://www.areas​asso.fr/
to reduced predicted environmental concentrations. image​s/resul​tats%20ess​ais%20sim​ul/broch​ure_10ans​_essai​s_PC_ruiss​
ellem​ent.pdf. Accessed 15 Nov 2017
Supplementary information 3. Areas (2005) Réduction du ruissellement de printemps sur pomme de
terre—effet des micro-barrages. http://www.areas​-asso.fr/image​s/resul​
Supplementary information accompanies this paper at https​://doi.
tats%20ess​ais%20sim​ul/AREAS​-2005-11.pdf. Accessed 07 Feb 2020
org/10.1186/s1230​2-020-00362​-1.
4. Areas (2007) Pomme de terre. Limiter la formation du ruissellement en sol
limoneux. Simulation de pluie (30 mm) juin 2007 Saint-Jouin-Bruneval.
Additional file 1. Additional tables. http://www.areas​-asso.fr/wp-conte​nt/uploa​ds/2016/11/8-pomme​s-de-
terre​.pdf. Accessed 07 Feb 2020
Sittig et al. Environ Sci Eur (2020) 32:86 Page 10 of 10

5. Beltman WHJ, Horst MMS ter, Adriaanse PI, Jong A de, Deneer JW (2014) 29. Olivier C, Goffart JP, Baets D, Xanthoulis D, Fonder N, Lognay G, Bar-
FOCUS_TOXSWA manual 4.4.2: Users Guide version 4. WOt-technical thélemy JP, Lebrun P (2014) Use of micro-dams in potato furrows to
report 14. Wettelijke Onderzoekstaken Natuur & Milieu, Wageningen reduce erosion and runoff and minimise surface water contamination
6. Aurbacher J, Krimly T, Billen N (2010) Querdammhäufler verringert Boden- through pesticides. Comm Appl Biol Sci 79(2):1–10
und Wasserverluste—ein Gerät zur Erosionsverringerung im Kartoffelbau. 30. PPDB (2016a): Pesticide Properties DataBase. University of Hertfordshire:
Landwirtschaftliches Wochenblatt/Hessenbauer/Ausgabe Süd https​://sitem​.herts​.ac.uk/aeru/ppdb/en/Repor​ts/331.htm. Accessed 04
7. CIPF (2013) Erosiebestrijding in maïs. Internal communication Bayer Crop- Dec 2017
Science, cooperation with Centre indépendant de promotion fourragère 31. PPDB (2016b): Pesticide Properties DataBase. University of Hertfordshire:
(Louvain-la-neuve, Belgium) 04.03.2015. Personal communication, not https​://sitem​.herts​.ac.uk/aeru/ppdb/en/Repor​ts/424.htm. Accessed 04
published Dec 2017
8. ECPA and Tessella Support Services plc. (2008) Surface Water Assessment 32. PPDB (2017): Pesticide Properties DataBase. Data interpretation; Univer-
eNabler (SWAN) v 1.1.4 sity of Hertfordshire: http://sitem​.herts​.ac.uk/aeru/ppdb/en/docs/5_1.pdf.
9. EFSA (2006) Conclusion on the peer review of metribuzin. EFSA Sci Rep Accessed 04 Dec 2017
88:1–74 33. Reichenberger S, Bach M, Skitschak A, Frede HG (2007) Mitigation strate-
10. EFSA (2008) Conclusion on the peer review of aclonifen. EFSA Sci Rep gies to reduce pesticide inputs into ground- and surface water and their
149:1–80 effectiveness; A review. Sci Total Environ 384:1–35
11. EFSA (2008) Conclusion regarding the peer review of the pesticide risk 34. Salem HM, Valero C, Muñoz MÁ, Gil-Rodríguez M (2015) Effect of inte-
assessment of the active substance fluazinam. EFSA Sci Rep 137:1–82 grated reservoir tillage for in situ rainwater harvesting and other tillage
12. EFSA (2011) Conclusion on the peer review of the pesticide risk assess- practices on soil physical properties. Soil Till Res 151:50–60
ment of the active substance terbuthylazine. EFSA J 9(1):1969 35. SETAC (2017) Alix A, Brown C, Capri E, Goerlitz G, Golla B, Knauer K, Laabs
13. EFSA (2013) Conclusion on the peer review of the pesticide risk assess- V, Mackay N, Marchis A, Alonso Prados E, Reinert W, Streloke M, Poulsen V
ment of the active substance tembotrione. EFSA J 11(3):3131 (eds.). MAgPIE—mitigating the risks of plant protection products in the
14. EFSA (2018) Scientific report of EFSA on the “repair action” of the FOCUS environment. From the two-part SETAC Workshop Mitigating the Risk of
surface water scenarios. EFSA J (under development). https​://www.efsa. Plant Production Products in the Environment. https​://www.setac​.org/
europ​a.eu/sites​/defau​lt/files​/consu​ltati​on/consu​ltati​on/Draft​_EFSA_scien​ store​/ViewP​roduc​t.aspx?id=90064​89. Accessed 09 Sept 2017
tific​_repor​t_FOCUS​.pdf. Accessed 29 Apr 2020 36. Suárez LA (2006) PRZM-3, A model for predicting pesticide and nitrogen
15. Elhakeem M, Papanicolaou T (2009) Estimation of the runoff curve num- fate in the crop root and unsaturated soil zones: user’s manual for release
ber via direct rainfall simulator measurements in the State of Iowa, USA. 3.12.2.
Water Resour Manag 23:2455–2473 37. Sui Y, Ou Y, Yan B, Xu X, Rousseau AN, Zhang Y (2016) Assessment of
16. EPA (2016) Pesticide in water calculator user manual for versions 1.50 and micro-basin tillage as a soil and water conservation practice in the black
1.5. Dirk F. Young, U.S. Environmental Protection Agency, Washington, DC soil region of Northeast China. PLoS ONE 11(3):e0152313
17. ESA (1973) Endangered Species Act of 1973. Codified 16 U.S.C. § 1531 et 38. Tang R, Zhu B, Katou H (2012) A review of rapid transport of pesticides
seq from sloping farmland to surface waters: processes and mitigation strate-
18. FOCUS (2001) FOCUS surface water scenarios in the EU evaluation pro- gies. J Environ Sci 24(3):351–361
cess under 91/414/EEC. Report of the FOCUS working group on surface 39. Truman C, Nuti R (2010) Furrow diking in conservation tillage. Agric Water
water scenarios, EC Document Reference SANCO/4802/2001-rev.2 Manage 97:835–840
19. FOCUS (2015) Generic guidance for FOCUS surface water Scenarios, ver- 40. UBA (2018) Federal German Environmental Protection Agency—Umwelt-
sion 1.4, May 2015. EU Document bundesamt. https​://www.umwel​tbund​esamt​.de/en/topic​s/soil-agric​ultur​
20. EFSA (2016) Conclusion on the peer review of the pesticide risk assess- e/land-a-preci​ous-resou​rce/erosi​on#textp​art-9. Accessed 11 Jan 2018
ment of the active substance linuron. EFSA J 14(7):4518 41. UCL (2012) Lutte contre le ruissselement et l‘érosion en culture de maïs
21. Goffart JP, Poulet V, Olivier C (2013) Study of the effects of tied ridges on sur terres en pente. Cooperation of UCL, LSM, CIPF, SPW and DGO3.
water runoff and its consequences for potato crop. Potato Agrophysiol Personal communication, not published
15:55–61 42. FIFRA (1980) Federal Insecticide, Fungicide, and Rodenticide Act. Codified
22. Harris BL, Krishna JH (1989) Furrow diking to conserve moisture. J Soil 7 U.S.C. § 136 et seq
Water Conserv 44(4):271–273 43. USDA (2004) United States Department of Agriculture, Natural Resources
23. Hawkins R (1975) The importance of accurate curve numbers in the Conservation Service; Estimation of direct runoff from storm rainfall. Part
estimation of storm runoff. Water Resour Bull 11(5):887–891 630 Hydrology—National engineering handbook, chapter 10
24. Hawkins R (1993) Asymptotic determination of runoff curve numbers 44. Young DF, Fry FF (2014) PRZM5. A model for predicting pesticide in
from data. J Irrig Drain E 119(2):334–345 runoff, erosion, and leachate: user manual. USEPA/OPP 734F14002. Office
25. ter Horst MMS, Adriaanse PI, Boesten JJTI (2009) Interpretation of the of Pesticide Programs U.S. Environmental Protection Agency
mitigation of runoff in the FOCUS surface water scenarios as described in
the FOCUS L&M report. Alterra, Alterra-Report, Wageningen, p 1794
26. Jones OR, Baumhardt RL (2003) Furrow dikes. In: Trimble S (ed) Encyclo- Publisher’s Note
pedia of water science. CRC Press, Boca Raton Springer Nature remains neutral with regard to jurisdictional claims in pub-
27. Mockus V (1949) Estimation of total (and peak rates of ) surface runoff for lished maps and institutional affiliations.
individual storms. Exhibit A of Appendix B, interim survey report, Grand
(Neosho) River Watershed, USDA Soil Conservation Service
28. Oliveira PTS, Nearing MA, Hawkins RH, Stone JJ, Rodrigues DBB, Pana-
chuki E, Wendland E (2016) Curve number estimation from Brazilian
Cerrado rainfall and runoff data. J Soil Water Conserv 71(5):420–429