Application of the Zürich-methodology for risk assessment concerning the inspection of pesticide application equipment in use according to article 8 (3) of Directive 2009/128/EC at the example of Germany

Anwendung der Zürich-Methode zur Risikobewertung hinsichtlich der Kontrollpflicht von in Gebrauch befindlichen Pflanzenschutzgeräten nach Artikel 8 (3) der Direktive 2009/128/EC am Beispiel von Deutschland

Fig. 1. Risk matrix according to Nohl and Thiemecke (1988).

The advantage of the Zürich-methodology concerning the risk assessment of PAE is that the risk assessment can be reduced to those technical parameters which are the focus of the inspection of PAE in use. This means that parameters being fraught with risks could be eliminated by inspection for that pesticide application equipment being obliged to inspection.

In order to use this approach for the risk assessment of PAE in use the extent of damage was discharged by a qualitative analyses of equipment components being part of the inspection (acc. EN 13790) and their impact in case of technical disorder on human health and the environment. Therefore, each category of PAE has to be judged about the impact of their different components by using qualitative measures. A first attempt was done by Ganzelmeier (2012) and presented during SPISE IV Workshop (Tab. 1). He used the following qualitative measures and quantified them by using a point system: ++ = 20 points, + = 15 points, 0 = 10 points, – = 5 points and -- = 0 points. Afterwards, he formed the sum of each category and ordered the groups by size. In this way the axis describing the probability of occurrence was discharged within the risk matrix (cf., Fig. 1).

The probability of occurrence is normally figured out by taking the number of incidents of each group of PAE into account. Since there are no such statistics available on a national level of all member states (SPISE, 2014) this lack of information has to be solved by taking the number of different PAE used in practice into account since these numbers should be proportional to the frequency of incidents. The numbers of PAE in use (Fig. 2) were defined by Ganzelmeier (2012) into six different levels (level 1 = 1,000 PAE, level 2 = 2,000 PAE, level 3 = 5,000 PAE, level 4 = 10,000 PAE, level 5 = 20,000 PAE, level 6 = 50,000 PAE). An unanswered question during SPISE IV was how to define the risk tolerance line. Therefore, Wegener (2015) expanded the matrix by multiplying each category of PAE (1–8) with each probability of occurrence level (1–6). Afterwards, in order to have an equal treatment of all PAE considered, the legal exemptions mentioned within article 8 (3) of the SUD concerning the necessity of inspections were drawn into the matrix (Fig. 2, green and red boxes). At the end the risk tolerance was generally defined by taking into account the highest acceptable risk which was calculated within the green box (Fig. 2, purple box).

Fig. 2. Risk tolerance based on Sustainable Use Directive (Wegener, 2015).

Fig. 3 shows the general risk matrix established based on the aforementioned considerations. Within the red zone one can find all cases within the matrix being either above the risk tolerance level of 12 or being within the group of PAE where inspection is mandatory due to the legal statements of the SUD. The green zone represents those cases which may be exempted from inspection if operators are trained, whereas the yellow zone represents those cases where different time tables and inspection intervals can be applied.

Fig. 3. Risk matrix for the need of inspection of different categories of PAE (Wegener, 2015).

Within a discussion during SPISE Technical Working Group meeting in Braunschweig (Germany) in March 2015 some more suggestions were made in order to improve the methodology and to make it universal for all member states. It was agreed

Concerning the first point the methodological improvement can be made by adopting the sum of each category of PAE from Tab. 1 into the first line of the matrix, representing the first level of probability of occurrence. Then, these figures can be multiplied within the matrix by the number of levels considered in total.

The second and third point can be solved by relating the country-specific number of PAE of each group to the total area under cultivation for each country. This approach considers differences between member states concerning their individual structures of the agricultural sector within the EU. It follows the hypothesis that a state with a high amount of small scale farms having more PAE per e.g. one million of hectares of agricultural used land has a higher risk for incidents compared to a state with a lower amount of PAE per one million of hectares. The definite scale for the probability of occurrence levels and its linear division can only be determined finally when all of the statistical data-sets for EU 27 are collected.

Fig. 4. Risk matrix for the need of inspection of different categories of PAE at the example of Germany.

Based on the described procedure and figures the presented risk assessment for the case of Germany would allow different time tables and inspection intervals for the PAE categories “Additional/low scale use”, “not used for spraying” and “hand-operated” according to article 8 (3) of the SUD.

The presented methodology takes those technical parameters into account, which are components being checked by inspection (e.g. Wegener, 2015). This approach limits the area of consideration concerning the risks to that one which can be suppressed by a technical inspection. Furthermore, it puts the same criteria to all different categories of PAE so that all of them are treated in an equal way concerning the question if exemptions from inspection are possible or not.

The risk assessment presented here is a qualitative one. It is lacking in accuracy at different steps of the approach due to subjective measures or due to a lack of information which are not existent and could maybe only be roughly estimated. But, the example of Germany demonstrates that the presented Zürich-methodology is applicable if Member States can provide (estimated) national figures about the different groups of PAE in use. The question is, if a more sophisticated approach would really come up with another ranking of the PAE categories within the extent of damage?

What is needed in any case is that an expert panel confirms the qualitative judgement made during the process of the determination of the extent of the damage and that they give a written statement about the specific evaluation of each point which clearly informs about why and how the estimations were made. Unfortunately, there are no notations available justifying the assessment made by Ganzelmeier (e.g. Tab. 1). Furthermore, there are two options for the classification of the figures used for the probability of occurrence-levels which have to be discussed by the experts. One possibility is to have a general risk tolerance line being valid for all European member states. This can be defined not until all numbers of different categories of PAE for all member states are collected. The country with the highest amount of either handheld or knapsack sprayers per one million of hectares would define the total number of probability of occurrence levels within the matrix being necessary and also define the general European risk tolerance line. Another option would be to have country specific risk tolerance lines being discharged at national levels as presented in this study.