Introduction

Pest regimes in modern agriculture and horticulture are in a process of constant change (Kolar and Lodge 2001). The occurrence of new pests is facilitated by changes in climate (Hellmann et al. 2008), or by changes in production and cultivation methods, which especially allow airborne pests to spread swiftly and establish in new regions when favourable conditions are met. Another driver in this process is local and international trade and travel, which spreads less mobile and non-airborne pests into new regions (Kolar and Lodge 2001). Greenhouse production of tomato in Germany and its pest regime is not exempt from this. Currently the greenhouse cultivation area of tomato in Germany measures 385 ha (Destatis 2020). Among the pests that occur in tomato production in Germany, the tomato russet mite Aculops lycopersici (Tryon) is currently spreading and establishing on tomato production sites across the country.

The tomato russet mite Aculops lycopersici (Acari: Eriophyoidea), an eriophyoid mite, is considered a pest of several Solanaceae crops (Perring and Farrar 1986). A. lycopersici is currently found throughout the world in both tropical and temperate regions. Before 1999 A. lycopersici was rarely reported as a pest on tomato in Germany (Merz 2020). In recent years the economic impact of eriophyoid mites such as Aculus scchlechtendali, Calepitrimerus vitis and A. lycopersici has increased worldwide (Duso et al. 2010), as well as in Germany in the case of A. lycopersici (Merz 2020). A. lycopersici causes bronzing and a russeted appearance of leaves and stem as it feeds on surface cells, leading to the death of leaves and even complete plants since they no longer are able to photosynthesise (Royalty and Perring 1988). In contrast to several other eriophyoid mites, A. lycopersici has a superficial lifestyle and does not induce or inhabit galls on plant tissue (Van Leeuwen et al. 2010).

Several factors make A. lycopersici a problematic pest. It has a high reproduction rate, and at less than 0.2 mm in length, is very small in size (Haque and Kawai 2003). There are no plant protection products authorised for use specifically against A. lycopersici in tomato cultivation in Germany (BVL 2023). Authorised products against mites in general have been shown to be limited in efficacy, or are expected to be limited in efficacy due to them being contact acaricides (Vervaet et al. 2021). Although a number of studies have shown that certain predatory mites, for example Amblyoseius fallacis and A. swirskii (Brodeur et al. 1997; Park et al. 2010, 2011) feed on A. lycopersici, the practical implementation of biological control is limited (van Houten et al. 2013). More recent studies show very promising results for different predatory mites under semi-practical conditions but have yet to be confirmed with trials in practical cultivation (Pijnakker et al. 2022a, b; Vervaet et al. 2022; Castañé et al. 2022). In practice, recognition of A. lycopersici infestation coincides with symptom recognition on plants. Early symptoms such as light chlorosis on leaves, or light grey and brown shades on the stem are easily overlooked. Later, more obvious symptoms may be misdiagnosed. For instance stem and leaf browning might be mistakenly attributed to the fungus Phytophthora infestans (Crüger et al. 2002) in practice and by people unexperienced with the pest A. lycopersici. As yet there are no efficient A. lycopersici monitoring methods available for practical tomato cultivation (Pfaff et al. 2020).

Whenever new or invasive pests occur and negatively impact the production of food, in this case tomato production in greenhouses, information on the frequency of occurrence and case specific data from farms are essential for estimating economic damage potential. When farms with A. lycopersici presence are identified, a closer look at cultivation techniques and factors such as substrate, crop rotation, climatic conditions, applied pesticides or introduced beneficial arthropods might reveal intervention strategies that can be exploited to reduce the impact of A. lycopersici. The experience of farmers may help to better understand the pest dynamic of A. lycopersici, and could aid in the development of efficient countermeasures. To obtain this information, the presented survey of farms with tomato production was carried out.

Materials and methods

Survey details

A survey was developed and integrated into the professional survey platform https://umfrageonline.com and the domain “tomatenschaedlinge.org” was created to redirect participants to the survey on umfrageonline. In December 2018, a link to the survey was forwarded to the official plant-protection consultation services of the federal states, to private plant protection consultants and to grower organisations, with the request to disseminate the survey to tomato growers. In this way, an undocumented number of tomato-producing farms were contacted via E-Mail and invited to fill out the questionnaire. Aside from being offered the possibility to receive information on the results of the survey, no form of payment was offered to growers for participating in the survey. Due to a rather small number of participants by the end of February 2019, the decision was made to switch to phone interviews. The contact addresses of growers were obtained using google search engine. The following search terms were used: “Tomatenproduktion in [federal state]” or “Tomatenanbau in [federal state]” for each of the 16 federal states of Germany to achieve a spatial distribution of participants across Germany.

Criteria for integrating farms in the analysis:

  • Production of tomato located in Germany

  • At least 50 m2 tomato cultivation area

  • Production for commercial purposes

  • The questionnaire was completed (all mandatory questions answered)

The questionnaire consisted of five parts:

  • General information on participant and farm

  • Crops, cultivars, cultivation methods

  • Confirmed pests and diseases of tomato on the surveyed farm

  • Impact of and measures against A. lycopersici

  • Agreement on use of data for scientific purposes only

Determination of the sample size

Prior to the study, it was considered that with the workforce conducting the study, it would be realistic to reach up to 100 valid participants. This number also took into account the fact that many factors could potentially limit the number of growers who might take part, for instance the fact that it may be difficult or time-consuming to identify and locate potential farms that could participate, that without monetary incentives there may be a limited motivation for the invited growers to participate, and finally it was recognised that it may be difficult to determine the representativeness of the recruited participants beyond a rough estimation. In the light of this, additional questions were included in the questionnaire with the ability produce insight into farmers perception and strategies for the pest of interest that go beyond a quantitative analysis approach.

Statistical analysis

Statistical analysis was conducted where feasible; Fisher exact tests were used for categorical data, a nonparametric Wilcoxon rank sum test to test for dependence between tomato production area and A. lycopersici presence, and a linear model was used to check whether there was a correlation between production area and length of break between tomato seasons. An exact binomial test was used to test whether an initial infestation increased the likelihood that there would be continuous A. lycopersici presence in subsequent consecutive years (Clopper and Pearson 1934). All functions used in these analyses are part of the statistical software R (R Core Team 2021; version 4.1.0). For correction of multiple testing in the Fisher exact tests the “fdr”-method (Benjamini and Hochberg 1995) was applied.

Results

Acquisition of participants and metadata on participating farms

A total of 83 tomato producing farms had responded to the invitation to participate by the 5th of August 2019. Of the 29 online respondents, 17 were excluded for not having completed the survey. Of the 54 farms contacted via phone, 38 farms agreed to participate and all of them completed the survey. 12 valid online respondents plus 38 valid phone respondents resulted in a total of 50 valid participants becoming subject to the following analyses, half of what initially was aimed for. For the growers contacted via E-Mail, the calculation of a response rate is not possible as the total number of recipients contacted is not known. For the growers contacted via phone the response rate was 70.4%. Germany is divided into 10 regions by the first digit of the postal code and into 99 regions by the first two digits of the postal code. The distribution of participating farms separated by first digit is shown in Table 1. In Fig. 1, separation is based on the combination of first and second digit of the postal code. On the first digit-level it was possible to cover all 10 areas with participants.

Table 1 Postal code areas with quantity of participants
Full size table
Fig. 1
figure 1

Map of Germany, divided into 99 regions by the first two digits of the postal code. If A. lycopersici was found at least in one of the participating farms in a region, the region is marked grey. If it was found in none of the participating farms in one region, it is marked white. Black regions indicate that there were no participants from this region (note this does not mean, that there are no tomato-producing farms in this particular region). White only means, that the respective pest was not found in the participating farms in the particular region, not that it does not occur in any of this region’s farms

Full size image

The 50 farms that participated in the survey represent 3.5% of the total number of 1448 farms with tomato production in Germany in 2019 (Destatis 2020). Together the participating farms account for 131.8 ha tomato production area. The total area used for tomato production in Germany in 2019 was 385.63 ha, thus the surveyed farms account for 34.2% of this total area (Destatis 2020). The mean tomato production area of the 50 participating farms was 2.64 ha. The median production area was 0.2250 ha.

Of the 50 participating farms, 27 (54%) were producing integrated (i.e. committed to the guidelines of integrated pest management) with an average cultivation area of 47.457 m2 and 23 farms (46%) were producing organic (at least following Regulation (EU) 848/2018), with an average cultivation area of 1593 m2. The sizes of the participating farms is visualised in Fig. 2.

Fig. 2
figure 2

Sizes of the participating farms displayed in m2 on the y-axis divided by production type displayed on the x-axis. To improve visualisation, the y-axis has been log-transformed

Full size image

Of the 50 participants, 29 participants (58%) stated that they had attended an apprenticeship in horticulture or agriculture. Twenty participants (40%) stated that they had completed a university degree in horticulture or agriculture. Four participants (8%) stated that they had both an apprenticeship and a university degree and five participants (10%) stated that they had not finished a degree in horticulture or agriculture.

Of the 50 participants 48 (96%), stated the sources that they use to obtain information about plant protection. Experience exchange with other growers, and specialist literature were each stated 43 times (89.6%), followed by websites on plant protection with 41 mentions (85.4%), and plant protection courses which received 35 mentions (72.9%). Eight participants (16.7%) stated that they rely on decision-support-systems (DSS) for plant protection. Thirty-one participants (64.6%) indicated that they were interested in the utilisation of DSS, and 17 (35.4%) stated that they were not interested in using DSS.

Employees on participating farms

It was assumed, that the employee background, familiarity and experience with and in tomato farming might have an influence on detection and / or treatment success against A. lycopersici. 28 participants (46%) stated that the majority of employees doing cultivation work are permanently employed. Of these 28 participants, 21 participants (75% of 28) stated that the employees are regularly offered training sessions on the topic of plant protection. 21 participants (42%) stated that the majority of employees doing cultivation work are seasonal employees that had already worked on the farm previously. Of these 21 participants, 14 participants (66.7% of 21) indicated that regular training sessions on the topic of plant protection are offered to the employees. Only one participant (2%) stated that most of the employees doing cultivation work were seasonal employees that had not worked on the farm previously, and that no training sessions on the topic of plant protection were offered to employees. Fisher exact tests were used to test whether there was a correlation between A. lycopersici incidence and employee demographics. No significant correlation was found between the two.

Relative importance of pests and diseases on participating farms

Participants were asked to name the most important pests (Table 2) and diseases (Table 3), in both cases results are displayed separately for farms with A. lycopersici presence and farms without A. lycopersici presence. Looking at most important pests A. lycopersici reached the highest relative importance in farms with A. lycopersici presence as well as in the overall ranking (farms with and without A. lycopersici presence combined). Tuta absoluta was considered important on 11 of 32 farms (34.4%) with A. lycopersici presence, making T. absoluta the third most important pest in this particular group of farms. On the contrary, none of the nine farms without A. lycopersici considered T. absoluta important. For the group of farms without A. lycopersici presence, seven from nine farms (77.8%) consider whiteflies important, making it the most important pest for this group of farms with a relative importance of 50%.

Table 2 Farmers were asked to name the three most important pests in their tomato cultivation and rank them in decreasing order
Full size table
Table 3 Farms were asked to name the three most important diseases in their tomato cultivation and rank them in decreasing order
Full size table

With regard to diseases, Botrytis cinerea was the most important disease among farms with A. lycopersici presence mentioned by 13 of these farms (40.6%), whereas for farms without A. lycopersici presence Phytophthora infestans was the most important disease (Table 3). P. infestans was the only disease according to a Fisher exact test (followed by a correction for multiple testing) that showed a significant difference in importance between the two groups of farms, farms with, and farms without A. lycopersici presence.

Initial occurrence and persistence of A. lycopersici on participating farms

Thirty-three farms reported that A. lycopersici was present at some time on their farm in the five years preceeding 2019. 32 of those farms were able to report the year that A. lycopersici was first noted. Of those 32 farms, 26 farms reported that the first occurrence was between 2014 and 2018, with a peak of nine reports of first occurrences in the year 2018 (Fig. 3).

Fig. 3
figure 3

First year of A. lycopersici occurrence on the farms that reported A. lycopersici presence. The y-axis shows the count of the farms and the x-axis shows the year

Full size image

On 24 of the 33 farms (72.7%) A. lycopersici was present in every year following the year of the first occurrence i.e. only nine farms (27.3%) reported A. lycopersici-free seasons after the year of first occurrence. An exact binomial test revealed that an initial infestation increased the chance for continuous A. lycopersici presence in all consecutive years (p = 0.01). The production area of these nine farms ranged from 500 to 81,000 m2. The mean production area was 16,855 m2, and the median production area was 1200 m2. Seven of these nine farms (77.8%) were heated during the colder months, and the cultivation break between tomato sets on these farms ranged between one and 29 months.

Yield impact of A. lycopersici

Of the 33 farms affected by A. lycopersici, 12 farms (36.4%) reported a negative impact on yield despite plant protection measures. Nine of those farms (75%) reported a specific yield loss. The yield loss reported by these farms ranged between 0.5 and 15%, and on average amounted to 5.89%. 21 farms (63.6%) reported no negative impact on yield considering plant protection measures taken.

Farm and cultivation parameters and possible links to A. lycopersici presence

In Fig. 4, a large tomato cultivation area and a short cultivation break seem to co-occur. A nonparametric Wilcoxon rank sum test revealed that there was a dependency between A. lycopersici presence and cultivation area (p < 0.00). A linear model showed that cultivation break was a significant predictor for cultivation area (p < 0.00) with a coefficient of −0.1068 at an adjusted R2 of 0.2799. A. lycopersici did not exclusively occur on farms with short cultivation breaks, but all 20 of the farms (40%), with a cultivation break between tomato sets of less than 3 months, with one exception, reported A. lycopersici presence. Similarly, all 23 farms (46%) with a cultivation area of 4800 m2 or larger, with one exception, reported A. lycopersici presence. The latter is also reflected by the total combined cultivation area of the 33 farms with A. lycopersici presence of 129.09 ha which account for 33.3% of the German tomato production area. The combined cultivation area of the 17 participating farms without A. lycopersici presence sums up to 2.71 ha, representing approximately 0.7% of the German tomato cultivation area. The mean cultivation area of the participating farms with A. lycopersici presence was 3.911 ha and of those without was 0.159 ha.

Fig. 4
figure 4

A. lycopersici presence on the different farms. Each symbol represents one farm. The tomato production area is displayed on the y-axis and the cultivation break between tomato sets on the x-axis. To aid visualisation of small values, the y-axis has been log-transformed. The numbers next to the symbols indicate the IDs of the specific farms. 45 of 50 farms are displayed; five farms with production areas between 250 and 5000 m2 did not state the length of their cultivation break

Full size image

Table 4 shows that the farm production type–whether the farm was integrated or organic—was a statistically significant predictor for A. lycopersici presence according to a Fisher exact test for independence (p < 0.00).

Table 4 A. lycopersici presence/absence on integrated and organic farms
Full size table

Fifteen of the 29 farms (51.7%) growing tomato in natural soils, and 19 of the 21 farms (90.5%) not growing in natural soils reported presence of A. lycopersici in their production systems (Table 5). The mixture of natural soil, compost and coir substrate was considered a natural soil in this analysis. According to a Fisher exact test, A. lycopersici presence depended on whether or not the tomatoes were grown in natural soil (p < 0.00).

Table 5 Percentage of substrate types shown separately for farms with (n = 33) and farms without (n = 17) A. lycopersici presence in addition to the total frequency of A. lycopersici presence (n = 50)
Full size table

Plant residues are removed at different time points and intervals in the participating farms (Table 6). A Fisher exact test did not reveal any significant difference regarding A. lycopersici occurrence both before and after correction for multiple testing.

Table 6 Frequency of removal of plant residues for farms with (n = 32) and without (n = 17) A. lycopersici presence and the combined total (n = 49)
Full size table

There was no significant effect on the presence of A. lycopersici depending on whether farms cultivate in a crop rotation or not (Table 7).

Table 7 Presence and absence of A. lycopersici on farms that cultivate tomato in a crop rotation and farms that do not cultivate in a crop rotation
Full size table

Of the 24 farms that grow tomato in a crop rotation with other crops, 21 farms provided information on the rotation crops (Table 8). None of the mentioned rotation crops acted as significant predictors for the presence of A. lycopersici.

Table 8 Frequency of crops grown in a rotation with tomato, shown separately farms where A. lycopersici was present (12 of 13 farms reported their rotation crops) and farms where A. lycopersici was not present (9 of 11 farms reported their rotation crops)
Full size table

Participants were asked whether they had reared fresh plants in the last 5 years on the farm, or whether they purchased them externally. Of the 33 farms with A. lycopersici presence, 23 farms (69.7%) received plants from external nurseries, six (18.2%) had reared fresh plants on the farm, and four farms (12.1%) had both on-farm reared and purchased plants in the last five years. Of the 17 farms without A. lycopersici presence seven farms (41.2%) had received plants from external nurseries, 6 (35.3%) had reared fresh plants on-farm, and 4 farms (23.5%) had both on-farm reared and purchased plants in the last five years. A Fisher exact test revealed that there was no difference between farms in terms of the presence/absence of A. lycopersici, depending on whether the farm had received plants from a nursery or had reared fresh plants on-farm in the last five years (p = 0.14).

Heating during the colder months was a statistically significant predictor of A. lycopersici presence according to a Fisher exact test for independence (p < 0.00; Table 9).

Table 9 A. lycopersici presence/absence on farms that do and farms that do not heat during the colder months
Full size table

Most of the farms that did not heat during the colder months reported first symptoms in August and September, around one month later compared to the farms that heat. Farms that heat reported the first symptoms of the season in July and August (Fig. 5).

Fig. 5
figure 5

Month when first A. lycopersici symptoms of the year were usually observed on farms with A. lycopersici presence (participants were able to name multiple months). The information is provided for both farms where the greenhouse was heated during the colder months and farms where the greenhouse was not heated

Full size image

Utilised countermeasures against A. lycopersici infestation

Farms with A. lycopersici presence were asked to select known countermeasures against A. lycopersici from a list, indicating those they implement in their own control strategy against the pest. Each countermeasure was used at 14 (42.4%) to 18 (54.5%) of the 33 farms (Table 10). No significant differences were found between farms that reported that A. lycopersici caused significant yield loss, and those that did not report yield loss.

Table 10 The countermeasures that farms with A. lycopersici presence have taken to combat the pest
Full size table

Participants had the opportunity to describe specific countermeasures or strategies they have used against A. lycopersici in more detail. Ten participating farms (20%) supplied a free-text answer. Answers usually consisted of specific combinations of countermeasures that had been explicitly asked about. One participant mentioned an additional (not previously asked about) measure—“herbal mixtures”. Another farm mentioned, that the first and strongest symptoms usually occur in areas that are most exposed to sunlight. Detailed answers can be viewed in Table S2 in the supplementary information.

Of the 15 farms that have applied acaricides, nine farms supplied specific detail about which acaricides they applied. Abamectin-based products were used by all nine farms that provided this detailed information. Spirodiclofen, potash soap and Azadirachtin were each applied on one farm.

Of the nine farms that applied predatory mites specifically against A. lycopersici, A. swirskii was chosen most often; five farms applied this mite. Phytoseilulus persimilis, Amblyoseius cucumeris, Amblyoseius barkeri and Amblyoseius californicus were each chosen by one farm.

Besides predatory mites specifically introduced against A. lycopersici, a total of 14 different Arthropods, and one nematode species were introduced on participating farms. Data on this are provided in Table S1 in the supplementary information for farms with and without A. lycopersici presence.

The intensification factors that A. lycopersici presence is shown to be dependent on are all more prevalent in integrated rather than organic farms (Table 11).

Table 11 Factors that A. lycopersici occurrence was statistically dependent on, shown separately for integrated and organic farms
Full size table

Discussion

Survey metadata

The survey that was carried out focussed on pests and diseases in tomato production. At a later stage during questioning the questionnaire revealed that the main focus of the survey was the pests A. lycopersici and Tuta absoluta. In this study, pest-specific data were only presented for A. lycopersici.

Fifty participants took part in the survey between January and August 2019. With this limited sample size and due to a low response rate for some of the non-mandatory survey questions there was insufficient statistical power to run quantitative analyses—to model dependencies and derive reliable conclusions about several of the potential explanatory factors for A. lycopersici incidence and damage levels. This study does, however, provide several interesting findings from a qualitative angle.

The fact that only 3.5% of the 1448 tomato producers in Germany participated and at the same time 34.2% of the 385.63 ha German tomato production area is covered by those 3.5% reveals, that mainly farms with considerably larger tomato production area than the average farm participated in the survey. This imbalance needs to be taken into account when interpreting the results. 38 of the 50 participants (76%) included in the analysis were contacted via phone. The phone numbers of these tomato producing farms were obtained via google search engine. This means there is a selection bias for farms that maintain a web page with unknown effect on the results. Those factors in consequence limit the representativeness of the results. Having acknowledged this limitation, the displayed data is useful giving as it does, a unique insight into German commercial tomato production facing A. lycopersici infestation which, to date, is not available in a structured and published format.

It was possible to avoid regional clusters of participants in this survey. The 50 participants were spread over all ten postal code areas (first digit of the German postal code), although, display of participants by the two-digit postal code reveals several areas where there were no participating farms. These areas might not have any professional protected tomato cultivation, or it may simply be that no farms in this area were contacted. Despite the fact that there are some areas where no farms participated, the map visualises that A. lycopersici occurs throughout Germany from south to north and west to east–its presence is not restricted to one specific region (Fig. 1).

Relevance of A. lycopersici as a pest in German tomato cultivation

This study focusses on the farmers perception. For all results displayed it must be considered, that they are based on the assumption that the participants answers are valid. The number of participants that experienced first A. lycopersici occurrence on their farms culminates towards the end of the surveyed window from the year 2004 to 2018 (Fig. 3). It is possible that A. lycopersici may have gone unnoticed for some time due to there being less experience with this pest in practical cultivation in the past when symptoms appeared late and were of minor nature. However, in the more devastating A. lycopersici incidences in which sometimes whole crops have been lost, it is unlikely that farmers would not have identified that A. lycopersici was the cause. Twenty-four of the 33 farms (72.7%) that experienced an A. lycopersici occurrence, experienced A. lycopersici presence and damage in every year following the year of first occurrence; initial infestation significantly increased the likelihood that A. lycopersici would continuously present in consecutive years. This shows, that within-farm eradication attempts are either not conducted at all, are conducted in an inefficient way, or they are simply not possible with the given circumstances in the farms that face continuous A. lycopersici presence. These results indicate that outbreaks in subsequent years could be a consequence of an initial infestation rather than a result of new independent migration events. However, a continued infestation through the use of fresh plants from external nurseries or other entrances with plant material or tools and packaging cannot be excluded as a potential yearly infestation source.

A. lycopersici reached the highest relative importance in the group of farms with A. lycopersici presence and also in the overall ranking of relative pest importance amongst all participating farms (Table 2). A. lycopersici was closely followed by whiteflies and by Tuta absoluta. “Importance” here refers to the plant protection effort exerted on the specific farm against the specifically named pest. The high plant protection effort demanded by A. lycopersici indicates a significant economic relevance for tomato producers and at the same time that there likely is room for improvement of A. lycopersici management in practice. In most cases the high plant protection effort seems to have prevented the farms from experiencing significant yield losses caused by A. lycopersici. Only 12 of 33 farms (36.4%) reported that yields were negatively affected by A. lycopersici. Unfortunately give that only nine participants reported a specific yield loss it is not possible to reliably quantify yield loss.

Due to the limited number of participants and/or answers, it was not possible to identify significant differences in the importance of specific pests (Table 2) or diseases (Table 3) with the exception of P. infestans, when the two groups of farms—with and without A. lycopersici presence—were compared. There were no mentions of T. absoluta as being important on farms without A. lycopersici presence in contrast to reported high relative importance on farms with A. lycopersici presence. Additionally, P. infestans was significantly more important on farms without A. lycopersici, compared to farms with A. lycopersici. This could mean that A. lycopersici favours some conditions similar to those favourable for T. absoluta and opposed to those favourable for P. infestans.

Intensification factors

Farms with large production areas and short cultivation breaks between tomato sets more often reported A. lycopersici presence (Fig. 4). Production area and cultivation break, among other factors included in this survey, can be categorised as intensification factors. As shown for production area and cultivation break, intensification factors tend to correlate or are mutually dependent upon one another. Mutual dependence for instance is the case for heating in cold months and short cultivation breaks as heating is usually only required when cultivation takes place in colder months, and cultivation in colder months usually only takes place when the cultivation break between tomato sets is short. Solely looking at large production area, it is questionable whether a large production area itself has an influence on A. lycopersici incidence. The described correlation likely is a result of the correlation between some of the intensification factors. A. lycopersici incidence was significantly higher in farms that heated during the colder months (Table 9). A. lycopersici requires temperatures of around 25 °C to reach its highest reproduction rate (Haque and Kawai 2003). This likely explains why first A. lycopersici symptoms are noticed earlier in the season on farms that heat in the cooler months (Table 9). Since only seven farms in the group without heating reported the months in which first symptoms were noticed, the conclusiveness of the presented data on heating in relation to A. lycopersici incidence is limited.

Growing in artificial substrate in special bags allows for precise water and fertiliser dosage to achieve optimal plant growth and high yields (IVA 2017). Since substrate bags, as opposed to natural soil, can easily be changed and renewed, problems with soilborne diseases are minimised. This allows farmers to disregard crop rotations and grow crops such as tomato over multiple seasons in a row in specialised greenhouses (IVA 2017). However, cultivation without crop rotation and with minimal breaks between tomato sets, as cultivation in substrates such as rock wool allows, comes at a cost. It might favour A. lycopersici survival by providing almost a year-round presence of the host crop. Organic growing associations usually only permit cultivation in natural soil (Bioland e.V. 2020). At the same time, due to soilborne diseases wider crop rotations are realised in farms that cultivate in natural soil that naturally result in a larger cultivation break / more time without a suitable host plant present. As mentioned, some of the factors overlap, and it is not possible to derive if and how severely the growing medium (Table 5) or substrate affect A. lycopersici incidence and persistence. To answer this question detailed studies on A. lycopersici survivability in soil, on plant residues, on greenhouse structures at varying temperature and humidity or on the effect of rotations with specific crops are needed. In summary, the data accumulated with this survey support the assumption that A. lycopersici favours one or more of the often-correlating intensification factors investigated in this survey: (i) short breaks between tomato seasons, (ii) heating in cold months, (iii) cultivation in non-natural soil, and (iiii) large cultivation area. Since these factors were more prevalent in integrated farms that participated as compared to the organic farms (Table 11), an explanation is provided as to why A. lycopersici incidence is significantly higher in the participating integrated farms (Table 4).

Plant protection measures against A. lycopersici

The nine farms that reportedly achieved A. lycopersici-free seasons after previous seasons with A. lycopersici presence (Fig. 3) were thoroughly checked for similarities in production factors and plant protection measures but no factors could be identified. The absence of a key countermeasure on those farms could mean that successful on-farm eradication relies on the creation of unfavourable conditions for A. lycopersici, both during and between the growing seasons.

There were no significant differences in specific countermeasures taken against A. lycopersici between farms on which A. lycopersici negatively affected yields, compared to those on which it did not (Table 10). This means that a standard effective countermeasure could not be identified amongst the participating farms. It is possible that not only the type of measure, but also the early detection and fast reaction on the initial outbreak is of importance. This could be traced back to the lack of reliable early detection methods for this pest (Pfaff et al. 2020).

The chance for participants to describe custom strategies and countermeasures against A. lycopersici did not reveal any novel methods not already reported or published in the literature. Strategies mostly consisted of repeated treatment with acaricidal substances, use of predatory mites, the removal of infested plant material or a combination of all three approaches. One participant responded to this question by stating that the first and strongest symptoms occur in the areas that are most exposed to sunlight. Assuming that those are the tomato plants most likely experiencing drought stress, this confirms previous findings that there is stronger population growth on plants that are in a state of drought stress (Pfaff et al. 2020). Naturally, the greenhouse areas most exposed to sunlight are the warmest and this favours A. lycopersici which has shown to have its peak population growth at around 25 °C (Haque and Kawai 2003).

During the survey window, the products Vertimec Pro and Agrimec Pro—both containing Abamectin were the only products specifically authorised for use against A. lycopersici in Germany. This explains why they were chosen by nine of the 15 farms that applied acaricides. Efficacy of Abamectin against A. lycopersici has been shown in the past (Royalty and Perring 1987; Kashyap et al. 2015). The other three compounds applied against A. lycopersici, Spirodiclofen, potash salt and Azadirachtin each mentioned by a different single farm were authorised in tomato, but not specifically against A. lycopersici, although a side effect on A. lycopersici is possible and likely the reason why they were mentioned. Abamectin is harmful to several predatory mites (Alhewairini and Al-Azzazy 2021). The negative effects on beneficial arthropods likely explain why Abamectin treatments often are restricted by farmers to A. lycopersici infection nests, or why local Abamectin treatments are combined with broader Sulphur treatments e.g. described by three farms. A broad variety of beneficial arthropods were introduced on the participating farms. Even if there were further acaricidal compounds available against A. lycopersici they would need to be highly specific in targeting A. lycopersici to not interfere with the established regime of beneficial arthropods in commercial tomato cultivation.

Among the participating farms, A. swirskii was the beneficial arthropod most often introduced specifically against A. lycopersici. A. swirskii was introduced on five of the participating farms. A. swirskii predates all life stages of A. lycopersici (Park et al. 2010). However, A. lycopersici has the ability to seek refuge from predators between trichomes on tomato plants (van Houten et al. 2013) and in doing so is likely able to limit the effectiveness of introduced predators to an uncertain extent.

Conclusion

This study provides a detailed picture of 50 tomato producing farms and how they are affected by A. lycopersici. Yearly, the number of farms where a first A. lycopersici occurrence is reported, has increased between 2005 and 2018 amongst participating farms and A. lycopersici incidence is not concentrated to certain regions in Germany. A. lycopersici was the pest with highest relative importance on the participating farms and 22 of 23 farms with a cultivation area of 4800 m2 or more report the presence of A. lycopersici. Repetition of the survey to detect possible changes in relative importance and on the farms affected, would clarify whether these findings are consistent over time, or if the status of A. lycopersici in Germany is still subject to change. Several intensification factors (1. Heating in cold months, 2. Large cultivation area, 3. Short break between tomato seasons, and 4. Not cultivating in natural soil) statistically favoured A. lycopersici occurrence, but autocorrelation prohibited the identification of a causal link to specific factors. Initial infestation of A. lycopersici significantly increased the chance for continuous presence in consecutive years. Detailed trials on A. lycopersici survivability and population dynamics under varying environmental conditions could help provide causal links to some of the aforementioned factors. Plant protection strategies in different combinations often consisted of broad treatments of sulphur, local abamectin treatments, removal of infested plant material or introduction of a wide variety of beneficial arthropods. None of the countermeasures could be identified as providing better or lasting control of A. lycopersici with the data gathered in this survey. Therefore, efficacy trials under practical conditions are advisable.