Background

The use and non-target ecological impacts of pesticides have gained a lot of public attention globally. Intensive use of chemical pesticides is generally linked to arable agriculture. In systems with a low proportion of arable agriculture, compared to more permanent grassland or forestry cover, the non-target impact of pesticides could be considered negligible. Measures such as the EU Sustainable use of Pesticides Directive (Directive 2009/128/EC) have led to the development of National Action Plans on pesticides use. These plans propose a number of actions for a more controlled use of pesticides. However, these measures are mainly focussed on human safety and effects on water bodies. Multiple recent studies have shown potential effects of pesticides on ES providers such as pollinators and biocontrol agents (Chagnon et al. 2015, Geiger et al. 2010, Power 2010); however, in countries with little information available on the current state of pesticide application, assessing the realised impacts on ES providers is difficult. Additionally, we are only starting to understand factors such as compound half-life in soils and detrimental effects of long-term exposure to sub-lethal pesticide dosage (Desneux et al. 2007). We argue therefore that especially in highly dynamic landscapes with ongoing land-use change, the effect of pesticides on ESs is potentially underestimated. Using the example of neonicotinoid application in the Republic of Ireland (hereafter Ireland), we explore important current knowledge gaps regarding pesticide application, as well as their effects on ES providers.

Neonicotinoids in agriculture

Although a range of pesticides (or plant protection products, PPPs) are commonly used in modern agricultural systems to ensure and maximise production, the potential effects of a specific group of insecticides, the neonicotinoids, on bees has fuelled widespread public and political concern (Sánchez-Bayo 2014). Neonicotinoid pesticides have been widely adopted since the 1990s because their systemic mode of action (i.e. the pesticide is taken up by the plant and expressed in all tissues) is considered more efficient than foliar spraying, and they have low toxicity to vertebrates. However, a specific concern with these pesticides is their widespread prophylactic use when applied routinely as seed dressings to crops. In this case, the pesticide dissolves into the soil water and is taken up by the crop, surrounding plants or washed into aquatic systems (Krupke et al. 2012, Main et al. 2014, Starner and Goh 2012), and this can result in widespread exposure to non-target insects, as well as organisms in higher trophic groups (Hallmann et al. 2014). Responding to concerns about potential impacts on bees, the EU implemented a ban (European Commission 2013) in 2013 on the use of three neonicotinoid pesticides (clothianidin, imidacloprid and thiamethoxam) for seed treatment, soil application (granules) and foliar treatment on plants attractive to bees and cereals (except in greenhouses and for winter cereals). The ban is currently under review.

The recently published European Academies Science Advisory Council report (EASAC 2015) has reinforced concerns about harmful effects of neonicotinoids on bees and other pollinators, but also other organisms which provide ESs important for agriculture. Three key ecosystem functions which deliver beneficial services to agriculture, but whose providers may be adversely affected by neonicotinoids are: pollination, pest population regulation via natural enemies (henceforth termed biocontrol), and various soil processes performed by decomposers, including soil formation, nutrient cycling and soil carbon storage (Box 1). The value of pollination to agriculture alone is thought to be about €153 billion per year (Gallai et al. 2009). Although there is a large body of evidence from laboratory-based and natural field experiments which demonstrate the negative effects of neonicotinoids on bees (see Box 1), the wider consequences of neonicotinoid use for ecosystem service provision are not well understood.

Neonicotinoid pesticides are not the only threats to ES provision in agricultural systems. Other PPPs have direct and indirect effects on ES providers. For example, herbicides are widely used in both grassland and arable cropping systems to control weeds which are toxic or unpalatable to livestock, or which compete with crop plants for resources. By removing weeds and reducing plant diversity in agricultural fields, resources for herbivorous insects are reduced. This includes floral resources essential to pollinating insects, exacerbating the pressures the latter face in agri-ecosystems. Fungicides may not have a direct toxic effect on insects which provide ESs, but can interact synergistically with insecticides to negatively affect them (Pilling and Jepson 1993, Schmuck et al. 2003).

Box 1

Pollination

Pollination services are considered one of the best studied ESs to date. Pollination is crucial for global food production, with about 75 % of crops traded globally depending on some degree of pollination (Klein et al. 2007). With recent declines in pollinators documented across the world (e.g. Biesmeijer et al. 2006, Cameron et al. 2011), the drivers of decline have received a lot of research attention (Potts et al. 2010). Current science suggests that multiple stressors promote decline, although the use of neonicotinoids is a contributing factor (Goulson et al. 2015). A number of studies provide evidence for the negative effects of neonicotinoids on the behaviour and fitness of bumble bees and solitary bees (e.g. Bryden et al. 2013, Gill and Raine 2014, Laycock et al. 2012, Rundlöf et al. 2015, Whitehorn et al. 2012).

Biocontrol

Numerous predatory taxa provide natural pest control through consumption of pest species. Utilising and managing populations of natural enemies is an important component of Integrated Pest Management (IPM), reducing the need for chemical pest control, which should only be applied as a last resort. The use of insecticides to control pests can have adverse impacts on populations of their natural enemies (Geiger et al. 2010, Hanson et al. 2015), potentially exacerbating pest-outbreaks. Pesticides affecting non-target organisms can thus significantly reduce agricultures natural resilience to pests, leading to increased pesticide use both at a cost for farmers and the other ESs. A number of studies have suggested harmful effects of neonicotinoids on beneficial species, through consumption of contaminated prey, however additional research is required (Cloyd and Bethke 2011, Douglas et al. 2014, Poletti et al. 2007).

Soil processes

The soil fauna is an integral part of soil formation, nutrient cycling and soil carbon sequestration, and therefore directly affects soil fertility. While there are few studies on the effects of neonicotinoids on soil fauna, some recent studies suggest detrimental effects on non-target soil taxa and a potential impact on soil functioning (reviewed in Chagnon et al. 2015). However it is still unclear if they are caused by directly by the pesticides or indirectly by the reduction of potential prey species (Peck 2009, Wang et al. 2012).

Mapping current pesticide use

Public availability and quality of information recorded on PPP application vary between and within countries (Thomas 2008). Using Ireland as an example we show how the lack of detailed pesticide information as well as poor understanding of land-use dynamics can lead to a potential underestimation of neonicotinoid exposure to the environment. The most recent documents on national level pesticide status are the Pesticide Usage Surveys published by the Pesticide Control Centre (PCS) for grasslands and fodder crops (PCS 2006), and for arable crops (PCS 2007) using data from 2003 and 2004, respectively. The only neonicotinoid mentioned in the surveys is imidacloprid, of which 1.6 tonnes have been used on arable crops in 2004. In the only recent study on the status of pesticide use in Ireland, Zhao et al. (2013) applied a GIS-supported approach to create a profile of the six most common pesticides used (four herbicides and two fungicides). Their approach combines the Pesticide Usage Survey data with data from the Central Statistics Office (CSO) on crop distribution, to estimate pesticide usage on county level. The results show a clear aggregation of high pesticide input areas along the east coast, correlating well with the area of high output arable agriculture. However, Zhao et al. (2013) did not consider neonicotinoids or any other insecticides. This lack of current data makes any assumption regarding the potential impacts of neonicotinoids on non-target organisms very difficult if not impossible.

The overall impact of pesticides in Ireland is considered negligible due to the relatively low area of arable crops. The total area of arable land in Ireland has remained relatively constant over the past decade with 392.5 ± 23.9 kha year−1 between 2000 and 2013 (CSO 2014), and simple extrapolation of pesticide application from crop data suggests a relatively limited area of pesticide application (9.9 ± 0.8 % of the total arable and pasture area and 5.6 ± 0.3 % of the total area of Ireland). However, land-use change from and to arable land may lead to a further spatial extent of potential pesticide pressure. Zimmermann et al. (2016) used high spatial and temporal resolution land-use data from the Land-Parcel Identification System (LPIS) to assess spatial and temporal dynamics of arable lands in Ireland between 2000 and 2012. Their research showed strong underlying land-use dynamics, leading to an area of 737.3 kha of land that has been reported as arable land for at least one year in the observed timeframe, due to constant shifts between grassland and arable lands. As a result the area of land potentially exposed to heavy PPP use over the 12 year timeframe could be as much as 18.6 ± 0.6 % of the total arable and pasture area or 10.5 % of the area of Ireland.

Impacts of pesticides on ES providers

Without detailed spatial and temporal resolution on pesticide usage, it is difficult to predict impacts on ES providers. Zhao et al.’s (2013) study only calculated pesticide usage in a county scale, this makes the assessment of local impacts very difficult to assess. These limitations can potentially underestimate pesticide pressures on a regional or national level. To better understand pesticide application in countries such as Ireland, we suggest that land-use dynamics are taken into account, with a special focus on:

  1. (i)

    Spatial heterogeneity of agriculture within counties

  2. (ii)

    Temporal heterogeneity of agricultural practices

Improving knowledge on heterogeneity can lead to a significant improvement of understanding pesticide pressures. Detrimental effects on ES providers can be the result of magnitude and the spatial extent of PPP exposure, as described in detail in the following section.

Magnitude

While Zhao et al. (2013) show that pesticide input correlates well with the proportion of arable land with each county, the majority of Irish agriculture is grassland-based for livestock rearing. Even counties with a high proportion of arable land still have a substantial amount of pasture/silage based agriculture leading to a non-uniform spatial distribution of pesticide input and potential pressures on ES providers. Therefore the impacts on non-target species may be spatially limited, but significantly larger due to localised, ongoing high concentrations of pesticides. While only a small proportion of the active ingredient is lost to drift during the application, lethal effects on honeybees nearby have been observed (Girolami et al. 2013, Girolami et al. 2012, Marzaro et al. 2011, Tapparo et al. 2012), and deposition of active substance in field margins could be measured (Krupke et al. 2012). Furthermore, non-target species may be exposed to neonicotinoids through consumption of guttation drops from treated plants (Girolami et al. 2009, Tapparo et al. 2011). The spread of neonicotinoid pollution in these areas may therefore be much larger than assumed (Nuyttens et al. 2013).

Extent

As shown by Zimmermann et al. (2016) a strong underlying dynamic is present in Irish arable systems, indicating ongoing land-use change between arable and grassland systems throughout the country. Therefore, land currently used as pasture may have been exposed to pesticide application in the recent past. Research shows that about 90 % of the active ingredient enters the soil where it either remains or enters the water cycle. Estimates on the half-life in soils vary significantly from less than a year to over 1000 days (Goulson 2013). Therefore, it can be assumed that the ongoing land-use dynamic can potentially lead to ongoing low-level exposure of non-target organisms in areas that are currently not used for arable agriculture.

Both magnitude and extend of neonicotinoid exposure can have potential impacts on ES. It is likely that short-term high-level inputs affect services differently than long-term low-level inputs, however depending on the service either input regime might be equally detrimental.

We propose a number of research options that would improve the understanding of pesticide pressures, these include:

  1. (a)

    Use high spatial and temporal resolution land-use data to assess the true extent of cropland over space and time to better assess areas of potential pesticide pressure.

  2. (b)

    Analyse the longevity of pesticides in areas recently changing from cropland. While basic research has already been conducted showing that neonicotinoids can persist in soils for years (Goulson 2013), the effects of ongoing land-use change between cropland and non-cropland land-use are still unknown.

  3. (c)

    In areas of current, recent and past pesticide application we suggest studies under field conditions to assess potential lethal and sub-lethal effects on non-target species.

  4. (d)

    The exposure of non-target species through pathways other than consumption of pollen and nectar, for example aerial drift or consumption of guttation drops from treated plants requires further research. Especially in areas with a low proportion of flowering crops these pathways may be of importance.

  5. (e)

    The link between effects of neonicotinoids on ES providers and the ES provided is still poorly understood. To further understand these impacts on ES we suggest more detailed studies focussing directly on ES provided in agricultural ecosystems.

Changes in agricultural policy

Farming practices continuously experience change. Globally agriculture is showing a trend of expansion and intensification to meet both rising food and bioenergy demands; however, policy decisions can significantly influence land-use patterns (Tilman et al. 2011). Examples are the changes to the EU Common Agricultural Policy (CAP), which may result in significant land-abandonment mainly in the pastoral sector (Renwick et al. 2013). In contrast, the recent abolishment of the milk quota, as well as the impending Food Harvest 2020 directive (DAFM 2010) in Ireland are expected to lead to an expansion and intensification of pastoral agriculture, i.e. the beef and dairy sector. Potential effects on pesticide application are complex and require additional research. Policy in Ireland may lead to a reduction in pesticide application, as the area of arable land is expected to contract, however economic considerations need to be taken into account. For example further intensification of grasslands may include increased use of PPPs on grasslands, and while the majority of neonicotinoid application is in the form of seed coating, it is possible to use them in grasslands against potential pest insects (Goulson 2013).

Conclusions

In conclusion, we argue that the existing information on the application of neonicotinoids, but also other PPPs, is insufficient to make any statement regarding magnitude or extent of their application in highly dynamic systems, such as those in Ireland. Furthering the understanding of this impact requires not only better availability of regularly updated, spatially accurate data on neonicotinoid application, but also research on drift during and after application, as well as persistence in vegetation and soils. Additionally, the interactions between neonicotinoids and ES providers are still poorly understood, especially the sub-lethal effects on non-target organisms and their capacity as ES providers, and this requires ongoing research. The research is especially important in the light of current changes to the structure of global agriculture to meet challenges of sustainability, and at the same time meeting the increasing global food and bioenergy demands and the potential impacts on ES providers. The recent EU ban on three neonicotinoid-based PPPs provides a potential opportunity to assess the full economic trade-off between the pest-control potential of the respective PPPs and the detrimental effects on ES providers. As PPPs are an integral part of maintaining high yields in current conventional agriculture, this assessment as well as comparison with other pest-control measures such as IPM is necessary to inform farmers and authorities on the full trade-offs, especially in financial terms.