Four Decades of Common Vole (Microtus arvalis Pallas 1778) Population Outbreaks in NW Spain: Transition from Environmentally Harmful Practices to Sustainable Integrated Pest Management (IPM)

Abstract

The common vole is one of the mammalian pests causing more agricultural damage in Europe. Since the late 1970s, this species has invaded the Duero valley in NW Spain, colonizing ca. 5 million ha of agricultural areas of the valley in about 20 years. Once settled in agricultural landscapes, the species experienced cyclic population outbreaks causing crop damages. The major vole population outbreak of 2006–2007 was managed by the Regional Government (Junta de Castilla y León, JCYL) mainly through large-scale application of anticoagulant rodenticides (ARs) and widespread destruction of field margins, natural vegetation patches, and crop stubbles by burning. These actions caused serious damage to regional agrarian biodiversity, including small game species. The coordinated action of scientific institutions and environmental NGOs, with the support of the main Spanish hunting association at a critical time, led to a progressive shift in pest management strategies during subsequent outbreaks, promoting the adoption of biological control and other management techniques causing less environmental damage. Finally, JCYL implemented an IPM program mainly based on biological control, good farming practices, and habitat management. This program has been increasingly adopted in recent years, leading to a marked reduction in chemical control and the complete elimination of burning as a tool of management. Over this period, the scientific knowledge of the species’ ecology has expanded substantially, providing key insights for the development and refinement of IPM strategies. Here, we synthesize this body of knowledge and provide additional recommendations to further improve the current IPM program, which may serve as a model for rodent pest management in other regions worldwide.

1. Introduction: Rodent Pests Around the World and Current Trends in Their Control

Rodents are considered among the most damaging agricultural pests worldwide, affecting multiple crop types and agricultural systems on all five continents [1,2,3]. Rodents may cause significant crop losses, as well as damage to stored food, seeds, or agrarian infrastructures [4]. In the case of Europe, an Arvicoline rodent, the common vole, is considered as one of the most crop-damaging vertebrates on the continent [5] and will be the focus of this review. Agricultural rodent pests are described in old texts since biblical and Aristotelian times, but they have probably been a problem since the start of human agricultural activities in the Neolithic [5,6,7]. Furthermore, rodents may act as reservoirs, spillover agents, or transmitters of multiple diseases, making them a serious public health issue as well [2,8,9,10].

A logical consequence of this widespread old problem has been the development of a millenary war against rodent pests that started early in the history of agriculture by using a plethora of lethal control methods, with particular preference for trapping and poisoning [2,11,12]. More recently, with the advent of anticoagulant rodenticides (ARs) by the mid-XX century, the use of chemical control on a large scale experienced a huge upsurge. These toxins became the main rodent control method around the world, both as Plant Protection Products (PPPs) in agriculture and as biocides in urban or suburban environments [2,4,13,14]. Thousands of tons of toxic baits with ARs have been released yearly around the world, but as early as the late 1970s it was recognized that they are highly bioaccumulative and toxic for many non-target species that could consume the baits intended for rodents, so they are currently considered among the most relevant contaminants globally in both terrestrial and aquatic ecosystems [13,14,15]. There is even concern about the implications of ARs in public health, because they may contaminate game species, livestock, or crops consumed by humans, as well as aquifers and rivers [13,14,16]. Furthermore, these toxic compounds are slow-acting by definition, so rodents may remain alive and active for days, weeks, or even months after consuming the toxic baits. During this period, they may be particularly vulnerable to predation due to sublethal neurological damage affecting mobility or escape ability [14,17,18]. Thus, many rodent predators or scavenging species often suffer secondary intoxication with these toxic products by consuming live rodents or their carcasses, so they are considered particularly vulnerable to this contaminant source [13,19,20,21].

Additionally, the massive and long-lasting use of ARs has induced widespread genetic resistance to the toxic in rodent populations in many areas of the world, reducing the efficacy of their application [2,13]. Finally, agricultural treatments with ARs may require repeated applications to be effective [22], becoming too expensive to be profitable when the rodent pests affect less valuable crops or are simply unaffordable for farmers in poorer countries [23].

In summary, the main rodent control method used up to recently in Europe and still employed in many other areas of the world, chemical control with ARs, has multiple problems. Currently it is widely recognized that complementary or alternative control methods must be developed and applied to allow a substantial reduction in the use of ARs [2,23]. Given the accumulated evidence about the environmental damage caused by the use of ARs to control rodent pests in agriculture, their use as PPPs has been banned in the EU since 2021 [21]. Their removal from the list of authorized biocides is also likely [24], underscoring the urgent need to develop new control strategies, particularly in this part of the world.

In fact, even before the discovery of the environmental damage and application problems that ARs (and other pesticides) may have, two crucial ideas were developed to advance into a new strategy for controlling agricultural pests, both environmentally sustainable and effective. First, the IPM concept, which dates back to the early XX century [12] and has experienced a continuous growth in practical application and development since then [25]. IPM can be simply defined as a strategy to control pests by combining all available methods besides pesticides, such as biological control, habitat management, selective trapping, barriers, good farming practices or emerging control techniques (e.g., fertility control or repellents), with the aim of reducing the use of chemical control as much as possible [12,25,26,27]. It is based on a simple fact: to date, there is no any panacea in rodent control (including chemical control), so a combination of control methods works better than any single one [12,25]. Furthermore, IPM focuses more on preventive measures than on controlling fully developed outbreaks, and stresses the need for monitoring and ecological research of the pest species, continuous scientific evaluation of the efficacy of the control strategy, and application of adaptive management [25].

Second, the concept of Ecologically Based Pest Management (EBPM), which is particularly well developed for the case of rodent pests [23]. This strategy is based on a simple principle: the more we know about the ecology and behavior of a pest species, the better we will be able to design an effective and environmentally sustainable control program [23,26,27]. The aims and working strategy are similar to IPM but stress the strong need for good ecological knowledge about the pest species, allowing the development of well-founded IPM programs [23].

In this manuscript we review the scientific and technical knowledge produced over the last decades about common vole pests in NW Spain. We report the rapid and substantial shift in control programs that has taken place in less than 20 years. This transition was driven by collaboration among scientists, environmental NGOs, local authorities, and farmers, supported at a critical time by the main hunting association in Spain, which together have pressured authorities in charge of pest control to move towards an ecologically based IPM program. We believe this is a good example that can be a reference for similar problems in other parts of the world. Finally, we critically evaluate the current status of that IPM strategy, proposing further evolution for the near future.

2. The Origin of the Problem: Recent Arrival of Vole Pests to Castilla y León, NW Spain (CyL Hereafter)

The southwestern limit of the common vole range in the Palearctic lies in Spain, reaching there the southernmost latitudes where the species can be found [28]. Common voles in the Iberian Peninsula were considered a subspecies (M. a. asturianus), a taxonomic singularity that has received support from recent genetic analyses [29]. Common voles in Spain could originally be found only in pasturelands of elevated lands in the northern half of the country (Pyrenees and mountains surrounding northern plateau; Figure 1), while to the north of the Pyrenees, this species occupies mainly grasslands and agricultural landscapes in lowlands [28]. Common voles were probably favored by the Neolithic expansion of agriculture in Central Europe, expanding their range over lowlands from glacial refugia, with larger European rivers acting as the main barrier to genetic dispersal, generating several distinct lineages [29]. However, after the last glaciation, the common vole in Spain became a species associated with montane pasturelands, which functioned as relicts of their original habitat in the peninsula. These pasturelands are similar to the extensive meadows that prevailed in the area during glaciations, and mountains acting as refuges in Spain for northern-dwelling organisms is a pattern shared with many animal and plant species in the Mediterranean area [30]. However, after the last glaciation, the central agrarian plateau (Duero river valley; Figure 1), with lower altitude and a pretty different landscape and climate, was apparently not occupied by the species [31], and this has been confirmed by recent genetic analyses [32]. However, in the late 1970s the common vole was detected in agrarian lowlands of the northern plateau for the first time [33], an area about 5 million ha wide that was fully colonized in just about the next 20 years (Figure 2) [34]. The species invaded the Duero valley through tributary river valleys from the NE and S, and the Duero river and tributaries acted mainly as a barrier to genetic dispersal [32,35,36].

The historical absence of common voles from central lowlands in CyL was most probably explained by unfavorable climatic conditions in the area, with long summer droughts typical of Mediterranean climates causing a virtual absence of fresh green herbaceous vegetation during summer. The scarcity of the main food and cover from aerial predators caused by summer drought probably created a too hostile environment for the species [37]. However, large-scale land use changes occurring by the 1970s, associated with agricultural intensification spreading at that time all over Europe, improved environmental conditions for voles in NW Spain. More specifically, the expansion of common voles was clearly associated with the increase in the area covered by “green crops”, mainly alfalfa (Medicago sativa), the optimal habitat for the species in agricultural land, both in our study area and in other European countries [22,38,39,40,41,42], but also to the expansion of other irrigated crops [37]. Another relevant factor contributing to explaining the common vole invasion in CyL was the large increase in average crop-field size associated with the still ongoing land-consolidation program in Spain [37].

Finally, extensive livestock roaming reduces the presence and abundance of common voles [43,44,45]. In a British study, vole abundance in exclosures unused by sheep herds was 1.5–2.5 times greater than in fields with the same herbaceous habitat but used by sheep [43]. Similar results were found in a long term exclosure experiment, reporting both a similar reduction in average vole abundance and a reduction in the amplitude of cycles [44]. In their original distribution area in Spain, common voles were virtually absent from trapping plots with the presence of cattle but were quite abundant in exclosures (2-year average of 37 voles/100 traps/24 h) [45]. Shortly before the onset of vole expansion, the presence of extensive livestock strongly declined within the species’ original distribution area. This reduction was primarily due to rural abandonment and the switch to intensive farming systems, which likely facilitated an increase in vole abundance in source populations [37]. In contrast, no effect of climate change was found in the same period, because some trends in seasonal climatic parameters could be favorable (e.g., increase in winter temperature and fall precipitation), but others the opposite (e.g., warmer and dryer summers) [37]. Recent climate change models predict a clear contraction in the future range of the common vole in Spain, where main populations would remain in montane areas of northern Spain or even just in the northern fringe of the country, in the Cantabrian Mountains and the Pyrenees (Figure 1) [46].

In summary, the invasion of Duero valley by common voles can thus be interpreted as a clear consequence of agricultural intensification and changes in livestock abundance, as pointed out by regional experts too [47]. However, it does not appear to be driven by climate. On the contrary, it has occurred in the opposite direction to that expected under climate change, expansion instead of contraction, underscoring that the effects of climate change on species may frequently operate indirectly through habitat modification. This finding highlights the relevance of land use in mitigating potential climate-driven shifts in species distributions.

3. Common Vole Population Outbreaks in CyL and Other Areas

As soon as the common vole was first found in lower areas of the Duero valley, reports of crop-damaging population outbreaks appeared too [35,48]. These population outbreaks probably favored the expansion of the species [34,36,48], which is supported by recent genetic analyses showing that population growth is associated with intense effective dispersal [49,50].

Recurrent population outbreaks of common voles in Spain looked similar to the cyclic demography of many northern populations of this and other vole species [35,36,51,52]. Briefly, vole numbers suffer regular oscillations in numbers, with peak (outbreak) years, followed by a rapid population crash (usually less than one year), a low phase with extremely low abundance, and a growth phase lasting several years [52]. An analysis combining information from available scientific publications, a technical agrarian publication from the Spanish Ministry of Agriculture, and news about vole pests appearing in the main regional newspaper concluded that, following their arrival in the Duero valley, common vole population outbreaks occurred cyclically, on average every five years. Once the colonization process was completed, outbreaks affected all provinces in the region (Figure 1, Figure 2 and Figure 3), with six major events recorded in 1978, 1983, 1988–89, 1993, 1997, and 2007) [34]. However, there was a period of ten years without any significant outbreak after the 1997 event, probably due to a persistent drought aborting the peak, as recently reported in the 2016–2017 outbreak [34,53,54].

Within the large agrarian area invaded by voles in NW Spain, the region known as Tierra de Campos (hereafter TC; Figure 1) has always been the one where vole outbreaks apparently reach the highest densities, causing most of the crop damage at the regional level [55,56]. Population cycles in this area have the highest amplitude (difference in abundance between peak and low phases of the cycle) and shorter duration (2–3 years) when compared with regional-scale outbreaks that occur every five years on average [57], showing traits similar to cycles in central and northern Europe (e.g., population crash in full breeding season during the decline phase) [38,58]. Interestingly, TC has the most deforested landscape with intensive agriculture in CyL, along with a land use highly favorable for voles (see Section 7). This kind of widely deforested agrarian area is typically affected by rodent pests in many areas of the world [59]. Finally, TC is also the area where voles cause more public health problems, as they act as spillover hosts of Francisella tularensis, the pathogen causing tularemia disease in humans. This zoonotic disease was first reported in Spain during the 1997 common vole outbreak in CyL. Notably, human tularemia epidemics fully coincide with vole outbreak years, and human cases are concentrated in TC [34,60,61,62]. This pathogen has multiple hosts, particularly in aquatic environments, as well as many infection routes in humans. Clinical manifestations vary depending on the route of bacterial invasion, affecting different organs, including a typhoid form with severe symptoms. Overall mortality in affected humans is about 2% when properly treated, rising to 5–15% without intervention.

The most recent and comprehensive work on vole population dynamics in CyL, based on 11 year time series of abundance data obtained from 40 study sites well distributed throughout the large agrarian area affected by vole outbreaks, detected a more varied suite of demographic patterns [55]. Most sites (n = 33) showed second-order population dynamics (regular cyclicity, delayed density dependence) with the two temporal patterns mentioned above (2–3 or 4–5 years between peaks) [55]. The remaining sites (n = 7) had first-order population dynamics (no cyclicity, temporally irregular outbreaks). Sites with cyclic demography were located mainly in the core area affected by vole outbreaks, showing periodicity of 2–3 years, particularly in TC and nearby sites. In contrast, longer cycles of 4–5 years occurred in more peripheral areas and across all provinces. Sites where cyclic demography was not found were also mainly peripheral. Demographic patterns were mainly explained by density dependence (21 sites), weather (11 sites), or a combination of both (8 sites). As reported in earlier studies on vole demography in Spain, precipitation was a crucial factor promoting outbreaks, something expected for a strictly herbivorous species requiring well-developed herbaceous vegetation and living in a Mediterranean area with highly irregular precipitation patterns [63,64]. However, for many populations weather (mean temperature and accumulated precipitation in each study area) was not apparently related to demography, most probably because of the role that land use changes (expansion of optimal cultivated “green crops”) have played in improving habitat suitability independently of weather, as reported above to explain the invasion. Interestingly, the variation in demographic patterns found in CyL is quite similar to that reported for agrarian lands in France, where this variation has been linked to habitat traits at the landscape level [39,65,66] (see Section 7). However, at the continental level, the variation reported in demographic patterns of this species is even larger, from mainly stable populations to erratic irruptive outbreaks or cyclic outbreaks every 2–10 years (see discussion in [64]).

Finally, a remarkable pan-European synchrony in common vole outbreaks in 2019, a year with simultaneous peaks in ten countries from Spain to Germany, has been reported [67]. This synchronization across European countries probably also occurred in 2006–2007 [68] or 2016–2017 (Figure 1 in [67]), so this may be a regular phenomenon that has remained overlooked until recently (but see [69]). This synchronization could even be geographically larger, since common vole population peaks have been reported for 2014 and 2019 (regional peaks in CyL) in the middle Volga region, Russia [58]. A pan-European synchrony of changes in demographic parameters of vole cycles has been reported indeed [70], as well as a continental analysis of time series also reporting spatial synchronization in population peaks for several rodent species [69]. Spatial synchronization of outbreaks is a common phenomenon at the local or regional level (see [71] for TC) and could be explained by dispersal, predation, or other ecological factors acting at a small spatial scale [67]. However, a pan-European synchronization of population dynamics is likely the result of common environmental factors acting at such a large scale, what is known as the Moran effect [72]. The best candidate factor to explain such large-scale synchrony is weather, which may show pseudo-cyclic events (ENSO or NAO oscillations) related to cyclic demography or that have suffered changes caused by climate change affecting vole demographic parameters [69,70]. Weather is considered a useful tool to predict variability in regional outbreak risk in Germany [73]. In the case of CyL, the 5-year cycle at the regional level [34] was well correlated with weather variables, showing pseudo-cyclicity too [74].

Overall, Caminero et al. [55] confirm that there is some ability to predict when an outbreak may occur, particularly for TC, the critical area where pest management is more necessary. This is a crucial aspect for the implementation of a good IPM program based on preventive measures instead of acting when the population is at a peak and we may expect a natural deep and fast decline soon, a major criticism of old-style management (see Section 6 and Section 7 and [68,75]).

4. The 2007 Outbreak: Environmental Impact, Tularemia Epidemics, and the Origin of a Change

After 10 years without relevant vole population outbreaks in CyL, one of the most significant known occurred in 2007, affecting more than 3 million ha of agrarian land in all provinces of the region (Figure 3), causing relevant crop damage (estimated to be about 30 million € by the main agrarian organization in the region) and promoting the most important tularemia epidemics reported to date in Spain [5,34,76]. Field samplings in TC, the most affected area, reported nearly 1000 voles/ha in optimal habitat (alfalfa fields) [60], a very high density, although still far from maximum values estimated for the species in Central Europe (>2000 voles/ha [67]). Voles not only invaded crops, but also urban parks and gardens, and it was common to see them running through villages in the area or crossing roads. These observations reflect both the intense dispersive behavior typical of many rodent species during outbreaks and probably the high competition for settlement in optimal habitats due to high population density, forcing individuals to move to unusual sites in what has been called “pathological” dispersal [77,78].

Vole pest management in CyL has been focused on AR use since the earliest outbreaks and has consistently persisted during all subsequent events [34]. The regional government (JCYL) is legally responsible for pest control. The usual management practice involved distributing bags of AR-treated baits (cereal grain or pellets) to farmers reporting crop damage (JV, pers. obs. in 1989 and 1994), and a similar program was implemented again in 2007. Common vole abundance started to be high in TC since autumn 2006, and by spring 2007 there was already a clear invasion of cereal fields [60,79]. In March 2007, JCYL started the first control campaign in TC by distributing toxic grains with chlorophacinone (a first-generation AR) over ca. 20,000 ha of crops with machinery (fertilizer spreaders), raising complaints from environmental NGOs that filed the first lawsuit in court as well as a complaint to the European Commission [68]. By July 2007 the outbreak had already extended to all provinces, and JCYL started a second control campaign affecting at least 100,000 ha of croplands and 37,000 km of culverts and other field margins. This time the baits were disposed of within short sections of corrugated pipes, trying to reduce the access of non-target species to the toxic, with not much success (Figure S1). During this second campaign, fire was also extensively used on stubble and field margins as a supposed control method, alongside cleaning of vegetation in field margins [68]. Finally, a third campaign was developed between February and April 2008 (with vole populations already in the decline or crash phase [68,75,79]), with similar management, but this time using bromadiolone, a second-generation AR (SGAR) that causes more problems of secondary intoxication in predators [13,68].

Environmental damage caused by AR use to control vole outbreaks in CyL before 2007 is poorly understood, except for a raptor species particularly prone to poisoning, including secondary poisoning with SGARs, such as the Red Kite (Milvus milvus Linn. 1758) [80,81,82,83,84]. The first national survey of Red Kite was conducted in 1993–1994 [85], coinciding with one of the most important vole outbreaks in CyL and extensive AR use [34]. This region is the most important area for the Red Kite in Spain and Europe, as it houses the largest breeding populations in the country. CyL is also the main wintering area for the largest populations of the species at the north of the Pyrenees, particularly German birds, which holds tens of thousands of birds during mid-winter [85,86]. During the 1993–94 survey, coinciding with the release of ARs, several cases of massive mortality of red kites (up to tens of individuals) were recorded, and similar reports were found for previous outbreak years [85]. The massive use of ARs during winter in CyL may at least partially explain the changes in population trends of the breeding German population over the last decades (see Annex S2 in [86]). With respect to 2007–2008 treatments, breeding Red Kite populations in areas of CyL where the three AR treatments were applied drastically declined (−42%) between 2004 and 2008, while populations in areas where there was no vole pest and therefore no AR treatments increased by 28.5% in the same period [87]. It is important to remark that during vole outbreaks many raptor species and individuals gather in large numbers to exploit such abundant prey [75], including wintering red kites that are more abundant in CyL during vole peak years [85], which enhances the problem for predator populations if rodenticides are released when vole abundances are maximum [81,82,83].

Red kites could be particularly affected by AR treatments, but many other non-target animals have also suffered primary or secondary intoxication from AR treatments in CyL, including endangered steppe birds, as well as pigeons or other small game species that could potentially be consumed by humans [21,68,88,89,90,91]. Small Passerine birds, particularly prone to consuming cereal grains, such as granivore species in the family Alaudidae, were also commonly found poisoned, and their populations suffered a marked decline in TC with no subsequent recovery, in particular the Calandra lark (Melanocorypha calandra Linn. 1776) (Jubete, F. unpub. data). A highly relevant case was that of the Iberian hare (Lepus granatensis Rosenhauer 1856) that suffered a dramatic decline in the areas most treated with ARs in 2007 (Figure 3) and was among the most commonly found non-target victims of ARs (40% of hares found dead in CyL had the toxic [68]). Poisoning, perhaps added to mortality caused by tularemia, virtually extirpated the species over large areas of the region, and hare hunting had to be completely stopped for more than one hunting season. The crash of hare populations was a major concern for hunters and rural people in the region, who also started to demand changes in vole management to protect game and hunting activity, including increased support for the biological control program that would soon be started by GREFA and IREC (see next section).

Finally, it is important to mention that rodenticide consumption interacted with tularemia disease, increasing the overall mortality rate of voles [60]. The release of large amounts of AR-treated baits over croplands in TC caused the presence above ground of a large number of dead voles with the disease during cereal harvesting, which in turn produced the only known tularemia epidemics in humans where the prevailing route of infection was inhalation, most probably due to high bacterial contamination in the aerosols produced by harvesting machinery [60,92]. Obviously, this was another important reason to reduce the extensive use of ARs and rethink the application timing of chemical control.

During that year of intensive vole pest management on a large scale and a relevant tularemia epidemic, there was considerable social concern in rural areas, continuous complaints by farmers and environmental NGOs, and a common presence of the conflict in the regional or even national media [34,79]. This relevant conflictive situation was also reported and deeply discussed in scientific journals [34,68,93,94] and in one of the main agrarian media in Spain [79]. However, 2007 was also the year when JCYL started to take the problem more seriously, creating a scientific committee to analyze the problem and guide management, contacting national and regional experts on rodents and pest control, and starting a monitoring program, setting the basis for upcoming changes based on science and environmental conservation applied to pest control. Finally, the 2007 vole outbreak promoted an alliance between environmental NGOs and scientists, also supported by hunters at the critical time of the 2014 pest, to promote the change in management we report in the next sections.

5. The Biological Control Program

Biological control of insect pests dates back as far as 300 AD in China (predatory ants were used in citrus orchards), and today it is a well-developed technique in multiple agricultural systems [12,95]. Early in the long history of the fight against rodent pests, the role of predators in reducing rodent abundance has been acknowledged, dating back at least to the use of cats as domestic allies and to the Roman recommendation of ferrets to control rabbits [2,12,95]. However, managing rodent predators on the large scale required by agriculture may be challenging, so the development of biological control has been relatively recent, but it is extending notably around the world as a useful tool within IPM and EBPM strategies [2,25,59,95,96]. Predation is one of the main regulatory factors of rodent abundance and could even contribute to explaining the generation of cycles, particularly influencing the crash phase [52,97]. Even the fact that rodent pests are particularly common in highly deforested agricultural landscapes could be due to the relative scarcity of predators in these deeply simplified habitats [59]. This could also be the case in CyL, given that common vole outbreaks have rarely been reported in the original montane area where the species lived, with more forested, varied, and less perturbed natural habitats than in the Duero valley, particularly compared to TC. High-density vole outbreaks in the original area have been absent even though a recent decline in extensive livestock presence improved the habitat for voles there even more (see Section 2 above) [37,45,47,64]. Thus, given that rodent pests are particularly problematic in deforested agrarian areas with few predators, increasing predator density in these kinds of landscapes could help to regulate rodent numbers. The scarcity of many avian predators in these deforested landscapes may be a direct consequence of the lack of adequate nesting habitat, so providing nest boxes for efficient vole predators such as barn owls (Tyto alba Scopoli 1769) or common kestrels (Falco tinnunculus Linn. 1758) has started to spread across many sites around the world since the second half of the XX century [23,59,96]. In their recent review of the application of this technique around the world, Labuschagne et al. [98] identified 2086 scientific papers dealing with avian predation as a rodent pest control system, which gives a good picture of the notable expansion that this technique is having around the world. These authors concluded that biological control may be effective in reducing rodent abundances 5.4-fold on average in the few studies reporting information about efficacy, but this efficacy was highly variable depending on the study system, for reasons poorly understood, and there are few statistically robust tests of efficiency, particularly in the long term [98].

In NW Spain, we have verified that common vole abundance in a montane area of Segovia province is clearly lower than in the Duero valley, with oscillations of lower amplitude, and that an artificially increased population of kestrels through provisioning of nest boxes was a factor clearly regulating vole abundance [64]. These results obtained in the original distribution area of the pest species stimulated the development of an experimental program promoting biological control in the agrarian areas suffering from the pest. Conversations between scientists involved in the early project and the wildlife conservation NGO GREFA revealed both common interests and the possibility of a potentially fruitful and synergic collaboration. The project started in 2009 with the installation of nest boxes for barn owls and common kestrels in three municipalities of CyL that were among those having more serious problems with voles during the 2007 outbreak. In these villages, city mayors and farmers were willing to avoid future use of ARs and test new ways to control voles. One of the sites was at the core of TC (Boada-Capillas, BC hereafter), another was in the western edge of TC with a slightly more forested landscape (San Martín de Valderaduey, SM), and the last one had a more complex and forested landscape out of TC in the central Duero valley (Villalar de los Comuneros, VC). In every study area, 100 nest boxes were installed on wooden poles at field margins, half with a design for barn owls and half for common kestrels (Figure S2; see details in [99]). Almost simultaneously with this experiment in CyL, the regional government of Navarra (NE Spain) started a similar project in the Ebro valley, where common voles have also arrived recently and were causing the first crop damages [100].

Rural people in the first three study areas in CyL had a good acceptance of the project, sometimes even a better perception of the efficacy of nest boxes than we scientists had, as we could detect in our informative meetings with farmers after vole outbreaks. Other nearby villages began to request nest boxes from GREFA. This NGO published successive editions of a management booklet for farmers produced in collaboration with scientific partners [101] and another edition for primary school students. Finally, GREFA has organized hundreds of dissemination activities for farmers over 16 years (2009–2025) (see S3). In this way, the initial seeds in the three areas of CyL became demonstration projects that facilitated the scaling up of the initiative. As early as 2012, the official organism in charge of technical management of agriculture in CyL (Instituto Tecnológico Agrario de Castilla y León, ITACYL hereafter) also joined the project, generating new experimental areas with nest boxes in coordination with GREFA.

The overall number of boxes for common kestrels and barn owls has reached 2420 in 77 municipalities of CyL at the date of writing this text, 2120 of which are under monitoring (Figure 4). Nest box provisioning continues to grow, and boxes have been recently installed for the first time in the ninth province of CyL (Soria, Figure 1 and Figure 4), the only one without boxes until now. Furthermore, since 2014, 262 nest boxes for the little owl (Athene noctua Scopoli 1769) have also been installed in CyL. Finally, nest boxes have been installed in other regions of Spain where other rodent species also cause problems, reaching now a total of 2922 nest boxes at the national level.

The first evaluations of the efficacy of biological control in the four study areas revealed results highly coincident with Labuschagne et al. [98]: the efficacy of biological control at the landscape level varied between almost negligible (BC, central TC) and highly successful in Navarra and VC, with intermediate results in SM [99,100,102]. Even in the area with the poorest performance at the landscape level, nest boxes reduced considerably vole abundance at the local level in the surroundings of their location (estimated vole abundance at distances less than 180 m from occupied boxes was almost half that at distances greater than 540 m) [102], even in one of the areas with the worst vole outbreaks since 2007 during a peak year (Osorno, Palencia province) [103]. It is important to note that during the first years of the project in CyL, nest boxes were readily taken by kestrels but more slowly by owls (Figure 5), and most probably even better results could be expected with higher barn owl occupancy [102], which has reached higher values during the last years (Figure 5). Nest box occupancy was clearly related to vole cycles, with maximum occupancy rates during peak years (up to 2.8 and 1.2 pairs/km² for kestrels and barn owls, respectively) and minimum during low years (as low as 0.1 and 0 pairs/km² for kestrels and barn owls, respectively) (Figure 5). Interestingly, the largest difference between maximal and minimal occupancy was found in TC, where the amplitude of vole cycles is also the highest. Finally, we found an increasing trend in nest box occupancy over the 15 years with data (Figure 5), in particular for barn owls. Besides variable barn owl occupancy between study sites, the variability in efficacy observed between study areas likely stems from the concomitant influence of variations in weather, overall predation pressure, and the spatial extension of the vole pest among study areas [100]. Furthermore, effects of previous use of ARs on raptor populations and site-specific landscape attributes (agrarian practices, crop distribution, and landscape structure) play a critical role in modulating these outcomes [98,99,100,102]. The three latter factors, amenable to management, will be examined in detail in Section 7 and Section 8.

Overall, biological control has been a useful tool for vole IPM in Spain, with good performance at the local level, although highly variable at the landscape level, with no relevant negative effects on non-target species identified thus far [100]. On the contrary, this program is contributing to restoring populations of two raptor species that had been seriously affected by AR poisoning in the study areas, in a sort of compensatory management [102]. Spain has been fully incorporated into the increasing and long list of regions or countries in the world applying this emerging rodent control technique, from California vineyards to Florida sugarcane, from SE Asian rice fields to Chilean forests, including several Mediterranean and nearby countries [104]. Nest boxes for avian rodent predators can thus be considered a case of biological control that should be considered in rodent IPM programs because they may provide a relevant service to farmers, but not disservices, may have negligible side effects on non-target species, have lower costs, and require less effort than chemical control [105,106,107].

6. The End of ARs and Fire as Tools to Control Vole Pests in CyL

After the 2007 outbreak, AR distribution and burning of stubbles and field edges became settled as the main management tools against vole pests and were also applied again in the mid-density outbreak occurring in TC during 2011–2012 [75,102], again raising complaints by environmental NGOs. Five environmental NGOs prepared a report about the extent of fires and the environmental damage caused by their use in TC [108]. This report demonstrated how fires not only affected stubbles or field margins in more than 150 municipalities over 5000 km² and >10,000 km of field edges, but also riparian vegetation in 9200 km of rivers and streams (Figure 6) [108]. This would sum up to about 3700 ha of riparian vegetation and 4100 ha of culverts and other field edges. These fires were clearly against recently implemented EU regulations prohibiting stubble burning and strictly regulating other uses of fire in agriculture, with the aim of reducing CO₂ emissions. Importantly, these actions were implemented without any scientific evidence supporting that fire was an effective vole control technique, and there were multiple doubts about its real utility (see Section 7, Section 8 and Section 9 below) [108].

In 2013, vole numbers rose steadily again, this time at a regional level, so a regional-scale outbreak could potentially be expected by 2014 [54,55,71]. During the last weeks of 2013, ITACYL acquired 144 tons of cereal grain with bromadiolone to be distributed among farmers in the affected areas. The scientific sector reacted again by making public a document signed by 143 scientists opposing the new massive release of rodenticides [109]. A technical document with detailed arguments against this campaign of AR and fire use was prepared by a team of scientists expert in voles and conservation of agricultural ecosystems, with the supporting signatures of the main environmental NGOs in Spain (WWF Spain, SEO/Birdlife, SECEM, GREFA, Ecologistas en Acción, FIRE) and CyL (ACENVA, ANPA, ASCEL, Asociación Zamorana de Ciencias Naturales, Colectivo Azalvaro, Fondo Refugio de las Hoces del Riaza, Grupo Naturalista Ornitológico Alauda, Salamanca Natural) as well as the Spanish Hunting Federation [110]. The document was sent to the JCYL and to the National Ministry of Agriculture. Furthermore, new EU regulations on AR use as PPP required introducing the baits within vole burrows, which was an unprofitable and hard task in the most affected crops, alfalfa, often cultivated in large parcels (tens of ha) that were full of vole burrows. Despite all this, ITACYL tried to distribute the toxic bait, but it was not accepted by many farmers due to the difficulty of application and the social concern mentioned above regarding the effects of ARs on game species and other fauna. However, other farmers took and used the toxic, which was released into the environment in amounts enough to reach kestrel nestlings in nest boxes of two study areas, inducing a reduction in their body condition [90].

By that year (2014), many EU countries had already retired ARs from the list of authorized PPPs, including those with the oldest and most serious problems with voles, such as Germany, where use of ARs in agriculture was stopped after 2007 due to detection of the toxic product in hares. Those countries turned to zinc phosphide use, a fast-acting toxic product that causes fewer problems of secondary intoxication in predators [111], but may still have relevant negative environmental effects on non-target species [112]. Finally, in 2015, the Spanish Ministry of Agriculture definitely retired ARs from the list of authorized PPPs for rodent control, replacing them with aluminium phosphide, a strong toxic product highly volatile in contact with humidity and thus dangerous to handle, so only specialized and registered enterprises are allowed to use it.

A new event of rapid vole population growth was detected in 2016, affecting mainly TC and nearby areas, with moderate maximum abundances (similar to 2011–2012), most probably because a persistent drought since autumn 2016 aborted the expected peak in 2017 [53,54,55,71]. ITACYL requested from the Ministry of Agriculture a special permit to use the toxic bait unused in 2014, which was granted and later recognized that “only” 19.1 tons of bait were distributed mainly in municipalities of TC and nearby areas [113]. A new campaign for fire use was also started. The NGO Ecologistas en Acción, with the support of WWF Spain, filed another lawsuit with a technical report prepared by the first author of this manuscript. ITACYL sent to court a critical response to that report, and plaintiffs sent the corresponding response. Given the technical dispute, the court decided to convene an in-person trial with the presence of the authors of the reports from both sides. The trial took place on 8 March 2018 and the sentence was favorable for the plaintiffs with respect to fire use, which has not been authorized anymore as a vole control system. The sentence was unfavorable in what respect to chemical control, because ARs were still authorized as PPP in some EU countries, and the only authorized alternative at that time (aluminium phosphide) was too expensive and risky for such large-scale use.

Coinciding with the next regional outbreak (2019) [55,67], ITACYL made public a new “Strategy of integrated management of risks derived from the presence of common voles in the territory of Castilla y León” [114] (see Table S1) that represented a radical change in the management of this pest. The strategy supported biological control as a fundamental preventive tool of the IPM program, did not mention the use of fire, included a wide toolkit of habitat management and good farming practices (see Section 7 and Section 8 below), and considered chemical control with authorized PPP (thus excluding ARs) as a last resort to be used in exceptional circumstances after the failure of previous actions [114] (Table S1). ITACYL never distributed toxic baits again, so the cost of chemical control should rely on farmers. In autumn 2019, the Ministry of Agriculture finally incorporated zinc phosphide as PPP to be used by farmers, since it is less dangerous than aluminium phosphide. Zinc phosphide requires an acidic environment to release the toxic gas, such as that found in the stomach; thus, the release of the gas outside the body of an animal ingesting the bait would be extremely rare, and handling is much safer than for aluminium phosphide. So even in the case of using chemical control, the toxic product authorized now would have much less environmental problems than ARs.

After a long series of vole outbreaks over 12 years and conflicts even reaching court, a good IPM strategy considering EBPM was finally in place. Thus, in the last vole outbreaks (2019 and 2024 at regional level, 2021 in TC), for the first time in decades, ITACYL did not distribute rodenticides and fire was not used anymore. Fire use in agriculture was almost eradicated before 2007, but the massive authorizations associated with vole control during several consecutive outbreaks encouraged farmers again to use this tool. Consequently, illegal fires have still been detected during last vole outbreaks, but the extent of the practice has been markedly reduced. Perhaps a new generation of farmers will be required to fully terminate these illegal practices. Similarly, we know some farmers still bought and used zinc phosphide baits, but there is no available information to know how much rodenticide has been released. However, with our long field size experience in CyL and permanent contact with farmers there, we think rodenticide use has been much more moderate than in previous outbreaks. This is something expected when farmers make calculations about the cost of the rodenticide compared with damages to crops: in the case of cereal, a suboptimal habitat for voles, crop damages caused by voles are usually below 5%, making the treatment unprofitable [115,116].

7. The Relevance of Habitat for Management and Crop Damages Caused by Vole Outbreaks

The original natural habitat of the common vole consists of open deforested grasslands, while in more bushy or forested areas, it is replaced by its sister species, the field vole (Microtus agrestis Linn. 1761) [28] or, in the case of Spain, by the ecologically equivalent and highly related cryptic species, the Mediterranean field vole (Microtus lavernerdii Crespon 1844) [117]. Agriculture has generated extensive deforested areas often occupied by herbaceous crops, replicating the original habitat of the species and favoring its expansion since the Neolithic [29].

Common voles spend most of their lives within large communal burrows with complex structures that tend to expand over time; so older burrows may reach large sizes (up to 30 m² in total subterranean area occupied, 70 m long, or 58 external entrances [38,118]). The persistence of these crucial refuges and breeding sites clearly promotes abundance of the species during the growth phase of the cycle [38,119]. As a consequence, on one hand, regular tilling is one of the best ways to control the abundance of the species [120,121], and is a central tool in the new IPM strategy [114] (Table S1). On the other hand, stable grounds unperturbed for several years pose a risk for pest development. In this sense, permanent or long-standing herbaceous crops and field margins are the main habitats with long term unperturbed soils available for voles in intensive agricultural landscapes and can be considered a major source of the pest. Field margins are often the only possible origin of colonization, given the scarcity of other permanent patches with herbaceous vegetation in these intensive agrarian ecosystems [37,38,42,57,71,122].

The main permanent crop in our study areas is alfalfa, which, consistent with observations across Europe, constitutes the optimal habitat for common voles. This perennial herbaceous habitat provides a continuous and abundant supply of high-quality food and effective cover against aerial predators (except during and after harvesting, see [102]). Furthermore, the absence of tillage for 3–10 years ensures soil stability, allowing these crops to function as a source of voles that colonize adjacent crops and field margins [22,35,37,38,39,40,41,42,57,65,66,110,119,120]. Alfalfa is the crop suffering the most damage; indeed, it can experience almost 100% loss in the more extreme cases, when almost the entire area of a parcel is occupied by burrows, along with other crops unable to recover from vole damage caused during early growth (e.g., peas or sunflower) [115,119,123,124]. The optimal habitat provided by alfalfa promotes fast growth of vole populations, making the control of vole numbers in alfalfa a challenging task, both for biological and chemical methods [22,102,125], but several agronomic actions have been proposed to help biological control (see detailed report in discussion of [102]), and some of them are already considered in the new IPM strategy designed by ITACYL [113] (see Table S1). Interestingly, changes in vole abundance in alfalfa and field margins occur simultaneously at landscape level, supporting the notion that alfalfa fields serve as a source of dispersing voles using the field margins network [42,56,126,127,128]. Furthermore, in our study area, there are two distinct management systems for alfalfa: dry regime or irrigated. The dry regime is mainly used in TC, where there is a regional old variety well adapted to this management and climate. In the rest of CyL flooding or sprinklers irrigate alfalfa, but dry alfalfa is a rare crop [129]. One of the best ways to control vole numbers in alfalfa fields is flooding, which is a crucial management within the IPM of vole pests in these crops [130,131], considered in the new IPM strategy [114] (Table S1), making irrigated alfalfa a hostile habitat when compared to dry alfalfa. Additionally, dry alfalfa has slower growth and thus fewer cuts per year, providing protection against aerial predators over longer periods [102].

When managing vole pests in alfalfa fields it is important to consider that this crop benefits other nearby crops providing pollination and insect pest control services, thus reducing the overall need for pesticide applications [126]. Alfalfa in CyL is also an important habitat providing food and shelter to many vertebrate and invertebrate species, including endangered steppe birds that find both food and nesting sites [126]. The increase in the area occupied by dry alfalfa in TC between 1998 and 2008 has been a major factor promoting the growth of the largest population of Great Bustards (Otis tarda Linn. 1758) at world level [129], but this change may have promoted vole outbreaks since 2007 too and past pest management based on ARs also negatively affected this species [68,91]. It is important to remark that world Great bustard populations have been declining over recent decades [132], so conservation of the largest population in CyL is particularly relevant. Consequently, the EU has included alfalfa cultivation in rotation systems as an agri-environmental measure indeed and several agronomic actions and good farming practices have been proposed, both to enhance biological control of vole pests, and to minimize the negative impacts of alfalfa management on biodiversity, maximizing their conservation value [102,126]. However, alfalfa can also act as an ecological trap when early harvesting destroys nests or incubating adult birds, something that commonly happens, particularly in irrigated fields where the first cut tends to be earlier [126], so delaying that first cut would be favorable for wildlife conservation. In contrast, for vole management, earlier and numerous cuts are recommended [114] (see Table S1). Unfortunately, the first alfalfa cut in TC may destroy a high number of nests of vole predators breeding on the ground in alfalfa fields: Common harrier (Circus pygargus Linn. 1758), Hen harrier (Circus cuyaneus Linn. 1766), and Short-eared owl (Asio flammeus Pontoppidan 1763) (as many as four nests of harriers and three of owls in just one 15–20 ha parcel; Jubete et al. unpub data.). A careful balance between promoting management useful for vole control and the conservation of vole predators should be implemented, most likely by developing programs for raptor monitoring and preservation of vegetation around nests during harvesting. Finally and interesting for a global reflection about alfalfas, most alfalfa production in CyL is aimed at export outside the EU (mainly to Arab countries and China), but the increase in the area covered by alfalfa in CyL to satisfy that external demand is promoting vole outbreaks, which management has caused biodiversity damage within the EU, particularly to the European Red kite population [86,133].

Besides alfalfa, there are other long-standing herbaceous crops that greatly promote vole abundance at landscape level in our study area and suffer damage, namely forage crops such as vetch (Vicia sativa Linn. 1753) and sainfoin (Onobrychis viciifolia Scop.) [38]. Another crop with a crucial role promoting vole outbreaks is winter rape (Brassica napus Linn. 1753), an herbaceous crop that provides refuge and abundant high-quality food during the critical winter months, and that also serves as a source habitat promoting invasion of adjacent crops during the next spring [134,135]. The area covered by this crop has been steadily increasing in CyL over the last decades (9-fold increase in cultivated area between 2007 and 2022) and more recently, as a result of the war in Ukraine [136].

In contrast, traditionally managed cereals, which are the dominant crops in the areas suffering vole outbreaks in CyL (mainly barley Hordeum vulgare Linn. 1753 and wheat Triticum spp.), are considered a suboptimal habitat for voles. The annual cycle of agrarian disturbances in these crops, which cause sudden and deep habitat changes, precludes permanent settlement and population growth (tilling and summer harvesting with no subsequent growth of vegetation until autumn) [38,39,40,42,65,66,120,121]. In most years, cereal plots are rarely occupied by voles beyond a few meters adjacent to field margins, but in vole outbreak years, cereal plots can be widely occupied far from the field edges [122]. Invasion of cereal crops during early stages of plant growth may cause little damage, as cereals are grasses that adapt well to grazing, re-growing vigorously when facing herbivory. Therefore, damage to cereals is often negligible, well below 2% of expected production [115]. In fact, an old farming style in many areas of Spain included moving extensive sheep herds into recently sprouted cereals or alfalfa after the last autumn cut. The crops did not suffer any significant damage, were fertilized by sheep dung, and the sheep had access to high quality food in a win-win agreement between farmers and shepherds. Given that livestock roaming reduces vole abundance (see Section 2), this management measure would also help to control common voles at the critical time of early cereal growth or on the crop holding maximum abundances in autumn. It is also considered in the new strategy implemented by ITACYL [113] (see Table S1). However, when vole numbers within cereal parcels are high during late spring and summer, damages in cereal plots can be very high too, even >90% of loss [123]. Paradoxically, in CyL, vole outbreak years tend to coincide with an overall high cereal production, because both the pest and the crop respond in a similar way to high spring precipitation, a key factor promoting cereal production in a Mediterranean climate [74,137]. A remarkable example is 2007, a year that was among those with the highest cereal production in CyL despite the development of one of the most important vole pests known to date. Thus, although voles may cause serious damage to cereal crops at local level, at regional level, damages to cereals are negligible [74].

Finally, the advent of low-tillage management of cereal crops has been another change favoring the development of vole outbreaks because, under this management regime, cereal fields become patches with less disturbed ground, allowing the development of long term vole burrows and thus dense populations that are never detected in traditionally managed cereals [38,53,138]. This modern management of cereals is also extending notably over CyL, particularly in TC. Surface, shallow tillage of non-tillage parcels is recommended as a preventive measure to reduce pest development risk in the new IPM strategy implemented by ITACYL [114] (Table S1).

As reported above (Section 5), we have found marked differences in the efficacy of biological control between study areas, and this may largely be explained by differences in landscape structure, crop composition and management of alfalfas (

Table 1. Study areas where efficacy of biological control has been evaluated, ordered from highly successful (Navarra) to negligible effect at landscape scale (BC). We show which study area had the maximum, minimum, and intermediate areas covered by different crops favorable for voles, as well as a similar ranking for landscape heterogeneity and the management regime of alfalfa cultivation.

Study Area	Alfalfa	Alfalfa Management	Other Crops	No-Tilling Cereal	Natural Vegetation	Landscape Heterogeneity
Navarra	Mid	Flooding	Min	Min	Max	Max
VC	Mid	Sprinklers	Mid	Min	Max	Mid
SM	Min	Dry	Mid	Mid	Mid	Min
BC	Max	Dry	Max	Max	Min	Min

). The study area in central TC (BC), where efficacy of biological control was negligible at landscape scale, had the “perfect storm” to promote vole outbreaks: the highest areas covered by non-irrigated alfalfa, other vole-producing crops (forage, rape) and non-tilling cereal in the most deforested and homogeneous landscape of all study areas (Table 1). On the other hand, the area where biological control was more successful, Navarra, the opposite was found: moderate area covered by alfalfa under flooding regime, low presence of other crops favorable for voles and maximum landscape heterogeneity with abundant natural vegetation near crops (Table 1). Furthermore, in the area where biological control worked better in CyL (VC, Table 1), farmers used to till irrigated alfalfas shortly before or during vole outbreaks, enhancing the efficacy of biological control [102].

Heterogeneous agricultural landscapes reduce the risk of rodent outbreaks, most probably because they hold higher diversity and abundance of rodent predators, enhancing the effectiveness of biological control [139,140,141]. A key component of landscape heterogeneity is field size [142]. As pointed out above (Section 2), the increase in field size occurred in CyL most likely favored the expansion of common vole into the Duero Valley, and Czech researchers have recently demonstrated that common vole density increases with field size [142]. Therefore, reducing field size may be added to the toolkit of any IPM program for common voles, and the recent support by EU to reduce field size in order to increase landscape heterogeneity to protect biodiversity conservation may have an additional value as a rodent pest control system [142]. In this line, the new IPM strategy in CyL includes promoting the conservation or expansion of islands or strips of natural vegetation within large crop fields [114] (Table S1). This technique is also promoted by GREFA in some of the study areas of the biological control program (VC, Villafruela, Campo de San Pedro and Valdefinjas), as a complementary measure to restore biodiversity in intensive agrarian habitats.

8. Field Margins and Common Weasels (Mustela nivalis Linn. 1766) as Key Targets for Management

The most common habitat for voles in agrarian landscapes is obviously the one that occupies the least total area: field margins, less than 5% of the total agricultural area [122]. Field margins work as a dispersal network and main source of crop invasion during outbreak years [56,122]. They also act as a refuge for voles that survive when agrarian management causes perturbations in crop fields (harvesting and tilling), so there appears to be a regular exchange of individuals between field margins and crops depending on perturbations in both habitats [49,120,121,122]. Despite the small area covered by field margins, they make up a wide network giving access to every single crop field in the landscape, which promotes simultaneous invasion of many fields during outbreak years. There is a wide variety of field margins in terms of width, vegetation cover and typology (e.g., culverts, riversides or edges between crop fields) which means a high variability in habitat quality for voles too (e.g., [38,42,71,122]), but margins of alfalfa fields use to hold the highest vole abundance [42]. Although field margins may remain undisturbed for many years, in other cases, they are managed to remove vegetation and stabilize slopes, particularly when they are culverts [71]. Given the crucial role that field margins have in the invasion process during outbreaks these linear habitats have been a major target of pest management at least since the 2007 outbreak, when ARs and fire were applied over thousands of km of margins (Section 4 and Section 6 above). Management of field margins has persisted through the most recent vole outbreaks by removing vegetation and disturbing ground in thousands of km of road and track culverts with motor graders. This management rarely respects the scarce natural woody vegetation, even though these jobs are always promoted by regional and provincial administrations, and the ITACYL strategy of 2019 specifies this requirement because of its importance for the EBPM [114] (Table S1).

Interestingly, field margins are almost the only habitat used in agrarian landscapes by the main specialist terrestrial predator of voles acting within burrows, the common weasel. Weasels are also very likely the main predator of vole litters, and could therefore play an important role in regulating vole numbers [143,144]. In Italian agrarian landscapes with similar habitats and climate than our study areas, home ranges of weasels determined by radio-tracking have a reticular shape matching the spatial distribution of field margins and they use only a small fringe around the margins within crop fields [144]. This pattern of habitat use has also been confirmed in our study areas of CyL by trapping, finding that weasels used mainly used well-vegetated field margins and were rarely caught outside them [145]. Weasels may regulate vole numbers not only by directly preying on individuals but also by creating a “landscape of fear”, by inducing considerable stress and disruption of the social structure of vole populations, ultimately reducing breeding activity and recruitment [146,147,148,149]. When predation risk from both avian predators and weasels is high, voles face a behavioral dilemma, because margins with good vegetation cover protect them from aerial predation, but are the main habitat where weasels search for prey [150], as well as the main refuge for other terrestrial predators such as lizards or snakes.

Coupled population cycles between weasels and voles have been reported in CyL, matching what has been found in other cyclic vole populations in the Palearctic [57]. Weasels are almost undetectable by trapping during the low and growing phases of vole cycles, but reach maximum densities during peak phase (8–12 captures/100 traps/24 h) and then experience a fast population crash similar to voles [57]. However, weasel population growth was only slightly delayed with respect to the growth of vole populations, so weasel predation would not be a delayed-density dependent factor producing the cycle in the prey species, and weasels could not be considered as drivers of the cycle in NW Spain [57].

Available research about the role that field margins play in the development of vole outbreaks has produced highly contrasting results. Some authors argue that higher areas covered by field margins, in general, or areas covered by margins with high quality for voles (wider, thick vegetation) promote vole outbreaks [111,151]. Other authors, however, found no significant effect of agri-environmental flower strips on the growth of vole populations outside the strips [152] or even found lower invasion of fields near wide margins with well developed vegetation [153,154]. Little is known about the reasons behind these contrasting results.

In NW Spain, several lines of evidence support that weasels can be a relevant regulatory factor of vole numbers in relation to field margins:

−

The total area covered by margins in a given study plot 12.5 ha wide was positively related to vole invasion of crop fields, but the area covered by riverside margins was associated with lower vole invasion [38]. Most field margins in our study areas were narrow strips (usually <1 m width, often <50 cm width) with sparse vegetation that are commonly colonized by voles, but rarely used by weasels, which select highly vegetated margins [145]. In contrast, riverside margins were wide strips (often 2–6 m) with full cover of thick vegetation, excellent habitat for weasels. Thus, the best hypothesis to explain contrasting effects of quality of field margins on vole abundance is the differential use by weasels and their regulatory effect on vole numbers [38].

−

At the peak phase of the cycles, vole abundance was higher in margins with high cover and short vegetation, while margins with low cover but taller vegetation, usually associated with higher stability in margins, were related to lower peak abundance [71]. The most parsimonious hypothesis to explain this result is again differential use of margins by weasels depending on vegetation and their role regulating vole abundance in margins.

−

In a study specifically designed to test the degree of invasion of cereal fields depending on the traits of field margins during a vole outbreak year, again wide and well-vegetated margins produced lower vole invasion [155].

−

In cyclic vole populations, individuals of large size, particularly males, appear in vole populations during the final growth and peak phases, what has been termed the “Chitty effect”. The Chitty effect was found in vole cycles of CyL; however, for the first time, we found an effect of habitat on the probability of appearance of these large-sized individuals: alfalfa and narrow field margins had the highest abundance of large voles, while large-sized voles were rarely found in riverside margins [54]. Once again, the habitat more used by weasels had reduced populations of large-sized voles, while habitats hardly used by weasels had high abundance of them.

In summary, our results support that weasels may be an important regulatory factor of vole abundance in undisturbed, well-vegetated field margins, reducing the probability of invasion of adjacent crops. Consequently, the best management to be applied in field margins would be to promote this kind of undisturbed margins. Paradoxically, the new strategy by ITACYL promotes many actions to promote presence and abundance of vole predators while also removing vegetation from field margins [114] (see Table S1). In fact, during last three vole outbreaks, with this strategy already on its way to being implemented, ITACYL has managed thousand of km of culverts during and after vole peaks, when these actions are neither profitable nor advisable, and this is the main point to be improved in future management of vole pests. It is important to note that field margins have a crucial importance for agrarian biodiversity, given they are among the few patches of natural vegetation in modern, intensive agrarian landscapes, providing other relevant services to agriculture, such as pollination or insect pest control (e.g., [156,157,158,159]). Consequently, the implementation of such strategy would be favorable not only for vole control, but also for the conservation of agrarian biodiversity promoting ecosystem services important for agriculture. This would include favoring game species of economic importance, such as the red-legged partridge (Alectoris rufa Linn. 1758), which has in field margins the most common and successful nesting habitat in agricultural landscapes of Spain [160].

9. Concluding Remarks and Suggestions for Future Research and Adaptive Management

Environmental NGOs and the hunting sector are usually opposing stakeholders with quite different opinions about wildlife management, if not declared enemies. Most likely, this is an increasing trend in the current polarized world full of internet trash [161]. Paradoxically, both sectors may have clear common interests, at least in agricultural ecosystems: enhancing overall biodiversity in agrarian landscapes will usually entail increasing numbers of game species. The case of vole outbreaks in CyL is a notable example of collaboration among those opposing sectors to defend common interests by successfully demanding changes in agricultural management from the authorities in charge of pest management. Scientists have an important but challenging role promoting this collaboration, by providing balanced and well-grounded recommendations or opinions [161]. Common pressure from the three sectors will always be more effective than working independently. Hunters and environmentalists should always set aside their differences, becoming collaborating “dear enemies” when necessary and favorable for both sectors, paying more attention to scientific facts, instead of beliefs [161].

Past vole pest management in CyL developed large-scale actions that were harmful to the environment when vole populations were at their peak or in decline phases of the cycles, so when it was no longer needed to take action because vole populations were already declining quickly or near to do so soon. Interestingly, since the earliest reports by the Spanish Ministry of Agriculture about vole pest management, it was clearly recognized that “voles disappear in a similar way from areas treated or untreated with rodenticides” [34,74,79,162], obviously, because treatments were applied during the peak or decline phases. Therefore, management during peak/decline phases has chronically been a clear waste of resources, causing unnecessary environmental damage, in particular increasing the poisoning risk of predators, the best natural allies in vole control [75]. Vole outbreaks should always be treated in a preventive way avoiding actions with full developed pests [83,163]. It should never be forgotten that declines in the populations of rodent predators caused by chemical control might create a vicious circle impairing the problem and increasing future needs for more aggressive treatments [100,164].

Unfortunately, management of the main habitat source of voles that invade crops—field margins—has been still performed during peak/decline phases in road or track culverts during last outbreaks. Certainly, culverts require regular maintenance works, but as detailed above, the utility of this management to control vole numbers or crop invasion is highly dubious, if not counteracting, because management in field margins can simultaneously cause abandonment of the margin by weasels and invasion of adjacent crops by voles [165,166]. Furthermore, voles recolonize managed field margins long before weasels [166]. Increasing the frequency of maintenance works in culverts beyond what is needed to keep the infrastructure in good shape should be avoided, particularly during peak/decline phases of vole cycles. On the contrary, the promotion of wide and well-vegetated margins should be more clearly implemented in the ITACYL vole control strategy [114] (Table S1). JCYL should use the main tool that governments in EU countries have, the CAP (Common Agricultural Policy), to achieve extending field-margin preservation and development, at least as an agri-environmental measure.

Forecasting the arrival of the peak in a vole cycle may be possible with current knowledge and monitoring areas established by ITACYL [55]; although precise determination of peak time may be sometimes tricky from scientific or management points of view, because density during vole peaks can be highly variable [71]. However, there is a complementary predictive tool available in the form of the Chitty effect mentioned above (see Section 8). When large-sized voles appear in the vole population under monitoring, we can be sure of the soon arrival of the peak phase and thus the subsequent quick decline in numbers, and this happens in both high- and mid-density outbreaks [54].

All available research summarized above indicates that the area covered by “vole-producing” crops, particularly alfalfa, promotes high-density, large-scale vole outbreaks difficult to control by any method. Thus, reducing the area covered by alfalfa at the landscape level would be the easiest way to reduce vole pest risk. This is precisely what happened in the Czech Republic, one of the countries in Central Europe that has had more problems with vole pests since ancient times [111]. The national area cultivated with alfalfa in that country declined more than 50% between 1995 and 2016 and “definitely, the problem [of vole pests] is not so burning as before” [167]. The burning problem is now in CyL, where area covered by alfalfa more than doubled between 1960 and 2000 [37]. In CyL, programs to promote a reduction at landscape level in the area cultivated with alfalfa or other “vole-producing” crops would be particularly important where voles cause more problems, in TC, and they should be seriously considered by local farmers and JCYL.

Innovative control methods in field margins should be tested (e.g., [128]). During the 2007 outbreak in CyL, farmers used “water traps” grounded where voles fell and died drowned, reporting high catch success. By testing this method, we observed that traps in margins, and particularly corners of alfalfa fields certainly caught many voles, but unfortunately, also many insects and other small mammals [168]. Aulicky et al. [111] report a similar system with ditches and holes that has been tested in the Czech Republic. Preventive regular management of margins of alfalfa fields with livestock could also be an interesting option potentially reducing effective dispersal from the main source populations through the field-margin network (see Table S1), probably combined with preservation of well vegetated margins useful for weasels in the fields around alfalfa.

Additional research on the factors explaining high variance in efficacy of biological control, at large scale and long term, should be developed. The coordination between GREFA and ITACYL, now collaborating in the biological control program and with a large-scale vole monitoring system working [55], would allow such crucial investigation.

The biological control program should extend further in TC, the most conflictive area in CyL, both adding new municipalities and increasing nest box density where this could allow a local increase in predator density. In the current situation, the area within TC with nest boxes is still much lower than the area without them (compare Figure 1B and Figure 4) and we know that there is intense effective vole dispersal during the growth phase of the cycle across TC [49]. Therefore, voles in TC could be in a metapopulation system with sink areas of increased predation pressure but larger source areas without it, rendering biological control at landscape scale ineffective. The idea would be to test if spatially extending continuous areas with maximum possible increased predation pressure may improve the efficacy of the system at landscape level.

Finally, what we feel may be one of the future keystones in vole management, the weasels, are unfortunately poorly known yet. Additional studies on demography, ecology, and behavior, including experimental work, may be of crucial importance. The weasel captive breeding facilities managed by GREFA and the starting project about refuges for weasels previously tested with captive animals may be a good start to such a program [169] (see Table S1).