Rodent management and cereal production in Asia: Balancing food security and conservation

Rodents present a major problem for food security in Asia where smallholder farming families are particularly vulnerable. We review here recent developments in the biology and management of rodent pests in cereal cropping systems in Asia. The past decade has seen a strong focus on ecologically-based rodent management (EBRM), its adoption in ﬁ eld studies signi ﬁ cantly increased rice yields (6 – 15%) and income ( > 15%) in seven Asian countries. EBRM principles have also been successfully applied to maize in China. We provide case studies on EBRM in Cambodia, on interactions between rodent pests and weeds, and on the importance of modi ﬁ ed wetlands for biodiversity and rodent pest management. Knowledge on post-harvest impacts of rodents is increasing. One research gap is the assessment of human health impacts from a reduction of rodent densities in and around houses. We identify 10 challenges for the next decade. For example, the need for population modelling, a valuable tool missing from our toolbox to manage rodent pests in cereal systems. We also need to understand better the interactive effects of cropping intensi ﬁ cation, conservation agriculture and climate change. Finally, new management approaches such as fertility control are on the horizon and need to be considered in the context of smallholder cereal farming systems and mitigating health risks from zoonotic diseases associated with rodents.


INTRODUCTION
In his summing up of an international conference on rodent biology and management in 2006, C.J. Krebs 1 presented the following message: '...progress in ecologically based rodent management in agricultural settings in Asia and Africa is truly impressive. If there is a general message it is that we need a varied toolbox for rodent pest problems. ' Despite this report of impressive progress in our knowledge of managing rodent agricultural pests in developing countries, rodents still present a major problem for food security at a global level. 2 The focus of this review is Asia where smallholder farmers are particularly vulnerable. 3 The impacts on food security are even more pronounced in years when parts of Asia experience episodic population outbreaks of rodent pests. 4 Ecologically-based rodent management (EBRM), a widely accepted paradigm for rodent pest management, builds on ecological, taxonomic and socio-economic studies to develop integrated ecologically-based approaches to rodent pest management. [5][6][7] A solid understanding of the species composition and the biology of the pest species in a specific agro-ecosystem enables the identification of optimal times, location and scale of action to develop management strategies that are cost-effective and minimize environmental harm by reducing reliance on rodenticides.
In this paper we review developments in our understanding of the biology and management of rodent pests in cereal cropping systems in Asia over the past decade. We cover progress in both pre-and post-harvest management of rodents. Most of the research focus has been on rice production systems, therefore, we consider developing trends in rice production that are likely to add further dimensions to future studies of rodent pest management. The inter-linking of rodents and other pests and diseases of rice, plus the likely impact of rodents or the perceptions by farmers that may impede progress on the need to better manage rice landscapes to firstly, mitigate greenhouse gas emissions, and secondly, as important wetlands to preserve biodiversity, will be considered. We build on previous reviews that capture the 'impressive' progress of EBRM in agricultural systems 8,9 since its reinvigoration in 1999. 10 We will not cover rodent-borne zoonoses, which is a rapidly developing field of epidemiology after languishing until the late 1990s. Diseases such as leptospirosis, murine typhus, human plague, Lassa fever and scrub typhus are indeed major concerns in rural communities in Asia and Africa, 11 and a recent outbreak of pneumonic plague in Madagascar in 2017-2018 has led to a 'call for action'. 12,13 We simply make two points. Firstly, that the economic impact of the effect of rodent zoonoses on smallholder rural communities is not well documented and requires much more research effort. Second, that well designed studies that combine rodent ecology, crop production and epidemiology of potential rodent zoonoses in human populations in an agricultural landscape are a priority. Such research would support a laudable effort to include an agro-ecological approach to protect crops from vertebrate pests in an extended 'One Health' concept because of limited but promising evidence that such an approach reduces viral zoonoses. 14

RICE AND RODENTS -AN IMPORTANT FOOD SECURITY ISSUE IN ASIA
Rice is the staple crop for most Asian countries 15 and rodents are a major pest in both lowland irrigated and upland rainfed rice cropping systems. 16,17 In the lowland systems that are the rice bowls of Asia, there is pressure to increase the intensity of rice production and to reduce post-harvest losses because of the imperative to provide food security to a rapidly growing human population. Indeed, FAO 18 estimates that rice production would need to be increased by 24% from 2005/2007 levels to provide food security for the projected 9 billion people in 2050.

Lowland irrigated rice systems
Intensification of annual cropping by transitioning from one crop of rice to two crops, two crops to three crops, or two crops of rice followed by a non-rice crop (e.g., maize or pulses), are all likely to increase the impacts of rodent populations on crop yields given the ability of the major rodent pest species in Asia to increase their annual breeding output when high quality food is available for more of the year. This includes Rattus argentiventer (most SE Asian countries), [19][20][21][22] Rattus tanezumi (Philippines) [23][24][25][26][27] and Bandicota bengalensis (Myanmar and South Asia). [28][29][30][31][32] We review the literature on rodent ecology and management, including socio-economic studies, in cereal production in Asia over the past decade (Table 1). We selected this time frame because the last major review of the impacts of rodent pests in Asia, be it in the context of rodent population outbreaks globally, was in 2010. 56 We do not review specific post-harvest studies because a comprehensive review has been recently published (see Brown,et al. 55 ). However, in section 2 we will provide some insights into a path forward for post-harvest management of rodent pests in intensive cereal production regions.
In this section we provide insights to what has emerged from the studies over the past decade. We consider some of the studies in more detail as case studies because of their interesting findings. An encouraging development is that there has been a stronger emphasis on obtaining quantitative estimates of economic impacts of rodents in rice cropping systems. In most cases these estimates are based on replicated field research with appropriate controls. The adoption of EBRM produced significant increases in yield (typically 6-15%) and/or increases in income (>15%) ( Table 1). There are two studies where the focus was on areas of high rodent impacts and the application of integrated EBRM options. In both instances, the yield and economic benefits from applying these actions are very encouraging. In Indonesia, Herawati and Purnawan 52 reported a mean yield increase from 1.6 to 6.4 t/ha for seven farmers within a 40 ha rice production area where best management practices for rodent pests were implemented. In Cambodia, Stuart, et al. 54 reported yield increases of 20 to 32% and increased income of 53 to 169%. We will consider the Cambodian study in more detail as a case study.
Although it is encouraging to have quantitative data on the benefits of EBRM from five Asian countries, some reported only on yield increases, crop losses or economic benefits. For R. argentiventer, crop losses to rice estimated immediately prior to harvest are known to under-estimate yield loss by 3-6 times. 57,58 We encourage future studies on economic impacts of rodent management actions to pay careful attention to the relationship between in-crop losses and estimated yield reduction, and to develop where possible estimates of benefit-cost ratios. Nevertheless, the published literature over the past decade indicate that the economic benefits are robust across countries and for different pest species (Table 1). In rice crops, Bandicota species are the main pests in Myanmar and Bangladesh, R. argentiventer in Indonesia, Vietnam and parts of the Philippines (but not Luzon), Rattus rattus complex in Laos, R. tanezumi in the northern part of the Philippines, and a combination of species in Cambodia and Thailand. There are reports from most of these countries in Table 1.
The intensification of rice-rice and rice-'other crop' systems in Asia, particularly Southeast Asia, often leads to asynchronous cropping within a local geographic area. Rodent pest species increase their reproductive output in response to the presence of a patchwork of crops in rice-cropping systems. 41,59,60 Two papers from Vietnam highlighted that the intensification of rice cropping in the Mekong River Delta, where three rice crops are often grown in one year, is another plus for more damage by rodent pest species. 38,46 In both instances the authors recommend that only two rice crops be grown each year, with a different crop grown when the third rice crop (Vu3) is grown. In addition, previous studies 17, 21 highlight the need to avoid asynchronous cropping of rice if possible.

Case study 1: ecologically-based rodent management (EBRM) in Cambodia
In Cambodia, an applied research study by Stuart, et al. 54 successfully brought together lessons learned from the previous 20 years of EBRM research in Southeast Asia to effectively manage rodent pests of rice, whilst at the same time reducing the use of hazardous rodenticides and electric fencing to protect crops from rodents. The authors of this study began by holding focus group discussions in two villages with rice farmers who had indicated that rodents were their main pest of concern. As previously demonstrated, 35 effective participatory adaptive research on EBRM includes discussions with farmers on understanding the rodent situation across the village rice-growing landscape, and then involving the farmers in decision-making to select the appropriate rodent management practices to implement and evaluate. These meetings were then followed by farmer participatory research involving the integration of several management methods that had been previously demonstrated elsewhere, for example, the Trap Barrier System (TBS). 61 The methods were carefully developed based on the ecology of the main pest species, Yesfour agricultural cooperatives; initially two control cooperatives but these quickly adopted community management actions. Analyses therefore were based on before and after EBRM implementation, Palis, et al. 35  the local conditions and the farmers' willingness to test them at a community level (5 ha) as most farmers had field sizes of less than one hectare. In one village, two rice crops were planted per year, one rain-fed crop in the wet season and another irrigated crop in the dry season. The latter was planted on a lake margin as the lake waters receded. EBRM in this village involved the use of a linear TBS (LTBS) along the edge of the rice fields and carefully targeted anticoagulant rodenticide use. In the second village, rice was grown intensively with three crops per year and the EBRM approach involved a community TBS plus trap crops (CTBS) as well as a LTBS alongside potential rodent refuge habitat. Other important EBRM activities such as synchronous planting, community rat hunting and field sanitation at key times were also conducted in both village treatment areas.
As a result of the EBRM approaches that were tailored to the local conditions, rodent damage was reduced by 84-99% and incomes were increased by 53-169% as compared to nontreatment farmers who continued business as usual. The study was thus a great success, and many farmers adopted the technology following the end of the project. However, most of these farmers decided to use a TBS around their entire field rather than strategically place LTBS at targeted locations or implement CTBS. The authors were informed by the farmers (Stuart, pers. obs.) that this approach ensured the greatest level of protection to their rice as the risk of rodent damage was high enough to compensate for the added costs of more fencing. This, after all, was the approach they had previously applied when using electric fencing that also surrounded their whole crop. Thus, even though the study is a good example of the economic benefits of EBRM, it highlights the challenges of changing farmer mind-sets and conducting EBRM at a community level. However, the response is highly positive because the adoption of EBRM replaced the illegal action of using electricity to protect their crop from rodents. The use of electricity presents a lethal health risk to humans, livestock and companion animals.

Interactions between non-pest and pest rodent species
There is evidence that native rodent species may be able to compete with invasive rodent species in less-disturbed natural ecosystems, especially if the native species became residents first in a particular location. 62 A number of recent studies conducted in Northern Luzon, Philippines, indicate that non-pest native rodent species will recolonize regenerating forest habitat, displacing potential pest non-native species. 48,63,64 Non-native species (R. tanezumi and R. exulans) were confined to heavily disturbed habitats, where the diversity and abundance of native rodent species was low. These findings suggest that these invasive species are unlikely to become established in intact forests where native rodent species are present because it appears that 'natives have competitive superiority'. 64 This idea merits further investigation.

Case study 2: the interactions between rodent pests and other cereal pests
In Asia, research on the management of pests of cereal crops has an extensive literature covering insect, disease, weed and rodent pests. 65 We have not included studies on post-harvest impacts in rice because they were reviewed in Brown, et al. 55  wileyonlinelibrary.com/journal/ps examined potential benefits of integrating rodent ecology and management with the management of other cereal pests. We present a case study of rodent and weed interactions in intensive rice cropping systems as an example of why such interactive research is important.
The impacts of rodents on rice crops and food security have been discussed in section 1. Weeds arguably have the highest contribution into losses to cereal production globally. 67,68 Herbicides and hand weeding provide management options for smallholder farmers, however, crop losses for rice remain around 10%. 67 There have been undocumented reports that where weed infestation of crops is high, rodent densities also tend to be high. 69 Likewise, where there has been high rodent damage to rice that opens the crop canopy, weed infestations appear to be high. In a replicated study in the fields of farmers in Myanmar, there were four treatments: rodents were excluded but not weeds, weeds were managed but not rodents, both rodents and weeds were managed, and neither were managed. Each of the 16 plots (4 treatments x 4 replicates) were 0.25 ha. The study was conducted across both dry and wet cropping seasons. 28 The study reported additive negative effects of rodent and weeds on crop yield, with the combined effect highest in the dry season. A clear recommendation from these findings is the need for concurrent weed and rodent management, especially during the early crop stages of rice production. 28 Such an interaction between weeds and rodents may not be unexpected but prior to this study, quantitative data were lacking.
Further research is urgently needed to quantify the interactions between rodents and other cereal crop pests, so that cost effective integrated management options can be developed and promoted. Why the urgency? The trends for increased intensification of rice cropping, moving from one rice crop to two, and two to three per year, is highly likely to increase the impacts of pests on yield and indeed rodent pests may exacerbate the losses caused by other types of pests and diseases. Moreover, increased impacts of cereal pests associated with climate change 70 and extreme climate events 41 are likely to provide unexpected interactions between insect, weed, and rodent pests and also crop diseases. The better we understand these possible interactions then we will be in a stronger position to anticipate them and to provide integrated management recommendations to assist smallholder farmers.

Widespread geographic adoption of EBRM, including lowland non-irrigated cropswheat and maize
Given that rice is by far the main cereal staple in Southeast Asia, 15 it is not surprising that most of the published literature on rodent management in agricultural systems has concentrated on rice (Table 1). In the decade beginning in 2009, there have been an impressive number of papers published on EBRM in China. Part of this focus is related to an increased awareness in China of the risks associated with environmental pollution and threats to nontarget biodiversity from the use of pesticides such as rodenticides. 50,71 The trap-barrier system (TBS) was initially promoted by Lam 72 to protect irrigated rice crops in Malaysia and later was modified to include an early planted 'trap-crop' to attract rodents to the multiple-capture traps set into the rectangular fence/barrier to protect rice-cropping systems in Indonesia, 61 the Philippines, 73 Vietnam 74 and elsewhere in Southeast Asia. 8 In China, wheat and maize are also important cereals. The TBS in a rectangular and linear form has been used successfully in wheat, maize, rice and other crops, resulting in increases in grain yield.
Apparently, the TBS is widely adopted in cropping systems in China but most of the publications are in Chinese (see Wang,et al. 50 ). This development of greater adoption of EBRM of rodent pests in China and in non-rice crops is indeed a positive development. Moreover, the TBS has been successfully adopted as a tool for EBRM in rice crops in Eastern Africa. 75 Although the geographic spread of EBRM over the past decade is impressive, and impacts at a local scale have been reported, 34,35,54 quantitative data on the number of farmers who have adopted EBRM and the area of coverage at a district, region or provincial level is limited. More such data are required to influence policy makers to increase support for EBRM.

Upland rainfed systems
Over the past decade there has been little progress except for better understanding of the life cycle of specific species of bamboo and the role they may play in rodent outbreaks. However, the lessons learned from the outbreaks in Chittagong in Bangladesh, Mizoram in India, and Rhakine and Chin States in Myanmar are well covered in Singleton, et al. 56 In Bangladesh, there has been a first study to look at management options of rodent pests in the Chittagong Hill tract region. 76 This is to be commended given the logistics for field work in this region is challenging, and because many of the farmers rely on subsistence farming.
In the upland rice habitats of Laos, there has been one study of the social-economic aspects of rodent pests via farmer interviews, 77 and another impressive study that reported positive impacts from a combination of EBRM activities and the use of bio-control using the protozoan Sarcocystis singaporensis. 45 In upland cereal systems, rodent impacts generally are episodic and severe. 16 Long term studies are desperately required in these farming systems to provide a data set that can be used to develop predictive models of rodent outbreaks. Long term data sets have provided the basis for models that have been successfully used in agricultural landscapes to assist with the timing of management of common voles, Microtus arvalis, in Germany, 77 and house mice, Mus domesticus, in Australia. 78

POST-HARVEST IMPACTS OF RODENTS -A WAY FORWARD
In this paper, we have not systematically reviewed post-harvest impacts of rodents on cereal crops in Asia. Until the early 2000s there were very few publications on post-harvest losses. And these few papers provided general qualitative descriptions on losses and recommended management actions (e.g., Meyer 79 ) rather than quantitative data obtained from replicated studies with appropriate controls. Encouragingly, after a long lapse in research on this topic, there have been several quantitative studies in Southeast Asia and elsewhere since the 2000s. A recent review by Brown, et al. 55 provides a detailed coverage of recent studies on post-harvest losses by rodents to cereal crops in Asia and recommended management actions.
A clear message from recent studies is that for a particular region, generally the rodent pest species in grain stores differ from those in the field. Again, this emphasizes the need to understand the population ecology of the particular rodent species that are causing a problem. We cannot simply extrapolate from our knowledge of the population and breeding dynamics of a major pest species of rodent that causes pre-harvest losses to manage rodents that are feasting on the freshly harvested cereals in nearby grain stores. Effective management of post-harvest rodent losses to cereals in stores include improved hygiene in and around the stores and nearby buildings, improved rodent proofing of grain stores that is affordable to smallholder farmers, community programs to highlight the best time to manage rodents (usually related to the breeding and spatial ecology of the pest species and the timing of harvest), and an awareness of social and religious factors that may influence the involvement of the farming community in proposed management actions. We provide two examples of the latter. Firstly, studies in Bangladesh and Myanmar indicate that the sharing within a community of kill-traps on a rotational basis can significantly reduce rodent losses. 80 However, in countries such as Myanmar where Buddhism is a dominant religion, the killing of animals is not readily accepted by farming communities. Second, a study in South Africa indicated that promoting a landscape of fear for rodents by encouraging predators such as cats and dogs, restricted the movements of rodents (decreased their access to grain stores), although the population density of the rodents was not affected. 81 In countries such as Indonesia where the Muslim religion is dominant, promoting dogs in villages would not be acceptable.
The review by Brown, et al. 55 highlighted the significant negative impact that rodent losses post-harvest can have on food security for smallholder communities. The progress of the past 10-15 years has been impressive in assessing the level of damage and the risks to food security. More now needs to be done to evaluate critically the effectiveness economically, environmentally and socially, of recommended management actions. One important component that is lacking from post-harvest studies is a quantitative assessment of the human health impact from a reduction of rodent densities in and around houses in rural communities in Asia. Given that rodents are carriers of many zoonoses 8 then we would expect there to be major health benefits.

IMPORTANCE OF MODIFIED WETLANDS FOR BIODIVERSITY -WHAT DOES IT MEAN FOR RODENT MANAGEMENT?
The essence of the United Nations Sustainable Development Goal 2 is to reduce hunger by producing sufficient food in a sustainable manner. This in turn links with Goal 15 which emphasizes the restoration of terrestrial ecosystems and halting biodiversity loss. 82 In Asia, there are 52 million ha of lowland irrigated rice. 15 Wetland habitats are at grave risk globally and as a result there are great concerns on the rates of loss of biological diversity. In 2018, the RAMSAR Convention on wetlands highlighted the loss of 35% of wetlands globally since 1970 1 , and the Living Planet Report 2020 by the World Wildlife Fund 83 estimated that since 1970, there has been a 45% loss in vertebrate biodiversity in Asia and Oceania, and that more sustainable agricultural production is an important intervention to redress biodiversity loss. Rice agricultural systems provide important human-modified wetlands for wildlife, 84-87 particularly in Asia. The challenge then is to increase rice production using methods that are environmentally, socially, and economically sustainable, which importantly includes effective stewardship of water use, crop margins and non-agricultural lands in the associated landscape.
4.1 Case study 3: what does increased water productivity mean for the management of rodent impacts in rice cropping systems? Alternate Wetting and Drying (AWD), also known as controlled or intermittent irrigation, is a technique developed to increase water productivity (amount of food produced per unit of total water input) in lowland irrigated rice, and to save on water for other purposes, or for further distribution along the irrigation system. 88,89 Just as importantly, the implementation of AWD that allows the rice field to dry but maintain sub-soil moisture within 15 cm of the surface, leads to substantial reductions in methane gas emissions by as much as 73% in the dry season and 21% during the wet season. 90 Rice farmers in Southeast Asia are hesitant to adopt this watersaving technology for fear the practice will lead to increased rodent pest activity, consequently exacerbating yield loss. Results of a study, which examined the effects of AWD on the population dynamics, habitat use, and damage levels inflicted on rice crops by Rattus argentiventer in Indonesia, and R. tanezumi in the Philippines, demonstrate that damage levels on standing rice crop were not affected by the water management scheme used, as shown by replicated damage assessments done on AWD and control fields. Rodent activity and movement, examined using spooland-line tracking, also was not influenced by water level. 53 AWD also had no significant effect on the breeding performance and population dynamics of these species.
4.2 Impact of rodent pests on efforts to promote ecoengineering and conservation agriculture in intensive ricecropping systems Like the AWD situation, farmer perceptions are important to maintain ecosystem benefits of crop margins. Firstly, the promotion of eco-engineering through growing plants that produce nectar along the banks of rice fields is used to increase biodiversity and thus attract predatory insects and arthropods. The rationale is that these predators will reduce the population density and impact of insect pests of rice. 91 The promotion of ecological intensification requires dense cover on the rice banks. Farmers can be hesitant to implement this practice because they perceive that such cover favors rodents (My Phung NT, pers. Comm.). If these habitats are targeted for community action for rodent management prior to crop establishment, which is recommended, 74 then a combination of early rodent management and the planting of nectarproducing plants can not only promote biodiversity but also has the potential to manage both rodent and insect pest populations.
Secondly, the edges of rice fields and associated riparian habitats favor non-pest rodent species as well as amphibian populations that have been documented to provide positive eco-system services in Asian rice cropping systems. 13,92,93 Again, we need to avoid perceptions that associated riparian habitats (e.g., reed beds, irrigation canals, small dams and banks of rice fields) will provide a benefit to rodent populations that cannot be effectively managed. A balanced approach is required. This may not be too large a problem given farmers in Luzon in the Philippines reported that they perceive amphibians as beneficial because they eat pest insects and that a healthy frog and toad population provides a positive indicator that the cropping system is healthy for humans. 94 Conservation agriculture in developed countries has altered the habitat use of some agricultural rodent pests, which then affects management recommendations. 95 Conservation agriculture is being promoted in Asia, but little is known about how this will influence the population dynamics and subsequent impact of rodent pests. For example, no-till agriculture, which conserves soil and water, can allow rodent pest populations to increase due to the accumulation of crop residue (which provides additional food source and insulation from climate extremes) and the continuity of intact burrow systems in the cereal-legume fields of Washington, USA. 96 No-till and high biomass mulch caused insect and rodent pests to overwinter in conservation agriculture plots in South Africa, which would then feed on maize seeds and seedlings. 97 Wheat farmers in India reported an increase in rodent depredation in zero-tillage crops. 98 These examples indicate a possible conflict between conservation agriculture and rodent pest management. We urgently need field research on conservation agriculture and the effects of such practices on pest rodent population dynamics. We contend that if potential conflicts are anticipated then appropriate management actions can be developed.

RECOMMENDATIONS FOR THE WAY FORWARD
In an agricultural context in Asia there are insect and weed pests that potentially have a significantly larger impact on rice production than rodents. However, as pointed out by Schiller, et al. 99 rodent pests, particularly in outbreak years, are often of high concern to farmers because they are recognized as a pest that they have the least control over. In upland rainfed systems, managing rodent population outbreaks is still a huge challenge, however, in non-outbreak years, and in lowland systems, there has been good progress in developing strategies to reduce rodent damage.
There are some management recommendations that apply to most if not all systems. These include synchrony of cropping (planting within 2 weeks of neighbors), 60 conducting community management prior to the onset of breeding and when rodents are aggregated in specific habitats whilst the fields are in fallow, improving hygiene around villages and reducing spilled grain post-harvest. Our review further confirms that rodent management approaches are dependent on the respective ecology and population dynamics of the main agricultural pests, on particular rice systems, on coordinated community action, and on the culture and beliefs of rural communities in different countries. Progress towards tailoring management for specific systems and rodent species continues to make impressive progress. As highlighted by Krebs, 1 for progress to continue along this path we require a 'toolbox' of management options and this requires researchers to embrace current and future trends in agricultural production and to communicate clear messages on how rodent populations are likely to respond to these actions, and to obtain objective evidence from replicated field studies in the fields of farmers to provide advice on management strategies.
We identify the following challenges to tackle (not in priority order) to assist smallholder farmers in Asia to manage their rodent problems. From our interactions with colleagues in Africa, many of these challenges also apply with EBRM showing promise to help redress the rodent impacts on cereal production pre-and postharvest. 6 Some of these challenges not surprisingly overlap with a recent global review that identified 10 essential questions for furthering our understanding of key population ecology processes associated with population cycles and outbreaks of small rodent populations. 100 • We require better estimates of crop and post-harvest losses to rodents, and economic analyses of EBRM, including livelihood benefits. This information is essential if policy makers are to be convinced of the need to broaden crop protection practices so that rodent management is more strongly integrated with recommendations for the management of agricultural pests and diseases. • There is still a paucity of knowledge on the biology, behavior and population ecology of most rodent pest species in developing countries despite the impressive progress over the past 20 years. In an agricultural context, the current generation of rodent biologists in Southeast Asia come with a background mainly in entomology. In Asian countries, the curricula for plant protection are usually dominated by entomology, plant pathology and, to a lesser extent, weed biology. It is rare for courses to be available on vertebrate biology. We need to continue to advocate for the importance of managing rodent impacts in cereal systems and mentor young scientists with an interest in rodent biology and management. • There is an urgent need for long-term data sets that would provide a solid foundation for developing predictive models for rodent outbreaks. These predictions are important in regions that suffer episodic outbreaks of rodent populations and need to trigger timely community management actions. Rodent management should be pro-active rather than reactive thus timely interventions are neededpopulation modelling is an important tool for our toolbox. • We need to explore species interactions between invasive and native rodents. Is it possible to foster 'biotic resistance' in the fight against invasive rodents? Can 'rewilding' to restore more natural habitat within agricultural landscapes reduce rodent pest populations through interspecific competition with native rodent species? 64,101 • Rodent communitiesthere is an important gap in our knowledge both in developed (see Parsons, et al. 102 ) and developing countries on what is happening at the rural-urban interface and cropping systems-forest interface. • Community action is the key to effective rodent management in Asian cropping systems where most farm holdings are less than 5 ha. It is encouraging to report that in our review of research from 2010 to 2020 there has been an impressive number that has a strong sociological component (Table 1). Capacity building of local extension specialists is vital to support smallholder farmers in developing effective community EBRM actions. 34,35 • We need to be able to anticipate how rodents will respond to changes in intensive production to meet increased food demands and therefore evaluate the sustainability of such systems. For example, three rice crops per year in the Mekong River Delta, Vietnam, lead to increased rodent impacts yet three crops have been reported as less profitable overall than two crops per year. 103 • Too little is known about conservation agriculture and the positive or negative effects of such practices on pest rodent population dynamics. • New management approaches are on the horizon. One example is fertility control of rodent pests. [104][105][106] Another, although controversial, 107 is synthetic gene drive technology that may provide an opportunity to substantially influence the sex ratio of a rodent population. 108 As well as potential socio-political opposition to the environmental release of gene drive technology to manage rodent pests, another essential consideration is how will wild rodent populations in agricultural landscapes respond to new methods of control. Will all species respond the same? • Climate change: A 46-year study of small mammal populations in the Boreal forest in northern Canada indicates that there have been long term shifts in the relative abundance of the four main species and marked changes in the energy flow in the herbivore populations contributed by the small mammals, and that these may be related to changes in the climate. 109 Unfortunately, we do not have long term monitoring of rodent populations in agricultural systems in Asia to assess effects of climate change at this level. However, the change of cropping systems associated with changes in climate and how rodent pest species respond to these changes needs careful study. An example is the Mekong River Delta where saline intrusion up the delta during the dry season is considerable and currently three provinces have greater than 80% of their cropland under salinity risk. 110 If salinity intrusion, coupled with flooding events, continues to increase because of climate change and sea level rise then that will likely lead to increased intensification of cropping of cereals and hence a likely increase in rodent impacts. Extreme weather effects can have indirect effects on rodent populations through changes in crop management. For example, Cyclone Nargis in the Ayeyarwaddy delta, Myanmar, devastated more than 738 000 ha of rice land. 59 Farming communities recovered at different rates and this led to an increase in asynchronous cropping and 15 months after the cyclone, there was a massive rodent population outbreak. 59 After such major climatic events, coordinated efforts are needed to avoid asynchronous planting. More quantitative data on the direct and indirect effects of climate change on rodent population dynamics is desperately needed to convince policy makers of the potential risks of rodent outbreaks for food security.