20 Apr 2017: Swanepoel LH, Swanepoel CM, Brown PR, Eiseb SJ, Goodman SM, et al. (2017) Correction: A systematic review of rodent pest research in Afro-Malagasy small-holder farming systems: Are we asking the right questions?. PLOS ONE 12(4): e0176621. https://doi.org/10.1371/journal.pone.0176621 View correction
Rodent pests are especially problematic in terms of agriculture and public health since they can inflict considerable economic damage associated with their abundance, diversity, generalist feeding habits and high reproductive rates. To quantify rodent pest impacts and identify trends in rodent pest research impacting on small-holder agriculture in the Afro-Malagasy region we did a systematic review of research outputs from 1910 to 2015, by developing an a priori defined set of criteria to allow for replication of the review process. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. We reviewed 162 publications, and while rodent pest research was spatially distributed across Africa (32 countries, including Madagascar), there was a disparity in number of studies per country with research biased towards four countries (Tanzania [25%], Nigeria [9%], Ethiopia [9%], Kenya [8%]) accounting for 51% of all rodent pest research in the Afro-Malagasy region. There was a disparity in the research themes addressed by Tanzanian publications compared to publications from the rest of the Afro-Malagasy region where research in Tanzania had a much more applied focus (50%) compared to a more basic research approach (92%) in the rest of the Afro-Malagasy region. We found that pest rodents have a significant negative effect on the Afro-Malagasy small-holder farming communities. Crop losses varied between cropping stages, storage and crops and the highest losses occurred during early cropping stages (46% median loss during seedling stage) and the mature stage (15% median loss). There was a scarcity of studies investigating the effectiveness of various management actions on rodent pest damage and population abundance. Our analysis highlights that there are inadequate empirical studies focused on developing sustainable control methods for rodent pests and rodent pests in the Africa-Malagasy context is generally ignored as a research topic.
Citation: Swanepoel LH, Swanepoel CM, Brown PR, Eiseb SJ, Goodman SM, Keith M, et al. (2017) A systematic review of rodent pest research in Afro-Malagasy small-holder farming systems: Are we asking the right questions? PLoS ONE 12(3): e0174554. https://doi.org/10.1371/journal.pone.0174554
Editor: Mathew S. Crowther, University of Sydney, AUSTRALIA
Received: September 26, 2016; Accepted: March 10, 2017; Published: March 30, 2017
Copyright: © 2017 Swanepoel et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data are available on researchgate with the DOI:10.13140/RG.2.2.24795.77609.
Funding: This research was funded by a European Union 9th European Development Fund grant from the African Caribbean and Pacific Science and Technology Programme (FED/2013/330-223), a grant from the United Kingdom’s Department for International Development (AgriTT/894), a grant from the Sasol Agriculture Trust (South Africa), and International Foundation for Science (SE) - D/4984-2 to LHS. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare that no competing interests exist.
Competing interests: The authors have declared that no competing interests exist.
World hunger and food insecurity are principally linked to poverty [1, 2]. The majority of the rural poor are dependent on farming where 90% of farm sizes are less than two hectares . Furthermore, up to 80% of undernourished people live in countries where the majority of farming occurs under such small-holder farming practices . As such, small-holder farmers are the backbone of global food security . While rural small-holder farmers face several social and environmental problems in food production , agricultural pests are a major factor in yield gaps pre-harvest and losses post-harvest [7, 8]. Rodent pests are problematic in terms of agriculture and public health since they can inflict considerable economic damage [9, 10], because of their abundance, diversity, generalist feeding habits and their high reproductive output .
In Asia, under traditional rice farming systems, rodents typically cause chronic losses to rice in the order of 5–10% per annum , while episodic population outbreaks cause severe losses that place at risk the food security of entire communities . In Tanzania, chronic pre-harvest losses to maize are around 15%, while damage at sowing and to seedlings can exceed 40% . Population irruptions of Mastomys natalensis have caused yield losses up to 48%, and during acute outbreaks, damage has reached 80–100% of sowing and seedling stages in maize [14, 15]. In Kenya, rodents have caused losses of 20–30% to maize crops, with 34–100% during rodent outbreaks . In parts of South America, native rodents cause crop damage varying between 5 and 90% of total production . Post-harvest losses caused by rodents in grain stores add around 5–14% each year to the losses of small-holder families [18, 19].
Meerburg et al.  conservatively estimated that if rodent losses to rice production could be reduced by 5% this would save 70 million tonnes of rice, which is sufficient to provide the annual food consumption for almost 280 million people in developing countries, i.e. enough to feed 34% of the total undernourished people in the world. Thus, controlling rodent numbers to prevent subsequent losses remains one of the key strategies to secure long-term food security, agro-ecological sustainability and economic development, especially among small-holder farmers. Rodents are also an important public health issue [20, 21], but this aspect is not covered in this review.
Several methods are employed globally to control rodent pests , even though definitive and sustainable solutions seem currently unattainable through poor application of improved management strategies and limited technology development . The paradigm of Ecologically-Based Rodent Management has gained momentum over the past 20 years as an alternative, effective and sustainable rodent control concept [23, 24]. Several research and community-based development programmes have addressed several aspects of EBRM and its implementation in rural communities [24–27]. These projects have been successful in raising awareness  about the seriousness of rodent pests, implementing several rodent control campaigns, as well as highlighting the challenges related to sustainable and effective control . While some reviews have emphasised the importance of rodent pests and their control in the developed world , and others have summarised the impact and control of pest rodents in southeast Asia small-holder agricultural areas [9, 12, 28], no study thus far has reviewed the impacts of rodents on small-holder farming in Africa. This is unfortunate for two reasons, first improving food security in the African-Malagasy small-holder industry can have large outcome effects on community well-being. Secondly, several rodent pest control actions are often suggested, with little research of the effectiveness of such approaches. For example, evaluating the effectiveness of intervention actions often requires complicated replicated experimental designs to test hypotheses about rodent population dynamics and associated crop damage [23, 25, 29]. It is, therefore, important to assess the current and historical approaches to pest rodent research in Africa to enable researchers to evaluate the effectiveness of current research approaches and to develop more appropriate research protocols if needed. Furthermore, such a review will highlight current and historic trends of rodent pest research, and its impact on small-holder agriculture and will enable researchers and governments to evaluate current and future control strategies. The aims of this study are firstly to establish the current state of knowledge for rodents in the agricultural sector, with an emphasis on Africa and Madagascar, by systematically reviewing research on rodents including their occurrence, damage and control and, secondly, to highlight important research gaps at continental and national scales.
Material and methods
Selection of studies
Our initial intention was to collate all relevant studies in a meta-analysis on rodent damage in the agricultural sector and effectiveness of practiced control measures. However, very few publications have focused on this subject. Hence, instead of a full meta-analysis, we compiled a review of published publications (and reports) in this review, to present the status quo on rodent research in agricultural systems in the Afro-Malagasy context, highlighting gaps and future research needs. We focused on the spatial and temporal distribution of rodent studies, extent of rodent damage, and the methods and effectiveness of their control.
To quantify rodent pest impact and identify trends in rodent pest research on small-holder agriculture in the Afro-Malagasy context we conducted a systematic literature review covering the period 1910–2015. We developed an a priori defined set of criteria to allow for replication of the review process  and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA statement and Checklist) guidelines in recording publications excluded or included during screening stages (Fig 1 and S1 Table; [31, 32]. We first searched the Web of Science TM database and focussed on peer-refereed journal publications and book chapters since these are systemically accessible and normally indicative of scientific progress in a particular field. We then used a snowball process where we scanned the references of appropriate publications to discover additional publications, reports and grey literature . Finally, we supplemented the publications from the various co-authors’ personal libraries.
The review was geographically limited to continental Africa and Madagascar and we only selected publications that described the effect of rodent pests on small-holder agriculture. This was achieved by searching online databases for publications that contained the following keywords in the topic: ‘Rodent pest damage’, ‘Africa’ and/or ‘Madagascar’. Abstracts of relevant studies were screened using the ‘metagear’ R package  to compile a list of appropriate publications matching the review criteria. We finally only retained publications for which we could source the full text version for quantitative data analysis (S1 List). Search history data retrieved from Web of Science can be found in S2 List and S1 Web of ScienceTM saved search (AU file).
Each paper was categorised using the following criteria: i) review publications that summarised knowledge regarding rodent pests in the Afro-Malagasy context; ii) field studies reporting rodent damage and/or ecology in agriculture; iii) knowledge, attitudes and practice (KAP) that interviewed local residents regarding rodent pests and associated damage (including traditional questionnaire studies–we classed all questionnaires and KAP as ‘interviews’). We designated more than one research category to a particular study (e.g. interview [KAP] studies were often paired with field trials;  when overlapping occurred.
We extracted the following information from each publication where available: spatial data (country and location of study sites), temporal data (publication date), size of study area, rodent ecological data (rodent species, densities, species composition), and agricultural data (crops and crop stages). To assess rodent impact and management approaches we extracted several crop impact matrixes. Extracting crop impact data was however complicated by several factors. First, the correlation between rodent crop damage and subsequent yield losses are not necessarily linear . This is important since some crops compensate for early crop damage (e.g. during seedling stage) and subsequent yield losses would not necessarily correlate to these early damage estimates . Secondly, the estimation and quantification of rodent crop impact varied considerably between different studies, crops, cropping stages and field methods. In order to standardize and improve the credibility of extracted crop loss data we restricted our analysis to crop loss data collected during field studies. We excluded crop loss data reported in reviews (since these are essentially a repeat of field data) and interview studies. We also restricted crop loss data to mature crop stages and storage and highlight that impact during the seedling stage should be viewed as damage and not necessarily crop losses. However, since levels of seedling damage almost equal subsequent crop losses, seedling damage is a good proxy for crop loss (e.g. for temperate crops, 10% seedling losses resulted in 9% crop loss at harvest; .
We considered all crops and standardized data extraction to only two growth stages (seedling, mature [defined as all stages]) plus storage losses. We used both ranges in crop losses (e.g. 10%-20% losses) and point estimates in presenting rodent crop impact. We further grouped crop losses into 0–20%, 20–50% and >50% bins. We followed this approach since a large proportion of crop losses were reported as ranges (e.g. up to 20% loss), rather as point estimates. We grouped publications into two broad research themes, those that emphasised applied or basic research. Applied research included publications that actively investigated methods to control rodent pest damage and populations, or at least related rodent abundance to damage levels, which included: 1) before and after intervention studies [damage and/or rodent densities]; 2) damage management publications [e.g. relating rodent abundance to crop damage, or comparing control methods]; and 3) mathematical modelling of crop damage to inform management. Basic research included publications that only reported results, and did not relate rodent abundance or crop damage levels to management interventions, which included: 1) rodent ecology [e.g. population ecology, movement ecology, genetics]; 2) crop damage/yield loss estimates; 3) interview studies; and 4) and reviews.
Effect sizes like Hedges’d or ln(R) are often used to quantify the direction and magnitude of experimental impact . However, none of the studies in this review qualified as replicated studies and we, therefore, could not estimate within-study variance. Furthermore, only seven studies (4%) reported before and after intervention estimates. We, therefore, used ln(Xe/Xc), where Xe and Xc are the mean estimates for intervention and control, respectively, to estimate manipulation effect. ln(Xe/Xc) > 0 means that intervention resulted in a higher response (e.g. higher density of rodents post intervention), ln(Xe/Xc) < 0 indicated a negative response (e.g. lower rodent density) while 0 indicates no response . We extracted values for Xe and Xc from graphs, tables or text.
We used the Chi-Square test to investigate effect of country on research theme, research theme per country and number of studies per country , the Mann-Kendall to test for a significant temporal trend in research publications , the t-test to test for significant mean differences between paired treatments and Kruskal-Wallis for difference in crop losses . All data analysis was done in R , and we used the ‘metagear’ package  to draw the PRISMA flow diagram and ‘Kendall’ package for analysis of research trend . We used QGIS 2.18.3  to produce the distribution map and took political boundaries (Africa and Madagascar) from www.Naturalearthdata .
Temporal and spatial
A total of 162 publications (Fig 1) met the criteria of being suitably focussed on rodent pest damage in Afro-Malagasy agricultural systems. During 2003 there were an unusually high number of publications recorded (n = 15), resulting from the publication of a conference proceedings dedicated to this topic (“Second International Conference on Rodent Biology and Management”, . No publications were found prior to 1960, and we subsequently on use the period post 1960 in analysis and graphs. Using publication date as an indication of a temporal trend, a mean of 3.44 (± 1.66) publications were published per annum, with a significant but small increase in publication rate over the time period (Mann-Kendall test: tau = 0.369, p < 0.0001; Fig 2).
Whilst rodent pest research has been conducted across much of the African continent (32 out of 48, countries (including Madagascar) registered at least one relevant study), there was a disparity in the number of studies per country (χ218 = 205.72, p < 0.0001; Fig 3). Research was focused in nodes in East and West Africa. Specifically four countries; Tanzania (24.69%), Ethiopia (8.64%), Nigeria (8.64%), and Kenya (8.02%), accounted for 50% of all rodent pest research on the continent (Fig 3). The high research intensity in Tanzania is due to the Pest Management Centre at the Sokoine University of Agriculture, with a strong focus on rodent pests in agricultural systems. Similarly, individual studies per country clustered around established research sites (Fig 3).
To extract quantitative data and assign an appropriate research theme to each paper, we had to have access to the full text of each paper. However, for 38 out of 162 (23%) publications we could not source full texts, due to limited publication as unpublished theses or in local (and often discontinued) journals. These publications were, therefore, excluded from all subsequent analyses but are reported in S1 List, and only the remaining 124 full text publications were used for further analysis.
There was a significant difference in the number of publications in the different categories (χ24 = 68.28, p < 0.0001), with the majority of studies being field trials (44.72%), followed by reviews (43.96%), interview studies (8.13%), mathematical models (8.13%) and a combination of field trials and interview studies (4.07%). Since the majority of publications originated from Tanzania (26.02%), we benchmarked research from the rest of Afro-Malagasy countries to Tanzania. There was a disparity in the research themes addressed by Tanzanian rodent pest publications compared to publications from the rest of the Afro-Malagasy areas (χ220 = 194.19, p < 0.001). Research in Tanzania had a more applied focus (50% of publications) compared to the rest of the Afro-Malagasy agricultural systems that had a more basic research approach (92% of publications; Fig 4). Furthermore, Tanzanian basic research was dominated by ecological studies of pest rodents (31% of publications; Fig 4). In contrast, research from the rest of the Afro-Malagasy countries focused on interview studies (17% of publications), while review studies (41% of publications) played an important role in providing rodent pest information. Other Afro-Malagasy studies also tended to focus on damage assessments (20% of publications) and rodent pest ecology (13% of publications; Fig 4). Applied research appears to be more established in Tanzania with damage intervention research (pre- and post-intervention studies (16%), mathematical modelling (16%) and management research (19%) contributing to the high levels of rodent pest research (Fig 4).
Rodent species involved
A total of 127 rodent species were detected during pest related research in the Afro-Malagasy agricultural systems (S2 Table). However, pest research was dominated by only 10 species (Table 1) of which Mastomys spp. (57% of studies), Arvicanthis spp. (34%) and Rattus spp. (31%; Table 1) were the most important species.
Extracting meaningful information on crop losses caused by pest rodents proved to be challenging. Nonetheless, 34% of studies provided some data on crop damage. From these studies, we excluded the review and interview studies since these did not always describe their method of damage estimation, which included several interview studies where damage were quantified by severity, rather than giving damage estimates.
There were large variations in crop loss estimates between the different growth stages and storage, as well as between losses reported as range values or point estimates (Fig 5A & 5B). Median crop losses were significantly different between crop stages and storage (H = 23.25, df = 2, p<0.0001) and the highest losses were recorded during the seedling stage (Fig 5B). The largest proportion of seedling loss studies (100%) reported losses below 50%. Crop losses during maturity varied considerably (0%-50%; Fig 5A); however the majority of losses (68%) fell between 20%-50%. Median mature crop losses (15.9%; Fig 5B) were lower than median storage losses (7.9%; Fig 5B). We found no significant difference in losses for the different crops (H = 4.32, df = 4, p = 0.365; Fig 6), probably due to large variation in crop loss estimates. Overall rodent damage was reported for 46 different crop species (S3 Table); however, maize dominated publications (22%) followed by rice (8%) and wheat (7%). In terms of cropping systems, durable commodities dominated research (56% of publications), followed by root vegetables (10%) and fruit trees and vegetables (7% respectively).
Range of crop losses (a) and median point estimates for crop losses (b) as reported for different crop stages and storage. Asterisks denote significant differences, whiskers minimum and maximum values, box plot indicate third and first quantile and median. n = number of data points.
Rodent control and interventions
Various methods were used to control rodent pests, which could be divided into four broad categories: habitat management, chemical control, predation, and trapping. The most important actions to control rodent pests was habitat management (25%), followed by trapping (15%), chemical control (17%), predation (15%; which included dog [2%] and cat predation [%] and avian predation), and other (22%), which include acoustic scaring, glue, guarding, sanitation, praying and charms. Habitat modification included burning of grass, ploughing, inter-cropping and fencing.
We could only source seven studies (8%) that investigated the efficacy of a management intervention on rodent pests and damage. However, all these studies lacked proper replication preventing us to draw appropriate conclusions about the effectiveness of management interventions, even though individual studies suggested that all intervention actions resulted in declining rodent pest damage and abundance (Table 2). For example, ploughing (n = 3 studies) had a significant but small negative effect on rodent abundance, which declined 1.45-fold after ploughing (t-test, mean ln (Xe/Xc) ± 95% CI = -0.34 ± 0.52, t  = 4.19, p = 0.03).
This analysis has highlighted that pest rodents do indeed have a significant negative effect on Afro-Malagasy small-holder farming communities. Even though damage estimates varied considerably between crop stages, in the methods used to estimate damage, and in geographical scope, there appears to be considerable support that total crop losses due to rodent pests remain around 15%. Interestingly, our observed crop loss estimates closely concur with simulated rodent grazing models for Australian crop systems showing a mean yield loss = 12.4% . These losses are much higher than damage levels farmers are willing to tolerate, usually around 5% , which emphasises the large impact rodent pests can have on agricultural production. Rodent damage varied between growth stages, and damage during the seedling stage can be extensive . We found that the majority of studies reported seedling losses in excess of 50% damage, which indicated that the greatest rodent pest impacts often occurs during seed emergence, particularly for maize . High seedling losses have also been observed to lead to high crop losses for both rice  and wheat . It has been suggested that the percentage seedling loss is a good predictor of subsequent crop losses (e.g. 10% seedling losses resulted in 9% crop loss at harvest; .
Whilst we could extract some meaningful crop loss data, estimated losses from the majority of studies was not insightful. For example, some studies did not distinguish between rodent damage and other agricultural pests [49, 50]. Other studies report on general agricultural losses that occurred during planting, harvesting, and post-harvest processes, of which rodents were only one of many agents [50, 51]. Extracting estimated losses from interview studies was particularly problematic since these studies were often only interested in perception of damage, or severity . Respondents also were not always able to estimate losses , suggesting that estimated losses based on interview data might be significantly biased.
Standardized methods for estimating rodent impact were not generally followed by the majority of studies. Whilst the robustness of a standardized approach in estimating rodent damage has been developed for maize , not all studies employed these methods. Furthermore, standardized methods in estimating damage during other crop stages and storage seem not to be well established. Unless damage estimates are standardized for crop stages, it remains difficult to make meaningful conclusions and comparisons between estimates, interventions and countries. Furthermore, relating damage caused to crops to subsequent yield loss was not adequately addressed in the majority of the publications. Since not all damages relate to yield losses  it is imperative to establish the relationship between crop damage and losses at various cropping stages.
An important relationship exists between crop damage and rodent density [35, 54]. As such, crop damage will vary by rodent abundance; thus monitoring rodent abundance should be an important variable in initiating control actions . Field studies and simulation work has established this relationship as sigmodal where increased damage is observed when rodent densities at the seedling stage increase above 20 animals/ha for Mastomys natalensis  and around 100 animals/ha for the house mouse Mus domesticus . However, some studies have highlighted confusion between rodent damage rates and germination failure rates  and most studies rarely take into account germination failure rates in assessing rodent damage at the seedling stage. Given the importance in the relationship between rodent abundance and damage, it is surprising that almost no studies have investigated these relationships to understand when rodent damage exceeds acceptable levels. Such information is of vital importance in setting rodent pest monitoring protocols and determining the initiation of intervention actions.
Effectiveness of rodent control
There was a paucity of research related to the impact and effectiveness of management actions. Habitat management or modification was the most used method in trying to control rodent abundance and damage. While our analyses showed that ploughing produced meaningful and detectable reductions in rodent abundance, the robustness of such a conclusion is limited by low sample sizes and lack of replication. Similarly, communal trapping has been shown to be an effective method to reduce rodent density to acceptable levels [56, 57]. Such non-chemical control methods are important since several studies have highlighted the limitations of relying on chemical rodent control, especially for resource poor farmers . For example, chemical control may reduce numbers initially, but surviving animals will compensate with higher survival and breeding success [58, 59]. This is important since several mathematical models have highlighted that control efforts are relatively ineffective when interventions occur at high rodent abundances, which is normally when chemical control is applied [60, 61]. The effectiveness of chemical control is also affected by the network of habitat patches that characterises small-holder farming communities , where depleted patches can quickly be recolonised from surrounding areas. Rodent control in such areas should follow a meta-population approach where rodent control is coordinated among several patches to limit individual patch recolonization . We found no published and available studies that evaluated coordinated control strategies, highlighting the lack of coordinated efforts in rodent control. Our results here concur with others that have indicated that current control methods by small-holder farmers for rodent pests seem to be inadequate [23, 63] and where there has been success in Africa (e.g.  and Asia  there has been a strong emphasis on community campaigns .
Results from the predation studies were inconclusive. The majority of predation studies focussed on birds of prey, which have elsewhere produced some positive effects on rodent pest control . However, our review found little to no support that increased avian predation reduced rodent populations to acceptable levels. The only observed effect detected was higher peak densities and faster population growth rates in the absence of predators . These results concur with a recent meta-analysis on the effect of predation on prey populations , which suggest that rodents adapt their foraging behaviour according to predation risk, which has indeed be shown in M. natalensis . These results here are however in contrast to a recent world-wide meta-analysis, which found a detectable negative effect of avian predation on rodent pests . This discrepancy highlights the paucity of studies that makes it difficult to calculate effect sizes needed to estimate the magnitude of intervention effects . Lastly, some studies suggested the use of domestic cats as a rodent control method. While domestic cats have been cited as rodent control agents , their effectiveness in rodent pest control is debateable . In contrast, domestic cat diet seems to be dominated by native sylvatic animals (mammalian, reptilian and avian) which could actually impede rodent pest control, and care should be taken in advocating cats as effective pest control agents . Furthermore, cats preying on rodents can be infected with Toxoplama gondii and in turn pose a significant health risk to humans .
Trends in rodent pest research and species involved
Our analysis has highlighted that even though there has been a recent increase in the number of publications on rodent pest research, on an annual basis the number of papers are still few. Furthermore, research was highly restricted to only just a few African countries or sites within countries. There may be several reasons for low interest in rodent pest research related to agriculture. First, research is normally undertaken by researchers holding academic posts where pest control research may not be considered as a field producing high impact publications, a metric used to gauge academic performance . Secondly, our review highlighted that well-funded large cross country collaborative projects (e.g. STAPLERAT; ; EcoRat;  are scientifically productive and may be the most appropriate avenues to establish long term projects which are required to produce robust estimates of abundance of rodents, and their damage and losses, with adequate spatial and temporal replication.
We observed a discrepancy in research themes between Tanzania and the rest of Africa. Tanzania generated the most research output due to the Pest Management Centre at Sokoine University of Agriculture, which also host the IRPM & BTD (African Centre of Excellence for Innovative Rodent Pest Management and Biosensor Technology Development). Tanzania therefore focussed more on applied research, compared with the rest of Africa that focussed on basic research. Researchers in Tanzania were ideally positioned to study the long history of rodent outbreaks in Africa, with specific focus into rodent outbreaks and its impact on crops [75, 76]. Over the 40 years Tanzanian researchers have invested heavily in understanding the ecology of its most important rodent pest, Mastomys natalensis , and with such understanding were able to investigate various interventions and modelling approaches to predict and manage crop losses [61, 77–80]. Of greatest concern is the fact that research regarding long-term rodent management is completely lacking from the rest of Africa.
One issue that is rarely considered is the balance of managing rodent pest species in an agricultural context whilst concurrently conserving those rodent species that occur in these landscapes and are not significant agricultural pests. In Africa, only about 5% of rodent species (up to 20 pest species out of 381 rodent species) are significant agricultural pests [59, 81]. The non-pest species often provide an important ecosystem service (see  for a review) and therefore need to be protected.
Conclusions and recommendations
The overall objective of pest research is to reduce crop losses by achieving long-term sustainable population suppression of the key pest species [83, 84]. As such, the focus area of many studies has been population ecology, and our analysis highlights the significant challenges ecologists face in understanding population regulation, or at least how to manipulate it with management. Nonetheless, we highlight several recommendations that we feel will improve future studies and the evaluation of interventions:
- We encourage researchers to adopt a ‘meta-analytic’ thinking framework when establishing intervention/management related research . We suggest that researchers report effect statistics (and confidence intervals; see Nakagawa and Cuthill 2007 for details) which will quantify the size of experimental effects (e.g. mean difference). Standardised effect statistics are also dimensionless, which means independent studies (or treatment actions) can be compared. Furthermore, adopting this approach will facilitate future meta-analysis. For example, we found that studies rarely reported vital information such as sample sizes and standard deviations needed for meta-analysis.
- We encourage researchers and funding organisations to establish and fund long-term studies . Our analysis shows that only once a firm foundation has been established on the population ecology of the dominant species, other important aspects like management and community ecology can be successfully developed [75, 85].
- We suggest that researchers adapt a unified or standardised approach to quantify rodent crop damage at the different cropping stages (e.g. [53, 86]. However, what is equally important is to relate crop damages to crop losses (yield losses). As such, more research is needed to accurately translate crop damages to crop (yield) losses, especially for different crops and cropping stages. Finally, it is also important to detail how crop damage was identified and if crop damage can be separated between different pest species.
- We highlighted the paucity of research relating rodent density or abundance to crop damage . This relationship is a key factor in determining if management intervention is needed, and yet little information pertaining to this relationship exists.
- Rodent populations are often reported as capture success or minimum number of individuals alive, which can essentially be described as indices . Ecologists need robust data to make informed management decisions , and we suggest that population data be analysed and reported in robust analytical frameworks (e.g. mark-recapture studies (88)).
- There needs to be more empirical treatment-control studies with replicates that investigate management actions on rodent pest populations and associated crop losses. Ideally, these should be linked with household surveys conducted before and after management was implemented.
- Finally, we suggest that ecologically-based rodent management will benefit from linking rodent population ecology with food webs [85, 89]. For example, the effectiveness of predation in reducing rodent populations is often investigated in isolation (or in a single species). However, by measuring these processes using food webs or ecological networks it is possible to establish a quantifiable yardstick to judge the structure and function of an ecosystem. This can guide future research as to how these networks may provide an ecosystem service like predation.
Our analysis shows that there are some well-recognised impacts of rodent pests on agricultural production, and that there are generally consistent estimates in the amount of crop damage occurring due to rodents, albeit based on a limited number of studies from a limited number of African countries. However, despite the high perceived losses gathered from interviews, backed up by a few systematic trials, replicated studies investigating sustainable rodent control solutions are few and far between. These gaps in research continue to have a significant impact on the development of ecologically-based rodent management strategies. Socio-economic benefits of sustainable rodent management are urgently required to improve food security for small-holder farmers in Africa and elsewhere where hunger, poverty and rodent pests are entwined.
S2 Table. List of rodent genera detected in rodent pest research in African agricultural systems from 1960–2015.
S3 Table. List of different crops and cropping system as impacted by rodent pests in African agriculture (1960–2015).
S1 List. Complete list of all publications used in the review–Publications in bold did not have full texts available at time of review.
We thank the anonymous reviewers for valuable comments which increased clarity of the manuscript.
- Conceptualization: LHS CMS PRB SRB PJT GRS HL AM.
- Data curation: LHS CMS.
- Formal analysis: LHS.
- Funding acquisition: SRB LHS.
- Investigation: LHS.
- Methodology: LHS CMS.
- Project administration: LHS CMS SRB.
- Resources: LHS CMS PRB SJE SMG MK FK HL TAMM RHM PM EFM AWM AM LSM GRS PJT VS SRB.
- Software: LHS.
- Supervision: LHS CMS SRB.
- Validation: LHS CMS.
- Visualization: LHS.
- Writing – original draft: LHS CMS.
- Writing – review & editing: LHS CMS PRB SJE SMG MK FK HL TAMM RHM PM EFM AWM AM LSM GRS PJT VS SRB.
- 1. Dixon JA, Gibbon DP, Gulliver A. Farming systems and poverty: improving farmers' livelihoods in a changing world: Food & Agriculture Org.; 2001.
- 2. Runge CF, Senauer B, Pardey PG, Rosegrant MW. Ending hunger in our lifetime: food security and globalization: Intl Food Policy Res Inst; 2003.
- 3. Tscharntke T, Clough Y, Wanger TC, Jackson L, Motzke I, Perfecto I, et al. Global food security, biodiversity conservation and the future of agricultural intensification. Biol Conserv. 2012;151(1):53–9.
- 4. WorldBank. World Development Report 2008. Agriculture for Development, Washington, DC. 2007.
- 5. Chappell MJ, LaValle LA. Food security and biodiversity: can we have both? An agroecological analysis. Agriculture and Human Values. 2011;28(1):3–26.
- 6. Baiphethi MN, Jacobs PT. The contribution of subsistence farming to food security in South Africa. Agrekon. 2009;48(4):459–82.
- 7. Oerke E-C. Crop losses to pests. The Journal of Agricultural Science. 2006;144(01):31–43.
- 8. Abrol DP. Integrated pest management: current concepts and ecological perspective: Academic Press; 2013.
- 9. John A. Rodent outbreaks and rice pre-harvest losses in Southeast Asia. Food Security. 2014;6(2):249–60.
- 10. Singleton GR, Belmain SR, Brown PR. Rodent outbreaks: an age-old issue with a modern appraisal. In: Singleton G, Belmain S, Brown P, Hardy B, editors. Rodent outbreaks: Ecology and impacts: IRRI; 2010. p. 1–8.
- 11. Buckle AP, Smith RH. Rodent pests and their control: CABI: Wallingford, UK.; 2015.
- 12. Singleton GR. Impacts of rodents on rice production in Asia. IRRI discussion paper series. 2003;45:1–30.
- 13. Singleton GR, Belmain SR, Brown PR, Aplin KP, Htwe NM. Impacts of rodent outbreaks on food security in Asia. Wildl Res. 2010;37(5):355–9.
- 14. Mwanjabe PS, Leirs H. An early warning system for IPM-based rodent control in smallholder farming systems in Tanzania. Belg J Zool. 1997;127.
- 15. Mwanjabe PS, Sirima FB, Lusingu J. Crop losses due to outbreaks of Mastomys natalensis (Smith, 1834) Muridae, Rodentia, in the Lindi Region of Tanzania. Int Biodeterior Biodegrad. 2002;49(2–3):133–7.
- 16. Taylor KD. An Outbreak of Rats in Agricultural Areas of Kenya in 1962. East Afr Agric For J. 1968;34(1):66–77.
- 17. Rodriguez JE. Roedores plaga: un problema permanente en América Latina y el Caribe. FAO, Oficina Regional para America Latina y el Caribe, Santiago, Chile. 1993.
- 18. Belmain SR, Htwe NM, Kamal NQ, Singleton GR. Estimating rodent losses to stored rice as a means to assess efficacy of rodent management. Wildl Res. 2015;42(2):132–42.
- 19. Htwe NM, Singleton GR, Maw PP. Post-harvest impacts of rodents in Myanmar; how much rice do they eat and damage? Pest Manag Sci. 2016.
- 20. Meerburg BG, Singleton GR, Leirs H. The Year of the Rat ends—time to fight hunger! Pest Manag Sci. 2009;65(4):351–2. pmid:19206089
- 21. Gratz NG, Arata AA. Problems associated with the control of rodents in tropical Africa. Bull W H O. 1975;52(4–6):697. pmid:1085224
- 22. Capizzi D, Bertolino S, Mortelliti A. Rating the rat: global patterns and research priorities in impacts and management of rodent pests. Mammal Rev. 2014:n/a-n/a.
- 23. Singleton GR, Leirs H, Hinds LA, Zhang Z. Ecologically-based management of rodent pests–re-evaluating our approach to an old problem. In: Singleton GR, Leirs H, Hinds LA, Zhang Z, editors. Ecologically-based Management of Rodent Pests Australian Centre for International Agricultural Research (ACIAR), Canberra.; 1999. p. 17–29.
- 24. Makundi RH, Massawe AW. Ecologically based rodent management in Africa: potential and challenges. Wildl Res. 2011;38(7):588–95.
- 25. Brown PR, Tuan NP, Singleton GR, Ha PTT, Hoa PT, Hue DT, et al. Ecologically based management of rodents in the real world: Applied to a mixed agroecosystem in Vietnam. Ecol Appl. 2006;16(5):2000–10. pmid:17069390
- 26. Belmain SR, Dlamini N, Eiseb SJ, Kirsten F, Mahlaba T, Makundi RH, et al. The ECORAT project: developing ecologically-based rodent management for the southern African Region. Int Pest Control. 2008;50(3):136–8.
- 27. Taylor PJ, Arntzen L, Hayter M, Iles M, Frean J, Belmain SR. Understanding and managing sanitary risks due to rodent zoonoses in an African city: beyond the Boston Model. Integrative Zoology. 2008;3(1):38–50. pmid:21396050
- 28. Singleton GR, Hinds LA, Krebs CJ, Spratt DM, editors. Rats, mice and people: rodent biology and management. Canberra: ACIAR Monograph No. 96; 2003.
- 29. Salo P, Banks PB, Dickman CR, Korpimäki E. Predator manipulation experiments: impacts on populations of terrestrial vertebrate prey. Ecol Monogr. 2010;80(4):531–46.
- 30. Tranfield D, Denyer D, Smart P. Towards a methodology for developing evidence‐informed management knowledge by means of systematic review. British journal of management. 2003;14(3):207–22.
- 31. Moher D, Liberati A, Tetzlaff J, Altman DG, group. TP. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med. 2009;6(7):e1000097. pmid:19621072
- 32. Lajeunesse MJ. Facilitating systematic reviews, data extraction and meta-analysis with the metagear package for r. Methods in Ecology and Evolution. 2016;7(3):323–30.
- 33. Willis-Shattuck M, Bidwell P, Thomas S, Wyness L, Blaauw D, Ditlopo P. Motivation and retention of health workers in developing countries: a systematic review. BMC Health Services Research. 2008;8(1):1–8.
- 34. Kasso M. Pest Rodent Species Composition, Level of Damage and Mechanism of control in Eastern Ethiopia. International Journal of Innovation and Applied Studies. 2013;4(3):502–11.
- 35. Brown PR, Huth NI, Banks PB, Singleton GR. Relationship between abundance of rodents and damage to agricultural crops. Agriculture, ecosystems & environment. 2007;120(2):405–15.
- 36. Myllymaki A. Denmark-Tanzania Rodent Control Project—Final Report (unpublished), Rodent Control Centre, Morogoro, Tanzania. 1989.
- 37. Nakagawa S, Cuthill IC. Effect size, confidence interval and statistical significance: a practical guide for biologists. Biological Reviews. 2007;82(4):591–605. pmid:17944619
- 38. Franke TM, Ho T, Christie CA. The Chi-Square Test Often Used and More Often Misinterpreted. American Journal of Evaluation. 2012;33(3):448–58.
- 39. Pettitt AN. A non-parametric approach to the change-point problem. Applied statistics. 1979:126–35.
- 40. Crawley MJ. Statistical computing: an introduction to data analysis using S-plus. Chichester: John Wiley & Sons Ltd; 2002.
- 41. RDevelopmentCoreTeam. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org/. 2012.
- 42. McLeod AI. Kendall: Kendall rank correlation and Mann-Kendall trend test. R package version 2.2. https://CRAN.R-project.org/package=Kendall. 2011.
- 43. QGIS Development Team, 2016. QGIS Geographic Information System. Open Source Geospatial Foundation. URL http://qgis.osgeo.org.
- 44. Natural Earth Data. 2016. Available from http://www.naturalearthdata.com/downloads/10m-cultural-vectors/10m-admin-0-countries/. Accessed March 2016.
- 45. Mwanjabe PS, Leirs H. An early warning system for IPM-based rodent control in smallholder farming systems in Tanzania. Belg J Zool. 1997;127:49–58.
- 46. Mulungu LS, Makundi RH, Massawe AW, Leirs H. Relationship between sampling intensity and precision for estimating damage to maize caused by rodents. Integrative Zoology. 2007;2(3):131–5. pmid:21396028
- 47. Mulungu LS, Lagwen PP, Mdangi ME, Kilonzo BS, Belmain SR. Impact of spatio-temporal simulations of rat damage on yield of rice (Oryza sativa L.) and implications for rodent pest management. Int J Pest Manag. 2014;60(4):269–74.
- 48. Brown PR. The effect of simulated house mouse damage to wheat in Australia. Crop Prot. 2005;24(2):101–9.
- 49. Funmilayo O. Mammals and birds affecting food production and storage in Nigeria. 1980.
- 50. Tefera T. Post-harvest losses in African maize in the face of increasing food shortage. Food security. 2012;4(2):267–77.
- 51. Tyler PS, Boxall RA. Post harvest loss reduction programmes: A decade of activities-what consequences? Tropical Stored Products Information. 1984;50:4–13.
- 52. Meheretu Y, Sluydts V, Welegerima K, Bauer H, Teferi M, Yirga G, et al. Rodent abundance, stone bund density and its effects on crop damage in the Tigray highlands, Ethiopia. Crop Prot. 2014;55:61–7.
- 53. Mulungu LS, Makundi RH, Leirs H. Robustness of techniques for estimating rat damage and yield loss in maize fields. ACIAR Monogr Ser. 2003;96:224–8.
- 54. Mulungu LS, Makundi RH, Leirs H, Massawe AW, Vibe Petersen S, Stenseth NC. The rodent density-damage function in maize fields at an early growth stage. In: Singleton GR, Hinds LA, Krebs CJ, Sprat DM, editors. Rats, mice and people: rodent biology and management. Canberra: ACIAR; 2003. p. 301–3.
- 55. Mulungu LS, Makundi RH, Leirs H, Massawe AW, Vibe-Petersen S, Stenseth NC. The rodent density-damage function in maize fields at an early growth stage. In: Singleton GR, Hinds LA, Krebs CJ, Spratt DM, editors. ACIAR Monograph Series No 96, 564p. 962003. p. 301–3.
- 56. Makundi RH, Oguge NO, Mwanjabe PS. Rodent pest management in East Africa—an ecological approach. In: Singleton GR, Hinds LA, Leirs H, Zhang Z, editors. Ecologically-based management of rodent pests1999. p. 460–76.
- 57. Belmain SR, Meyer AN, Timbrine R, Penicela L. Managing rodent pests in households and food stores through intensive trapping. ACIAR Monogr Ser. 2003;96:440–5.
- 58. Stenseth NC, Leirs H, Mercelis S, Mwanjabe P. Comparing strategies for controlling an African pest rodent: an empirically based theoretical study. J Appl Ecol. 2001;38(5):1020–31.
- 59. Singleton GR, Brown PR, Jacob J, Aplin KP. Unwanted and unintended effects of culling: A case for ecologically‐based rodent management. Integrative zoology. 2007;2(4):247–59. pmid:21396042
- 60. Leirs H, Skonhoft A, Stenseth NC, Andreassen H. A bioeconomic model for the management of Mastomys natalensis mice in maize fields. ACIAR Monogr Ser. 2003;96:358–61.
- 61. Davis SA, Leirs H, Pech RP, Zhang Z, Stenseth NC. On the economic benefit of predicting rodent outbreaks in agricultural systems. Crop Prot. 2004;23(4):305–14.
- 62. Russell JC, Mackay JWB, Abdelkrim J. Insular pest control within a metapopulation context. Biol Conserv. 2009;142(7):1404–10.
- 63. Eisen RJ, Enscore RE, Atiku LA, Zielinski-Gutierrez E, Mpanga JT, Kajik E, et al. Evidence that rodent control strategies ought to be improved to enhance food security and reduce the risk of rodent-borne illnesses within subsistence farming villages in the plague-endemic West Nile region, Uganda. Int J Pest Manag. 2013;59(4):259–70. pmid:26500395
- 64. Palis FG, Singleton GR, Brown PrR, Huan NH, Umali C, Nga NTD. Can humans outsmart rodents? Learning to work collectively and strategically. Wildl Res. 2011;38(7):568–78.
- 65. Singleton GR, Flor RJB. Sociology and Communication of Rodent Management in Developing Countries. In: Buckle AP, Smith RH, editors. Rodent Pests and Their Control. Wallingford, UK: CABI; 2015. p. 295–314.
- 66. Paz A, Jareño D, Arroyo L, Viñuela J, Arroyo B, Mougeot F, et al. Avian predators as a biological control system of common vole (Microtus arvalis) populations in north-western Spain: experimental set-up and preliminary results. Pest Manag Sci. 2013;69(3):444–50. pmid:22517676
- 67. Vibe-Petersen S, Leirs H, Bruyn LD. Effects of predation and dispersal on Mastomys natalensis population dynamics in Tanzanian maize fields. J Anim Ecol. 2006;75(1):213–20. pmid:16903058
- 68. Mohr K, Vibe‐Petersen S, Lau Jeppesen L, Bildsøe M, Leirs H. Foraging of multimammate mice, Mastomys natalensis, under different predation pressure: cover, patch‐dependent decisions and density‐dependent GUDs. Oikos. 2003;100(3):459–68.
- 69. Labuschagne L, Swanepoel LH, Taylor PJ, Belmain SR, Keith M. Are avian predators effective biological control agents for rodent pest management in agricultural systems? Biol Control. 2016;101:94–102.
- 70. Wodzicki K. Prospects for biological control of rodent populations. Bull W H O. 1973;48(4):461. pmid:4587482
- 71. Fitzgerald BM. Is cat control needed to protect urban wildlife? Environ Conserv. 1990;17:168–9.
- 72. Loss SR, Will T, Marra PP. The impact of free-ranging domestic cats on wildlife of the United States. Nat Commun. 2013;4:1396. pmid:23360987
- 73. Eyre-Walker A, Stoletzki N. The assessment of science: the relative merits of post-publication review, the impact factor, and the number of citations. PLoS Biol. 2013;11(10):e1001675. pmid:24115908
- 74. Leirs H, Makundi RH, Davis S. The present issue of the Belgian Journal of Zoology contains the proceedings of the 9th International African Small Mammal Symposium (ASMS), held at the Sokoine University of Agriculture, Morogoro, Tanzania from 14–18 July 2003. Belg J Zool. 2005;135:3.
- 75. Mulungu LS, Ngowo VD, Makundi RH, Massawe AW, Leirs H. Winning the fight against rodent pests: recent developments in Tanzania. Journal of Biological Sciences. 2010;10(4):333–40.
- 76. Leirs H, Sluydts V, Makundi RH. Rodent outbreaks in sub-Saharan Africa. In: Singleton G, Belmain S, Brown PR, Hardy B, editors. Rodent Outbreaks: Ecology and Impacts2010. p. 269–80.
- 77. Leirs H, Verhagen R, Verheyen W, Mwanjabe P, Mbise T. Forecasting rodent outbreaks in Africa: an ecological basis for Mastomys control in Tanzania. J Appl Ecol. 1996:937–43.
- 78. Leirs H. Management of rodents in crops: the Pied Piper and his orchestra. ACIAR Monogr Ser. 2003;96:183–90.
- 79. Leirs H. Population ecology of Mastomys natalensis (Smith, 1834): implications for rodent control in Africa. A report for the Tanzania-Belguim joint rodent research project (1986–1989) 1994.
- 80. Leirs H, Verhagen R, Verheyen W, editors. The use of rainfall patterns in predicting population densities of multimammate rats, Mastomys natalensis. Proceedings of the Fourteenth Vertebrate Pest Conference 1990; 1990.
- 81. Taylor PJ, Downs S, Monadjem A, Eiseb SJ, Mulungu LS, Massawe AW, et al. Experimental treatment-control studies of ecologically based rodent management in Africa: balancing conservation and pest management. Wildl Res. 2012;39(1):51–61.
- 82. Dickman CR. Rodent-ecosystem relationships: a review. In: Singleton G, Hinds L, Leirs H, Zhang Z, editors. Ecologically-based Management of Rodent Pests. 113–133: ACIAR Monograph No. 59 Australian Centre For International Agricultural Research: Canberra.; 2003.
- 83. Singleton GR, Leirs H, Schockaert E. Integrated management of rodents: a southeast Asian and Australian perspective. Belg J Zool. 1997;127(Supplement 1):157–69.
- 84. Baker SE, Johnson PJ, Slater D, Watkins RW, Macdonald DW. Learned food aversion with and without an odour cue for protecting untreated baits from wild mammal foraging. Appl Anim Behav Sci. 2007;102(3–4):410–28.
- 85. Krebs CJ. One hundred years of population ecology: successes, failures and the road ahead. Integrative zoology. 2015;10(3):233–40. pmid:25664990
- 86. Aplin KA, Brown PR, Jacob J, Krebs CJ, Singleton GR. Field methods for rodent studies in Asia and the Indo-Pacific, Australian Centre for International Agricultural Research Monograph. Australia: Australian Centre for International Agricultural; 2003.
- 87. Anderson DR. The need to get the basics right in wildlife field studies. Wildl Soc Bull. 2001;29(4):1294–7.
- 88. Hayward MW, Boitani L, Burrows ND, Funston PJ, Karanth KU, MacKenzie DI, et al. Ecologists need robust survey designs, sampling and analytical methods. J Appl Ecol. 2015;52(2):286–90.
- 89. Memmott J. Food webs: a ladder for picking strawberries or a practical tool for practical problems? Philosophical Transactions of the Royal Society B: Biological Sciences. 2009;364(1524):1693–9. pmid:19451120