title / Development of a decision support system for ecologically sound rabbit
management / DEFRA
project code / VC0232
Department for Environment, Food and Rural Affairs CSG 15
Research and Development
Final Project Report
(Not to be used for LINK projects)
Two hard copies of this form should be returned to:Research Policy and International Division, Final Reports Unit
DEFRA, Area 301
Cromwell House, Dean Stanley Street, London, SW1P 3JH.
An electronic version should be e-mailed to
Project title / Development of a decision support system for ecologically sound rabbit
management
DEFRA project code / VC0232
Contractor organisation and location / Central Science Laboratory
Sand Hutton
YO41 1LZ
Total DEFRA project costs / £ 365,067
Project start date / 01/04/02 / Project end date / 31/03/04
Executive summary (maximum 2 sides A4)
To tab in this section press the tab key and the Control key together
Press the DOWN arrow once to move to the next question.
CSG 15 (Rev. 6/02) 3
Projecttitle / Development of a decision support system for ecologically sound rabbit
management / DEFRA
project code / VC0232
1. Rabbit damage to crops is a major economic problem for agriculture in the UK and rabbit numbers continue to increase as the effects of myxomatosis wane. Improvements in our ability to predict the effectiveness of different management methods under varying conditions are required to allow sound and flexible advice to be developed regarding optimal management strategies for particular agricultural contexts.
2. A substantial knowledge base has accrued through considerable Defra investment in rabbit research. This project seeks to integrate this knowledge into a decision support system for RDS wildlife advisors in their advisory and statutory work regarding rabbit damage in lowland agricultural landscapes, offering guidance on optimised management approaches.
3. Underpinning the system are ecologically sound models of rabbit population biology and rabbit damage to crops in terms of the relationships between rabbit numbers and yield loss.
4. In order to predict future losses, in a given problem setting, we need to be able to extrapolate from existing static models of the relationships between fixed numbers of rabbits and crop damage to more complex settings which are dynamic, in that rabbit numbers are allowed to fluctuate naturally and the animals have choices available to them in terms of foraging opportunities. The required research has been completed at CSL’s unique field research station in Hampshire where replicated studies can be conducted with complete control over agronomy and experimental design.
5. The dynamic damage studies suggest that we can allow for population changes by expressing damage as a linear function of cumulative rabbit grazing hours from crop emergence. This is how the decision support system now converts future predictions regarding rabbit numbers into predicted yield loss estimates. As a result of RDS user evaluation of the system, pro rata extrapolations from winter wheat to other crops are now included in the options available. These extrapolations need to be treated with care in the absence of scientifically rigorous crop damage studies for crops other than winter wheat, grass grown for silage and spring barley.
6. Previous results from our studies of enclosed populations suggest that juvenile recruitment can be impaired in low density populations, known in the ecological literature as an “Allee effect”. A possible mechanism is interaction between plant sward height and density dependent processes. Specifically, increased sward height could lead to reduced food quality, increased predation and increased thermoregulatory challenge to juveniles. Increased population density, above some critical threshold, could maintain low sward height and thus enhance juvenile survival. This possibility was examined in a study where both population density and sward height were manipulated, the latter by mowing.
7. There was no evidence that mowing led to enhanced juvenile recruitment. It was thus considered unnecessary to allow for the putative “Allee effect” in the rabbit population model underpinning the decision support system. However, given that future studies may uncover evidence for such processes, a critical threshold density function has been incorporated into the model which can be invoked by the user to reduce juvenile recruitment at an increasing rate the further the population size falls below the threshold.
8. The original rabbit population model, underpinning the decision support system, represented variation in productivity and mortality rates with respect to age, sex, season and density for closed populations only. Movements between populations, which may influence recovery rates following control operations, were thus not allowed for. This knowledge gap was addressed here by quantifying rabbit migration in the context of active control measures against populations in lowland agricultural landscapes.
9. This showed that the great majority of rabbits rarely disperse despite being given the opportunity of moving into depopulated areas. Specifically, only 8% of tagged animals moved into areas with artificially low population densities areas over a one-year period. The majority of dispersing individuals were males. The mean distance moved by male rabbits was 273m compared to 225m by females. Movements between populations are thus unlikely to have a major influence on population recovery rates following control operations. Nevertheless, migration has now been incorporated into the rabbit population model, using the rates recorded here, and users can specify the number of unmanaged adjacent populations present that might provide sources of immigration into the population that is the focus of concern and subject to control. The decision support system thus now functions for both closed and open rabbit populations
10. Fundamental to the development of the decision support system is the ability to relate the extent of problems, such as crop loss, to the size of rabbit populations. A simple and reliable method for assessing the size of the problem in any given lowland agricultural setting is thus required. Previous studies developed a validated method based on sight counts. Nevertheless, this technique requires a series of counts to be made at night along field margins and is therefore time consuming. This study thus sought to develop and validate a simple and robust daytime census method based on assessing rabbit signs during a single visit to a problem site. Overall, the study suggested that the measurement of activity signs during a single daytime visit does have the potential to either estimate rabbit numbers or at least offer an index of abundance, provided that different techniques are adopted for cereal and non-cereal fields. An approach based on systematic assessment of rabbit droppings is recommended for non-cereal sites and one based on assessment of rabbit scrapes for cereal sites. The specifications for the recommended census methods have been defined as part of the user information for the decision support system with instructions on how to use these census methods to initiate the system for a particular problem scenario.
11. The decision support system for ecologically sound rabbit management, named the Rabbit Management Adviser, or RabMan, now provides a reliable tool for predicting future numbers of rabbits in a given lowland agricultural setting, based on estimates derived from simple census methods; the potential costs of the damage those rabbits will cause in a specific agricultural setting; and cost-benefit analyses of management options. This has been subject to evaluation by users. The user requested revisions have been incorporated and a fully functional decision-support system made available to RDS wildlife advisors
- These advisors are seen as the key focus for technology transfer of the extensive scientific knowledge base via the decision-support system. In the future, additional work may be required to cover new issues such as novel crops, management methods or agronomic practices. It is also likely that, as the system is put into practice, users will identify desirable additional functions or options. Once the system has been successfully established, in the challenging practical context of users who are already expert in this field, then there will be the potential for diversifying the target users to consultant agronomists, growers and farmers to maximise the technology transfer opportunities offered by this unique system.
CSG 15 (Rev. 6/02) 3
Projecttitle / Development of a decision support system for ecologically sound rabbit
management / DEFRA
project code / VC0232
Scientific report (maximum 20 sides A4)
To tab in this section press the tab key and the Control key together
Press the DOWN arrow once to move to the next question.
CSG 15 (Rev. 6/02) 3
Projecttitle / Development of a decision support system for ecologically sound rabbit
management / DEFRA
project code / VC0232
1. Introduction
The rabbit population in the UK continues to increase in size as it recovers from the waning effects of myxomatosis. In the mid-1980s economic losses to agriculture arising from rabbit damage were estimated at £100 million per annum and are now likely to be considerably higher. Under the 1954 Pest Act, Defra has enforcement responsibilities for the obligations of occupiers regarding rabbit control, and must therefore ensure that cost-effective, environmentally sensitive and humane management strategies are available. These strategies need to be informed by rabbit population biology, the costs of damage and the effectiveness and costs of different management methods. Considerable research has been carried out over the past twenty years on a variety of these issues so that we now have extensive knowledge of rabbit biology and management in lowland agricultural landscapes in the UK. This research has reflected five key areas. Firstly, developing an understanding of rabbit population biology and the ecological factors influencing rabbit abundance. Secondly, measuring the relationships between rabbit abundance and levels of damage to key agricultural crops. Thirdly, developing methods of measuring rabbit abundance. Fourthly, measuring the effectiveness of various management methods. Finally, developing new humane and environmentally sensitive approaches to reducing rabbit damage. Much of this work has been published in the peer reviewed scientific literature while the main means of technology transfer to date has been popular articles in the agricultural press and the production of advisory leaflets. Modern personal computer technology, however, offers the prospect of integrating this vast knowledge base into a scientifically robust rabbit management decision-support system. This system will integrate the extensive existing data set on rabbit biology, agricultural damage and management to provide guidance on costed, scientifically rigorous, coherent rabbit management strategies for a broad spectrum of the main problem settings in lowland agricultural landscapes. The model of rabbit population biology that underpins the system (Smith 1997), has been validated and, in general, makes reliable predictions for future rabbit numbers given different control scenarios (Smith 2001). However, there are four key gaps in our understanding of rabbit biology that need to be filled for a fully functional decision support system to be realised. Firstly, we need to be able to link our rabbit population model to one of the relationship between rabbit numbers and crop damage, in order to predict future losses in a given problem setting. Research thus needs to be completed that will enable extrapolation from our static models of the relationship between rabbit numbers and crop damage (Mckillop et al. 1996, Dendy et al. 2003, Dendy et al. in press) to more complex settings which are dynamic, in that rabbit numbers are allowed to fluctuate naturally and the animals have choices available to them in terms of foraging opportunities. Such dynamic systems are more representative of the real world. It is thus essential for the decision-support system to include such dynamic damage models if it is to be used with confidence under field conditions. Secondly, rabbit population biology sometimes appears to exhibit what is known as the ”Allee effect” (Allee 1931). Here at low densities, population growth is constrained which could explain why rabbit populations sometimes suddenly irrupt, perhaps having crossed some critical threshold. Identifying threshold levels, beneath which populations would be naturally held in check, could be vital to enhanced management strategies. Thirdly, the population model used to underpin the system represents a closed population reflecting variation in productivity and mortality rates with respect to age, season and density. Thus movements between populations, in the form of migration, are not considered. However, immigration may influence the rate of recovery of local populations that have been subject to control. Fourthly, in order to set the population model running for a particular problem setting, it is essential that a reliable estimate of initial rabbit population size be obtained. CSL has developed a validated night count census method to assess the actual number of rabbits utilising a particular tract of agricultural land (Poole et al. 2003). However, this technique requires a series of counts to be made at night and is therefore time consuming and impractical for everyday use at all sites where statutory concerns are raised. What is needed is a simple and robust daytime census method, which can be conducted quickly by advisors during a single site visit, so that a suite of census methods, offering varying degrees of resolution, appropriate for different levels of severity of problem, is available to enable full exploitation of the decision-support system as a tool to inform statutory and advisory services. This project seeks to address each of these gaps in our knowledge and complete the integration of information in the form of a fully functioning decision support system for RDS wildlife advisors.
2. Development of rabbit damage models
A key concept in the development of the decision support system is linking predictions about future rabbit numbers, provided by the rabbit population model for a particular setting, with predictions about the damage that these will cause to crops grown there. We thus need to develop models of the relationships between rabbit numbers and damage to particular crops. In the past yield loss due to rabbit grazing was estimated by protecting areas from grazing and comparing the yields of protected and unprotected areas within fields (e.g. Bell et al. 1999, Church et al. 1953, Crawley 1989, Gough 1955, Gough & Dunnet 1950). However, in the absence of information on the number of rabbits grazing fields in these studies it was not possible to relate yield loss to actual rabbit numbers. The unique facilities at the CSL field ecology research site in Hampshire offer complete control over both the agronomy and numbers of rabbits in large enclosures enabling, for the first time, losses to be related to rabbit numbers. Average yield losses to cereal were 1% rabbit -1 ha -1 for winter wheat (McKillop et al. 1996), 0.8% rabbit -1 ha -1 for grass grown for silage (Dendy et al. 2003) and 0.5% rabbit -1 ha –1 for spring barley (Dendy et al. in press). These damage models allowed us to paramaterise the decision support system but were based on static single-sex populations of fixed size feeding almost exclusively on the study crop. In the real world rabbit populations are dynamic in that they can change through both productivity and mortality. Furthermore, rabbits usually have alternative food sources available to them in field margins, hedgerows and woodland adjacent to the crop. We thus need to develop our static rabbit damage models so that they predict yield losses when rabbit numbers fluctuate naturally and choices are available in terms of foraging opportunities. The required studies were carried out on mixed-sex enclosed populations allowed to fluctuate naturally after establishment. The rabbits could choose to feed on the crop or on an adjacent grassed area where shelter was provided and they were encouraged to construct burrow systems.