Project
title / Developing a rabbit management expert system
/ DEFRA
project code / VC0220

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 / Developing a rabbit management expert system
DEFRA project code / VC0220
Contractor organisation and location / Central Science Laboratory
Sand Hutton
York YO41 1LZ
Total DEFRA project costs / £ 578,094
Project start date / 01/04/98 / Project end date / 31/03/02
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 (9/01) 3

Project
title / Developing a rabbit management expert system
/ DEFRA
project code / VC0220
  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.
  2. 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.
  3. A substantial knowledge base has accrued through considerable DEFRA investment in rabbit research which could be integrated, using modelling approaches, into a system for supporting RDS staff in their advisory and statutory work regarding rabbit damage.
  4. This project aims to produce a decision-support system enabling guidance on the choice and timing of control methods to be optimised for the conditions on individual holdings.
  5. Underpinning the system are models of rabbit population biology and rabbit damage to crops in terms of yield loss.
  6. The rabbit population model allows different types of control to be simulated in each month, and the resultant population size to be projected.
  7. Damage models, for static systems with constant rabbit numbers with no alternative foraging opportunities available, have been developed that result in yield losses of approximately 1% per rabbit per hectare for winter wheat and grass grown for silage, and 0.5% per rabbit per hectare for spring barley.
  8. Studies of dynamic damage systems, with fluctuating rabbit numbers and foraging choice, suggest that rabbit grazing pressure is highest on young growing cereal plants in early Autumn and early Spring. These data will allow more realistic models of rabbit damage to be incorporated into the completed decision support system.
  9. A prototype system, with a user-friendly front end, has been completed that integrates the population and damage models with information on the effectiveness and costs of various management options.
  10. The system can predict, for a given lowland agricultural context, future rabbit numbers under different management scenarios, estimate the amount of crop damage that will accrue, assess the effectiveness of various methods of reducing this damage and the benefits that would accrue, in terms of the potential increases in crop yields and their associated cash values.
  11. The inputs required from the user are relatively simple and include an assessment of current rabbit population size, the length and type of field margin over which damage can occur, crop type and growing season, and the management options available.
  12. Unified project VC0232 will now bring together the various elements of the current rabbit research programme, in terms of developing the decision support system with underpinning rabbit population and dynamic damage models (VC0220), understanding migration and density dependent population processes (VC0227) and developing simple census methods (VC0228), thereby completing the integration of information in the form of a fully functioning decision support system for use by RDS wildlife advisers by 1 April 2004.

CSG 15 (9/01) 3

Project
title / Developing a rabbit management expert system
/ DEFRA
project code / VC0220
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 (9/01) 3

Project
title / Developing a rabbit management expert system
/ MAFF
project code / VC0220

Introduction and Policy Rationale

The rabbit population in the UK has continued to increase in size as it recovers from the waning effects of myxomatosis. In the mid-1980s economic losses to agriculture were estimated at £100 million. 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 but only recently has the prospect of integrating this vast knowledge base been realised. This has been the focus of this project in terms of developing a computer based decision-support system that would assist in advising on the costs of damage in a given setting and the optimum choice of management methods and the timing of applying these. The aim is to provide RDS wildlife advisors with a powerful guidance tool that would predict the effectiveness of particular methods for reducing damage and the benefits that would accrue, in terms of the potential increases in crop yields and their associated cash values. The system also needs to be versatile so that further information, not only on the effectiveness of control but also on the costs of damage, could be incorporated as and when it becomes available. This would enable continued improvements of the quality and usefulness of the system to be implemented easily and cheaply. Furthermore the system has to be robust and comprehensively validated in the field.

Prior to the commencement of this research the only comparable work has been a more simple approach made by the New Zealand MAF (unpublished) in which the sole option was to examine the cost-effectiveness of poison baiting. The modelling research in this project is thus providing for the first time for rabbit management a comprehensive system, covering not only data on the costs and effectiveness of control methods, but also using predictions for the financial consequences of various management options to facilitate cost-benefit analyses. This comprehensive approach will be the first real attempt to produce such a model for any species. The previous project VC0212 has shown that rabbit density is related to habitat features (Trout et al. 2000), fecundity is dependent upon soil type (Trout & Smith 1995) and not population density (Trout & Smith 1998), different rabbit control methods produce different biases in the remaining population (Smith et al. 1995) and density-dependent mortality can change the optimum timing of control (Smith & Trout 1994, Smith 1997). In addition work has been published on theoretical aspects of data analysis (Smith 1995). The first stage of this project has been to use this detailed understanding of rabbit population biology to build a model of rabbit population dynamics, in which the consequences for rabbit population size of applying different types of management at different times of year can be predicted. These predictions have then been tested to validate the population model. The next phase has been to incorporate the costs of damage, as revealed by previous work (e.g. VC0214), whilst at the same time carrying out further damage studies to allow cross-over from the highly controlled conditions under which these initial data were collected to settings representative of typical agricultural damage. Finally, the costs of different management methods are to be incorporated to allow full cost-benefit analyses to be applied to determining optimal management strategies.

Ultimately the availability of the full decision-support system will enable informed and robust advice to be given that will lead to:

·  full utilisation and exploitation of the data generated by DEFRA’s considerable investment in rabbit research,

·  more cost-effective use of resources by growers and others when deciding upon control methods,

·  long-term reductions in rabbit numbers as a result of improved choice of control operations enabling farmers to deal better with critical problems,

·  improved ability to advise farmers, and others with rabbit problems, on how to tackle these hot-spot problems in a cost-effective, long-term manner.

Developing yield loss models to calculate per capita losses caused by rabbits

Prior to the commencement of this research, yield loss due to rabbit grazing was calculated by protecting areas of crop from grazing and comparing the yields of protected and unprotected areas within fields (Church et al. 1953, Crawley 1989, Gough 1955, Gough & Dunnet 1950). However, the number of rabbits grazing a field could not be accurately determined (but see projects VC0219 & VC0228) and therefore it was not possible to relate yield loss to rabbit numbers in those field studies. The unique facilities at the CSL rabbit research site in Hampshire, where we have complete control over both the agronomy and numbers of rabbits in large enclosures, has enabled for the first time, yield loss figures where losses were related to rabbit numbers at the individual level (McKillop et al. 1996). This work was extended to estimating damage at the level of static single sex populations of constant size under VC0214 and a key figure of 1% loss per rabbit per hectare derived for winter wheat. The first phase of the current project has been to extend these studies to derive models of the damage caused by static populations to spring barley and grass grown for silage (Milestones 01/02 and 02/02).

Static model of rabbit damage to spring barley

In a previous study (VC0412), where rabbits were given a choice of 7 cereal crops and oilseed rape on which to graze, spring barley was found to be the least favoured cereal crop, resulting in the smallest yield loss at harvest. Winter wheat was found to be more vulnerable to rabbit grazing. Potentially, farmers could reduce crop losses by using this information to make appropriate decisions about crop selection depending on the levels of rabbit grazing in their fields i.e. growing more spring barley in areas with higher rabbit populations. Nevertheless, spring barley is widely grown throughout the UK and significant yield loss can occur as a result of heavy rabbit grazing. A study of rabbit damage to a spring barley crop when grown alone was thus undertaken.

Methods

A spring barley crop (variety Alexis) was sown in February of each year (1996-1998) in six enclosures each of approximately 1 hectare in size, at the CSL rabbit research facility. Identical agro-chemical treatments were applied in each enclosure when required throughout the trial. Rabbit-proof fences, designed to prevent movement of rabbits between treatments, separated enclosures from each other. The fences were regularly checked throughout the course of the experiment and any damage repaired when necessary. Wooden rabbit-boxes were put into each of the enclosures containing rabbits to provide them with shelter. Each year the four enclosures to contain rabbits were randomly selected, as were treatment densities. Rabbit populations of appropriate density were established using adult male rabbits to ensure the population levels would remain constant. Each rabbit was fitted with a coloured, numbered rototag. Rabbits were placed into the trial about two months after the crop had been sown. A density of 18 rabbits per hectare was used every year to provide a standard measure of the effects of grazing between years; the other densities used were 11, 28 and 35 ha-2.

Every year small (4x4 metre) pens were erected in each enclosure to prevent rabbit activity so the effects of grazing on individual plants could be assessed. Each pen was erected in one of three rows using a randomisation plan to determine its position. Two sets of pens Control (C) and Grazed (G) were established. The control pens were erected at the start of the experiment, before any rabbits were placed in the enclosure and grazed pens were erected 6 weeks later, midway through the trial. The positions of the Always Grazed (AG) pens were clearly marked throughout the trial however fences were not actually erected, enabling rabbits to graze on the area throughout the trial.

At the time of pen erection crop samples were taken from each enclosure. A 0.5m stick was placed parallel to the line of crops (approximately 1m away from the site of each pen) and all the plants on either side of the stick were removed. The samples were used for damage assessment and dry weight determination.

The crop was harvested within each enclosure (excepting that within the fenced and marked pen areas) to give a total yield value for each enclosure. Grain samples were taken at this time to determine the moisture content. The grain from each enclosure was weighed together to give a total yield for each enclosure.

A Sampo small-plot combine was used to cut a 2m strip of crop from each pen and marked pen area. Samples of approximately 2000g of grain were taken from the crop removed from each strip to determine specific weight and 1000 grain weight. Quality analysis was also performed on the grain taken from within the pens.