Department for Environment, Food & Rural Affairs

Interim Report on the Effects of the Management of
Field Scale Releases of GM Herbicide Tolerant Crops on the
Abundance and Diversity of Farmland Wildlife (October 2000)

4 PROTOCOL DEVELOPMENT AND PROJECT MANAGEMENT

4.1 Objectives for the period March – September 2000

The project management objectives for the period of this report were as follows:

4.2 Project management and quality assurance

The project is large, and highly dispersed. The project therefore has to be integrated, in that procedures need to be applied consistently, and that the development of the project draws upon the substantial expertise available to us. It also has to be responsive to any changing requirements, including internal quality assurance.

Field survey teams are managed from seven survey centres. They are integrated, through the use of training events, quality evaluation exercises, common protocols, and the simple yet highly effective measure of maintaining frequent communication between teams and with the protocol leaders and project managers.

4.3 Staff training

The field staff were already trained to an appropriate level for the vast majority of the work they were expected to do. Training has therefore concentrated on developing an understanding of the precise methods and goals of the project, and on topping-up for particular individuals for particular methods. Much of this has taken place within the survey teams themselves, and with exchanges between teams. In addition, there were three major training events:

General field staff training exercise, Rothamsted, 7-8 March
This covered the full range of field protocols.

Gastropod field assessments, Rothamsted, 24 July 2000
This concentrated on field identification

Statistical workshop, Dundee, 26-27 September 2000
This meeting introduced statistical, data management and analysis issues, and began the process of considering how to report the results of the project.

4.4 Quality evaluation exercises

The first major field quality evaluation exercise was a tour of all field survey centres by one of the crop co-ordinators in July 2000. This involved working with survey teams, and holding discussions with them to identify inconsistencies and clarify uncertainties. In general, there did not appear to be any major inconsistency across field teams in the approach adopted for each of the observed protocols.

4.5 Collaboration with other monitoring exercises

Data are also being collected from some or all of the sites as follows:

We collaborate with each of these groups in providing access to field sites and, where required, facilitating contact with the farmers.

In addition, site visits are conducted by SCIMAC and CSL for ensuring that the crops are managed within the conditions of the consents; we are not involved in this process.

4.6 Re-assessment of field protocols

The project is managed to include regular re-assessments of the work programme. This re-assessment is particularly important now that the first full field season is almost complete, especially for the field protocols.

Some protocols have developed during the season (see also Section 2). Other protocols have developed as a result of early experience within the season.

While it is important that the protocols do not change so much that it becomes impossible to collect comparable data from across the whole project, there are nevertheless grounds for rethinking the protocols. They are:

The sampling intensity for a given protocol should be enough so that we are confident that we can separate the signal of any differences between treatments from the noise of random variation within treatments.

Here, the issues and changes concerning the different protocols are reviewed.

4.6.1 Crop assessment
Crop assessments provide a measure of density, crop height and phenological development. They have been simplified so that initial density is measured along drill lines and thereafter using quadrats.

4.6.2 Margin attributes
Small changes were made to the list of features to be noted at each sampling station. The aim was to ensure that they could be coded up uniformly for entry to the database, avoiding reliance on features recorded only on maps, which will be of little use in analysis.

4.6.3 Soil seed bank
Preliminary analyses of soil seed bank data revealed a satisfactory distribution of species and abundance across sites, as well as acceptable variation within sites (Section 3.1). There is therefore no statistical reason to change the protocol. One safety issue did arise, however. It became apparent that slurry and other wastes had been applied to some sites before sampling. Accordingly, an additional seed bank record sheet was prepared for the autumn 2000 samples. In this, the site recorders were asked to indicate whether farm yard manure animal slurry or human effluent had been applied to the field; the form allows those preparing the soil for testing to take the necessary precautions.

4.6.4 Weed seedling counts
The area counted in each quadrat was reduced from 0.25 m² (the area sampled in 1999) to 0.125 m²; crop sampling and cover estimates continued to be for the full 0.25m². This change had already been notified to the SSC and has minimal effects on sampling error because most of the variation in weed density within a field is between quadrats. Where weed densities exceed 100 individuals per 0.125 m², surveyors are permitted to make estimates based on one-half or one-quarter of this area.

4.6.5 Weed biomass
The main change to weed biomass sampling is that weeds are counted in a standard 0.125 m² quadrat at the same time. It has been found logistically more efficient to sort the vegetation fresh and to freeze plant material only in emergency. Where biomass of weeds is very large, so that the sample from each 1 m² quadrat cannot fit into a single medium-sized bag, surveyors are permitted to sample a smaller area, down to 0.25 m². The need for this is exceptional, but occasional very weedy fields are encountered.

4.6.6 Seed rain
A method for recording seed rain was tried for the first time in the 2000 season. The data are currently being processed and any methodological problems (such as collection in wet conditions) are being reviewed.

4.6.7 Subsequent vegetation
Recording of subsequent vegetation is by the standard protocol for counting weeds in the crop. It was found that by late June, spring-sown cereal crops had sometimes lodged, making movement difficult and reducing the accuracy of sampling. The plan for next season is therefore to make these counts in late May or early June.

4.6.8 Edge vegetation
Sampling intensity for edge vegetation was increased in 2000 to provide estimates of flower availability, presence of seeds and herbicide spray damage. Flowers and herbicide damage were sampled on 3 occasions at each of 9 stations beside each half-field, with separate records for a 10x1m strip in the boundary feature, the verge and the crop margin. Seed presence (3 occasions) and full species enumerations (1 occasion) were made at 3 stations round the field, also recording boundary, verge and crop margin. The protocols were developed, following recommendations from the SSC, in spring 2000 and will be discussed further with field surveyors at the end of October. The results of these discussions are not available at the time of writing.

4.6.9 Field margin gastropods
No current concerns.

4.6.10 Within-field gastropods
Overall gastropod counts were low during the field season. While causes may have included inefficient traps, or problems with always sampling under appropriate weather conditions (additional experiments are being analysed to explore this issue further), it is also not unlikely that these species were scarce within these crops during the summer. It is likely that an increase in trap numbers may be appropriate, for example by increasing the number of traps per sampling station to more than one.

4.6.11 Bees and butterflies
No current concerns. The requirement for good weather is inevitable.

4.6.12 Crop pests
Crop pests were initially sampled on the basis of presence / absence on 90 plants per half field; this was changed to sampling on a count basis on 45 plants per half-field.

4.6.13 Invertebrates on vegetation
No current concerns. The requirement for dry weather is inevitable.

4.6.14 Soil surface arthropods
There is concern that samples are large, which has resource implications because of the time required for sample identification.

4.6.15 Crop management data
Crop management data are being collected, but, in many cases farmers fail to appreciate the urgency of this information and are reluctant (or too busy) to send it immediately after a management event. This can cause mis-timings of surveys, and has the potential to risk staff safety if they visit a site too soon after some pesticide applications. At the end of the season completed management forms are sent to the farmer for verification and if necessary correction – again it is variable how quickly this is achieved. Methods to improve the ease of collection of this data are being sought both in terms of information to the farmers and the design of the forms/data sheets used.

4.7 Health and safety

The project has had only two safety problems reported. The first is the risk of staff unknowingly handling slurry and other wastes from seed bank samples; this risk is being addressed by revising the field record sheet (see Section 4.6.3).

The second problem is that there have been two minor injuries with the Vortis suction samplers for invertebrates. These will be reviewed during the late October meeting, and any actions will be agreed then (e.g. additional training, contact with the manufacturer).

There remains a potential concern that staff may enter fields too soon after the application of agrochemicals (Section 4.6.15). We appreciate the need to facilitate communication with farmers to ensure that crop management does not compromise staff safety.

[ Previous ] [ Contents ] [ Next ]

Published 2 January 2001
Genetically Modified Crop Farm-Scale Evaluations Index
Environmental Protection Index
Defra Home Page