Bekæmpelsesmiddelforskning fra Miljøstyrelsen, 105 – Ukrudtsstriglingens effekter på dyr, planter og ressourceforbrug

Ukrudtsstriglingens effekter på dyr, planter og ressourceforbrug

Summary

Mechanical weeding with a view to benefits for open-field fauna

This project investigated the impact of mechanical weeding on the wildlife contained in open fields together with the agricultural resource usage. Based on the results it is recommended that a maximum of two mechanical weeding operations are used in spring cereal crops, with the last not later than 30 days after sowing. Mechanical weeding can be used without significant impacts on energy budgets or gross margin, but at the current level of technology, and without changing rotations, it must be considered a supplement rather than a realistic alternative to herbicides on conventional farms.

The Bichel-committee considered the possibility for a reduction or cessation of pesticide use in Denmark and the associated benefits and problems. Subsequent research has demonstrated that mechanical weeding can be used as an alternative to herbicides in certain crops and in areas where environmental considerations require reduced pesticide use. Pesticide use has been an important factor in the decline of a range of European farmland bird species over the last 20 years, primarily via indirect effects on wild plants and arthropods. Mechanical weeding is rarely used in conventional farming, despite the interest to combine this with herbicide use to reduce pesticide usage. The effectiveness of mechanical weeding for weed control is well documented but there are few experimental investigations into its direct or indirect impact on birds and arthropods in fields.

The aim of the project was to quantify direct and indirect effects of mechanical weeding on skylark reproduction, a number of arthropod species, arable weeds, and agricultural resource usage through the use of field experiment and modelling approaches. An important element in the modelling was to consider the mechanical weeding’s long-term effect on wildlife. Agricultural resource usage is considered in order to compare overall wildlife impacts with energy usage for mechanical weeding. Experimental and modelling results were used to prepare recommendations for ‘wildlife friendly’ use of mechanical weeding together with its influence on agricultural management.

This project (2003-2006), was carried out on the Kalø Estate, by NERI departments of Wildlife Ecology and Biodiversity and Terrestrial Ecology, together with KVL Institute for Agricultural Sciences and DIAS, Foulum. Lektor Mogens Erlandsen (Institute for Biostatistics, Århus University) acted as a statistical consultant. The project is a result of the combination of two separate applications to the Danish EPA’s programme for pesticide research 2003-2005: 1) “Gør mekanisk ukrudtsbekæmpelse en forskel? Effekter på naturkvalitet, indikeret med jordlevende dyr, og økologisk bæredygtighed, indikeret ved det fossile energiforbrug i dyrkningssystemer med og uden brug af herbicider” and 2) ”Effekten af markstrigling/radrensning på sanglærke, hvirvelløse dyr og vilde planter”.

Between 50 and 100 % of the lark nests exposed to mechanical weeding are expected to be destroyed;
Timing of the individual weeding operations in the field determines the number of nests potentially exposed;
Under normal agricultural practices where mechanical weeding replaces herbicides, the risk of weeding affecting skylark reproduction in organic spring wheat or a conventionally managed spring barley or winter wheat is low.
W.r.t. skylark mechanical weeding in spring barley should not occur after approximately 10^th May (or 30 days after sowing), and in winter wheat it should not be used after approximately 10^th April.
Mechanical weeding carried out under normal farming practice impacted the modelled arthropods with the combined effect of three or four consecutive operations resulting in serious negative consequences for the populations.
Main conclusions from the results for the modelled species were that the negative consequences of mechanical weeding in cereal crops can be reduced by:
- Maximally using two mechanical weeding operations in spring cereal crops and not later than 30 days after sowing.
- Not to use weeding in winter cereal crops in autumn, and to only carry out one mechanical weeding operation in the spring before mid-April.
Earthworms, springtails and mites will probably not be affected by mechanical weeding in spring barley on clay soils. It is uncertain whether these results will hold for lighter soils.
Mechanical weeding can be used without significant negative impacts on energy budgets or gross margins.
Mechanical weeding is not considered to be a realistic alternative to herbicides on conventional farms, but rather as a supplement.
In the field experiment here, only the first weeding operation had an effect on the weeds and crop, the following applications were without effect.

Field Experiments 2004 & 2005

Fifty-nine breeding skylark attempts were made in 2004 on the two experimental fields, in 32 breeding territories. In 2005 there were 33 attempts in 20 territories (Table 5.1 and Figure 5.2 A).There were no significant differences between the number of ‘active’ nests (with eggs or nestlings) per territory as a result of mechanical weeding. There were no nesting attempts before the first mechanical weeding operation, and all in all weeding had no direct influence on the overall reproduction. In neither year was there a significant difference between mechanical weeded and un-treated areas in the number of eggs, nestlings, or fledged young per nest or per territory (Table 5.4 and Table 5.5).

None of the experiments showed significant effects of mechanical weeding on the number of earthworms, earthworm species, or earthworm weight (Figure 5.3). The effect on the numbers of springtails and mites was very varied (Table 5.6), and there was no clear effect of the different number of weeding operations (Figures 5.3-5.6). Tests for weeding vs no-weeding showed a statistically significant but very small negative impact on the number of springtails in October 2004, together with a positive effect on number of springtails in May 2005.

On average mechanical weeding reduced weed biomass by 33 %, but there was no clear difference between 1, 2, or 3 operations (Figure 5.7). The percentage reduction was independent of the weed community structure. Harvest was negatively affected by mechanical weeding, but again there was no difference between the numbers of operations used. Yield was reduced on average by 6 %. At a visual evaluation of the plots after the last weeding operation it was not possible to identify the treated from non-treated plots, which indicates that the weeding operation was relatively gentle.

In the field experiment, variation in weed biomass and species composition was so large that it was not possible to determine any general effect of mechanical weeding on weeds. A major cause of this variation was variation in soil type within the field.

For a cattle farm rotation the average diesel consumption will increase by at least 10 l/Ha, representing an average increase of 10 % (calculated on the basis of standard crop values from Figure 5.15).

Conversion to purely mechanical weed control will mean relatively more diesel consumption for pig and arable units compared to cattle farms (see Figure 5.16). Comparing the increase in combined direct and indirect energy consumption associated with conversion to mechanical weed control with the total energy consumption of the three farm types (Table 5.8) show the same tendency: cattle farms increase their overall energy consumption by 1-2%, whereas pig farmers increase by 2-3 % and arable farmers by 6-7 %. These effects are calculated for intensive mechanical weeding which is deemed necessary for a long-term and worst case example for weed control.

Model simulations of skylark, spider and beetle

The simulations showed that none of the standard mechanical weeding operations (i.e. 3 or 2 in spring barley and winter wheat) under the cattle farm simulation had any negative impact on the variables considered. In fact all operations were carried out before the period for the first skylark breeding attempt.

In monoculture spring barley (see section 3.4) there was no negative impact of the four weeding operations on the total population of skylarks (Figure 5.17 B). The effect of delaying the 4^th operation by 10 days until the 20/5 was a reduction of the overall population in all years (Figure 5.17 B). Further delaying of the operation until the 30/5 resulted in a greater reduction since the operation was then carried out when nests with both eggs and young could be destroyed (Figure 5.19).

In a landscape where only winter wheat is grown, the model skylark population did not survive (Figure 5.20). However, the latest mechanical weeding operation (20/4) has a powerful effect of reducing the mean number of fledglings produced and the overall population size (Table 5.12).

The impact of mechanical weeding on the spider was a 22 % reduction of the total population of the baseline cattle scenario (Figure 5.18 A). For the beetle, the impact was a 19 % reduction (Figure 5.18 B).

The isolated effect on spiders of the 2^nd to 4^th operation in spring barley is of the same magnitude (ca. 20 %) (Figure 5.21B; Table 5.13). The impact on the beetle was greater then the spider and highest with the first operation both in spring barely and winter wheat (Figure 5.22 A and C; Table 5.14).