6 Output from the sugarbeet model
 | 6.1 Introduction |
| 6.2 Materials and methods |
| 6.3 Results |
| 6.4 Discussion |
The content of this chapter was presented as a paper at the Second International Weed Control Congress in Copenhagen 1996 under the title ‘Simulation of herbicide use in a crop rotation with transgenic herbicide resistant sugarbeet’ by the authors K. H. Madsen, W. M. Blacklow and J. E. Jensen. The model was constructed during a five month study visit to University of Western Australia, Faculty of Agriculture in Perth.
6.1 Introduction
Background
In Denmark there is increasing public concern about the levels of pesticides used in agriculture, which has lead to an action plan to reduce pesticide use by 50% during the period from 1987 to 1997 (Haas, 1989; Thonke, 1991). This concern is, furthermore, raised in the debate over the merits of transgenic herbicide resistant crops (HRCs) in Danish agriculture. A specific concern is whether the introduction of HRCs will cause a consequential increase in herbicide use .
Concerns
It is expected that any weed population that is repeatedly treated with the same herbicides will change its composition towards species or biotypes with a higher natural tolerance to these herbicides. Furthermore, an extensive use of fewer herbicides and uncontrolled gene flow may lead to problems with resistant weeds. This may render the environmentally benign herbicides ineffective and leave the farmer with the older and not so environmentally benign herbicides if these are still available. These possibilities lead to environmental concerns about future herbicide use. Farmers need to know which combinations of resistance genes, crops, herbicide dose rates, application strategies and crop rotations will minimize the risk of losing the environmentally benign herbicides, because many of the environmentally malign herbicides may be banned in the future. Coupled to this requirement is the need to know strategies that will meet the requirements of the Danish Action Plan.
Present knowledge
Our present knowledge about the consequences of growing genetically modified HRCs on a large scale is limited because it is based primarily on small scale releases and on expected similarities with current agricultural practices.
Oilseed rape and sugarbeet
Two HCRs of current concern are oilseed rape and sugarbeet. Both species hybridize within the species as well as with wild relatives (Jørgensen & Andersen, 1994; Madsen, 1994). However, neither HRC had a greater competitive ability than non-transgenic cultivars (Fredshavn et al., 1995; Madsen, 1994). Consequently, herbicide-rates would need to be maintained in the HRCs.
Clearly, HRCs will have impacts on herbicide use. At this early stage in the release of HRCs, simulation provides a method to investigate these impacts. Specifically, the model considers herbicide use in glufosinate and glyphosate resistant sugarbeet in Denmark.
6.2 Materials and methods
STELLA
The model simulated the growth of crops and weeds in a rotation for sugarbeet and was programmed in STELLA (Peterson & Richmond, 1994). The model included four weed species (Chenopodium album, Stellaria media, Viola arvensis and Elytrygia repens). The weed species and volunteers regenerated in succeeding crops from a simulated seed bank (bank of propagules). Growth of the different species was modelled with logistic growth curves and competed for the site potential which was determined by the maximum biomass of the crop. The crop was sprayed with herbicides at a fixed biomass of weeds resulting in an efficacy of control dependent upon the weed species. Dicotyledonous (dicot) weeds in the barley and wheat crops were sprayed with a sulfonylurea herbicide. Grass weeds growing in cereal crops was controlled with a pre-harvest application of glyphosate when biomass of the grass weed exceeded a critical level at the end of the growing season.
Scenarios
The model simulated three scenarios for weed management in a rotation of sugarbeet - barley (Hordeum vulgare) - wheat (Triticum aestivum) - wheat rotation:
1. A non-transgenic sugarbeet sprayed with a mixture of metamitron, phenmedipham and ethofumesate for the control of the dicot weeds and a grass herbicide to control the grass weeds and cereal volunteers.
2. A glyphosate resistant sugarbeet sprayed only with glyphosate.
3. A glufosinate resistant sugarbeet sprayed only with glufosinate.
Input
Information for the model was from several sources (Madsen, Poulsen & Streibig, 1997):
| The crop rotation was constructed from agricultural statistics (Danmarks Statistik, 1994). |
| Selection of weeds was based upon a recent Danish survey (Andreasen, 1990). |
| Initial levels of weed seeds were estimated from a Danish survey (Jensen & Kjellsson, 1993). |
| Initial levels of volunteers from different crops were estimated from the literature and field trials with sugarbeets (Alstedgaard, 1994). |
| Herbicide doses were those currently recommended (Statens Planteavlsforsøg, 1997) and which were considered likely to be used in the future. |
Appendix 1.
For specific model equations and parameter values used in the sugarbeet model, see appendix 1.
6.3 Results
Herbicide use
The model simulated the growth of crops and weeds for the three scenarios of weed control (Fig. 6.1), and the simulated levels corresponded with field experiences. In sugarbeet the critical levels of weed biomass that initiated a spraying were low and, consequently, the several herbicide applications predicted by the model agreed with the practice of weed control in this crop.
Herbicide use accumulated over time (Fig. 6.2). The greatest amounts of herbicide were applied to sugarbeet and spring barley was only sprayed between one and four times in the different scenarios (Fig. 6.2). Introduction of HR sugarbeet mainly decreased the amount of herbicide applied to that crop and had only minor influence on the amounts of herbicides applied to other crops in the rotation (Fig. 6.2). Over 20 years the total amount of herbicides used in the rotation was almost halved through the introduction of the HR sugarbeets.
6.4 Discussion
Model behaviour
The model simulated growth of crops and weeds in the four-year rotation through five cycles of the rotation. Applications of herbicides were triggered when the weed biomass reached critical levels. The simulated levels and dynamics of crops and weeds agree with those observed in the field. Furthermore, the simulated sprayings and levels of herbicides used in the rotation without HR sugarbeet agree with those used currently in the field. It can, therefore, be concluded that the model provides a basis for comparisons in herbicide use when HR sugarbeets are introduced into the rotation.
Assumptions
The introduction of HR sugarbeet into the model showed reductions in the amounts, and frequencies of herbicide use. These results were obtained under the assumptions that the introduction of HR sugarbeet will not change the current rotations, weed flora and herbicide use in other crops of the sugarbeet rotation. These assumptions can, and should, be challenged. The current simulation model can be extended to explore these assumptions.
Dose-response curves
More information is needed about herbicide efficacy on the weed species, volunteers and hybrids occurring in the HRC. These could be included in the model by experimentally determining the dose-response curves for the weedy species sprayed with (A) the herbicide used in the transgenic crop and (B) the herbicide(s) currently used. This makes it possible to compare efficacy at different dose levels of herbicide A based on practical use of herbicide B (Madsen & Jensen, 1995).
Other rotations
The simulation model can be adapted to investigate herbicide use in other rotations of Danish agriculture. Sugarbeet rotations occupy only 10% of Danish arable land and, although herbicide use on this land is currently high (Madsen et al. 1997), reductions in herbicide use in other rotations through the introduction of other HRCs have the potential to assist in meeting the objectives of the Danish Action Plan for pesticide use. Again, however, the present assumptions in the model about unchanging practices and flora as well as herbicide efficacy in different weeds need to be investigated. To these ends simulation and experiments need to be carried out on concert.