Forside Nyt Viden Om DMU Kontakt os Søg In English
 

Integrating agricultural statistics and environmental pesticide models

Jesper Sølver Schou, Johnny M. Andersen, Peter B. Sørensen

NERI Technical report, no. 326

Abstract

The main objective of this report is to examine the methodology for linking a set of agricultural models with environmental pesticide-models used in the IMIS framework of the National Environmental Research Institute. The linkage has been tested and discussed by analysing the effects of the Agenda 2000 scenario on the risk of pesticide loading to groundwater.

The IMIS work in the agricultural area is concentrated on establishing a set of models, which can be used for integrated environmental economic analyses. In this context a NP-model is employed for model prediction of agricultural nitrogen and phosphorus emissions. The model describes agricultural production according to farm types and catchment areas. In order to perform scenario analyses where farmers response to changes in – among others – agricultural and environmental policies demonstrated, the NP-model is linked to an econometric sector model (ESMERALDA). ESMERALDA, in turn, is connected to the economic model ADAM, so that the socio-economic aspects of a specific course of development can be examined.

Neither ESMERALDA nor the NP-model includes a specification of pesticide use, which naturally plays a central role in the connection with the linkage to the environmental pesticide-model. Therefore the Pesticide Statistics of the Danish Environmental Protection Agency (Miljøstyrelsen) are used to provide an initial description of pesticide use in agriculture. The Pesticide Statistics include a distribution of pesticide use according to crop type based on sales figures. By connecting these statistics with the crop types in ESMERALDA and the NP-model, the pesticide use described in the Pesticide Statistics can be transferred to these models.

In connection with the scenario-analyses the results from ESMERALDA are transformed to the environmental pesticide-models in two steps. First, the changes in pesticide use resulting from changes in crop mix are taken into account, and then the changes in intensity of pesticide use for the individual crop types are incorporated. In this way, the scenario results from the environmental pesticide-models can be decomposed in a crop mix effect and an intensity effect. Compared to the Pesticide Statistics the description of pesticide use in the ESMERALDA model is presented in a very rough fashion, namely as the expenditures to pesticides per hectare on the crop type level. This simplifies the linkage of models, as it is assumed that the pesticide description in the Pesticide Statistics simply corresponds with the pesticide expenditures in ESMERALDA. However, for the scenario analyses this results in a number of limitations, as changes in pesticide use are not explicitly categorised according to single substances or even the general pesticide types (herbicides, fungicides, insecticides, and growth regulators). In the present analyses it is therefore assumed, that the distribution of the separate pesticides according to crop types remains unchanged and that a change in pesticide intensity appears as a uniform percentage change in the use of all pesticide types for a given crop.

In the scenario analyses carried out, the effects of the Agenda 2000 reforms of the European Union’s common agricultural policies are compared with the situation in 1995, where it is assumed that all measures of the Second Danish Action Plan for the Aquatic Environment (Vandmiljøplan II) are fully implemented. Thus, two hypothetical situations are compared, which, however, do reflect a realistic development in the agri-environmental policies. The two scenarios are compared by using a non-quantitative ranking of pesticide use based on differences in the pesticides’ dosage, the area where the pesticides are applied, and the pesticides adsorption to soil particles. The results primarily relate to a qualitative ranking of the risk of pesticide leaching. The overall result is that the Agenda 2000 scenario reduces the risk of pesticide leaching. Looking only at the effects resulting from changes in crop type composition shows a weak indication of the opposite result. However, because of the reduction in the intensity of pesticide use caused by the Agenda 2000 reform cuts in intervention prices for cereals, the overall effect is a reduction in the risk of pesticide leaching.

With the scenarios presented here it is the first time that the environment pesticide-models are employed for this kind of analyses. The possibility for assessing the certainty for the difference in ranking of two scenarios has emerged as a main point of discussion. Here it is suggested to focus on the number of important events, i.e. events that are ranked differently with regard to all parameters, and afterwards let that variable determine the significance of the results. The significance of scenario rankings using the environment pesticide-models is an obvious field for further development, and it is especially important in cases like the examined one, where the outcome is a result of contrasting effects.

The application of the Pesticide Statistics for description of pesticide use is another area for improvements. As mentioned, the advantage of using the Pesticide Statistics to initialise pesticide use in the NP-model and ESMERALDA-model is that this implies consistency between the applied figures. In its current form, however, the classification of crop types in the Pesticide Statistics does not corresponding with the classification used in the set of agricultural models. As a solution the land use in the NP-model has been readjusted, but on a longer perspective, a harmonisation of crop types could be an advantage.

Further, the Pesticide Statistics are based on pesticide sales figures, which are subsequently disaggregated to crop types and sprayed area on the basis of expert valuations. This means that the Pesticide Statistics are normative, and do not enable an analysis of variations in the applied dosage, e.g. in the form of reduced dosages in repeated treatments. Further, it is not possible to analyse local and regional differences in the pesticide use. This results in a substantial limitation of the application of the environment pesticide-model, because the model is not linear. The result, therefore, is sensitive to the fact if a certain pesticide is applied in one treatment at 1 litre per ha or in two treatments at ½ a litre per ha. As the Pesticide Statistics are based on disaggregated sales figures these two examples will be identical in the statistics, while on the other hand they will not necessarily lead to the same results in the environment pesticide-model. The same problem arises in connection with the scenario analyses, because of the aggregated description of pesticide use in ESMERALDA.

An improved description of pesticide use in agriculture could be carried out by establishing a recording of farmers pesticide use, instead of a disaggregation of sales numbers. A pilot study has been carried out at the Danish Institute of Agricultural and Fisheries Economics, which shows how such an assessment can be undertaken, based on reporting from the mandatory spraying records. This method is also recommended in the OECD and EUROSTAT’s guidelines for the collection of national pesticide usage statistics (OECD, 1999). Such statistics of pesticide use will significantly improve the basis for both environmental and economic analyses of agricultural pesticide use, and thereby – e.g. with the model linkages described here – further support the political decisions in this area.

 

Bedes citeret: Schou, J.S., Andersen, J.M. & Sørensen, P.B . (2000): Integrering af landbrugsdata og pesticidmiljømodeller - Integrerede MiljøInformationsSystemer (IMIS). Danmarks Miljøundersøgelser. 62 s. - Faglig rapport fra DMU, nr. 326. (online). http://faglige-rapporter.dmu.dk

Full report in pdf format (597 KB)

Dato for sidste opdatering 2002-07-10 Til toppen Tilbage Frem Send os dine spørgsmål og kommentarer Sitemap Links