Internship project: Impact of field characteristics on the allocation of agri-environmental schemes
by Meike Will
By Johanna Eichler
I spent six weeks during February and March 2022 as an intern at the POLISES working group. Unfortunately, I did not spend the internship in Leipzig but in home office in Osnabrück where I am studying for a Master's degree in System Science. Nevertheless, I have learned a lot and will talk about some of that now.
I worked in thethat aims to improve the design of EU rural policies to contribute to a more sustainable agricultural sector. The focus of my internship was on the agent-based model (ABM) to simulate farmers’ decision-making when choosing and allocating voluntary agri-environmental schemes (AES) on their fields. In the model three decision making steps are considered:
- First, the openness of the farmer towards specific AES, influenced for example by their intrinsic openness or access advisory services.
- Second, the selection of suitable fields for example, depending on the land use required for the AES.
- Third, comparing the accepted payment with the offered payment and selecting fields where AES should be adopted.
I was concerned with the second part of the third step: selecting the specific fields where a farmer allocated AES. Previously, the selection of fields was dependent on two characteristics: the soil quality and size of a field. Farmers would choose fields with the lowest soil quality and smallest size. Since for the German case study of the project, a regression analysis on the impact of 17 farm and field factors on the allocation of AES on fields was available, the plan for me was to implement these findings and see if and how the chosen fields would change.
Based on the regression analysis, I implemented nine additional field variables from this analysis and calculated one indicator for the suitability of the field I called “desirability”. Running the model with the previous and the new sorting mechanism and comparing the fields that were chosen showed that there was a difference in field allocation for the two mechanisms. The majority of fields was chosen for both but for some the sorting algorithm led to different results. The new sorting by desirability has better support by data and for this reason alone could be the preferred method. However, a large fraction of fields (in my sample around 80%) was chosen for both procedures. Soil-quality and field size might already be good proxies for the overall suitability of a field for AES adoption. Still, combining all factors into one indicator proved to simplify the implementation of the decision-making process in the model, which supports maintaining the new sorting mechanism.
Overall, I feel I got a real insight into research project work, taking part in meetings and talks and experiencing some challenges in “real-world” modelling first hand. Of course, I also improved data analysis and coding skills. Even remotely, I had great supervision and a really interesting time and would recommend an internship at the POLISES group for anyone interested in environmental research and its workflows.