Epsilon Archive for Student Projects

Abstract

Agricultural expansion, prompted by increasing global food demand, has fragmented, simplified, and homogenized the landscape in the last decades. The resilience and sustainability of croplands are at stake, resulting in a hight vulnerability to climate change, and having major impacts on environmental protection, biodiversity preservation and food security. As agricultural landscapes are becoming more and more homogeneous, it is essential to identify the impacts of forest edges on adjacent cropland and their provision of ecosystem services. Scientific research has shown a decrease in soil organic carbon (SOC) and total nitrogen (TN) concentration at a greater distance from hedgerows, but data and methodological studies along forest edges remain scarce.
This study was conducted to identify the spatial variability in yield across a rye field with forested edges at the Research Station of Bjertorp, Kvänum in Sweden. Ecosystem services chosen for assessment included carbon sequestration, nitrogen retention and crop production, studied through indicators such as soil organic carbon, total nitrogen, and yield, respectively.
The objectives of the study were multiple: first, to provide a methodological framework for collecting and analysing data on spatial variability in a field at the forest edge for SOC, TN and yield. Second, the ability of the Decision Support System for Agri-Technology Transfer (DSSAT) to predict this spatial variability was assessed. Finally, the identification of particular patterns in the spatial variability of each ecosystem service was discussed as well as their potential attribution to proximity to forest edges.
The methodology showed a good efficiency in the preparation of data, as the simulation was run in ArcMap and DSSAT without major inconveniences. The use of ArcMap required some knowledge of GIS but provided a useful spatial visualisation and interface for data processing. However, the predictions showed mixed results, with poor agreement between measured and predicted yield for the areas close to the field edges. A map of the yield prediction identified the need for improvements regarding the spatial variability resulting from the proximity to the forest edges. This lack of ability to depict in-field variations are assumed to be related to the lack of microclimatic data, an important factor defining the production potential. On the other hand, DSSAT forecasted a SOC accumulation in the long run with an assumed good agreement between observed and simulated data. It was difficult to capture the nitrogen cycle due to external input and crop uptake.
The collection of higher-quality data such as microclimate or site-specific soil data, as well as re-calibration, could improve the predictions accuracy. Agro-ecosystem models could also be improved to depict soil heterogeneities and tree-crop interactions.
The outcomes of this work could be used to run impact assessment studies and evaluate the effect of different landscape and land organization policies designed to increase the sustainability of food systems, resulting in new objectives of diversifying the landscape, broadening the ecosystem services to be promoted and respecting the environment.