Agrivoltaic systems potentials in Sweden: a geospatial-assisted multi-criteria analysis

Mohamed Elkadeem, Sebastian Zainali, Silvia Ma Lu, Ali Younes, Mohamed Ali Yousif Abido, Stefano Amaducci, Michele Croci, Jie Zhang, Tomas Landelius, Bengt Stridh, Pietro Elia Campana2023 · DOI

Abstract

Agrivoltaic (APV) systems represent an intelligent solution combining electricity production from solar photovoltaic technology with agricultural production, effectively avoiding land-use conflicts. Geographic Information System (GIS) technologies can support the implementation and spread of APV systems by identifying the most suitable areas using spatially explicit information concerning techno-agro-socio-economic criteria. In this study, we developed a procedure to identify and classify suitable areas for agrivoltaic systems in Sweden. An Ordinal Priority Approach (OPA) based multi-criteria decision-making algorithm was established to calculate the weights of the selected evaluation criteria through expert interviews. The results show that 8.55% of the Swedish territory, approximately 38,485 km², is suitable for installing APV systems. Through the deployment of vertically mounted agrivoltaic systems with bifacial photovoltaic modules, the land classified as "excellent" and "very good" could potentially provide about 207 TWh, significantly exceeding Sweden's 2021 electricity consumption of 143 TWh.

Introduction

As nations race toward net-zero, a "land-use tug-of-war" has emerged between renewable energy expansion and food security. In Sweden, agricultural land is legally protected as being of "national importance," making it difficult to permit traditional ground-mounted solar farms. This study addresses this conflict by evaluating Agrivoltaics (APV)—the co-location of solar panels and crops.

A key advantage of the proposed model is the focus on vertically mounted, bifacial PV modules. This setup only reduces effective crop area by 5–10%, allowing for 90–95% of traditional crop yields while generating significant power. This dual-use approach effectively bypasses the "food vs. fuel" legal barrier and supports the Swedish Environmental Code by increasing the economic viability of smallholder farms.

Methodology

The researchers employed a five-step GIS-MCDM (Geographic Information System - Multi-Criteria Decision Making) approach to map the national potential for APV systems.

Methodology flowchart for the geospatial-assisted multi-criteria analysis.

The process began with constraint mapping to exclude unsuitable areas, followed by criteria selection and weighting using the Ordinal Priority Approach (OPA). Unlike the more common Analytic Hierarchy Process (AHP), OPA is noted for its robustness in handling expert opinions to determine which factors matter most. Through this algorithm, the study found that Solar Irradiation (30.15%) is the dominant factor, followed by Aspect (17.3%) and Land Use (16.36%).

Datasets and Constraints

The study integrated several high-resolution datasets, including Corine Land Cover (CLC) 2018 for identifying agricultural land types and SMHI (2022) for seasonal precipitation. To derive the restriction map, several binary constraints were applied to "black out" unsuitable land.

Restriction criteria and technical thresholds for APV site suitability.

As shown in the restriction criteria, areas with steep slopes (>10%), northern aspects, or close proximity to airports, settlements, and protected areas were excluded. Technical thresholds for solar irradiation (GHI < 785 kWh/m²/year) and water stress were also established to ensure the viability of both energy production and agricultural yield.

Key Results

The analysis identifies that approximately 38,485 km² (8.55% of Sweden) is viable for APV. The suitability was classified into five categories, with 0.17% of the total area deemed "excellent" and 15% "very good."

  • Energy Surplus: Utilizing only the "excellent" and "very good" areas could provide 207 TWh/year, far exceeding Sweden’s 2021 electricity consumption of 143 TWh.
  • Installed Capacity: The total potential capacity for these top-tier areas is 201 GWp.
  • Regional Potential: Despite Sweden's high latitude, southern counties emerged as hotspots. Skåne showed the highest potential for "very good" land, while Västra Götaland had the highest overall potential for "good" land.

Conclusion

This geospatial roadmap demonstrates that APV can play a transformative role in Sweden's energy transition without compromising food security. However, some limitations remain. Due to a lack of high-resolution data on the distribution grid, "proximity to grid"—a major economic factor—was not included in the restrictions. Future work should also incorporate deep-dive economic modeling, such as Levelized Cost of Electricity (LCOE) for vertical bifacial systems, and sensitivity analyses to generalize the model for other Nordic regions.

Summary updated

Abstract

Agrivoltaic systems represent an intelligent solution combining electricity production from solar photovoltaic technology with agricultural production and avoiding land use conflicts. Geographic Information System technologies can support the implementation and spread of agrivoltaic systems by identifying the most suitable areas using useful spatially explicit information concerning techno-agro-socio-economic criteria. In this study, we have developed a procedure to identify and classify suitable areas for agrivoltaic systems in Sweden. An Ordinal Priority Approach based multi-criteria decision making algorithm is established to calculate the weights of the selected evaluation criteria through expert interviews. The land use data refers to the Corine Land Cover 2018 product.The results show that 8.55% of the Swedish territory, approximately 38,485 km2, is suitable for installing APV systems. Among this area, 0.17% is classified as "excellent", about 15% as "very good", about 72% as "good", about 13.1% as "moderate", and less than 0.1% as "poor". Through the deployment of vertically mounted agrivoltaic systems with bifacial photovoltaic modules, the total "excellent" areas can potentially supply 2.44 TWh against the electricity consumption in 2021 of about 143 TWh. On the other hand, the land classified as "excellent" and "very good" could potentially provide about 207 TWh, which is a much higher production capacity than the 2021 electricity consumption. The total potential installed capacity for "excellent" areas is 2.3 GWp, while for areas classified "excellent" and "very good" is 201 GWp.
BibTeX
@misc{elkadeem2023agrivoltaic,
  author = {Mohamed Elkadeem and Sebastian Zainali and Silvia Ma Lu and Ali Younes and Mohamed Ali Yousif Abido and Stefano Amaducci and Michele Croci and Jie Zhang and Tomas Landelius and Bengt Stridh and Pietro Elia Campana},
  title = {{Agrivoltaic systems potentials in Sweden: a geospatial-assisted multi-criteria analysis}},
  year = {2023},
  doi = {10.31223/x5g663},
}
Figure