Unraveling the crop yield response under shading conditions through the deployment of a drought index: A meta-analysis

Extensive meta-analyses have examined the effects of shading on vegetation growth and Crop Yields (CY) in Agrivoltaic (AV) systems. These studies have demonstrated a strong relationship between shading and crop performance. Certain varieties, such as Berries and Leafy Vegetables, thrive under shaded conditions, while Forage Crops remain largely unaffected. Conversely, other crops, including C3 Cereals, Grain Legumes, Fruits, and Root Crops, experience reduced yields when exposed to shade. Previous meta-analyses often neglected environmental factors such as temperature, evapotranspiration, and precipitation when evaluating the effects of shading on CY, making it difficult to fully understand how shading influences crop performance. This study seeks to address this research gap by integrating a drought index, the Standardized Precipitation Evapotranspiration Index (SPEI), for an improved meta-analysis on shading and CY across various crops. SPEI, encompassing information on potential evapotranspiration and precipitation is an effective indicator of moisture availability and accessible worldwide at a reasonable temporal and spatial resolution. Multiple Linear Regression (MLR) techniques are used to analyze various crop categories. From a policy perspective, the MLR models developed in this study can help policymakers make more accurate assessments of the impact of AV systems deployment on CY at both national and regional levels.

The results of the MLR models, both with and without the inclusion of the SPEI, were compared to evaluate the impact of shading on determining CY under different environmental conditions. Incorporating SPEI into the MLR models improved performance metrics across all crop categories with adequate sample sizes. The least improvement was observed for Fruits, with a marginal 0.01 gain in coeffiecient of determination (R²), while the most substantial improvement was seen in Berries, with a 0.32 increase. The analysis was reinforced by uncertainty quantification, which demonstrated that the predictability of CY improves when SPEI was included, as supported by a 95 % confidence level. In all crop categories, the MLR model exhibited increased certainty when SPEI was considered, compared to using shading alone as a determinant for CY in the uncertainty analysis. A minor improvement of 13 % was observed in Forage Crops, while a significant increase of 47 % was noted in Root Crops.