Unraveling the crop yield response under ... through the deployment of a drought index

Sultan Tekie, Sebastian Zainali, Tekai Eddine Khalil Zidane, Silvia Ma Lu , Mohammed Guezgouz, Jie Zhang, Stefano Amaducci, Pietro Elia Campana

Extensive research has explored the impact of shading on vegetation growth and crop yield (CY) under agrivoltaic (APV) systems. These studies have revealed a notable connection between shading and CYs, with certain crop varieties showing benefits from shadings e.g., Berries and Leafy Vegetables, Forage remaining largely unaffected, and some crops e.g., Cereals, Grain Legumes, Fruits, and Root crops experiencing reduced yields when subjected to shaded conditions. Previous studies often overlooked environmental factors such as temperature, evapotranspiration, and precipitation when assessing shading effects on CY, making it difficult to fully understand their impact on crop performance. This study seeks to address this research gap by integrating a drought index, known as the Standardized Precipitation Evapotranspiration Index (SPEI), into existing improved meta-analysis on shade and CY across various crops. SPEI, encompassing information on potential evapotranspiration, and precipitation is highly relevant to soil moisture, and accessible worldwide with a reasonable temporal resolution. Multiple linear regression (MLR) techniques are applied to analyse different crop categories. The MLR models’ results with and without incorporating SPEI are compared to assess the influence of shading on determining CY amidst varying environmental conditions. The inclusion of SPEI in MLR models resulted in improved performance metrics across all crop categories with good sample sizes, with the least and most significant improvements observed for Fruit (1% increase in model precision) and Berries (40% increase in model precision) respectively. 

The analysis is strengthened by uncertainty quantification, demonstrating how the predictability of CY improves with the inclusion of SPEI, supported by a 95% confidence level. The MLR model in all crop categories showed improved certainty when SPEI was factored in, compared to using shading alone as a determinant for CY in the uncertainty analysis. Similarly, incorporating SPEI into the uncertainty analysis of the MLR models enhanced the certainty levels across all crop categories, with the smallest improvement observed in Forage at 13% and the largest in Root Crops at a 50% increase.