Simulation of aeroponic systems with the AquaCrop model

The increase in population, scarcity of fertile lands and urbanization constitute a widespread challenge that is impacting food security worldwide. In response, Controlled Environment Agriculture (CEA) is becoming increasingly popular due to its focus on optimizing agricultural practices, resource management and space use (particularly in Vertical Farming), aiming to identify the most efficient methods for crop cultivation. Within the framework of CEA, soilless agriculture stands out, including techniques such as hydroponics, aeroponics, aquaponics and bioponics. This thesis aims to evaluate whether the AquaCrop model, designed for traditional soil-based agriculture, accurately simulates the growth of lettuce (Lactuca sativa) in an aeroponic system implemented at Politecnico di Torino by comparing the simulated dry biomass with the one measured experimentally. Experimental activity was carried out considering four growth cycles conducted in the laboratory, each under distinct climatic conditions. Normalized water productivity (WP*) in AquaCrop was calibrated to match the experimentally observed final dry biomass, allowing a comparison of simulated and measured dry biomass on intermediate days. Results reveal that given equal final dry biomass, the dry biomass in the aeroponic system exponentially increases in the final days of the growth cycle. In contrast, AquaCrop models a more steady and linear growth trend, resulting in a discrepancy in measurements on intermediate days of the cycle. This demonstrates the incapability of the model to simulate growth in an aeroponic setting. In contrast, a machine learning algorithm applied to the same aeroponic system, resulted in higher accuracy in simulating crop growth.