Dynamics of Lemon Crop Production in Tambo Grande, Piura: Implementation of Convolutional Neural Networks and Analysis of Risk Management Associated with Thermal Climatic Phenomena †

Lemon production in Piura, Peru faces climatic challenges such as droughts and El Niño. This study uses satellite indices (NDVI and NDWI) and convolutional neural networks (CNNs, specifically AlexNet) to analyze crop dynamics and health during extreme events. The results highlight the impact of extreme precipitation and temperatures, with reflectance levels as early indicators of deterioration. In 1998, El Niño reduced yield to 9.2 tons/ha, compared to 14 tons/ha in 2014. These variables provide key information for pruning, irrigation, and fertilization decisions, improving crop productivity.