Climate Forecasting as Support for Planting Calendar in North Aceh Using the Backpropagation Method

Sari

North Aceh Regency is the largest rice producer in the Aceh Province, yet it faces vulnerability to climate change. Data indicates that floods have damaged agricultural land, including rice fields, significantly affecting rice production. In response to this challenge, this research aims to develop a climate prediction model using artificial neural networks with backpropagation algorithms to support the creation of a planting calendar in North Aceh Regency. The research method involves collecting historical data related to climate conditions, such as rainfall and air temperature, which are then used as inputs to train and test the ANN model with backpropagation. The ANN model is configured to understand complex climate patterns and identify relationships between climate variables and the optimal planting season for various local crops. The results of this study show that for rainfall data, the best architecture is 12-75-1, with a learning rate of 0.2 and a momentum of 0.15. Meanwhile, for air temperature data, the best architecture is 12-25-1, with a learning rate of 0.1 and a momentum of 0.20. Based on the conversion of climate prediction data into a planting calendar, planting schedules for rice and corn are determined from September to December, soybeans from June to December, cassava and eggplants from April to October, sweet potatoes and tomatoes from February to October, chili peppers and shallots from June to October, long beans from May to September, and green beans can be planted throughout the year, while peanuts are not suitable for year-round planting.