Pengembangan Model Optimasi Irigasi untuk Mitigasi Global Warming Potential

Abstract

Peningkatan emisi gas rumah kaca (GRK) dari sektor pertanian, khususnya budidaya padi, menjadi tantangan utama dalam mitigasi perubahan iklim. Penelitian bertujuan mengembangkan model optimasi pengelolaan irigasi berbasis algoritma genetika untuk meminimalkan Global Warming Potential (GWP) pada sistem budidaya padi dengan pendekatan System of Rice Intensification (SRI). Penelitian dilakukan di greenhouse dengan empat skema irigasi: tergenang, basah, kering, dan mid-season drainage. GWP diukur berdasarkan emisi CH4 dan N2O, dengan hasil tertinggi pada rezim tergenang sebesar 2.285,53 kg CO2-eq/ha/musim dan terendah pada rezim kering sebesar -1.541,38 kg CO2-eq/ha/musim. Model prediksi GWP dikembangkan menggunakan Artificial Neural Network dengan algoritma Backpropagation, sedangkan optimasi pengelolaan air dengan Algoritma Genetika. Hasil model mampu mengidentifikasi skenario tinggi muka air yang optimal pada tiap fase pertumbuhan padi, yaitu awal 1,962 cm, vegetatif -1,363 cm, tengah 0,824 cm, dan akhir -8,788 cm. Kemudian dihasilkan nilai GWP 426,64 kg/ha/musim dan produktivitas 507,28 g/m2. Temuan ini menunjukkan bahwa pengelolaan air berbasis optimasi dapat menurunkan emisi GRK secara signifikan.

Increased greenhouse gas (GHG) emissions from the agricultural sector, particularly rice cultivation, pose a major challenge in climate change mitigation. This study aims to develop a genetic algorithm-based irrigation management optimization model to minimize Global Warming Potential (GWP) in rice cultivation systems using the System of Rice Intensification (SRI) approach. The study was conducted in a greenhouse with four irrigation schemes: flooded, wet, dry, and mid-season drainage. GWP was measured based on CH4 and N2O emissions, with the highest results in the flooded regime at 2,285.53 kg CO2- eq/ha/season and the lowest in the dry regime at -1,541.38 kg CO2-eq/ha/season. A GWP prediction model was developed using an Artificial Neural Network with a Backpropagation algorithm, while water management optimization was performed using a Genetic Algorithm. The model results were able to identify the optimal water level scenarios for each rice growth phase: initial 1.962 cm, vegetative -1.363 cm, mid-season 0.824 cm, and final -8.788 cm. The resulting GWP value was 426.64 kg/ha/season, and productivity was 507.28 g/m². These findings indicate that water management based on optimization can significantly reduce GHG emissions.