Spatial-based Rice Production System Engineering and System Dynamics Modeling for Mitigating Paddy Field Conversion in Tabanan Regency

Abstract

Rice field conversion threatens the sustainability of rice production and self-sufficiency in Tabanan Regency, one of Bali Province’s major rice-producing areas. This study aims to determine the rate of rice field conversion, evaluate the balance between rice production and demand, and assess its long-term impacts through dynamic system scenarios. Spatial analysis was conducted using Sentinel-2 imagery processed in Google Earth Engine with Random Forest classification for 2020–2024, followed by dynamic system modeling to simulate rice production and demand from 2020 to 2050. Classification accuracy was very high, with Kappa coefficients ranging from 0.90 to 0.94. Rice field area decreased by an average of 1.71% per year. Under the business-as-usual scenario, rice production is projected to decline gradually and fall below demand around 2038–2039. Model validation produced nRMSE values below 10%, indicating very good model performance. A 50% reduction in land conversion slowed the decline in production, while combining land protection with a 10% increase in rice productivity produced a more stable production trajectory and maintained a surplus through 2050. Therefore, sustainable rice self-sufficiency requires integrated policies that control rice field conversion, protect agricultural land, and improve productivity.