Scholarlink

<p>In rice-dependent regions of Asia—including Indonesia, where rice is a primary staple—urban expansion increasingly drives the conversion of irrigated rice fields, posing risks to food security and environmental sustainability. This study analyzed the vulnerability of rice fields to conversion in the Mamminasata Urban Area of Indonesia and identified the factors influencing this vulnerability using a machine learning approach based on the Random Forest Algorithm. Sentinel-2 satellite imagery from 2016 to 2024 was employed for land use classification and change detection, while various demographic and agronomic spatial variables served as predictors. Variable selection was performed using the Variance Inflation Factor (VIF) to minimize multicollinearity. Hyperparameter optimization was executed through RandomizedSearchCV with 5-fold cross-validation. The model was trained on balanced data using the Synthetic Minority Over-sampling Technique (SMOTE) and evaluated based on accuracy and classification metrics. The results indicated that most rice fields exhibit low vulnerability to conversion. However, significant areas with moderate to high vulnerability existed, particularly near rapidly developing urban zones. The primary factors influencing land conversion were assessed land value and distance to secondary roads, highlighting the dominant role of economic and infrastructure considerations over agronomic factors. The resulting vulnerability probability map serves as a crucial spatial decision-making tool for urban planners and policymakers. We recommend reinforcing land protection policies in designated agricultural zones to mitigate uncontrolled land conversion and promote sustainable urban development.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Probabilistic vulnerability analysis of converted rice fields in the mamminasata urban area

  • 期刊论文

  • 发表日期：

    2026年5月20日

查看全文
  • Ahmad Firman Ashari,
  • Risma Neswati,
  • Samsu Arif

摘要

In rice-dependent regions of Asia—including Indonesia, where rice is a primary staple—urban expansion increasingly drives the conversion of irrigated rice fields, posing risks to food security and environmental sustainability. This study analyzed the vulnerability of rice fields to conversion in the Mamminasata Urban Area of Indonesia and identified the factors influencing this vulnerability using a machine learning approach based on the Random Forest Algorithm. Sentinel-2 satellite imagery from 2016 to 2024 was employed for land use classification and change detection, while various demographic and agronomic spatial variables served as predictors. Variable selection was performed using the Variance Inflation Factor (VIF) to minimize multicollinearity. Hyperparameter optimization was executed through RandomizedSearchCV with 5-fold cross-validation. The model was trained on balanced data using the Synthetic Minority Over-sampling Technique (SMOTE) and evaluated based on accuracy and classification metrics. The results indicated that most rice fields exhibit low vulnerability to conversion. However, significant areas with moderate to high vulnerability existed, particularly near rapidly developing urban zones. The primary factors influencing land conversion were assessed land value and distance to secondary roads, highlighting the dominant role of economic and infrastructure considerations over agronomic factors. The resulting vulnerability probability map serves as a crucial spatial decision-making tool for urban planners and policymakers. We recommend reinforcing land protection policies in designated agricultural zones to mitigate uncontrolled land conversion and promote sustainable urban development.