Scholarlink

<p>The monitoring and management of environmental parameters depends substantially on Land Use Land Cover (LULC) prediction. Multiple frameworks have been created to forecast future LULC, with CA integrated models being the most well-liked models by researchers. This study uses six hybrid Cellular Automata (CA)-Markov Chain (MC) models to anticipate future LULC of the Noyyal River Basin (36,000 sq.km), TamilNadu, India. The purpose of this study is to compare and contrast these six CA integrated model’s performances in terms of LULC prediction as well as to predict the study area's LULC in 2050. LULC data from 2001 and 2011 were used as inputs, along with LULC drivers. The past LULC images were created using Landsat imagery using Support Vector Machine—supervised classification approach. The models are validated using 2021 imagery and kappa coefficient is evaluated. Based on the findings, the CA-MC-Weighted Normalized Likelihood (CA-MC-WNL) achieved the highest kappa for LULC prediction in the study region. This was followed by CA-MC-Support Vector Machine (CA-MC-SVM), which showed acceptable area deviation, and then by CA-MC-Weight of Evidence (CA-MC-WoE), CA-MC-Multi-Layer Perceptron (CA-MC-MLP), CA-MC-Logistic Regression (CA-MC-LR), and CA-MC- Random Forest (CA-MC-RF). The CA-MC-WNL forecast an equivalent trend in 2050, indicating that the Noyyal basin will experience rapid urbanization in subsequent decades, with a reduction in the area used for agriculture, water bodies, and forests and an increase in the area used for barren land and urban areas.</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

Assessing the Effectiveness of CA Integrated Models for Land Use Land Cover Prediction: A Comparative Analysis

  • 期刊论文

  • 发表日期：

    2025年9月12日

查看全文
  • Ajisha Stephen,
  • Manjula Renganathan

摘要

The monitoring and management of environmental parameters depends substantially on Land Use Land Cover (LULC) prediction. Multiple frameworks have been created to forecast future LULC, with CA integrated models being the most well-liked models by researchers. This study uses six hybrid Cellular Automata (CA)-Markov Chain (MC) models to anticipate future LULC of the Noyyal River Basin (36,000 sq.km), TamilNadu, India. The purpose of this study is to compare and contrast these six CA integrated model’s performances in terms of LULC prediction as well as to predict the study area's LULC in 2050. LULC data from 2001 and 2011 were used as inputs, along with LULC drivers. The past LULC images were created using Landsat imagery using Support Vector Machine—supervised classification approach. The models are validated using 2021 imagery and kappa coefficient is evaluated. Based on the findings, the CA-MC-Weighted Normalized Likelihood (CA-MC-WNL) achieved the highest kappa for LULC prediction in the study region. This was followed by CA-MC-Support Vector Machine (CA-MC-SVM), which showed acceptable area deviation, and then by CA-MC-Weight of Evidence (CA-MC-WoE), CA-MC-Multi-Layer Perceptron (CA-MC-MLP), CA-MC-Logistic Regression (CA-MC-LR), and CA-MC- Random Forest (CA-MC-RF). The CA-MC-WNL forecast an equivalent trend in 2050, indicating that the Noyyal basin will experience rapid urbanization in subsequent decades, with a reduction in the area used for agriculture, water bodies, and forests and an increase in the area used for barren land and urban areas.