ENVIRONMENTAL CONSEQUENCES OF AGRICULTURAL EXPANSION AND SETTLEMENT GROWTH ON FOREST STRUCTURE AND BIODIVERSITY PATTERNS

Abstract

Agricultural expansion and settlement growth significantly contribute to forest degradation, biodiversity loss, and ecosystem imbalance by altering forest structure and habitat connectivity. Traditional field survey-based biodiversity assessment methods provide limited spatial coverage and are inadequate for continuous large-scale environmental monitoring. Therefore, this research proposes an intelligent Deep Learning (DL)-based environmental monitoring model for analyzing the impact of agricultural expansion and settlement growth on forest structure and biodiversity patterns. The ChinaSAT Land Cover & Environmental Dataset (4000 rows and 15 columns) from Kaggle is utilized and preprocessed using Min-Max Normalization, Median Filtering, and Fmask. Feature extraction is performed using Principal Component Analysis (PCA). The Dilated Convolutional-tuned Efficient Recurrent Neural Network (DilatedCon-ERNN) is proposed to accurately analyze spatial-temporal environmental patterns for land-cover classification, forest degradation monitoring, and biodiversity prediction, while DilatedCon-ERNN for classification of land-use and land-cover and ERNN for predicting temporal biodiversity and forest-cover changes. Experimental results using Python 3.10 achieved an overall accuracy (OA) of 98.26%, 98.12% precision, and a 0.981 kappa coefficient and enable intelligent large-scale forest monitoring, biodiversity assessment, and ecological risk identification, supporting sustainable land management, conservation planning, and long-term environmental sustainability