Visual Image Design Based on Multi-sensor Machine Learning for Monitoring Plateau Lake Dynamics and Pasture Change

Area of Science:

• Environmental Science
• Remote Sensing
• Ecosystem Services Analysis

Background:

• Land use/land cover (LULC) changes significantly impact ecosystem services (ES) and resource sustainability.
• Monitoring these changes is crucial for effective environmental management, especially in semi-arid regions.

Purpose of the Study:

• To analyze multi-decadal LULC dynamics in the Sahiwal area from 1994 to 2024.
• To quantify the resulting changes in ecosystem service values (ESV).
• To assess the sustainability of water and vegetation resources.

Main Methods:

• Utilized Landsat time-series data for LULC classification.
• Incorporated Random Forests (RF) algorithm within Google Earth Engine (GEE) for systematic analysis.
• Calculated ESV based on LULC changes and assessed resource sustainability.

Main Results:

• Observed a decline in forest (0.229%) and vegetation (3.9%) areas, alongside a substantial increase in built-up areas (from 7.83% to 16.53%) between 1994 and 2024.
• Vegetation ESV initially increased then declined, while water body ESV showed variability.
• Achieved high overall accuracy (OA) and kappa (OK) values for LULC classifications across the study period.

Conclusions:

• Significant LULC changes necessitate urgent improvements in land and water management planning.
• Ecosystem restoration and policy interventions are vital for promoting nature-based solutions.
• Ensuring sustainable water resources in semi-arid regions requires integrated management strategies.