SAR image integration for multi-temporal analysis of Lake Manchar Wetland dynamics using machine learning

Area of Science:

• Environmental Science
• Remote Sensing
• Ecology

Background:

• Manchar Lake, Pakistan's largest freshwater wetland, faces ecological threats from climate change and human activities.
• Urgent, data-driven conservation strategies are needed for sustainable management of this vital ecosystem.

Purpose of the Study:

• To analyze dynamic changes in Manchar Lake's ecosystem from 2015 to 2023 using multi-sensor remote sensing.
• To provide a comprehensive understanding of wetland dynamics for informed management decisions.

Main Methods:

• Integrated Sentinel-1 Synthetic Aperture Radar (SAR) and Sentinel-2 multispectral imagery.
• Assessed wetland extent, vegetation health (EVI), and land-use patterns using spectral indices and topographic data (TWI).
• Achieved over 92% classification accuracy across all study years.

Main Results:

• Water body extent fluctuated: 318.5 km² (2015) to 397.0 km² (2019) then 369.9 km² (2023).
• Vegetation cover significantly increased from 405.5 km² (2019) to 1081.6 km² (2023), with EVI recovery.
• Agricultural areas expanded from 118.4 km² (2015) to 498.0 km² (2023).

Conclusions:

• Remote sensing effectively quantified Manchar Lake's complex ecosystem dynamics.
• Findings highlight the interplay of natural processes and human pressures, necessitating adaptive management.
• Urgent conservation strategies are crucial for the resilience of Manchar Lake.