Forecasting climate change effects on Saline Lakes through advanced remote sensing and deep learning.

Area of Science:

• Environmental Science
• Remote Sensing
• Climate Modeling

Background:

• Saline lakes are crucial ecosystems in arid regions.
• Understanding their long-term changes is vital for ecological conservation.
• Climate change poses a significant threat to these unique environments.

Purpose of the Study:

• To analyze long-term changes in saline lakes using spectral reflectance and physical-chemical properties.
• To integrate climate change modeling for predicting future shifts and ecological impacts.
• To enhance satellite image analysis for precise differentiation of lake components.

Main Methods:

• Super-Resolution Generative Adversarial Network (SRGAN) for image enhancement.
• Multiresolution Segmentation (MRS) for differentiating lake components.
• Cellular Automata (CA)-Markov modeling and Long Short-Term Memory (LSTM) for climate and salinity prediction.

Main Results:

• SRGAN and image classification improved identification of physical characteristics and prediction of chemical properties.
• CA-Markov modeling predicts a ~15% salinity increase by 2050.
• LSTM models show high accuracy (R² > 0.9) for predicting rising temperatures and evaporation.

Conclusions:

• Rising temperatures and evaporation will increase salinity, drying, and albedo in Chaka, Tuz, and Razzaza Lakes.
• RCP8.5 scenarios project significant increases by 2100, exacerbating evaporation and salinity.
• These changes threaten surrounding ecosystems, potentially accelerating desertification.