Related Experiment Video
Updated: Jun 19, 2026

Early Detection of Cyanobacterial Blooms and Associated Cyanotoxins using Fast Detection Strategy
Published on: February 25, 2021
Transforming global water cycle observations via synergistic AI and remote sensing
Zhaoyuan Yao1,2, Yaokui Cui1,2, Zhenong Jin3,4
1Institute of Remote Sensing and Geographic Information System, School of Earth and Space Sciences, Peking University, Beijing, China.
A new AI framework, BERTH, enhances global water cycle monitoring using remote sensing. It accurately tracks evapotranspiration, precipitation, soil moisture, and runoff at high resolution, outperforming traditional models.
Area of Science:
- Earth and Environmental Sciences
- Artificial Intelligence
- Remote Sensing
Background:
- Climate change accelerates the terrestrial water cycle, increasing water-related disasters.
- Existing remote sensing models struggle with high spatiotemporal resolution and complex hydrological dynamics due to empirical parameters and multistage workflows.
- Parameter complexity for fine-scale hydrological representation under diverse terrestrial conditions is prohibitive.
Purpose of the Study:
- To introduce a novel, end-to-end artificial intelligence (AI) framework, Bidirectional Encoder Representations from Transformers for Hydrology (BERTH).
- To enable global water cycle monitoring at 30-meter and daily scales by directly transforming remote sensing radiance.
- To overcome the limitations of current models in handling complex, accelerated water cycles with high accuracy.
Main Methods:
- Developed BERTH, an AI framework utilizing transformer architecture for hydrological applications.
- Trained BERTH on global remote sensing datasets.
- Directly mapped remote sensing radiance to key hydrological variables: evapotranspiration, precipitation, soil moisture, and runoff.
Main Results:
- BERTH achieves global water cycle monitoring at 30-meter resolution and daily scale.
- The AI framework accurately captures spatiotemporal dynamics across diverse terrestrial conditions.
- BERTH demonstrates superior accuracy compared to existing geophysical models.
Conclusions:
- BERTH represents a transformative paradigm for quantitative Earth science and AI integration in remote sensing.
- This framework offers a robust platform for next-generation investigations into global water resources and Earth system dynamics.
- The AI-driven approach provides a more effective solution for monitoring the water cycle amidst climate change challenges.
Related Concept Videos
The Water Cycle
Global Climate Change
Precipitation Gravimetry
In determining nickel by gravimetric analysis, a precipitant of ethanolic dimethylglyoxime is added to a hot nickel salt solution. This is quickly followed by the dropwise addition of dilute ammonia solution until precipitation occurs. A...
Precipitation Processes
Responses to Drought and Flooding