ORIGINAL RESEARCH article
Front. Environ. Sci.
Sec. Big Data, AI, and the Environment
This article is part of the Research TopicAdvanced Applications of Artificial Intelligence and Big Data Analytics for Integrated Water and Agricultural Resource Management: Emerging Paradigms and MethodologiesView all 6 articles
Multiscale Remote Sensing Methods for Monitoring Wetland Ecosystem Dynamics and Crop Development
Provisionally accepted
- Zhongyuan University of Science and Technology, Zhengzhou, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Understanding the interplay between wetland ecosystems and agricultural crop development is vital for sustainable water and food resource management amid climate variability. Emerging technologies in artificial intelligence (AI) and big data analytics now offer powerful tools to integrate multiscale remote sensing with ecosystem modeling. This study introduces a unified framework that combines remote sensing and AI-driven inference to monitor the spatiotemporal dynamics of wetland hydrology and crop phenology, aligning with modern approaches to integrated resource management. Traditional remote sensing methods often struggle to capture the temporal variability and complex dependencies in such ecosystems due to their reliance on static thresholds and single-modality data. To overcome these limitations, we propose an AI-enhanced methodology comprising two modules: the Graph-Augmented Attention Recommendation Network (GAARN) and the Multi-Perspective Preference Distillation (MPPD) strategy. GAARN fuses structural insights from environmental graphs with temporal patterns via attention-based encoders, enabling detailed mapping of land-water-vegetation transitions. MPPD incorporates semantic knowledge from ecological ontologies, meteorological data, and crop taxonomies to guide learning through consistency regularization and contrastive embedding alignment. Our dual-module framework offers robust interpretation of sparse observations, adaptive forecasting under climate variability, and scalable modeling of wetland-crop interactions. Validation across diverse agroecological zones reveals superior performance over conventional baselines in predicting vegetative indices, water extent changes, and crop growth stages. These results highlight the potential of our framework for advancing precision agriculture, wetland monitoring, and climate-resilient policy-making.
Keywords: Multiscale remote sensing, Wetland dynamics, Crop development, Graph attention networks, Knowledge distillation
Received: 09 May 2025; Accepted: 07 Nov 2025.
Copyright: © 2025 Lin, Hei, Wang and Zhang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence:
Xinhao Lin, szwechavnora@hotmail.com
Ang Zhang, angzhang@163.com
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.