Taming the heterogeneous dynamics in ocean chlorophyll-a concentration prediction with a deep learning model
This repository implements STD-Hunter method, a novel machine learning approach designed to facilitate long-term, large-scale prediction of chlorophyll-a (Chl_a) concentration. The core innovation is that STD-Hunter forms a predictive model for each task by integrating commonly shared basis models with the specific spatiotemporal characteristics of Chl_a. STD-Hunter strikes an optimal balance between effectively leveraging data and preserving idiosyncratic dynamics with an interpretable mechanism, offering a trustworthy tool for heterogeneous data exploration.
The chlorophyll-a (Chl_a) concentration and sea surface temperature (SST) data is publicly available in MODIS Aqua projects. We also provide the used data in MODIS data.
If you find the STD-Hunter package or any of the source code in this repository useful for your work, please cite:
Taming the heterogeneous dynamics in ocean chlorophyll-a concentration prediction with a deep learning model.
Fa Zhang, Hiusuet Kung, Fan Zhang, Zhiwei Wang, Can Yang#, and Jianping Gan#. 2025.
The python repository STD-Hunter is developed and maintained by Fa Zhang.
Please feel free to contact Fa Zhang (fa.zhang@connect.ust.hk), Prof. Can Yang (macyang@ust.hk), or Prof. Jianping Gan (magan@ust.hk) if any inquiries.