Method for identifying key driving factors of evolution of ecosystem structure and function in estuary and adjacent sea area

By employing two-stage collinearity decomposition, time-varying causal structure modeling, and reservoir causal enhancement model, the problem of identifying key driving factors under strong collinearity and high-dimensional sparse observation conditions in estuarine ecosystems was solved, achieving robust quantitative attribution and cross-regional adaptation.

CN122241015APending Publication Date: 2026-06-19BEIHAI FORECASTING CENT OF STATE OCEANIC ADMINISTRATION ((QINGDAO MARINE FORECASTING STATION OF STATE OCEANIC ADMINISTRATION) (QINGDAO MARINE ENVIRONMENT MONITORING CENT OF STATE OCEANIC ADMINISTRATION))
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
BEIHAI FORECASTING CENT OF STATE OCEANIC ADMINISTRATION ((QINGDAO MARINE FORECASTING STATION OF STATE OCEANIC ADMINISTRATION) (QINGDAO MARINE ENVIRONMENT MONITORING CENT OF STATE OCEANIC ADMINISTRATION))
Filing Date
2026-05-22
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing technologies cannot robustly explain and quantitatively attribute key driving factors of estuarine ecosystems under conditions of strong collinearity, nonlinear driving response relationships, and high-dimensional sparse observations.

Method used

Employing a two-stage collinearity decomposition framework, time-varying causal structure modeling, and a reservoir causal enhancement model, this study identifies key driving factors through multi-source long-term series data preprocessing, variance inflation factor analysis, wavelet coherence analysis, hidden Markov models, and information geometric geodesic distance ranking, combined with ecological mechanism knowledge graphs and elastic network stability selection.

Benefits of technology

It effectively eliminates linear collinearity, captures the dynamic causal structure of seasonal dependence, suppresses overfitting under high-dimensional sparse conditions, achieves robust quantitative attribution under sparse data, and provides cross-regional generalization ability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122241015A_ABST
    Figure CN122241015A_ABST
Patent Text Reader

Abstract

This invention provides a method for identifying key driving factors in the structural and functional evolution of estuarine and adjacent marine ecosystems, belonging to the field of ecosystem technology. This invention preprocesses multi-source long-term series data and constructs an ecosystem evolution index system. It utilizes a two-stage collinearity decomposition framework to achieve interpretable spatial contribution decomposition of original physical variables under strong collinearity conditions. Wavelet coherence analysis and adaptive dynamic time warping are used to extract time-varying time delays, and a time-varying causal structure is formed by modeling multi-state time delay transition probabilities using a hidden Markov model. This structure is then input into a causal enhancement artificial intelligence model in a reservoir for comprehensive inference, outputting contribution scores, stability scores, and confidence scores for each candidate driving factor. This method solves the technical problem of being unable to robustly and interpretably quantitatively attribute key driving factors in estuarine ecosystems under conditions of strong collinearity, nonlinear driving response relationships, and high-dimensional sparse observations.
Need to check novelty before this filing date? Find Prior Art