Organic carbon density prediction method and system based on soil profile database

By using a classification matrix and ensemble learning model based on a soil profile database, the problem of inaccurate organic carbon density prediction caused by missing bulk density data and soil type heterogeneity in the soil profile database was solved, and high-precision estimation of organic carbon density and carbon storage was achieved.

CN121687262BActive Publication Date: 2026-06-23INST OF AGRI RESOURCES & ENVIRONMENT GUANGDONG ACADEMY OF AGRI SCI
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
INST OF AGRI RESOURCES & ENVIRONMENT GUANGDONG ACADEMY OF AGRI SCI
Filing Date
2025-12-16
Publication Date
2026-06-23

AI Technical Summary

Technical Problem

Existing methods for predicting organic carbon density based on soil profile databases suffer from problems such as high missing rates of bulk density data, low prediction accuracy, and failure to consider soil type heterogeneity and depth effects, resulting in inaccurate estimation of organic carbon density.

Method used

By collecting soil profile data, the samples were divided into different groups according to pH value and texture parameters, a classification matrix was constructed, bulk density prediction models were trained, missing values ​​were imputed, and an organic carbon density transfer function was established. The organic carbon density of each depth layer was calculated by integrating support vector machine, Cubist model, random forest model and gradient boosting machine model.

Benefits of technology

It improves the accuracy of bulk density prediction and organic carbon density prediction at different depths, ensuring the integrity and accuracy of carbon storage estimation and solving the problem of low prediction accuracy for deep soil layers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121687262B_ABST
    Figure CN121687262B_ABST
Patent Text Reader

Abstract

The application relates to the technical field of data processing, and discloses a soil profile database-based organic carbon density prediction method and system. The method comprises the following steps: collecting soil profile data, obtaining soil attribute parameters and environmental parameters of each depth layer, and obtaining a profile data set; dividing samples into different groups according to pH values and texture parameters to obtain a classification matrix; training a bulk density prediction model in each group, interpolating the missing values of the bulk density, and obtaining complete bulk density data; establishing a transfer function, substituting the organic carbon content, the bulk density value and the layer thickness, and calculating the organic carbon density of each depth layer; and weighting and accumulating the organic carbon density of each depth layer according to the layer thickness to obtain the profile organic carbon storage. The application solves the problems that, in the prior art, a global unified model leads to low bulk density prediction accuracy, and the heterogeneity of soil types and the depth effect are not considered, resulting in inaccurate organic carbon density estimation.
Need to check novelty before this filing date? Find Prior Art