.png){kind=icon}
.png){kind=icon}
Return
Exploring Spatial Stratified Heterogeneity for Local Geometry Structures
.png){kind=icon}
PRE
AI
DOI：10.1080/24694452.2025.2564227.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Geometry structure modeling plays an increasingly important role in the era of artificial-intelligence-guided geoscience discovery. It poses challenges in detecting and quantifying spatial stratified heterogeneity, which is critical in capturing causal association and developing stratum-specific models in geoscience, as up-to-date methods mainly focus on value disparities and leave the corresponding spatial geometry structure aside. This article introduces the search tree structure in multiple-point geostatistics to model the local geometry structure in the entire study area and each stratum, quantifies the stratified heterogeneity degree of local geometry structure using the average Jensen–Shannon divergence between each stratum and the study area, and finally tests whether the existence of local geometry structure stratified heterogeneity is statistically significant based on simulations. Experiments on five illustrative and three publicly accessible real-life data sets show that the proposed measure effectively captures the stratified heterogeneity of local geometry structure missed by commonly used methods, such asq-statistic and normalized relative entropy. Additionally, the proposed method can effectively explore the trends of local geometry structure stratified heterogeneity changing with distance and spatial resolution. The code for this algorithm is available athttps://github.com/crafly/SH-package.
Keywords:
local geometry structure
multiple-point geostatistics
multiscale analysis
spatial stratified heterogeneity
Journal
IF:
2.9
Papers:
1.4K
Citations:
4.5K