Comparison of geostatistical and response surface methodology for estimating soil saturated hydraulic conductivity

Soil saturated hydraulic conductivity (Ks) is a critical parameter for modeling water and solute transport in soils. Conventional laboratory measurements of Ks are labor-intensive, costly, and susceptible to measurement errors, underscoring the need for more reliable estimation techniques. This stud...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Scientific reports Jg. 15; H. 1; S. 34103 - 14
1. Verfasser: Cheng, Lin
Format: Journal Article
Sprache:Englisch
Veröffentlicht: London Nature Publishing Group UK 30.09.2025
Nature Publishing Group
Nature Portfolio
Schlagworte:
631/158
639/166
704/158
704/172
704/2151
704/242
Accuracy
Agricultural management
Co-kriging
Environmental science
Humanities and Social Sciences
Hydraulics
Hydrology
Land use planning
Machine learning
multidisciplinary
Neural networks
Ordinary kriging
Physicochemical properties
Response surface methodology
Salinity
Saturated soils
Science
Science (multidisciplinary)
Soil erosion
Soil properties
Soil saturated hydraulic conductivity
Solute transport
Support vector machines
Water
Water resources
China
Henan China
Co-kriging
Soil saturated hydraulic conductivity
Response surface methodology
Ordinary kriging
ISSN:2045-2322, 2045-2322
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Soil saturated hydraulic conductivity (Ks) is a critical parameter for modeling water and solute transport in soils. Conventional laboratory measurements of Ks are labor-intensive, costly, and susceptible to measurement errors, underscoring the need for more reliable estimation techniques. This study systematically compares the performance of Ordinary Kriging (OK), Ordinary Co-Kriging (OCK), and Response Surface Methodology (RSM) for Ks estimation, thereby integrating geostatistical and statistical optimization frameworks. Soil samples were collected from 135 locations within the surface layer (0–30 cm), and Ks along with key soil physicochemical properties were determined. In the geostatistical domain, OK based on a spherical semivariogram (R 2  = 0.81; nugget/sill = 10.19%) yielded moderate predictive ability (R 2  = 0.70, RMSE = 3.62 mm day −1 , MAE = 10.02 mm day −1 ), whereas OCK employing an exponential cross-semivariogram (R 2  = 0.91; nugget/sill = 0.45%) substantially improved accuracy (R 2  = 0.85, RMSE = 3.21 mm day −1 , MAE = 9.43 mm day −1 ). By contrast, RSM achieved the highest predictive performance, with a quadratic model producing R 2  = 0.94 and Adeq Precision = 49.2. Optimization within the experimental range indicated a maximum Ks of 137.18 mm day −1 at 8.9% clay and 86% sand. Collectively, these findings demonstrate that while OK and OCK provide valuable insights into the spatial dependence of Ks, RSM offers superior predictive accuracy and practical applicability for optimizing soil hydraulic functions in water resources and agricultural management.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:2045-2322
2045-2322
DOI:10.1038/s41598-025-19820-y