Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
D5923-18 – Standard Test Method Technical Guide
ASTM D5923-18, “Standard Guide for Selection of Kriging Methods in Geostatistical Site Investigations,” provides a consensus-based framework for choosing appropriate kriging techniques in environmental and geotechnical studies. The guide emphasizes that selection should be based on study objectives, exploratory data analysis, and analysis of spatial variation.
📐 Commonly Used Kriging Methods in D5923-18
The standard focuses on several well-established kriging variants, which are least-squares estimators that account for spatial continuity. The table below summarizes the primary forms of kriging covered by D5923-18.
| 🟦 Method | 📐 Description | 🎯 Application |
|---|---|---|
| Ordinary Kriging | Estimates with unknown but constant mean. | General spatial interpolation. |
| Simple Kriging | Assumes known mean across domain. | When prior mean is available. |
| Lognormal Kriging | Apply to log-transformed data. | Skewed distributions like contaminants. |
| Universal Kriging | Accounts for spatial trends. | Data with regional drift. |
| Indicator Kriging | Estimates binary threshold probabilities. | Risk of exceeding limits. |
⚙️ Selection Criteria and Data Analysis
Selecting the appropriate kriging method requires careful data analysis. D5923-18 outlines that study objectives, exploratory data analysis, and spatial variation analysis guide method selection. The table below summarizes these criteria.
| 🟦 Criterion | 📐 Consideration | ⚡ Impact on Selection |
|---|---|---|
| Study Objectives | Estimation vs. uncertainty assessment. | Kriging for estimation, simulation for uncertainty. |
| Exploratory Data Analysis | Distribution, outliers, trends. | Informs transformation or trend removal. |
| Spatial Variation | Variogram modeling and stationarity. | Defines kriging model assumptions. |
💡 Exploratory data analysis and variography are critical before method selection to ensure data support the kriging assumptions.
📊 Conditions for Geostatistical Simulation
Kriging methods are not always appropriate. D5923-18 describes conditions where geostatistical simulation should be used, such as when preserving variance and extreme values is necessary. Simulation avoids the smoothing effect of kriging and is suitable for risk assessment.
⚠️ Kriging smooths local details. For applications requiring full distribution preservation, such as health risk assessments, geostatistical simulation is recommended.
❓ Frequently Asked Questions
🔍 Which kriging methods are discussed in D5923-18?
The standard covers ordinary kriging, simple kriging, lognormal kriging, universal kriging, and indicator kriging. Multivariate and space-time methods are not discussed.
💡 How does D5923-18 guide method selection?
Selection is based on study objectives, exploratory data analysis, and spatial variation analysis. The standard offers recommendations without prescribing a specific course of action.
⚡ When is simulation recommended over kriging?
Simulation is recommended when kriging is inappropriate, typically when the smoothing effect of kriging obscures critical variability or when uncertainty must be assessed.
📌 Does D5923-18 discuss non-geostatistical alternatives?
No, the standard does not discuss alternatives such as splines or inverse-distance techniques. It focuses solely on kriging and geostatistical simulation.