Skip to contentScholarGate
LibraryBookshelfDeskReview StudioAssistant
Sign in
On this page
IntuitionHow it worksWhen to use itStrengths & limitationsCommon pitfallsApplicationsFrequently asked🔒 Read the full methodSourcesRelated methods
Cite this pageSpotted an issue on this page? Report or suggest a fix →
Home›Spatial analysis›Local Kriging (Moving-Window Kriging)
Regression modelGIS / spatial

Local Kriging (Moving-Window Kriging)

Local (Moving-Window) Kriging · Also known as: moving-window kriging, local kriging interpolation, windowed kriging, neighborhood kriging

Local Kriging is a spatially adaptive geostatistical interpolation method that restricts each prediction to a moving neighborhood of nearby observations, fitting a variogram model locally within that window. This allows spatial covariance structure to vary across the study region rather than imposing a single global variogram, making it better suited to large or non-stationary spatial fields.

ScholarGate
  1. Regression model
  2. v1
  3. 2 Sources
  4. PUBLISHED
Cite this page →
Tools & resources
Download slides
Learn & explore

Read the full method

Members only

Sign in with a free account to read this section.

Sign in

Method map

The neighbourhood of related methods — select a node to explore.

Local Kriging
Co-krigingGeographically Weighted…Ordinary KrigingBayesian Kernel Density…Global Kriging

When to use it

Use Local Kriging when the spatial covariance structure changes across the domain (non-stationarity), when the dataset is large enough that a global variogram would mask important local patterns, or when different sub-regions have clearly different nugget or range characteristics. It is especially valuable in environmental monitoring, mining, and atmospheric sciences over large or heterogeneous areas. Do not use it when the dataset within any plausible window is too small (fewer than ~20 observations) to reliably fit a variogram, or when stationarity can be safely assumed — in those cases ordinary kriging is more stable and computationally cheaper.

Strengths & limitations

Strengths
  • Accounts for spatial non-stationarity by fitting variogram parameters locally rather than globally.
  • Adapts prediction weights to the local covariance structure, improving accuracy over heterogeneous regions.
  • Produces location-specific uncertainty estimates (kriging variance) that reflect local data density.
  • Scalable via parallel computation: each window is an independent kriging problem.
  • Retains the exact interpolation property of standard kriging — predictions pass through observed values.
Limitations
  • Requires sufficient observations within each window; sparse data cause unreliable local variogram fits.
  • Computationally expensive compared to global kriging because a variogram must be fitted and a kriging system solved at every prediction location.
  • Window size (bandwidth) is a tuning parameter with no universally optimal choice; too small a window yields noisy variograms, too large loses local adaptivity.
  • Discontinuities or artifacts can appear at window boundaries if the local variogram structure changes abruptly.
  • Automatic variogram fitting within small windows can fail or produce degenerate parameter estimates.

Frequently asked

How is Local Kriging different from ordinary kriging?

Ordinary kriging fits a single variogram to the entire dataset and uses global weights for every prediction. Local Kriging re-fits the variogram within a moving neighborhood around each target point, so the spatial covariance model can differ from one prediction location to another. This makes it more flexible in non-stationary settings but more computationally demanding.

How do I choose the window size?

A common rule of thumb is to include at least 20–50 nearby observations per window. In practice, cross-validation (leave-one-out or k-fold) is used to compare prediction errors across candidate window sizes or neighbor counts. Software packages like gstat in R provide utilities for this sensitivity analysis.

Can Local Kriging handle categorical or binary data?

Standard Local Kriging is designed for continuous spatial variables. For categorical outcomes, indicator kriging — applied locally — is the appropriate extension, though it requires fitting separate indicator variograms for each category within each window.

Is Local Kriging the same as Geographically Weighted Regression (GWR)?

No. GWR is a regression model that estimates locally varying relationships between a response variable and predictors using a spatial kernel weighting scheme. Local Kriging is a pure interpolation method based on spatial covariance (variogram) structure with no predictors beyond the spatial locations of observations. They address different questions and rely on different theoretical frameworks.

What software supports Local Kriging?

The gstat package in R implements moving-window kriging. GSLIB (Geostatistical Software Library) and its descendants support local approaches in Fortran-based workflows. Commercial packages such as ISATIS and Leapfrog Geo also provide local kriging functionality.

Sources

  1. Haas, T. C. (1990). Kriging and automated variogram modeling within a moving window. Atmospheric Environment, 24(7), 1759-1769. DOI: 10.1016/0960-1686(90)90508-K ↗
  2. Goovaerts, P. (1997). Geostatistics for Natural Resources Evaluation. Oxford University Press. ISBN: 9780195115383

How to cite this page

ScholarGate. (2026, June 3). Local (Moving-Window) Kriging. ScholarGate. https://scholargate.app/en/spatial-analysis/local-kriging

Related methods

Co-krigingGeographically Weighted RegressionOrdinary Kriging

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

  • Co-krigingSpatial analysis↔ compare
  • Geographically Weighted RegressionSpatial analysis↔ compare
  • Ordinary KrigingSpatial analysis↔ compare
Compare side by side →

Referenced by

Bayesian Kernel Density EstimationGlobal Kriging

Similar methods

Local Ordinary KrigingGlobal KrigingLocal Universal KrigingGlobal Ordinary KrigingOrdinary KrigingSpace-Time KrigingKrigingSpace-Time Ordinary Kriging

Related reference concepts

Gaussian Process ModelsCross-ValidationSpatial Point ProcessesDensity EstimationK-Means ClusteringQuadratic Discriminant Analysis

Spotted an issue on this page? Report or suggest a fix →

ScholarGate — Local Kriging (Local (Moving-Window) Kriging). Retrieved 2026-07-20 from https://scholargate.app/en/spatial-analysis/local-kriging · Dataset: https://doi.org/10.5281/zenodo.20539026
Quick facts
Originator
Haas, T. C.
Year
1990
Type
Spatial interpolation (local variant)
DataType
Georeferenced continuous measurements (point data)
Subfamily
GIS / spatial
Related methods
Co-krigingGeographically Weighted RegressionOrdinary Kriging
ScholarGate

A content-first reference library for research methods — what each one is, how it works, and where it comes from.

Open data (CC-BY)

Explore

  • Library
  • Search the library…
  • Browse by field
  • Fields
  • Journey
  • Compare
  • Which method?

Reference

  • Subjects
  • Atlas
  • Glossary
  • Methodology
  • Philosophy

Your tools

  • Bookshelf
  • Desk
  • Chat

Company

  • About
  • Pricing
  • Contact
  • Suggest a method

Entries are compiled from published sources for reference. Verifying the accuracy and suitability of any information for your own use remains your responsibility.

© 2026 ScholarGate · A research-method reference library
  • Privacy
  • Cookies
  • Terms
  • Delete account