KrigeKit#
KrigeKit is a Python interface to a high-performance Fortran kriging and Sequential Gaussian Simulation engine, parallelised with OpenMP.
Install the precompiled package from PyPI:
python -m pip install krigekit
The binary wheel includes the Fortran engine; no compiler is required.
Get from installation to your first kriging map in five minutes.
Full reference for every class, method, and convenience function.
Task-oriented walkthroughs for ordinary kriging, co-kriging, SGSIM, and more.
Gallery of runnable scripts — ordinary kriging, co-kriging, SGSIM, and more.
What KrigeKit does#
Capability |
Notes |
|---|---|
Ordinary and simple kriging |
Point and block support |
Co-kriging |
Multiple variables, Linear Model of Coregionalisation |
Universal kriging / KED |
External drift variables |
Variogram analysis and fitting |
Experimental clouds, directional averages, nested models, LMC fitting |
Score transforms |
Normal-score / uniform quantile transforms for kriging and SGSIM |
Sequential Gaussian Simulation |
Reproducible random paths, multi-realisation |
Space-time kriging |
Sum-metric and product-sum ST covariance models |
Spatially Varying Anisotropy |
Per-block variogram (SVA mode) |
Multiple Indicator Kriging / SIS |
Categorical variables; uniform, proportional, or independent cross-variogram strategies |
Cross-validation |
Leave-one-out |
Kriging weight reuse |
Store and replay weights for fast value updates |
OpenMP parallelism |
Thread count controllable per |
Why KrigeKit?#
KrigeKit is designed for workflows that need Python usability with a compiled-Fortran backend. Its Fortran core handles large grids, SGSIM realisation paths, space-time systems, and OpenMP scheduling in a single library. The Python layer is a thin ctypes wrapper — no heavy dependencies, no JIT compilation step.