Merging Climate Station Records: Inverse-Variance Weights, Correlated Errors and Merged Uncertainty
Several weather stations measure the same temperature with different noise, and their errors may be correlated. Set the number of stations, their quality spread, the error correlation and a biased record, and watch the best weights, the merged uncertainty and a live Monte Carlo test against the plain average. A Blender-built station network with numerical checks of the generalised least-squares formulas.
Several weather stations measure the same temperature with different noise, and their errors may be correlated. Set the number of stations, their quality spread, the error correlation and a biased record, and watch the best weights, the merged uncertainty and a live Monte Carlo test against the plain average. A Blender-built station network with numerical checks of the generalised least-squares formulas.
Several weather stations measure the same temperature with different noise, and their errors may be correlated. Set the number of stations, their quality spread, the error correlation and a biased record, and watch the best weights, the merged uncertainty and a live Monte Carlo test against the plain average. A Blender-built station network with numerical checks of the generalised least-squares formulas.
3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install