Embedded Files
Tools: Geostatistical Wizard, Create Chart, Cross Validation
The purpose of this project was to learn to apply three interpolation techniques to predict continuous surfaces from point data. This project focuses on precipitation data across the contiguous United States and explores the differences between interpolation methods, their strengths and weaknesses, and the impact of key parameters on prediction accuracy.
In this project, I explored three interpolation techniques: Inverse Distance Weighting, Polynomial (local and global), and Radial Basis Function.
Out of the three techniques, I would choose the Radial Basis Function to be the best precipitation interpolation surface. This surface is smoother than the IDW surface and it doesn’t have the concentric patterns the IDW surface has. It also doesn’t have interpolation issues that create exaggerated values, like the polynomial function. Typical precipitation patterns have smooth transitions influenced by water bodies, terrain, and weather fronts, which I believe is being captured well by the radial basis function surface. This surface also has the lowest RMS error out of the three surfaces, which indicates a higher level of modeling accuracy.
Page updated
Google Sites
Report abuse