The method uses Junkins weighting functions to achieve global continuity and smoothness to high order. It uses a new adaptive polynomial selection technique to optimize the memory footprint. The GRACE 360×360 GGM03C spherical harmonics model is fit using the new method. Several truncated versions are also fit.
The coefficients for the archived Fetch models and runtime Fortran evaluation libraries are found here: LINK
Details are given in the following paper:
Nitin Arora and Ryan P. Russell. Efficient Interpolation of High-Fidelity Geopotentials, Journal of Guidance, Control, and Dynamics, Vol. 39, No. 1 (2016), pp. 128-143. https://doi.org/10.2514/1.G001291
The models, memory footprint, and expected typical speedups over spherical harmonics for typical low altitude applications are given below:
Fetch Model | Resolution | Memory (MB) | Speedup over SH |
F399m33v1 | 33×33 | ~121 | ~8 |
F399m70v1 | 70×70 | ~360 | ~25 |
F399m156v1 | 156×156 | ~898 | ~200 |
F399m360v1 | 360×360 | ~2360 | ~3000 |