Core interpolation module for converting between gridded forecast data and station observations. Implements multiple methods (nearest neighbor, linear, cubic, IDW, Cressman) with support for global longitude wrapping, multi-dimensional coordinates (member, level, time, dtime), and efficient vectorized operations. This module encapsulates a common and non-trivial data engineering pattern in meteorology and could be extracted as a standalone interpolation library with minimal adaptation.

Utility module for defining colormaps and contour levels tailored to meteorological variables (temperature, precipitation, wind, humidity, etc.). Provides a unified function `def_cmap_clevs` to create custom colormaps from lists of colors and levels, and element-specific presets. The module handles missing values, colorbar extensions, and supports both continuous and discrete color schemes. It is self-contained and reusable in any meteorological plotting context.

Implementation of the Fractions Skill Score (FSS) for spatial forecast verification. FSS compares forecast and observed binary fields at multiple neighborhood scales, providing a scale-dependent skill measure. The module includes convolution-based smoothing, thresholding, and score computation. It is a well-known spatial verification technique, and the implementation is self-contained (using numpy/scipy) and could be reused in other verification frameworks.