Netcdf analysis

Skill cxcscmu/SkillLearnBench/skills/b1-one-shot-claude-opus-4-6/temperature-simulation/netcdf-analysis

[COLM'26] SkillLearnBench is the first benchmark for evaluating continual learning methods that automatically generate agent skills.

Install

npx -y skills add cxcscmu/SkillLearnBench --skill netcdf-analysis

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

Reading and analyzing GLM NetCDF output with Python netCDF4 and pandas for RMSE evaluation.

SKILL.md

1.2 KB, 282 tokens by cl100k_base, as published. Nobody here has run it

NetCDF Analysis for GLM Output

Reading GLM Output

import netCDF4 as nc
import numpy as np
import pandas as pd

ds = nc.Dataset('output/output.nc')
temp = ds.variables['temp'][:]  # masked array [time, layers]
z = ds.variables['z'][:]        # height above bottom [time, layers]
time_var = ds.variables['time']
times = nc.num2date(time_var[:], time_var.units)

Extracting Temperature at Specific Depths

GLM uses variable layer heights. For each timestep:

lake_depth = 25  # from morphometry (crest_elev - min(H))
for t in range(len(times)):
    valid = ~temp[t].mask if hasattr(temp[t], 'mask') else np.ones(temp.shape[1], bool)
    depths_from_surface = lake_depth - z[t, valid]
    temps = temp[t, valid]
    # Interpolate to desired depth

RMSE Calculation

# Merge on exact datetime and rounded depth
# rmse = sqrt(mean((obs - sim) ** 2))

Key Notes

GLM z is height from lake bottom; depth = lake_depth - z
Round depths to nearest integer for matching
Use exact datetime matching (no nearest-time)

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.