glm-basics

A good starting point for GLM calibration tasks. Use to inspect glm3.nml, confirm how GLM runs, and identify the relevant files before moving on to calibration and output evaluation.

SKILL.md

GitHub

GLM Basics Guide

Overview

GLM (General Lake Model) is a 1D hydrodynamic model that simulates vertical temperature and mixing dynamics in lakes. It reads configuration from a namelist file and produces NetCDF output.

Suggested Skill Flow

For benchmark calibration tasks, use the GLM skills in this order:

  1. glm-basics: inspect files, confirm run mechanics, and identify the task's editable/configuration surface
  2. glm-calibration: search for a physically passing parameter set within the allowed scope
  3. glm-output: compute the final exact task metrics and write /root/metrics.json

This flow tends to work better than jumping straight into calibration or reporting without first checking the run setup and file layout. Once calibration has found a configuration that already passes every task metric, stop searching and move directly to glm-output rather than spending more runs polishing an already passing point.

Running GLM

cd /root
glm

GLM reads glm3.nml in the current directory and produces output in output/output.nc.

Input File Structure

File Description
glm3.nml Main configuration file (Fortran namelist format)
bcs/*.csv Boundary condition files (meteorology, inflows, outflows)

Configuration File Format

glm3.nml uses Fortran namelist format with multiple sections:

&glm_setup
   sim_name = 'LakeName'
   max_layers = 500
/
&light
   Kw = 0.3
/
&mixing
   coef_mix_hyp = 0.5
/
&meteorology
   meteo_fl = 'bcs/meteo.csv'
   wind_factor = 1
   lw_factor = 1
   ch = 0.0013
/
&inflow
   inflow_fl = 'bcs/inflow1.csv','bcs/inflow2.csv'
/
&outflow
   outflow_fl = 'bcs/outflow.csv'
/

Modifying Parameters with Python

import re

def modify_nml(nml_path, params):
    with open(nml_path, 'r') as f:
        content = f.read()
    for param, value in params.items():
        pattern = rf"({param}\s*=\s*)[\d\.\-e]+"
        replacement = rf"\g<1>{value}"
        content = re.sub(pattern, replacement, content)
    with open(nml_path, 'w') as f:
        f.write(content)

# Example usage
modify_nml('glm3.nml', {'Kw': 0.25, 'wind_factor': 0.9})

Common Issues

Issue Cause Solution
GLM fails to start Missing input files Check bcs/ directory
No output generated Invalid nml syntax Check namelist format
Simulation crashes Unrealistic parameters Use values within valid ranges

Best Practices

  • Always backup glm3.nml before modifying
  • Run GLM after each parameter change to verify it works
  • Check output/ directory for results after each run