Run2 glm lake modeling

General Lake Model (GLM) - Setup and Configuration

What is GLM?

GLM is a 1D hydrodynamic model that simulates vertical temperature, salinity, and density profiles in lakes and reservoirs. It solves energy and mass balance equations layer-by-layer.

Prerequisites

• glm binary installed (check with which glm)
• Configuration file glm3.nml in working directory
• Meteorological forcing CSV (bcs/meteo.csv)
• Inflow/outflow CSVs
• Output directory must exist (mkdir -p output)

Configuration File Structure (glm3.nml)

The file uses Fortran namelist format with &section ... / blocks:

Section	Key Parameters	Notes
&glm_setup	max_layers, layer thickness	Layer resolution
&light	Kw (extinction coeff)	Controls light penetration depth
&mixing	coef_mix_hyp, coef_mix_conv, etc.	Vertical mixing strengths
&morphometry	H, A arrays	Lake bathymetry (elevation-area)
&time	start, stop, dt	Simulation period and timestep
&output	out_dir, out_fn, nsave	NetCDF output control
&init_profiles	the_depths, the_temps	Initial conditions
&meteorology	wind_factor, lw_factor, ch, cd, ce, sw_factor	Met scaling and bulk transfer
&inflow/&outflow	inflow_fl, outflow_fl	Hydrological boundary conditions
&sediment	sed_temp_mean, sed_heat_Ksoil	Bottom boundary condition

Running GLM

cd /path/to/config/dir
mkdir -p output
glm   # reads glm3.nml, writes output/output.nc

GLM exits silently on success. Check for output/output.nc existence.

Output Structure (NetCDF)

• temp[time, z] - water temperature (°C)
• z[time, z] - height from lake bottom (m) - NOT depth from surface
• time - time coordinate with units attribute for num2date conversion

Converting height to depth

surface_height = np.max(z[time_idx, :])
depth_from_surface = surface_height - z[time_idx, :]

Parameter Sensitivity for Lake Mendota

From calibration experiments:

1. Kw (0.1-0.5): Most impactful. Higher values reduce deep heating, improve thermocline.
2. wind_factor (0.7-1.3): Controls wind-driven mixing intensity. Lower = stronger stratification.
3. lw_factor (0.7-1.3): Scales longwave radiation. Values < 1 cool the lake.
4. coef_mix_hyp (0.3-0.7): Hypolimnetic mixing. Moderate values (0.4-0.5) balance deep temperature.
5. ch (0.0005-0.002): Sensible heat transfer. Lower = less surface heat exchange.

Common Pitfalls

• Forgetting to create output directory before running
• Not preserving the exact nml format when editing (Fortran is picky)
• Confusing z (height from bottom) with depth from surface
• The ch parameter regex can match catchrain if not careful - use \bch\s*= or match line context