Skip to content

Modular compositional learning model for water quality

Notifications You must be signed in to change notification settings

LimnoDataScience/WQ_MCL

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Water Quality - Modular Compositional Learning (WQ-MCL)

authors

Robert Ladwig, Bennett McAfee, Paul C Hanson

overview

Repository to run Modular Compositional Learning for water quality research questions. The modularized backbone of the methdology relies on 1D-AEMpy, which runs a vertical one-dimensional aquatic ecosystem model (AEM) for water temperature, dissolved oxygen and organic carbon (dissolved and particulate as well as labile and refractory) dynamics using the general equation form of:

$A \frac{\partial C}{\partial t} + w \frac{\partial C}{\partial z} - \frac{\partial}{\partial z}(A K_z \frac{\partial C}{\partial z}) = P(C) - D(C)$

Water temperature and heat transport are simulated using an eddy-diffusion approach in which the turbulent eddy diffusivity coefficients are parameterized based on the gradient Richardson number. To ensure stability, we apply the implicit Crank-Nicolson scheme for the diffusive transport. Production and consumption terms of the water quality dynamics (dissolved oxygen and organic carbon) are simulated using a modified Patankar Runge-Kutta scheme to ensure mass conservation and to prevent unrealistic negative values. Net primary production is acting as a boundary condition based on vertical light limitation, integrated total phosphorus concentrations, and water temperature. Convective wind mixing is parameterized based on an integral energy approach.

To fetch the latest contents of 1D-AEMpy, run

git submodule update --init --recursive

About

Modular compositional learning model for water quality

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published