B36. MAM verification and evaluation using a box model

Poster Title	Verification and evaluation of the aerosol microphysics parameterization in E3SM using a box model
First Author	Jian Sun (Unlicensed)
Topic	software tools, verification
Affiliation	NGD Architecture Software and Algorithm, ACME-SM: A Global Climate Model Software Modernization Surge
Link to document

Title

Verification and evaluation of the aerosol microphysics parameterization in E3SM using a box model

Authors

Jian Sun (Unlicensed), Richard Easter (Unlicensed), Hui Wan Kai Zhang

Abstract

Atmospheric aerosol particles play an important role in the climate system, but the representation of aerosol processes in Earth system models still needs to be improved. The Modal Aerosol Module (MAM) used in E3SM provides a simplified but rather complete treatment of the aerosol processes. Based on assumptions on the size and composition of aerosols, the model solves different kinds of equations, including algebraic equations (e.g. for aerosol water uptake), ordinary differential equations (e.g. for trace-gas condensation), and/or partial differential equations (e.g. for sedimentation). Since many processes are involved and some of them are tightly coupled, we need to make sure that these equations are correctly formulated and accurately solved. Using the terminology from other computational sciences, we distinguish two types of testing: (1) verification, which checks whether the numerical methods and code implementation correctly solve the equations the model developer intended to formulate and solve, and (2) validation, which checks if the results reasonably represent the physical phenomena in the real world. However, testing MAM in the full E3SM is challenging, since the aerosol processes and other physical/dynamical processes have strong interactions.

In this effort, we use a box model to test MAM and its components in isolation from the resolved atmospheric dynamics and other parameterized processes (e.g., clouds and radiation). The box model is updated to the MAM version used in the current E3SM (i.e., they share the same code and data structure), so that we can verify the same code and merge any code improvements to E3SM more easily. We have developed offline drivers for aerosol microphysical processes including water uptake, nucleation, condensation, coagulation, aging, and redistribution of particles among modes of different size ranges. In addition to performing unit tests for individual aerosol processes, the box model is also used to test numerical convergence for time integration of multiple processes and the impact of operator splitting. We show two examples in this presentation: 1) verification of the numerical solution for trace-gas condensation in terms of time step convergence; and 2) evaluation of the aerosol water uptake parameterization in MAM by varying the ceiling values of relative humidity.