B13 Generalize MPAS-SI Algal Grazing for Multiple Algal Groups: Design Document
The Design Document page provides a description of the algorithms, implementation and planned testing including unit, verification, validation and performance testing. Please read Step 1.3 Performance Expectations that explains feature documentation requirements from the performance group point of view.
Design Document
In the table below 4.Equ means Equations and Algorithms, 5.Ver means Verification, 6.Perf - Performance, 7. Val - Validation, 
- competed, 
- in progress, 
- not done
Overview table for the owner and an approver of this feature
1.Description | generalize MPAS-SI algal grazing for multiple algal groups |
| 2.Owner | Nicole Jeffery |
| 3.Created | Nov 3, 2020 |
| 4.Equ | |
| 5.Ver | |
| 6.Perf | |
| 7.Val | |
| 8.Approver | Katherine Calvin (Unlicensed) |
| 9.Approved Date | |
| V2.0 |
Title: B13 Generalize MPAS-SI Algal Grazing for Multiple Algal Groups
Requirements and Design
E3SM CBGC Group
Date: 11/03/2020
Summary
In E3SMv1 MPAS-SI, the sea ice algal grazing closure term, which is linear in the algal growth rate, is appropriate for a coupled biogeochemical system consisting of a single algal group. However, the model consists of three algal groups – diatoms, small algae, and Phaeocystis sp. We overcame this deficiency in the v1 simulations by turning off grazing for the diatoms and increasing the grazing parameter to questionably high values (70%) for the small algae and Phaeocystis sp. groups, and then only publishing results that validated total chlorophyll-a, not the relative contributions. However, there are successful and simple treatments of grazing for multiple phytoplankton groups in ocean biogeochemical models that could be easily applied. In addition, improvements to the nitrogen cycling in MPAS-SI have reduced the primary production bias in the Arctic but exposed a chlorophyll-a bias that will undoubtably benefit from improved grazing.
Requirements
Requirement: Analysis of fully coupled simulations with appropriate flags activated and de-activated
- Implement the work of Dunne et al (2005) who treats grazing pressure from zooplankton implicitly such that it keeps pace with small algae, maintaining moderate to low production, while allowing for blooms of the diatom class. In effect, update the linear grazing term to a nonlinear grazing term that depends on the size classes characterized by the respective algal group.
- Compare sea ice primary production and chlorophyll in ten-year ocean-ice simulations with and without the Dunne grazing treatment. We are looking for high primary productivity without the poor chlorophyll-a biases of the current model.
- Some parameter tuning may also be required.
Date last modified: 11/03/2020
Contributors: Nicole Jeffery
Algorithmic Formulations
The original algal grazing formulation, graze(k), is:
graze(k) = fr_graze(k) * grow(k) (1)
for each algal group k, where grow(k) is the rate of accumulation of algal and fr_graze(k) is a set of fixed parameters.
The Dunne et al (2005) formulation for a coupled ecosystem with multiple algal groups generalizes (1) as:
graze(k) = fr_graze(k) * grow(k) (algal_N(k)/graze_conc)^graze_exp(k) (2)
where two additional parameters, graze_conc and graze_exp, could depend on the algal groups. In general, loss terms in the sea ice biogeochemical model are capped to 90% of the tracer concentration. We implement that here as well and modify (2) as follows:
graze(k) = min(max_loss* algal_N(k)/dt, fr_graze(k) * grow(k) (algal_N(k)/graze_conc)^graze_exp(k)) (3)
For parameter values, we use the implicit grazing exponents of Dunne et al (2005): for diatoms, graze_exp = 0.333, and for small phytoplankton, graze_exp = 1. We treat both small algae and Phaeocystis sp as small phytoplankton groups. The parameter, graze_conc, is assumed independent of the type of algal group and use 1.35 mmol N/m^3, again after a Dunne et al (2005).
Date last modified:
Contributors: Nicole Jeffery
Design and Implementation
All coding changes are made in ice_algal.F90 subroutine algal_dyn. Some small coding changes will be needed to ensure that algal nitrogen source and sink terms are consistent with the new grazing parametrization.
Date last modified:
Contributors: Nicole Jeffery
Planned Verification and Unit Testing
Verification and Unit Testing: restartability and parallelism
MPAS-SI already has unit testing for the BGC component. The regression test will fail, but restartability and parallelism should pass. All other tests will pass.
Date last modified:
Contributors: Nicole Jeffery
Planned Validation Testing
Validation Testing: Comparison of fully coupled run with base model and literature
At least two GCASE simulations of at least 10 years will be completed, one with and one without the new implementation. We will compare sea ice primary production and chlorophyll-a between runs and with observations (Jeffery et al 2020), particularly in the Arctic where we currently see biases. However we also need to verify that the fidelity of Southern Ocean primary production and chlorophyll-a concentrations is not degraded. Some tuning runs may be required. We will also verify that diatoms are the dominant sea ice primary producers.
References:
Dunne, J. P., R. A. Armstrong, A. Gnanadesikan, and J. L. Sarmiento (2005), Empirical and mechanistic models for the particle export ratio, Global Biogeochem. Cycles, 19, GB4026, doi:10.1029/2004GB002390.
Jeffery N et al. (2020). Investigating controls on sea ice algal production using E3SMv1.1-BGC. Annals of Glaciology 1–22. https://doi.org/10.1017/aog.2020.7
Date last modified:11/03/2020
Contributors: Nicole Jeffery
Planned Performance Testing
Performance Testing: Contrast timings of control and development GCASE simulations
Date last modified:10/13/2020
Contributors: Nicole Jeffery
We'll use the same GCASE runs to contrast computational expense. It is expected that there will be negligible additional expense from modifying the grazing parameterization.