I2.1 BeTR BGC modeling farm

Abstract

Soil biogeochemistry (BGC) modeling contributes uncertainty to earth system model predictions comparable to that of other components, such as clouds. This large uncertainty has been attributed to several factors, including forcing data, parameterization, model structure, and numerical implementations. However, while there are now many soil BGC model formulations used in the modeling community, a numerically robust framework does not exist to support a systematic evaluation of how these different soil BGC formulations will perform within the same land model. This then raises several risks in quantifying and constraining soil BGC modeling uncertainties, including: (1) attributing poor performance of a land model incorrectly to its soil BGC, although the problem may be associated with its vegetation component, or vice versa; (2) asserting a good soil BGC formulation as inappropriate even though its poor performance could be due to inappropriate numerical implementation; and (3) stating that different soil BGC models are incomparable because they are not implemented and parameterized in a consistent manner. The first two of these risks will cause a misplacement of parameterization uncertainty to soil BGC, and the three together present a barrier to the selection of which model formulation should be preferred in an Earth System Model, such as E3SM. Using the soil BGC reactive transport solver BeTR, which is successfully integrated in v1 of ELM (Tang and Riley, 2018), we demonstrate that BeTR’s polymorphism feature supports consistent implementations of different soil BGC models (including a single layer topsoil, single soil column, and fully coupled) with ELMv1. This ELMv1-BeTR “soil BGC model farm” allows one to evaluate different soil BGC formulations within ELM, while harmonizing uncertainties due to differences in forcing, numerical implementations, and simulation protocols. This ELMv1-BeTR hierarchical modeling capability is a valuable platform for soil BGC model development and field testing, and provides more comprehensive quantification and constraint of land soil BGC modeling in E3SM.