Premium
Modeling Transient Soil Moisture Limitations on Microbial Carbon Respiration
Author(s) -
Liu Yuchen,
Lawrence Corey R.,
Winnick Matthew J.,
Hsu HsiaoTieh,
Maher Katharine,
Druhan Jennifer L.
Publication year - 2019
Publication title -
journal of geophysical research: biogeosciences
Language(s) - English
Resource type - Journals
eISSN - 2169-8961
pISSN - 2169-8953
DOI - 10.1029/2018jg004628
Subject(s) - akaike information criterion , soil water , environmental science , soil respiration , soil science , carbon cycle , water content , dormancy , atmospheric sciences , moisture , range (aeronautics) , arid , soil carbon , biomass (ecology) , biological system , ecology , mathematics , meteorology , statistics , ecosystem , agronomy , biology , materials science , geology , physics , geotechnical engineering , germination , composite material
Soil microorganisms are known to survive periods of aridity and to recover rapidly after wetting events, with the ability to transition between a dormant state in dry conditions and an active state in wet conditions. Though this dynamic behavior has been previously incorporated into soil carbon respiration modeling frameworks, a direct comparison between this active‐dormant transition mechanism and a more simplified first‐order model has yet to be made. Here, we demonstrate the necessary extent of model complexity needed to reproduce transient carbon respiration rates obtained from a set of soil incubation experiments implemented over a range of soil depths and time intervals. Two approaches are tested, one uses simplified first‐order kinetics, whereas the other employs a transition between active and dormant biomass. The performance of each model is evaluated using an Akaike Information Criterion (AIC) based on the accuracy with which they reproduce an experimental dataset consisting of two sets of time series soil incubations collected across a range of time and depth resolutions. Based on the AIC evaluation and model‐data comparison, we conclude that a dormancy‐enabled model featuring two distinct microbial strategists performs best for the majority of the soil profile (above 108 cm) for both high and low depth resolution and sampling frequency, despite the added parameters required. In contrast, the first‐order model achieves better AIC scores when simulating our deepest soils (112–165 cm), where moisture fluctuations are expected to be less prevalent. These results guide how and where we choose to apply more cost intensive models.