Increased sensitivity of microbial respiration to soil water content in a fertilized boreal forest

Abstract

Projections of soil carbon (C) models are known to underestimate soil C stocks in boreal soils with higher nutrient status which is likely because they do not account for effects of soil nutrient status on the kinetics of microbial respiration and their sensitivities to environmental conditions. Here we evaluated the effects of long-term N addition (once per decade since 1960 until 2020) on soil heterotrophic respiration (Rh) and its dependence on soil temperature and moisture in an originally N limited boreal Scots pine (Pinus sylvestris) forest. We measured Rh, soil temperature and soil moisture biweekly during the vegetative seasons of 2021-2023 in both fertilized and control forests. We fitted Rh rates to soil temperature and moisture separately for the control and N fertilization treatment using parametric non-linear regression models and non-parametric machine learning (boosted regression tree) models. The functional dependencies of Rh were similar between fertilized and control forests for soil temperature but differed for soil moisture. In the N fertilized forest soil, Rh increased rapidly from dry conditions towards a soil moisture optimum followed by a clear reduction in wet conditions. In contrast, in the N limited forest soil, Rh mainly increased with soil moisture. The models based solely on temperature (assuming identical and non-limiting effect of moisture) predicted higher annual Rh than the models accounting for soil moisture effects. Thus, to avoid overestimation of soil CO2 emissions and underestimation of soil C stocks accumulation in fertile boreal soils, it is crucial to link the soil moisture dependencies in soil C models to nutrient status. The different Rh response to moisture between N limited and N fertilized soils could be related to different levels of enzyme activities and contrasting microbial traits found by other studies.