Nitrogen immobilization usually leads to nitrogen retention in soil and, thus, influences soil nitrogen supply for plant growth. Understanding soil nitrogen immobilization is important for predicting soil nitrogen cycling under anthropogenic activities and climate changes. However, the global patterns and drivers of soil nitrogen immobilization remain unclear. We synthesized 1350 observations of gross soil nitrogen immobilization rate (NIR) from 97 articles to identify patterns and drivers of NIR. The global mean NIR was 8.77 ± 1.01 mg N kg−1 soil day−1. It was 5.55 ± 0.41 mg N kg−1 soil day−1 in croplands, 15.74 ± 3.02 mg N kg−1 soil day−1 in wetlands, and 15.26 ± 2.98 mg N kg−1 soil day−1 in forests. The NIR increased with mean annual temperature, precipitation, soil moisture, soil organic carbon, total nitrogen, dissolved organic nitrogen, ammonium, nitrate, phosphorus, and microbial biomass carbon. But it decreased with soil pH. The results of structural equation models showed that soil microbial biomass carbon was a pivotal driver of NIR, because temperature, total soil nitrogen, and soil pH mostly indirectly influenced NIR via changing soil microbial biomass. Moreover, microbial biomass carbon accounted for most of the variations in NIR among all direct relationships. Furthermore, the efficiency of transforming the immobilized nitrogen to microbial biomass nitrogen was lower in croplands than in natural ecosystems (i.e., forests, grasslands, and wetlands). These findings suggested that soil nitrogen retention may decrease under the land use change from forests or wetlands to croplands, but NIR was expected to increase due to increased microbial biomass under global warming. The identified patterns and drivers of soil nitrogen immobilization in this study are crucial to project the changes in soil nitrogen retention.

This article studies a general joint model for longitudinal measurements and competing risks survival data. The model consists of a linear mixed effects sub-model for the longitudinal outcome, a proportional cause-specific hazards frailty sub-model for the competing risks survival data, and a regression sub-model for the variance–covariance matrix of the multivariate latent random effects based on a modified Cholesky decomposition. The model provides a useful approach to adjust for non-ignorable missing data due to dropout for the longitudinal outcome, enables analysis of the survival outcome with informative censoring and intermittently measured time-dependent covariates, as well as joint analysis of the longitudinal and survival outcomes. Unlike previously studied joint models, our model allows for heterogeneous random covariance matrices. It also offers a framework to assess the homogeneous covariance assumption of existing joint models. A Bayesian MCMC procedure is developed for parameter estimation and inference. Its performances and frequentist properties are investigated using simulations. A real data example is used to illustrate the usefulness of the approach.

Abstract. The concentration–carbon feedback (β), also called the CO2 fertilization effect, is a key unknown in climate–carbon-cycle projections. A better understanding of model mechanisms that govern terrestrial ecosystem responses to elevated CO2 is urgently needed to enable a more accurate prediction of future terrestrial carbon sink. We conducted C-only, carbon–nitrogen (C–N) and carbon–nitrogen–phosphorus (C–N–P) simulations of the Community Atmosphere Biosphere Land Exchange model (CABLE) from 1901 to 2100 with fixed climate to identify the most critical model process that causes divergence in β. We calculated CO2 fertilization effects at various hierarchical levels from leaf biochemical reaction and leaf photosynthesis to canopy gross primary production (GPP), net primary production (NPP), and ecosystem carbon storage (cpool) for seven C3 plant functional types (PFTs) in response to increasing CO2 under the RCP 8.5 scenario. Our results show that β values at biochemical and leaf photosynthesis levels vary little across the seven PFTs, but greatly diverge at canopy and ecosystem levels in all simulations. The low variation of the leaf-level β is consistent with a theoretical analysis that leaf photosynthetic sensitivity to increasing CO2 concentration is almost an invariant function. In the CABLE model, the major jump in variation of β values from leaf levels to canopy and ecosystem levels results from divergence in modeled leaf area index (LAI) within and among PFTs. The correlation of βGPP, βNPP, or βcpool each with βLAI is very high in all simulations. Overall, our results indicate that modeled LAI is a key factor causing the divergence in β in the CABLE model. It is therefore urgent to constrain processes that regulate LAI dynamics in order to better represent the response of ecosystem productivity to increasing CO2 in Earth system models.

No abstract

Abstract. Soil organic carbon (SOC) has a significant effect on carbon emissions and climate change. However, the current SOC prediction accuracy of most models is very low. Most evaluation studies indicate that the prediction error mainly comes from parameter uncertainties, which can be improved by parameter calibration. Data assimilation techniques have been successfully employed for the parameter calibration of SOC models. However, data assimilation algorithms, such as the sampling-based Bayesian Markov chain Monte Carlo (MCMC), generally have high computation costs and are not appropriate for complex global land models. This study proposes a new parameter calibration method based on surrogate optimization techniques to improve the prediction accuracy of SOC. Experiments on three types of soil carbon cycle models, including the Community Land Model with the Carnegie-Ames-Stanford Approach biogeochemistry submodel (CLM-CASA') and two microbial models show that the surrogate-based optimization method is effective and efficient in terms of both accuracy and cost. Compared to predictions using the tuned parameter values through Bayesian MCMC, the root mean squared errors (RMSEs) between the predictions using the calibrated parameter values with surrogate-base optimization and the observations could be reduced by up to 12 % for different SOC models. Meanwhile, the corresponding computational cost is lower than other global optimization algorithms.