Default Model Research Articles

Predicting Critical Heat Flux in Narrow Rectangular Channels: A CFD Approach for Subcooled Flow

The PALLAS-reactor is an advanced nuclear reactor designed for producing medical isotopes and for carrying out research on (medical) nuclear technology. RELAP5 is the primary tool for modeling the accident scenarios relevant to the licensing of this reactor. The Groeneveld Look-Up Tables (LUT) are the default model of RELAP5 for estimating Critical Heat Flux (CHF). Literature has shown that the LUT over-predict CHF inside narrow, rectangular channels such as the cooling channels of the core of the PALLAS-reactor. Empirical correlations such as that of Sudo–Kaminaga or its modified version by Kim et al. (SK-Kim) are better suited for these geometries. However, these correlations are typically implemented in one-dimensional (1D) System Thermal-Hydraulics codes whereby the quantities of interest are averaged-out over the control domains. Computational Fluid Dynamics (CFD) could capture three-dimensional effects that could help reducing unnecessary conservatism embedded in some 1D approaches (e.g. Sudo–Kaminaga or SK-Kim correlations). Therefore, CFD is to be employed to support the Deterministic Safety Analyses of the PALLAS-reactor if there are scenarios in which the SK-Kim correlation predicts CHF. Boiling flows inside round or annular vertical channels have been extensively modeled with CFD. However, attempts at modeling CHF in narrow, rectangular geometries are still scarce. Therefore, this work aims at developing and validating a CFD approach for predicting CHF in rectangular channels. Four combinations of turbulence interaction and turbulent dispersion models are investigated; the results are compared against experiments and the SK-Kim correlation; sensitivity analyses are conducted on the turbulence models as well as the mesh. Certain combinations of turbulence interaction and turbulent dispersion models promote the transport of steam towards the center of the channel, thus increasing CHF, while other cause the steam to remain confined at the wall, thereby lowering the CHF. This study provides several CFD options to estimate CHF depending on the required level of conservatism. Estimating CHF in a conservative manner is crucial for licensing the PALLAS-reactor and other reactors of similar design, thereby securing the production of medical isotopes in Europe and the rest of the world.

Autoregressive HMM resolves biomolecular transitions from passive optical tweezer force measurements

Optical tweezer (OT) single molecule force spectroscopy is a powerful method to map out the energy landscape of biological complexes and has found increasing applications in academic and pharmaceutical research. The dominant method to extract molecular conformation transitions from the thermal diffusion-broadened trajectories of the microscopic OT probes attached to the single molecule of interest is through hidden Markov models (HMM). In standard applications, the HMMs assume a white noise spectrum of the probes superimposed onto the molecular signal.Here, we demonstrate, through theoretical derivation, computer modeling and experimental measurements that this standard white noise HMM (wnHMM) misses key features of real OT data. The deviation is most pronounced at higher frequencies because the white noise model does not account for the over-damped nature of particle diffusion in an OT harmonic potential in aqueous environments. To address this, we derive how to incorporate autoregression (ar) between consecutive data points into an HMM, and demonstrate through modeling and experiment that such an arHMM captures real OT data behavior across all frequency ranges. Through analysis of real OT data we recorded on a single DNA hairpin undergoing folding and unfolding transitions, we show that the wnHMM extracts lifetimes that are at least a factor of 2 faster and less consistent than the arHMM results which match expectations and prior measurements. Overall, our work suggests that arHMM should be the default model choice for analysis OT single molecule transitions and that its use will improve the fidelity and accuracy of single molecule force spectroscopy measurements.

Cross-shore hydrodynamics and morphodynamics modeling of an erosive event in the inner surf zone

The phase-averaged and depth-integrated coastal morphodynamic model, XBeach-Surfbeat, was investigated for its capability of predicting the cross-shore hydrodynamics and morphodynamics in the inner surf zone by simulating the storm-induced berm erosion, sediment transport, and subsequent sand bar formation. By utilizing a comprehensive hydrodynamic and morphodynamic dataset measured in a large wave flume and high-fidelity 3D large-eddy simulation (LES) data, a rigorous model validation was conducted to assess its capability in predicting inner-surf zone hydrodynamics and to explore how the improved hydrodynamic performance impacts the predicted morphodynamics. Using the default model parameters of the model, the undertow was overestimated with the peak magnitude being 30%–35% larger in the inner surf zone. Combining Monte Carlo simulation, the optimum hydrodynamic calibration for the simulated undertow was achieved when the roller energy dissipation parameter (β) was maintained below 0.1, and the threshold water depth (hmin) exceeded 0.25 m. The calibrated undertow improved the morphodynamic predictions by reducing the excessive berm erosion (Event I) and sand bar growth in the inner surf zone (Event II). Further improved morphodynamic predictions were achieved by calibrating sediment transport parameters, including the onshore sediment transport coefficient (γua) and the bore interval coefficient (Tbfac) associated with turbulence-bed interaction. A consistent set of optimized model coefficients for the model is shown to be effective in simulating the entire erosive event (combined Events I and II). This study reveals that further improvement of the model's capability may require incorporating new parameterizations and physics, such as wave-breaking-induced turbulence and wave nonlinearity associated with sediment transport in the inner surf and swash zones.

Revised 4-Point Water Model for the Classical Drude Oscillator Polarizable Force Field: SWM4-HLJ.

In this work the 4-point polarizable SWM4 Drude water model is reparametrized. Multiple models were developed using different strategies toward reproduction of specific target data. Results indicate that no individual model can reproduce all the selected target data in the context of the present form of the potential energy function. The changes considered in the new models include, 1) variations in the gas phase dipole moment, 2) variations in the molecular polarizability, 3) variations of the distance between the oxygen and the M site, 4) variation of the oxygen Lennard-Jones (LJ) parameters, 5) introduction of a LJ potential to the hydrogen atoms, and 6) variations of the H-O-H angle. Detailed analysis is presented for 3 new water models from which a final model, SWM4-HLJ, is selected as the future default model for the Drude polarizable force field. The model maintains the gas phase dipole moment as the experimental value while the remaining listed terms were adjusted including a larger H-O-H angle (108.12). Compared to its predecessor, SWM4-NDP, the self-diffusion coefficient, water dimer properties, and water cluster energies are greatly improved. The temperature dependence of the density of the new model also performs better. Overall, the new SWM4-HLJ water model is a general improvement and a good balance between microscopic and bulk properties is achieved.

Explainable tuned machine learning models for assessing the impact of corrosion on bond strength in concrete

This study mainly aims to evaluate the bond strength of corroded reinforcements in reinforced concrete members. In this regard, a comprehensive dataset containing a total of 285 specimens was collected from previous experiments. All collected specimens, including normal concrete, were subjected to pull-out tests to ensure consistent results. The features evaluated are associated with both concrete and rebar characteristics, corrosion rate, and duration. Six machine learning (ML) models were used to assess the dataset: Decision Tree, Random Forest, Light Gradient-Boosting Machine, Gradient Boosting, Extreme Gradient Boosting, and Extra Tree. Hyperparameter tuning was conducted using grid search to optimize model performance and enhance predictive accuracy. Additionally, the Shapley Values technique was utilized to interpret the significance of the features on bond strength.The results show that Extreme Gradient Boosting and Extra tree methods outperformed the other models, with R2 score of 0.9 each and RSME of 2.21 and 1.87, respectively. Furthermore, tuned models resulted in more accurate performance than the default models. Evaluating the significance of studied features indicated that the elevated levels of corrosion were associated with a negative impact on bond strength. In addition, the corrosion rate is considered to be the most influential factor affecting the bond strength.

Employer Credit Checks: Poverty Traps Versus Matching Efficiency

Abstract We develop a framework to understand pre-employment credit screening as a signal from credit markets that alleviates adverse selection in labour markets. In our theory, people differ in both their propensity to default on debt and the profits they create for firms that employ them; in our calibrated economy, highly productive workers have a low default probability. This leads firms to create more jobs for those with good credit, which creates a poverty trap: an unemployed worker with poor credit has a low job finding rate, but cannot improve her credit without a job. This manifests as an endogenous loss in present-discounted wages that is typically taken as exogenous in quantitative models of consumer default. Banning employer credit checks eliminates the poverty trap, but pools job seekers and reduces matching efficiency: average unemployment duration rises by 2 days for high productivity workers and falls by 13 days for low-productivity workers.

In at the Deep End: Innovative approaches to engaging underserved communities in Northern England

Abstract Background People living in socio-economically deprived communities have greater health needs, but worse access and outcomes from care. A growing international movement of Deep End (DE) general practice networks established to promote health equity, address challenges posed by the inverse care law and advocate for increased research involvement. Researchers, community members and practitioners in the North of England are exploring ways of engaging with underserved communities and addressing low levels of inclusion in research. Methods DE research i) co-designing the North East and North Cumbria (NENC) DE network with practitioners to prioritise service investment, research activities and initiatives for local communities; ii) Deep End Research Alliance Yorkshire and Humber (DERA-YH) and Participatory Action Research to increase research capacity and reflect upon the evolution of a novel primary care Community Research Link Worker role (CRLWs). Results Prioritisation activities in the NENC DE network, uncovered the volume and complexity of clinical and social need, leading to the co-design and evaluation of initiatives to improve mental health care, increase childhood immunisations, tailor deprescribing, support training, and increase patient inclusion in research. In the DERA-YH, the CRLW model increased participation in research from ethnic minority communities regarding prostate cancer, lung health, dementia, and contraceptive services. Community leaders stated that sharing resources and power in the research process builds trust and interest in research in their communities. Conclusions DE networks in Northern England have driven innovative approaches to engage underserved communities in research, fostering empowering and enduring relationships to address health and care inequalities. Co-design methodologies and CRLWs integration into research teams had a transformative impact on the accessibility of research for communities and for the researchers’ positionality. Key messages • Deep End networks are advancing approaches to meaningful engagement and partnership working with underserved communities, to address health, care and research participation inequalities. • We recommend co-design by default and CRLW models to build capacity and embed reflective practice, mutual respect, and power sharing across the research team.

Partial hedging in credit markets with structured derivatives: a quantitative approach using put options

This study develops a novel method for mitigating credit risk through the use of structured derivatives, focusing in particular on the use of European put options as a strategic hedging tool. Inspired by the work of Merton (1974), our approach introduces the concept of default triggered by the stock price ST breaching a predefined barrier B. By establishing a distributional equivalence between an existing default model and P(ST<B) for a given time T, we demonstrate the potential for reducing the necessary capital allocation for a projected loss X(T) by partially hedging with a European put option. We formulate and solve an optimization problem w.r.t. a specific risk measure to determine the optimal strike price for the option, and our numerical analysis confirms a reduction in the Solvency Capital Requirement (SCR) in markets with and without jumps. Our findings provide (insurance) companies with a pragmatic approach to mitigating losses while maintaining their current risk management framework.

Who feels they contribute to U.S. society? Helping behaviors and social class disparities in perceived contributions.

Americans in lower (vs. higher) social class contexts are less likely to believe they contribute to society. Helping others by giving one's time is an important way of contributing to others that also varies with social class. Five studies (N = 7,326) investigated whether one source of the social class disparity in perceived contributions is a default model that considers helping distant others (i.e., bridging help, e.g., volunteering) as more of a contribution than helping close others (i.e., bonding help, e.g., caring for family members). In Study 1, Americans in lower (vs. higher) social class contexts perceived they contribute less to society (i.e., self-perceived contributions, Part A) and believed others perceive them as contributing less (i.e., metaperceived contributions, Part B). Studies 2-4 provide evidence for a default model of social good: Americans across social class contexts and even helpers themselves perceived bridging help as more of a contribution than bonding help, in part, because bridging help is perceived as reflecting more choice to help. With a representative sample (Midlife Development in the United States), Study 5 finds that Americans in lower (vs. higher) social class contexts engaged in relatively less bridging help and more bonding help. However, bridging help served as a stronger pathway to feelings of contributing than bonding help did. Together, these studies suggest that people in lower social class contexts may experience a psychological inequality, in part, because some of the forms of help that are most accessible, familiar, and practiced are widely perceived as less of a contribution. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Textual analysis and credit scoring: a new matrix factorization approach

Credit scoring models are important for financial institutions’ credit decisions. This study examined how variables are extracted from loan statements and whether textual variables can improve the accuracy of the default model. We used a combination of forward selection and non-negative matrix factorization to extract variables from loan statements. We also built a credit scoring model using both loan statement and numerical data. The results show that in the comparative analysis, the credit scoring model built using the optimal cut-off logistic regression model and the two types of data had the highest accuracy. Moreover, compared with the credit scoring model constructed using the deep learning method based on word vectors, the credit scoring model in this study had better interpretation. The regression analysis revealed that the variables from the loan statement have a significant effect on the default status.

Prolegomena to a genealogy of “society”

The article reconstructs the evolution of the semantics of one of the social sciences’ most important concepts, that of society. It follows two main systematizations, those of the societas and the societas civilis, respectively, through a millennia-long history rooted in Graeco-Roman civilization. This history exhibits remarkable continuity across time. Its modern adaptations survive to the present day. But the social sciences appear to know only the latter which, in the form of the nation-state, they adopted as their default model under the intellectual leadership of Parsons. Recent developments in the social world have rendered that model obsolete. This leaves the social sciences in dire straits, as they have failed, with one exception, to develop a persuasive alternative. The exception is Niklas Luhmann's systems theory of world society, a theory barely known outside small niche circles. The article ends on a brief speculative note about what might account for the odd tenacity of a theoretically untenable concept.

Imbalanced Data Parameter Optimization of Convolutional Neural Networks Based on Analysis of Variance

Classifying imbalanced data is important due to the significant practical value of accurately categorizing minority class samples, garnering considerable interest in many scientific domains. This study primarily uses analysis of variance (ANOVA) to investigate the main and interaction effects of different parameters on imbalanced data, aiming to optimize convolutional neural network (CNN) parameters to improve minority class sample recognition. The CIFAR-10 and Fashion-MNIST datasets are used to extract samples with imbalance ratios of 25:1, 15:1, and 1:1. To thoroughly assess model performance on imbalanced data, we employ various evaluation metrics, such as accuracy, recall, F1 score, P-mean, and G-mean. In highly imbalanced datasets, optimizing the learning rate significantly affects all performance metrics. The interaction between the learning rate and kernel size significantly impacts minority class samples in moderately imbalanced datasets. Through parameter optimization, the accuracy of the CNN model on the 25:1 highly imbalanced CIFAR-10 and Fashion-MNIST datasets improves by 14.20% and 5.19% compared to the default model and by 8.21% and 3.87% compared to the undersampling model, respectively, while also enhancing other evaluation metrics for minority classes.

Evaluation of railway noise calculation models using CNOSSOS-EU considering the effect of rail roughness

Railway noise assessment is an important tool in decision-making for urban planning, investment in transportation infrastructure and implementation of noise reduction strategies. Currently, the procedure is based on CNOSSOS-EU (Common Noise Assessment Methods in Europe), which was developed in accordance with European Directive 2002/49/EC. This method takes into account a variety of factors impacting rail noise levels, such as vehicle characteristics and track features. The purpose of this study was to investigate the influence of rail roughness on generated railway noise and to evaluate the accuracy of calculation models using CNOSSOS-EU. The study involved in-situ railway noise measurements at three different locations near the city of Zagreb, Croatia, and measuring rail roughness at each location. In order to eliminate the influence of other factors, the rail track structures at the measurement locations were similar, and an identical train configuration was used for the study. Finally, a comparison was made between calculation models based on default roughness values and models including values obtained from in-situ measurements. This approach enabled an assessment of the variations in CNOSSOS-EU calculation models taking into account actual in-situ rail roughness measurements.

Genome‐Enabled Parameterization Enhances Model Simulation of CH4 Cycling in Four Natural Wetlands

AbstractMicrobial processes are crucial in producing and oxidizing biological methane (CH4) in natural wetlands. Therefore, modeling methanogenesis and methanotrophy is advantageous for accurately projecting CH4 cycling. Utilizing the CLM‐Microbe model, which explicitly represents the growth and death of methanogens and methanotrophs, we demonstrate that genome‐enabled model parameterization improves model performance in four natural wetlands. Compared to the default model parameterization against CH4 flux, genomic‐enabled model parameterization added another contain on microbial biomass, notably enhancing the precision of simulated CH4 flux. Specifically, the coefficient of determination (R2) increased from 0.45 to 0.74 for Sanjiang Plain, from 0.78 to 0.89 for Changbai Mountain, and from 0.35 to 0.54 for Sallie's Fen, respectively. A drop in R2 was observed for the Dajiuhu nature wetland, primarily caused by scatter data points. Theil's coefficient (U) and model efficiency (ME) confirmed the model performance from default parameterization to genome‐enabled model parameterization. Compared with the model solely calibrated to surface CH4 flux, additional constraints of functional gene data led to better CH4 seasonality; meanwhile, genome‐enabled model parameterization established more robust associations between simulated CH4 production rates and environmental factors. Sensitivity analysis underscored the pivotal role of microbial physiology in governing CH4 flux. This genome‐enabled model parameterization offers a valuable promise to integrate fast‐cumulating genomic data with CH4 models to better understand microbial roles in CH4 in the era of climate change.

Sovereign risk and intangible investment

This paper measures the output and TFP losses from sovereign risk, considering firm-level intangible investment. Using Italian firm-level data, we show that firms reallocated from intangible assets to tangible assets during the 2011–2012 Italian sovereign debt crisis. This asset reallocation is more pronounced among small firms and high-leverage firms. This reallocation affects aggregate output and TFP. To explain the reallocation pattern and quantify the output and TFP losses, we build a sovereign default model incorporating firm intangible investment. In our model, sovereign risk deteriorates bank balance sheets, disrupting banks’ ability to finance firms. Firms with greater external financing needs are more exposed to sovereign risk. Facing tightening financial constraints, firms shift their resources towards tangibles because they can be used as collateral. We find that elevated sovereign risk explains 45% of the observed output losses and 31% of the TFP losses in Italy from 2011 to 2016.

An Explainable AI-Based Modified YOLOv8 Model for Efficient Fire Detection

Early fire detection is the key to saving lives and limiting property damage. Advanced technology can detect fires in high-risk zones with minimal human presence before they escalate beyond control. This study focuses on providing a more advanced model structure based on the YOLOv8 architecture to enhance early recognition of fire. Although YOLOv8 is excellent at real-time object detection, it can still be better adjusted to the nuances of fire detection. We achieved this advancement by incorporating an additional context-to-flow layer, enabling the YOLOv8 model to more effectively capture both local and global contextual information. The context-to-flow layer enhances the model’s ability to recognize complex patterns like smoke and flames, leading to more effective feature extraction. This extra layer helps the model better detect fires and smoke by improving its ability to focus on fine-grained details and minor variation, which is crucial in challenging environments with low visibility, dynamic fire behavior, and complex backgrounds. Our proposed model achieved a 2.9% greater precision rate, 4.7% more recall rate, and 4% more F1-score in comparison to the YOLOv8 default model. This study discovered that the architecture modification increases information flow and improves fire detection at all fire sizes, from tiny sparks to massive flames. We also included explainable AI strategies to explain the model’s decision-making, thus adding more transparency and improving trust in its predictions. Ultimately, this enhanced system demonstrates remarkable efficacy and accuracy, which allows additional improvements in autonomous fire detection systems.

XGBOOST HYPERPARAMETER OPTIMIZATION USING RANDOMIZEDSEARCHCV FOR ACCURATE FOREST FIRE DROUGHT CONDITION PREDICTION

Climate change and increasing global temperatures have increased the frequency and intensity of forest fires, making fire risk evaluation increasingly important. This study aims to improve the accuracy of predicting forest fuel drought conditions (Drought Code) by using the XGBoost algorithm optimized with RandomizedSearchCV. The research methods include collecting data related to forest fires, preprocessing data to ensure quality and consistency, and using RandomizedSearchCV for XGBoost hyperparameter optimization. The results showed that the optimized XGBoost model resulted in a decrease in Mean Squared Error (MSE) and an increase in R-squared value compared to the default model. The optimized model achieved an MSE of 0.0210 and R2 of 0.9820 on the test data, indicating significantly improved prediction accuracy for forest fuel drought conditions. These findings emphasize the importance of hyperparameter optimization in improving the accuracy of predictive models for forest fire risk assessment.

Comparative Analysis of Hepatitis C virus Genotype 1a (Isolate 1) using Multiple Regression Algorithms and Fingerprinting Techniques

Approximately 70 million people worldwide have been infected with Hepatitis C virus (HCV), presenting a critical global health challenge. As a member of the Flaviviridae family, HCV can cause severe liver diseases such as cirrhosis, acute hepatitis, and chronic hepatitis. The Hepatitis C virus (HCV) genome encodes a single polyprotein consisting of 3010 amino acids, which when processed contains 10 polypeptides derived from cellular and viral proteases. These include structural proteins such as core protein, E1 and E2 envelope glycoproteins, and nonstructural proteins such as NS1, NS2, NS3, NS4A, NS4B, NS5A, and NS5B. Nonstructural proteins will be released by HCV NS2-3 and NS3-4A proteases, however, structural proteins will be released by host ER signaling peptidases. co-translationally and post-translationally form 10 individual structural proteins: 5'-C-E1-E2-p7-NS2-NS3-NS4A-NS4B-NS5A-NS5B-3'. Despite extensive research, there are significant gaps in predictive and analytical approaches to managing HCV, particularly in understanding the polyprotein structure and its implications for drug discovery. This study addresses these gaps by employing machine learning techniques to analyze HCV polyprotein using various fingerprinting methods and regression algorithms. The data was sourced from the ChEMBL database, and fingerprinting techniques such as PubChem, MACCS, and E-State were utilized. Regression algorithms, including Gradient Boosting Regression (GBR), Random Forest Regression (RFR), AdaBoost Regression (ABR), and Hist Gradient Boosting Regression (HSR), were applied. Model performance was evaluated using R² and Adjusted R² metrics, comparing default models with those enhanced by hyperparameter tuning. Feature importance analysis was conducted to identify key features influencing model performance, aiding in model simplification. The results show that although hyperparameter tuning does not significantly improve the predictive power of a model, it can provide an insight into model optimization. In particular, the default model showed higher R² and Adjusted R² values across different fingerprinting techniques compared to models with hyperparameterized features. Gradient Boosting Regression (GBR) and Random Forest Regression (RFR) consistently performed well, with GBR showing the highest R² values when using PubChem fingerprints. Although there was no significant improvement through hyperparameter tuning, this study was able to find out the features that strongly influenced the model performance by conducting a feature importance analysis. This analysis helped simplify the model and highlighted the potential of machine learning in improving the understanding of HCV polyprotein structure. This research identifies optimal regression models and fingerprinting techniques, providing a strong framework for future drug discovery efforts aimed at improving global health outcomes. The research also shows that it is important to date to advance drug discovery using machine learning.

Disaster scenes as a foundation for emergency response planning

This article introduces the notion of ‘disaster scenes’. In this work, we draw on interview data collected from residents impacted by the Jagersfontein tailings disaster in South Africa in September 2022. Our methodological aim is to reconstruct the experiences of impacted people in the format of empirically derived disaster scenes. These scenes provide a credible source for modelling scenarios because they showcase the prior, immediate, response and long-term recovery phases of disaster management. Studies that provide empirical detail on the lived experience of disasters confirm the heavy burden placed on impacted people largely originating from institutional decisions. The overarching empirical narrative from this case and other cognate cases is that the absence of state institutions is a factor that likely contributes to disaster events. Our argument is that sharing information that includes models of state default can empower potentially affected groups to demand safeguards and greater transparency in relation to the operation and regulation of tailings facilities.

Implications of climate and litter quality for simulations of litterbag decomposition at high latitudes

Abstract. Litter decomposition is a vital part of the carbon cycle and is thoroughly studied both in the field and with models. Although temporally and spatially limited, litterbag decomposition experiments are often used to calibrate and evaluate soil models, coupled to land models, that are intended for use on large scales. We used the microbially explicit soil decomposition model MIMICS+ to replicate two high-latitude litterbag decomposition experiments of different spatial and temporal scales. We investigated how well the model represented observed mass loss in terms of the controlling factors of climate and litter quality and their relative importance with time. In addition to default model forcing, we used measured and site-specific model-derived microclimatic variables (soil moisture and temperature), hypothesizing that this would improve model results. We found that MIMICS+ represented mass loss after 1, 3, and 6 years well across a climatic gradient of Canadian sites but had more variable results for 1-year mass loss across a climate grid in southern Norway. In terms of litter quality, the litter metabolic fraction had more influence on modeled mass loss than the carbon-to-nitrogen ratio of the litter. Using alternative microclimate sources led to up to 23 % more mass remaining and down to 22 % less mass remaining compared to the simulations using default model inputs. None of the input alternatives significantly improved results compared to using the default model setup. We discuss possible causes for our findings and suggest measures to better utilize short-term field experiments to inform microbially explicit decomposition models.