Large Discrepancies Research Articles

Evaluation of WRF-Chem-simulated meteorology and aerosols over northern India during the severe pollution episode of 2016

Abstract. We use a state-of-the-art regional chemistry transport model (WRF-Chem v4.2.1) to simulate particulate air pollution over northern India during September–November 2016. This period includes a severe air pollution episode marked by exceedingly high levels of hourly PM2.5 (particulate matter having an aerodynamic diameter ≤ 2.5 µm) during 30 October to 7 November, particularly over the wider Indo-Gangetic Plain (IGP). We provide a comprehensive evaluation of simulated seasonal meteorology (nudged by ERA5 reanalysis products) and aerosol chemistry (PM2.5 and its black carbon (BC) component) using a range of ground-based, satellite and reanalysis products, with a focus on the November 2016 haze episode. We find the daily and diurnal features in simulated surface temperature show the best agreement followed by relative humidity, with the largest discrepancies being an overestimate of night-time wind speeds (up to 1.5 m s−1) confirmed by both ground and radiosonde observations. Upper-air meteorology comparisons with radiosonde observations show excellent model skill in reproducing the vertical temperature gradient (r>0.9). We evaluate modelled PM2.5 at 20 observation sites across the IGP including eight in Delhi and compare simulated aerosol optical depth (AOD) with data from four AERONET sites. We also compare our model aerosol results with MERRA-2 reanalysis aerosol fields and MODIS satellite AOD. We find that the model captures many features of the observed aerosol distributions but tends to overestimate PM2.5 during September (by a factor of 2) due to too much dust, and underestimate peak PM2.5 during the severe episode. Delhi experiences some of the highest daily mean PM2.5 concentrations within the study region, with dominant components nitrate (∼25 %), dust (∼25 %), secondary organic aerosols (∼20 %) and ammonium (∼10 %). Modelled PM2.5 and BC spatially correlate well with MERRA-2 products across the whole domain. High AOD at 550nm across the IGP is also well predicted by the model relative to MODIS satellite (r≥0.8) and ground-based AERONET observations (r≥0.7), except during September. Overall, the model realistically captures the seasonal and spatial variations of meteorology and ambient pollution over northern India. However, the observed underestimations in pollutant concentrations likely come from a combination of underestimated emissions, too much night-time dispersion, and some missing or poorly represented aerosol chemistry processes. Nevertheless, we find the model is sufficiently accurate to be a useful tool for exploring the sources and processes that control PM2.5 levels during severe pollution episodes.

Donors with repeated blood product discards for filtration problems, clots or hemolysis: Causes and follow-up.

Sanquin donor medicine department is informed when donations or their components are rejected. This can occur isolated or frequently. It is undesirable because the donations cannot be used and there may be an underlying medical cause. Based on regional approaches, a uniform procedure was developed. Information about whole blood, plasma- plateletpheresis donations from which one or more components were rejected for filtration time (>2 h), hemolysis or clots were extracted from blood bank information system. After rejection of two successive components or donations or total ≥3 the donor is contacted. Depending on the medical history and investigation by the family doctor, the donor carrier is re-evaluated. We looked for the causes of the discarded products and performed a survey among blood services regarding polices with discarded products. One or more components from 1742 of about 2.2 million successful donations (0.08%) were rejected. The highest percentage of rejection was seen in plateletpheresis (1.5%), all for clots. No underlying medical causes were found. 24 whole blood donors were found to have sickle cell trait (SCT) and were permanently deferred. The policies for follow-up after discarded products or acceptance of SCT donors vary between the 16 blood banks. Six organizations do not follow-up donors and seven accept SCT for blood or plasma donation. Informing donors with repeated discarded products avoids the non-use of donations. Causes of repeated discarded products can be found by follow-up of donors. The results of the survey indicate a large discrepancy in policies applied worldwide.

Estimating the loss of economic predictability from aggregating firm-level production networks.

To estimate the reaction of economies to political interventions or external disturbances, input-output (IO) tables-constructed by aggregating data into industrial sectors-are extensively used. However, economic growth, robustness, and resilience crucially depend on the detailed structure of nonaggregated firm-level production networks (FPNs). Due to nonavailability of data, little is known about how much aggregated sector-based and detailed firm-level-based model predictions differ. Using a nearly complete nationwide FPN, containing 243,399 Hungarian firms with 1,104,141 supplier-buyer relations, we self-consistently compare production losses on the aggregated industry-level production network (IPN) and the granular FPN. For this, we model the propagation of shocks of the same size on both, the IPN and FPN, where the latter captures relevant heterogeneities within industries. In a COVID-19 inspired scenario, we model the shock based on detailed firm-level data during the early pandemic. We find that using IPNs instead of FPNs leads to an underestimation of economic losses of up to 37%, demonstrating a natural limitation of industry-level IO models in predicting economic outcomes. We ascribe the large discrepancy to the significant heterogeneity of firms within industries: we find that firms within one sector only sell 23.5% to and buy 19.3% from the same industries on average, emphasizing the strong limitations of industrial sectors for representing the firms they include. Similar error levels are expected when estimating economic growth, CO2 emissions, and the impact of policy interventions with industry-level IO models. Granular data are key for reasonable predictions of dynamical economic systems.

A Preliminary Approach towards Rotor Icing Modeling Using the Unsteady Vortex Lattice Method

UAV rotors are at a high risk of ice accumulation during their operations in icing conditions. Thermal ice protection systems (IPSs) are being employed as a means of protecting rotor blades from ice, yet designing the appropriate IPS with the required heating density remains a challenge. In this work, a reduced-order modeling technique based on the Unsteady Vortex Lattice Method (UVLM) is proposed as a way to predicting rotor icing and to calculate the required anti-icing heat loads. The UVLM is gaining recent popularity for aircraft and rotor modeling. This method is flexible enough to model difficult aerodynamic problems, computationally efficient compared to higher-order CFD methods and accurate enough for conceptual design problems. A previously developed implementation of the UVLM for 3D rotor aerodynamic modeling is extended to incorporate a simplified steady-state icing thermodynamic model on the stagnation line of the blade. A viscous coupling algorithm based on a modified α-method incorporates viscous data into the originally inviscid calculations of the UVLM. The algorithm also predicts the effective angle of attack at each blade radial station (r/R), which is, in turn, used to calculate the convective heat transfer for each r/R using a CFD-based correlation for airfoils. The droplet collection efficiency at the stagnation line is calculated using a popular correlation from the literature. The icing mass and heat transfer balance includes terms for evaporation, sublimation, radiation, convection, water impingement, kinetic heating, and aerodynamic heating, as well as an anti-icing heat flux. The proposed UVLM-icing coupling technique is tested by replicating the experimental results for ice accretion and anti-icing of the 4-blade rotor of the APT70 drone. Aerodynamic predictions of the UVLM for the Figure of Merit, thrust, and torque coefficients agree within 10% of the experimental measurements. For icing conditions at −5 °C, the proposed approach overestimates the required anti-icing flux by around 50%, although it sufficiently predicts the effect of aerodynamic heating on the lack of ice formation near the blade tips. At −12 °C, visualizations of ice formation at different anti-icing heating powers agree well with UVLM predictions. However, a large discrepancy was found when predicting the required anti-icing heat load. Discrepancies between the numerical and experimental data are largely owed to the unaccounted transient and 3D effects related to the icing process on the rotating blades, which have been planned for in future work.

Empirical challenges in the study of employer associations and their representativeness

AbstractThe article examines the quality and appropriateness of the data available to measure firms' affiliation to employer associations (EAs). We find large discrepancies in affiliation rates obtained from the five different data sources available for France, leading us in particular to discard tax data. Focusing on survey data, we show that asking managers about affiliation to EAs in general or affiliation to a list of specific EAs can lead to large differences in affiliation rates, highlighting the importance of the framing of survey questions. We then provide methods to estimate an aggregate firm‐level affiliation rate from surveys covering workplaces with 11 or more employees. Exploiting (i) conflicting survey responses regarding EA affiliation between distinct establishments in the same firm and (ii) survey responses for firms that report paying contributions to EAs in their financial statements, we finally estimate the shares of employers that wrongly declare being or not being affiliated to EAs, and provide a rate of affiliation corrected for such errors. The implications for econometric analysis of the high observed error rates are discussed.

Personalized Federated Learning with Layer-Wise Feature Transformation via Meta-Learning

Federated learning enables multiple clients to collaboratively learn machine learning models in a privacy-preserving manner. However, in real-world scenarios, a key challenge encountered in federated learning is the statistical heterogeneity among clients. Existing work mainly focused on a single global model shared across the clients, making it hard to generalize well to all clients due to the large discrepancy in the data distributions. To address this challenge, we propose pFedLT , a novel approach that can adapt the single global model to different data distributions. Specifically, we propose to perform a pluggable layer-wise transformation during the local update phase based on scaling and shifting operations. In particular, these operations are learned with a meta-learning strategy. By doing so, pFedLT can capture the diversity of data distribution among clients, therefore, can generalize well even when the data distributions among clients exhibit high statistical heterogeneity. We conduct extensive experiments on synthetic and real-world datasets (MNIST, Fashion_MNIST, CIFAR-10, and Office+Caltech10) under different Non-IID settings. Experimental results demonstrate that pFedLT significantly improves the model accuracy by up to 11.67% and reduces the communication costs compared with state-of-the-art approaches.

Distributional Learning for Network Alignment with Global Constraints

Network alignment, pairing corresponding nodes across the source and target networks, plays an important role in many data mining tasks. Extensive studies focus on learning node embeddings across different networks in a unified space. However, these methods have not taken the large structural discrepancy between aligned nodes into account and, thus, are largely confined by the deterministic representations of nodes. In this work, we propose a novel network alignment framework highlighted by distributional learning and globally optimal alignment. By modeling the uncertainty of each node by Gaussian distribution, our framework builds similarity matrices on the Wasserstein distance between distributions and applies Sinkhorn operation, which learns the globally optimal mapping in an end-to-end fashion. We show that each integrated part of the framework contributes to the overall performance. Under a variety of experimental settings, our alignment framework shows superior accuracy and efficiency to the state-of-the-art.

Effect of increased ocean resolution on model errors in El Niño–Southern Oscillation and its teleconnections

AbstractDespite the complexity of the underlying processes, coupled climate models simulate fairly realistic El Niño–Southern Oscillation dynamics and teleconnections. However, there are many long‐standing errors that remain. We use the High Resolution Model Intercomparison Project suite of models to assess how El Niño–Southern Oscillation and its late‐winter teleconnection to the North Pacific changes when ocean resolution is quadrupled from 1° to 0.25°. We find that increased resolution eliminates large errors in the western extent of El Niño and La Niña sea‐surface temperature anomalies, and that there is some improvement in the asymmetry between El Niño and La Niña. In low‐resolution models, the teleconnections from El Niño and La Niña to the North Pacific are both underestimated and are centred too far west. With increased resolution, the position of the teleconnection is highly accurate during El Niño, but there is less improvement during La Niña. We find no significant improvements in teleconnection strength for either phase. Tropical mean‐state sea‐surface temperatures are found to be too cold by around 1°C throughout the central/eastern Pacific in low‐resolution models, but this bias is not present in high‐resolution models. Despite this, a large discrepancy between observed and modelled mean‐state tropical rainfall persists in high‐resolution models, which may limit improvements in the simulation of the La Niña teleconnection.

Costs of molecular adaptation to the chemical exposome: a focus on xenobiotic metabolism pathways.

Organisms adapt to their environment through different pathways. In vertebrates, xenobiotics are detected, metabolized and eliminated through the inducible xenobiotic-metabolizing pathways (XMP) which can also generate reactive toxic intermediates. In this review, we will discuss the impacts of the chemical exposome complexity on the balance between detoxication and side effects. There is a large discrepancy between the limited number of proteins involved in these pathways (few dozens) and the diversity and complexity of the chemical exposome (tens of thousands of chemicals). Several XMP proteins have a low specificity which allows them to bind and/or metabolize a large number of chemicals. This leads to undesired consequences, such as cross-inhibition, inefficient metabolism, release of toxic intermediates, etc. Furthermore, several XMP proteins have endogenous functions that may be disrupted upon exposure to exogenous chemicals. The gut microbiome produces a very large number of metabolites that enter the body and are part of the chemical exposome. It can metabolize xenobiotics and either eliminate them or lead to toxic derivatives. The complex interactions between chemicals of different origins will be illustrated by the diverse roles of the aryl hydrocarbon receptor which binds and transduces the signals of a large number of xenobiotics, microbiome metabolites, dietary chemicals and endogenous compounds. This article is part of the theme issue 'Endocrine responses to environmental variation: conceptual approaches and recent developments'.

Numerical Investigation of Thermo-Flow Characteristics of Tubes with Transverse Micro-Fins

The article presents the results of numerical studies of heat transfer and pressure drops in a channel with transverse micro-fins. The main aim of the study was to prepare the thermal and flow characteristics of such a channel for a variable longitudinal spacing of micro-fins. For the tested pipe with an internal diameter of D = 12 mm, the absolute height of the micro-fins was e = 0.243 mm, which is 2% of its diameter. The tests were carried out for turbulent flow in the range of Reynolds numbers of 5000–250,000 with the variable spacing of micro-ribs in the range of L = 0.28–13.52 mm, which corresponds to their dimensionless longitudinal distance, L/D = 0.023–1.126. For the studied geometries, the characteristics of the friction factor, ft(Re), and the Nusselt number, Nu(Re), are shown in the graphs. The highest values of Nu were observed for a spacing of L/D = 0.092 in the range of Re = 5000–60,000, while the lowest were observed for a geometry of L/D = 0.035 for Re = 60,000–250,000. The friction factors, however, were the highest for the two geometries L/D = 0.161 and L/D = 0.229 over the entire range of the tested Re numbers. A large discrepancy was observed between the friction factors calculated from the Colebrook–White equation (for irregular relative roughness depicted in the Moody diagram) and those obtained from simulations (for pipes with the same roughness height but regular geometry created by micro-fins). An analysis of the heat transfer efficiency of the tested geometries was also presented, taking into account the criterion of equal pumping power, i.e., the PEC (performance evaluation criteria) coefficient. The highest values of the PEC coefficient, up to 1.25–1.28, were obtained for micro-fin spacings of L/D = 0.069 and L/D = 0.092 in the Re number range of 20.000–30.000.

FAT: Frequency-Aware Transformation for Bridging Full-Precision and Low-Precision Deep Representations.

Learning low-bitwidth convolutional neural networks (CNNs) is challenging because performance may drop significantly after quantization. Prior arts often quantize the network weights by carefully tuning hyperparameters such as nonuniform stepsize and layerwise bitwidths, which are complicated since the full- and low-precision representations have large discrepancies. This work presents a novel quantization pipeline, named frequency-aware transformation (FAT), that features important benefits: 1) instead of designing complicated quantizers, FAT learns to transform network weights in the frequency domain to remove redundant information before quantization, making them amenable to training in low bitwidth with simple quantizers; 2) FAT readily embeds CNNs in low bitwidths using standard quantizers without tedious hyperparameter tuning and theoretical analyses show that FAT minimizes the quantization errors in both uniform and nonuniform quantizations; and 3) FAT can be easily plugged into various CNN architectures. Using FAT with a simple uniform/logarithmic quantizer can achieve the state-of-the-art performance in different bitwidths on various model architectures. Consequently, FAT serves to provide a novel frequency-based perspective for model quantization.

Burden of Stroke in Europe: An Analysis of the Global Burden of Disease Study Findings From 2010 to 2019.

While most European Regions perform well in global comparisons, large discrepancies within stroke epidemiological parameters exist across Europe. The objective of this analysis was to evaluate the stroke burden across European regions and countries in 2019 and its difference to 2010. The GBD 2019 analytical tools were used to evaluate regional and country-specific estimates of incidence, prevalence, deaths, and disability-adjusted life years of stroke for the European Region as defined by the World Health Organization, with its 53 member countries (EU-53) and for European Union as defined in 2019, with its 28 member countries (EU-28), between 2010 and 2019. Results were analyzed at a regional, subregional, and country level. In EU-53, the absolute number of incident and prevalent strokes increased by 2% (uncertainty interval [UI], 0%-4%), from 1 767 280 to 1 802 559 new cases, and by 4% (UI, 3%-5%) between 2010 and 2019, respectively. In EU-28, the absolute number of prevalent strokes and stroke-related deaths increased by 4% (UI, 2%-5%) and by 6% (UI, 1%-10%), respectively. All-stroke age-standardized mortality rates, however, decreased by 18% (UI, -22% to -14%), from 82 to 67 per 100 000 people in the EU-53, and by 15% (UI, -18% to -11%), from 49.3 to 42.0 per 100 000 people in EU-28. Despite most countries presenting reductions in age-adjusted incidence, prevalence, mortality, and disability-adjusted life year rates, these rates remained 1.4×, 1.2×, 1.6×, and 1.7× higher in EU-53 in comparison to the EU-28. EU-53 showed a 2% increase in incident strokes, while they remained stable in EU-28. Age-standardized rates were consistently lower for all-stroke burden parameters in EU-28 in comparison to EU-53, and huge discrepancies in incidence, prevalence, mortality, and disability-adjusted life-year rates were observed between individual countries.

Image Restoration via Frequency Selection.

Image restoration aims to reconstruct the latent sharp image from its corrupted counterpart. Besides dealing with this long-standing task in the spatial domain, a few approaches seek solutions in the frequency domain by considering the large discrepancy between spectra of sharp/degraded image pairs. However, these algorithms commonly utilize transformation tools, e.g., wavelet transform, to split features into several frequency parts, which is not flexible enough to select the most informative frequency component to recover. In this paper, we exploit a multi-branch and content-aware module to decompose features into separate frequency subbands dynamically and locally, and then accentuate the useful ones via channel-wise attention weights. In addition, to handle large-scale degradation blurs, we propose an extremely simple decoupling and modulation module to enlarge the receptive field via global and window-based average pooling. Furthermore, we merge the paradigm of multi-stage networks into a single U-shaped network to pursue multi-scale receptive fields and improve efficiency. Finally, integrating the above designs into a convolutional backbone, the proposed Frequency Selection Network (FSNet) performs favorably against state-of-the-art algorithms on 20 different benchmark datasets for 6 representative image restoration tasks, including single-image defocus deblurring, image dehazing, image motion deblurring, image desnowing, image deraining, and image denoising.

Uncertainty of Atmospheric Winds in Three Widely Used Global Reanalysis Datasets

Abstract Atmospheric winds are crucial to the transport of heat, moisture, momentum, and chemical species, facilitating Earth’s climate system interactions. Existing weather and climate studies rely heavily on the wind fields from reanalysis datasets. In this study, we analyze the uncertainty of instantaneous atmospheric winds in three reanalysis (ERA5, MERRA-2, and CFSv2) datasets. We show that the mean wind vector differences (WVDs) between the reanalysis datasets are about 3–6 m s−1 in the troposphere. The mean absolute wind direction differences can be more than 50°. Large WVDs greater than 5 m s−1 are found for 30%–50% of the time when the observed precipitation rate is larger than 0.1 mm h−1 over the eastern Pacific Ocean, Indian Ocean, Atlantic Ocean, and some mountain areas. The mean WVDs exhibit seasonal variations but no significant diurnal variations. The uncertainty of vertical wind shear has a correlation of 0.59 with the uncertainty of winds at 300 hPa. The magnitudes of vorticity and horizontal divergence uncertainties are on the order of 1 × 10−5 s−1, which is comparable to the mean values of vorticity and horizontal divergence. In comparison with some limited observations from field campaigns, the reanalysis datasets exhibit a mean WVD ranging from 2 to 4.5 m s−1. Among the three reanalysis datasets, ERA5 shows the closest agreement with the observations while MERRA-2 has the largest discrepancy. The substantial uncertainty and errors of the reanalysis wind products highlight the critical need for new satellite missions dedicated to 3D wind measurements.

Labour shortages amidst unmet demand for decent work

AbstractChapter 3 provides an assessment of the growing imbalances between the demand for and supply of labour. It documents large discrepancies between labour demand and available supply across advanced economies and key sectors that are slow to resolve and discusses both short‐term and secular factors that affect these imbalances. The chapter suggests a certain number of policy options, including comprehensive skills partnerships between countries to alleviate the global imbalances that exist between countries.

Testing CORDEX GCMs for Projecting Rainfall in Amhara, Ethiopia

Five CORDEX Global Circulation Models (GCMs): ICHEC-EC-EARTH, MIROC5, HadGEM2-ES, MPI-ESM-LR, and NorESM1-M were tested and validated for projecting rainfall in the Amhara regional state of Ethiopia. The GCMs were evaluated in terms of their performance during the historical period 1981-2020 and of the near-term, mid-term, and long-term future periods in three Representative Concentration Pathway (RCP) scenarios, RCP2.6, RCP4.5, and RCP8.5, across 71 grid points. Monthly observed rainfall data was used to compare the GCMs' performance and correct their biases using three non-parametric quartile mapping methods: robust empirical quartiles, empirical quartiles, and smoothing splines. The results show that HadGEM2-ES and MPI-ESM-LR had the best performance in the study area. The test and validation results for these two GCMs have come up with r = 0.8, NSE = 0.5-0.6, and RMSE = 64-70 mm/month. As there was a large discrepancy in historical and projected CORDEX rainfall data, bias correction was necessary, and the robust empirical quartiles method was found the best for the Amhara region. Compared to the historical, there will be a decrease in the monthly rainfall amount for the months of March, May, June, July, and October and an increase for the rest in all projected scenarios. The result concluded that using an ensemble of HadGEM2-ES & MPI-ESM-LR GCMs would better simulate the rainfall in the Amhara region

Sensitivity Assessment of Building Energy Performance Simulations Using MARS Meta-Modeling in Combination with Sobol’ Method

Large discrepancies can occur between building energy performance simulation (BEPS) outputs and reference data. Uncertainty and sensitivity analyses are performed to discover the significant contributions of each input parameter to these discrepancies. Variance-based sensitivity analyses typically require many stochastic simulations, which is computationally demanding (especially in the case of the large number of input parameters involved in the analysis). To overcome these impediments, this study proposes a reliable meta-model-based sensitivity analysis, including validation, Morris’ method, multivariate adaptive regression splines (MARS) meta-modeling, and Sobol’ method, to identify the most influential input parameters on BEPS prediction (annual energy consumption) at the early building design process. A hypothetical building is used to analyze the proposed methodology. Six statistical metrics are applied to verify and quantify the accuracy of the model. It is concluded that the cooling set-point temperature and g-value of the window are the most influential input parameters for the analyzed case study.

Independent Monte Carlo dose calculation identifies single isocenter multi-target radiosurgery targets most likely to fail pre-treatment measurement.

For individual targets of single isocenter multi-target (SIMT) Stereotactic radiosurgery (SRS), we assess dose difference between the treatment planning system (TPS) and independent Monte Carlo (MC), and demonstrate persistence into the pre-treatment Quality Assurance (QA) measurement. Treatment plans from 31 SIMT SRS patients were recalculated in a series of scenarios designed to investigate sources of discrepancy between TPS and independent MC. Targets with>5% discrepancy in DMean[Gy] after progressing through all scenarios were measured with SRS MapCHECK. A matched pair analysis was performed comparing SRS MapCHECK results for these targets with matched targets having similar characteristics (volume & distance from isocenter) but no such MC dose discrepancy. Of 217 targets analyzed, individual target mean dose (DMean[Gy]) fell outside a 5% threshold for 28 and 24 targets before and after removing tissue heterogeneity effects, respectively, while only 5 exceeded the threshold after removing effect of patient geometry (via calculation on StereoPHAN geometry). Significant factors affecting agreement between the TPS and MC included target distance from isocenter (0.83% decrease in DMean[Gy] per 2cm), volume (0.15% increase per cc), and degree of plan modulation (0.37% increase per 0.01 increase in modulation complexity score). SRS MapCHECK measurement had better agreement with MC than with TPS (2%/1mm / 10% threshold gamma pass rate (GPR)=99.4±1.9%vs. 93.1±13.9%, respectively). In the matched pair analysis, targets exceeding 5% for MC versus TPS also had larger discrepancies between TPS and measurement with no GPR (2%/1mm / 10% threshold) exceeding 90% (71.5%±16.1%); whereas GPR was high for matched targets with no such MC versus TPS difference (96.5%±3.3%, p=0.01). Independent MC complements pre-treatment QA measurement for SIMT SRS by identifying problematic individual targets prior to pre-treatment measurement, thus enabling plan modifications earlier in the planning process and guiding selection of targets for pre-treatment QA measurement.

Explicit ion modeling predicts physicochemical interactions for chromatin organization

Molecular mechanisms that dictate chromatin organization in vivo are under active investigation, and the extent to which intrinsic interactions contribute to this process remains debatable. A central quantity for evaluating their contribution is the strength of nucleosome-nucleosome binding, which previous experiments have estimated to range from 2 to 14 kBT. We introduce an explicit ion model to dramatically enhance the accuracy of residue-level coarse-grained modeling approaches across a wide range of ionic concentrations. This model allows for de novo predictions of chromatin organization and remains computationally efficient, enabling large-scale conformational sampling for free energy calculations. It reproduces the energetics of protein-DNA binding and unwinding of single nucleosomal DNA, and resolves the differential impact of mono- and divalent ions on chromatin conformations. Moreover, we showed that the model can reconcile various experiments on quantifying nucleosomal interactions, providing an explanation for the large discrepancy between existing estimations. We predict the interaction strength at physiological conditions to be 9 kBT, a value that is nonetheless sensitive to DNA linker length and the presence of linker histones. Our study strongly supports the contribution of physicochemical interactions to the phase behavior of chromatin aggregates and chromatin organization inside the nucleus.

Explicit ion modeling predicts physicochemical interactions for chromatin organization.

Molecular mechanisms that dictate chromatin organization in vivo are under active investigation, and the extent to which intrinsic interactions contribute to this process remains debatable. A central quantity for evaluating their contribution is the strength of nucleosome-nucleosome binding, which previous experiments have estimated to range from 2 to 14 kBT. We introduce an explicit ion model to dramatically enhance the accuracy of residue-level coarse-grained modeling approaches across a wide range of ionic concentrations. This model allows for de novo predictions of chromatin organization and remains computationally efficient, enabling large-scale conformational sampling for free energy calculations. It reproduces the energetics of protein-DNA binding and unwinding of single nucleosomal DNA, and resolves the differential impact of mono- and divalent ions on chromatin conformations. Moreover, we showed that the model can reconcile various experiments on quantifying nucleosomal interactions, providing an explanation for the large discrepancy between existing estimations. We predict the interaction strength at physiological conditions to be 9 kBT, a value that is nonetheless sensitive to DNA linker length and the presence of linker histones. Our study strongly supports the contribution of physicochemical interactions to the phase behavior of chromatin aggregates and chromatin organization inside the nucleus.