Model Run Research Articles

A Lightweight Convolutional Neural Network Based on Hierarchical-Wise Convolution Fusion for Remote-Sensing Scene Image Classification

The large intra-class difference and inter-class similarity of scene images bring great challenges to the research of remote-sensing scene image classification. In recent years, many remote-sensing scene classification methods based on convolutional neural networks have been proposed. In order to improve the classification performance, many studies increase the width and depth of convolutional neural network to extract richer features, which increases the complexity of the model and reduces the running speed of the model. In order to solve this problem, a lightweight convolutional neural network based on hierarchical-wise convolution fusion (LCNN-HWCF) is proposed for remote-sensing scene image classification. Firstly, in the shallow layer of the neural network (groups 1–3), the proposed lightweight dimension-wise convolution (DWC) is utilized to extract the shallow features of remote-sensing images. Dimension-wise convolution is carried out in the three dimensions of width, depth and channel, and then, the convoluted features of the three dimensions are fused. Compared with traditional convolution, dimension-wise convolution has a lower number of parameters and computations. In the deep layer of the neural network (groups 4–7), the running speed of the network usually decreases due to the increase in the number of filters. Therefore, the hierarchical-wise convolution fusion module is designed to extract the deep features of remote-sensing images. Finally, the global average pooling layer, the fully connected layer and the Softmax function are used for classification. Using global average pooling before the fully connected layer can better preserve the spatial information of features. The proposed method achieves good classification results on UCM, RSSCN7, AID and NWPU datasets. The classification accuracy of the proposed LCNN-HWCF on the AID dataset (training:test = 2:8) and the NWPU dataset (training:test = 1:9), with great classification difficulty, reaches 95.76% and 94.53%, respectively. A series of experimental results show that compared with some state-of-the-art classification methods, the proposed method not only greatly reduces the number of network parameters but also ensures the classification accuracy and achieves a good trade-off between the model classification accuracy and running speed.

Assessment of the Madrid region air quality zoning based on mesoscale modelling and k-means clustering

According to the Air Quality (AQ) Directive (2008/50/EC), European Member States should establish (and revise every five years) their own AQ zones, where the air quality and population exposure are homogeneous. In Madrid, there are currently seven AQ zones, which were determined based on administrative geographic and land use criteria in 2014. However, so far, there has been no standardized methodology to define an objective AQ zoning. In this study, a new methodology is applied to revise the AQ zones of the Madrid region by using the WRF-CMAQ modelling system with 1 km2 spatial resolution. All the relevant legal indicators of the main pollutants (NO2, O3, PM10 and PM2.5) were computed from an annual 1-h temporal resolution model run and aggregated at municipality level. Then, seven basic statistics (mean, interquartile range, etc.) are computed for each air quality indicator within each of the 179 municipalities of the Madrid region. A Principal Components Analysis (PCA) is applied to identify the most relevant clustering variables from all these statistics to subsequently apply a k-mean cluster analysis. The definition of the number air quality zones (clusters) is based on three methods: Elbow, Silhouette, and Gap statistic. Following this methodology, five zones for NO2, PM10, and PM2.5 and four zones for O3 are proposed. To assess the resulting zoning and to compare it with the current one, concentration distributions in each zone are visualized through boxplots. In addition, in order to confirm significant differences among the zones of both zonings, they are examined by two statistical tests: the Kruskal-Wallis and Dunn tests. Finally, the coverage and potential redundancy areas of the 47 existing air quality monitoring stations in the region are analysed for the two alternatives, confirming the suitability of the new air quality zoning proposed.

Understanding systemic disruption from the Covid-19-induced semiconductor shortage for the auto industry

Covid-19 has allowed us to study systemic disruptions that impact entire industries. This paper explores how disruptions start, propagate, and continue over time by examining the semiconductor chip shortage faced by the auto industry during the years following Covid-19 in 2020. First, we carried out a thematic analysis of 209 pertinent newspaper articles. The analysis resulted in a thematic model of such disruptions with the interplay of various factors leading to the prolonged disruption to the auto sector. Second, we present the results from a stylized supply chain planning model run at different times to show how disruptions propagate to the auto and other sectors, causing systemic shortages. Overall, we contribute to the supply chain risk literature by focusing on system disruptions impacting entire industries versus normal disruptions affecting a particular company’s supply chain.

Comparison of East Asian Summer Monsoon Simulation between an Atmospheric Model and a Coupled Model: An Example from CAS-ESM

In this study, the Chinese Academy of Sciences’ Earth System Model Version 2 (CAS-ESM2) and its atmospheric component were evaluated for the ability to simulate the East Asian summer monsoon (EASM), in terms of climatology and composites in El Niño decaying years (EN) and La Niña years (LN). The results show that the model can realistically simulate the El Niño Southern Oscillation (ENSO) annual cycle, the interannual variation, the evolution process, and the prerequisites of ENSO, but the trend of developing and decaying is faster than that of the observations. With regard to the climatological mean state in the EASM, the coupled model run can largely improve the precipitation and 850 hPa wind simulated in the atmospheric model. Moreover, the coupled run can also reduce the mid-latitude bias in the atmospheric model simulation. Composite methods were then adopted to examine performance in different phases of the ENSO, from a mature winter to a decaying summer. The atmospheric model can well reproduce the Western North Pacific Anomalous Anticyclone (WNPAC)/Western North Pacific Anomalous Cyclone (WNPC) during EN/LN well, but the westerly/easterly anomalies and the associated precipitation anomalies over the equatorial Central Eastern Pacific are somewhat overestimated. Compared with the atmospheric model, these anomalies are all underestimated in the coupled model, which may be related to the ENSO-related SST bias appearing in the Eastern Indian Ocean. Due to the ENSO and ITCZ bias in the historical simulations, the simulated ENSO-related SST and the precipitation anomaly are too equator-trapped in comparison with the observations, and the cold tongue overly extends westward. This limits the ability of the model to simulate ENSO-related EASM variability. For the subseasonal simulations, though atmospheric model simulations can reproduce the westward extension of the Western Pacific subtropic high (WPSH) in EN decaying summers, the eastward retreat of the WPSH in LN is weak. The historical simulations show limited improvement, indicating that the subseasonal variation in the EASM is still a considerable challenge for current generation models.

Peridynamics and surrogate modeling of pressure-driven well stimulation

In this work we use the peridynamics theory of solid mechanics to simulate fracture in an annular rock domain subject to an in-situ stress and create surrogate models that predict the area of the resulting cracks. Peridynamics is a non-local formulation of continuum mechanics that naturally accommodates material discontinuities . Furthermore, unlike other fracture modeling techniques there is no need to provide information about the crack path. We utilize the peridynamics code Peridigm and take a two-stage approach to fracture modeling. First an implicit solve is performed to compute the in-situ stress state. We then execute an explicit solve where a pressure loading designed to emulate fluid-driven hydraulic fracture is applied at the borehole and transmitted to the pre-stressed rock. We present results from polynomial and single and multi-level Gaussian process surrogate models constructed from a sampling study of the peridynamics model . The surrogates predict crack area given a measure of the in-situ stress anisotropy and rise time and amplitude of the pressure loading. These surrogates take a minuscule fraction of peridynamics model's running time to evaluate and are a step towards enabling advanced optimization and uncertainty quantification workflows that require many model evaluations.

Non-Hydrostatic Regcm4 (Regcm4-NH): Evaluation of Precipitation Statistics at the Convection-Permitting Scale over Different Domains

Recent studies over different geographical regions of the world have proven that regional climate models at the convection-permitting scale (CPMs) improve the simulation of precipitation in many aspects, such as the diurnal cycle, precipitation frequency, intensity, and extremes at daily—but even more at hourly—time scales. Here, we present an evaluation of climate simulations with the newly developed RegCM4-NH model run at the convection-permitting scale (CP-RegCM4-NH) for a decade-long period, over three domains covering a large European area. The simulations use a horizontal grid spacing of ~3 km and are driven by the ERA-Interim reanalysis through an intermediate driving RegCM4-NH simulation at ~12 km grid spacing with parameterized deep convection. The km-scale simulations are evaluated against a suite of hourly observation datasets with high spatial resolutions and are compared to the coarse-resolution driving simulation in order to assess improvements in precipitation from the seasonal to hourly scale. The results show that CP-RegCM4-NH produces a more realistic representation of precipitation than the coarse-resolution simulation over all domains. The most significant improvements were found for intensity, heavy precipitation, and precipitation frequency, both on daily and hourly time scales in all seasons. In general, CP-RegCM4-NH tends to correctly produce more intense precipitation and to reduce the frequency of events compared to the coarse-resolution one. On the daily scale, improvements in CP simulations are highly region dependent, with the best results over Italy, France, and Germany, and the largest biases over Switzerland, the Carpathians, and Greece, especially during the summer seasons. At the hourly scale, the improvement in CP simulations for precipitation intensity and spatial distribution is clearer than at the daily timescale. In addition, the representation of extreme events is clearly improved by CP-RegCM4-NH, particularly at the hourly time scale, although an overestimation over some subregions can be found. Although biases between the model simulations at the km-scale and observations still exist, this first application of CP-RegCM4-NH at high spatial resolution indicates a clear benefit of convection-permitting simulations and encourages further assessments of the added value of km-scale model configurations for regional climate change projections.

The Feasibility of Leaf Reflectance-Based Taxonomic Inventories and Diversity Assessments of Species-Rich Grasslands: A Cross-Seasonal Evaluation Using Waveband Selection

Hyperspectral leaf-level reflectance data may enable the creation of taxonomic inventories and diversity assessments of grasslands, but little is known about the stability of species-specific spectral classes and discrimination models over the course of a growing season. Here, we present a cross-seasonal dataset of seventeen species that are common to a temperate, dry and nutrient-poor calcareous grassland, which spans thirteen sampling dates, a week apart, during the spring and summer months. By using a classification model that incorporated waveband selection (a sparse partial least squares discriminant analysis), most species could be classified, irrespective of the sampling date. However, between 42 and 95% of the available spectral information was required to obtain these results, depending on the date and model run. Feature selection was consistent across time for 70 out of 720 wavebands and reflectance around 1410 nm, representing water features, contributed the most to the discrimination. Model transferability was higher between neighbouring sampling dates and improved after the “green-up” period. Some species were consistently easy to classify, irrespective of time point, when using up to six latent variables, which represented about 99% of the total spectral variance, whereas other species required many latent variables, which represented very small spectral differences. We concluded that it did seem possible to create reliable taxonomic inventories for combinations of certain grassland species, irrespective of sampling date, and that the reason for this could lie in their distinctive morphological and/or biochemical leaf traits. Model transferability, however, was limited across dates and cross-seasonal sampling that captures leaf development would probably be necessary to create a predictive framework for the taxonomic monitoring of grasslands. In addition, most variance in the leaf reflectance within this system was driven by a subset of species and this finding implies challenges for the application of spectral variance in the estimation of biodiversity.

A derivative-free optimisation method for global ocean biogeochemical models

Abstract. The skill of global ocean biogeochemical models, and the earth system models in which they are embedded, can be improved by systematic calibration of the parameter values against observations. However, such tuning is seldom undertaken as these models are computationally very expensive. Here we investigate the performance of DFO-LS, a local, derivative-free optimisation algorithm which has been designed for computationally expensive models with irregular model–data misfit landscapes typical of biogeochemical models. We use DFO-LS to calibrate six parameters of a relatively complex global ocean biogeochemical model (MOPS) against synthetic dissolved oxygen, phosphate and nitrate “observations” from a reference run of the same model with a known parameter configuration. The performance of DFO-LS is compared with that of CMA-ES, another derivative-free algorithm that was applied in a previous study to the same model in one of the first successful attempts at calibrating a global model of this complexity. We find that DFO-LS successfully recovers five of the six parameters in approximately 40 evaluations of the misfit function (each one requiring a 3000-year run of MOPS to equilibrium), while CMA-ES needs over 1200 evaluations. Moreover, DFO-LS reached a “baseline” misfit, defined by observational noise, in just 11–14 evaluations, whereas CMA-ES required approximately 340 evaluations. We also find that the performance of DFO-LS is not significantly affected by observational sparsity, however fewer parameters were successfully optimised in the presence of observational uncertainty. The results presented here suggest that DFO-LS is sufficiently inexpensive and robust to apply to the calibration of complex, global ocean biogeochemical models.

Decomposing oceanic temperature and salinity change using ocean carbon change

Abstract. As the planet warms due to the accumulation of anthropogenic CO2 in the atmosphere, the interaction of surface ocean carbonate chemistry and the radiative forcing of atmospheric CO2 leads to the global ocean sequestering heat and carbon in a ratio that is nearly constant in time. This ratio has been approximated as globally uniform, enabling the intimately linked patterns of ocean heat and carbon uptake to be derived. Patterns of ocean salinity also change as the Earth system warms due to hydrological cycle intensification and perturbations to air–sea freshwater fluxes. Local temperature and salinity change in the ocean may result from perturbed air–sea fluxes of heat and fresh water (excess temperature, salinity) or from reorganisation of the preindustrial temperature and salinity fields (redistributed temperature, salinity), which are largely due to circulation changes. Here, we present a novel method in which the redistribution of preindustrial carbon is diagnosed and the redistribution of temperature and salinity is estimated using only local spatial information. We demonstrate this technique in the NEMO ocean general circulation model (OGCM) coupled to the MEDUSA-2 biogeochemistry model under an RCP8.5 scenario over 1860–2099. The excess changes (difference between total and redistributed property changes) are thus calculated. We demonstrate that a global ratio between excess heat and temperature is largely appropriate regionally with key regional differences consistent with reduced efficiency in the transport of carbon through the mixed layer base at high latitudes. On centennial timescales, excess heat increases everywhere, with the North Atlantic being a key site of excess heat uptake over the 21st century, accounting for 25 % of the total. Excess salinity meanwhile increases in the Atlantic but is generally negative in other basins, consistent with increasing atmospheric transport of fresh water out of the Atlantic. In the North Atlantic, changes in the inventory of excess salinity are detectable in the late 19th century, whereas increases in the inventory of excess heat do not become significant until the early 21st century. This is consistent with previous studies which find salinification of the subtropical North Atlantic to be an early fingerprint of anthropogenic climate change. Over the full simulation, we also find the imprint of Atlantic meridional overturning circulation (AMOC) slowdown through significant redistribution of heat away from the North Atlantic and of salinity to the South Atlantic. Globally, temperature change at 2000 m is accounted for by both redistributed and excess heat, but for salinity the excess component accounts for the majority of changes at the surface and at depth. This indicates that the circulation variability contributes significantly less to changes in ocean salinity than to heat content. By the end of the simulation excess heat is the largest contribution to density change and steric sea level rise, while excess salinity greatly reduces spatial variability in steric sea level rise through density compensation of excess temperature patterns, particularly in the Atlantic. In the Atlantic, redistribution of the preindustrial heat and salinity fields also produces generally compensating changes in sea level, though this compensation is less clear elsewhere. The regional strength of excess heat and salinity signals grows through the model run in response to the evolving forcing. In addition, the regional strength of the redistributed temperature and salinity signals also grows, indicating increasing circulation variability or systematic circulation change on timescales of at least the model run.

Prediction of fluvial erosion rate in Jamuna River, Bangladesh

ABSTRACT River bank erosion of alluvial rivers is one of the major challenges for river management. This paper deals with the prediction of the fluvial erosion rate along the left bank of Jamuna River in Bangladesh using the excess shear stress method. Surface water modelling systems i.e. SMS and SRH-2D, were used to estimate the hydrodynamic effects on the riverbank along the 8.21 km reach of the river. Riverbank along the selected reach of Jamuna River consists of two layers of soil type. The upper layer soil is sandy silt with a median size of D 50 = 0.028 mm and the bottom layer is silty sand with D 50 = 0.167 mm. Results of the present analysis show that shear stress along the riverbank attains its maximum and initiates the bank erosion, when the flow rate is about 45,000 m3/s. Based on the results obtained from model run and empirical analysis, riverbank materials have been categorized as ‘erodible’ to ‘very erodible’, as far as erodibility parameters are concerned. Critical shear stress of bank soil is found to be varied between 0.15 and 0.22 Pa using erodibility co-efficient ranges between 5.05 and 6.03 cm3/N s. From this study, the maximum bank erosion rates have been estimated as 51.95 m/year to 69.82 m/year. However, average erosion rates have been estimated which range between 38.65 and 40.57 m/y. It is hoped that the results obtained from the study will be helpful in determining the riverbank stability for the implementation of appropriate river protective measures along the riverbank of Jamuna River.

Global sensitivity analysis in hydrodynamic modeling and flood inundation mapping

Flood simulation with two-dimensional hydrodynamic models is subject to different sources of uncertainties in model configuration, boundary conditions, and model parametrization. HEC-RAS 2D is a widely used hydrodynamic model. Here, we assess the sensitivity of the HEC-RAS 2D to its configuration factors and parameters. We evaluate the impacts of different model configuration factors, including the floodplain and channel roughness coefficients, terrain and mesh size, as well as river boundary conditions on the dynamic of water levels, maximum water level, and flood extent, and determine the importance of these contributing factors for reliable flood inundation modeling using both variogram and variance-based analyses. For our case study, we found that depending on the perturbation scale, the performance measure, and the predicted output, the sensitivity to input factors changes. While for simulation of water level dynamics, the DEM and mesh resolutions are the most important factors, for flood extent mapping the floodplain roughness coefficient and the upstream boundary condition are the key factors. In addition, our analyses indicated that for a reliable simulation of maximum flood water level, modelers need to spend resources on the calibration of floodplain roughness coefficient while using fine DEM and mesh resolutions. Here, we also investigate the role of computational time interval and mesh resolution on the model's run time and accuracy. Our results suggested that reducing the computational time interval has minimum impact, as it increases the model run time without much improving the model accuracy.

Metamodel-based generative design of wind turbine foundations

Wind turbines play an integral role in energy transition agendas. The optimized design of wind turbine foundations is a complex and intricate task that requires iterative running of computationally-intensive and time-consuming finite element models. However, given the popularity of these structures over the past two decades, there is a wealth of data from the designs of the past projects that can be used for the data-driven modeling of these structures. Given the demonstrated accuracy and success of metamodels as an alternative approach for other computationally-intensive simulation-based problems, this study aims to develop a generative-design framework for the optimization of wind turbine foundations using a metamodel, as a complementary step to more accurate finite element modeling, to reduce the overall design time without compromising the accuracy. To this end, first, the random forest method is used to develop a multi-output metamodel for the wind turbine foundations based on a set of historical data. Then, a metaheuristic method, i.e., NSGA II, is adopted to optimize the design process based on the developed metamodel. In a case study, a wind turbine foundation was designed using the proposed framework and the accuracy of the output was assessed in terms of the ultimate bending moment. The results of the case study indicate that the proposed method provides a significant time gain (i.e., 99.93%) without compromising the accuracy (i.e., 1.75% for the percent error). Besides, the conducted study also offers designers a better understanding of the importance of each design variable and how certain design variables influence the moment-rotation behavior of the wind turbine foundation

Building Decision Support Systems in Excel for Production and Distribution Planning: A Case Study

We develop a decision support system in Microsoft Excel that integrates production and distribution for a manufacturer of natural fiber-based products in North America. The production and distribution of the company’s products were optimized using a linear programming model, implemented in Excel. The spreadsheet dynamically adjusts the formulation to reflect the user’s current requirements, solves the optimization model in the background, and generates detailed managerial reports. In addition, it allows users to conduct what-if analyses by varying the number of plants and warehouses. It demonstrates the ability of a Linear Programming Model run on an Excel platform to provide the firm with an optimized production plan resulting in significant, cost savings since implementation.

Control simulation experiment with Lorenz's butterfly attractor

Abstract. In numerical weather prediction (NWP), sensitivity to initial conditions brings chaotic behaviors and an intrinsic limit to predictability, but it also implies an effective control in which a small control signal grows rapidly to make a substantial difference. The Observing Systems Simulation Experiment (OSSE) is a well-known approach to study predictability, where “nature” is synthesized by an independent NWP model run. In this study, we extend the OSSE and design the control simulation experiment (CSE), where we apply a small signal to control “nature”. Idealized experiments with the Lorenz-63 three-variable system show that we can control “nature” to stay in a chosen regime without shifting to the other, i.e., in a chosen wing of Lorenz's butterfly attractor, by adding small perturbations to “nature”. Using longer-lead-time forecasts, we achieve more effective control with a perturbation size of less than only 3 % of the observation error. We anticipate our idealized CSE to be a starting point for a realistic CSE using the real-world NWP systems, toward possible future applications to reduce weather disaster risks. The CSE may be applied to other chaotic systems beyond NWP.

Reduced Order Modeling of Deformable Tire-Soil Interaction With Proper Orthogonal Decomposition

AbstractIn this study, model order reduction of high-fidelity off-road mobility models is explored to address the computational intensity of nonlinear finite element deformable tire–soil interaction models. To this end, a model order reduction procedure for the tire–soil interaction model is developed with the proper orthogonal decomposition (POD), and it is integrated into the off-road mobility simulation framework, leveraging high-performance computing. The POD is, however, limited in that the modes are dependent on snapshot data collected during the running of a full order model, limiting the modes to being accurate only for the specific scenario from which they were collected. Due to this limitation, a method of mode adaptation through interpolation on a tangent space of the Grassmann manifold is investigated to allow modes to be predicted for cases in which a full order model has not been run. It is demonstrated by several numerical examples that the POD modes are effective at retaining predictive accuracy while reducing computational time. The results show that adapted POD modes are more capable of characterizing the behavior of the model than modes produced at a different value of the simulation parameter. The POD-based reduced order modeling approach is further extended to the full vehicle simulation on deformable terrain through the co-simulation coupling algorithm by leveraging the high-performance computing technique.

Using machine learning as a surrogate model for agent-based simulations.

In this proof-of-concept work, we evaluate the performance of multiple machine-learning methods as surrogate models for use in the analysis of agent-based models (ABMs). Analysing agent-based modelling outputs can be challenging, as the relationships between input parameters can be non-linear or even chaotic even in relatively simple models, and each model run can require significant CPU time. Surrogate modelling, in which a statistical model of the ABM is constructed to facilitate detailed model analyses, has been proposed as an alternative to computationally costly Monte Carlo methods. Here we compare multiple machine-learning methods for ABM surrogate modelling in order to determine the approaches best suited as a surrogate for modelling the complex behaviour of ABMs. Our results suggest that, in most scenarios, artificial neural networks (ANNs) and gradient-boosted trees outperform Gaussian process surrogates, currently the most commonly used method for the surrogate modelling of complex computational models. ANNs produced the most accurate model replications in scenarios with high numbers of model runs, although training times were longer than the other methods. We propose that agent-based modelling would benefit from using machine-learning methods for surrogate modelling, as this can facilitate more robust sensitivity analyses for the models while also reducing CPU time consumption when calibrating and analysing the simulation.

Particle-filter-based volcanic ash emission inversion applied to a hypothetical sub-Plinian Eyjafjallajökull eruption using the Ensemble for Stochastic Integration of Atmospheric Simulations (ESIAS-chem) version 1.0

Abstract. A particle-filter-based inversion system is presented, which enables us to derive time- and altitude-resolved volcanic ash emission fluxes along with its uncertainty. The system assimilates observations of volcanic ash column mass loading as retrieved from geostationary satellites. It aims to estimate the temporally varying emission profile endowed with error margins. In addition, we analyze the dependency of our estimate on wind field characteristics, notably vertical shear, within variable observation intervals. Thus, the proposed system addresses the special challenge of analyzing the vertical profile of volcanic ash clouds given only 2D high temporal-resolution column mass loading data as retrieved by geostationary satellites. The underlying method rests on a linear combination of height–time emission finite elements of arbitrary resolution, each of which is assigned to a model run subject to ensemble-based space–time source inversion. Employing a modular concept, this setup builds the Ensemble for Stochastic Integration of Atmospheric Simulations (ESIAS-chem). It comprises a particle smoother in combination with a discrete-grid ensemble extension of the Nelder–Mead minimization method. The ensemble version of the EURopean Air pollution Dispersion – Inverse Model (EURAD-IM) is integrated into ESIAS-chem but can be replaced by other models. As initial validation of ESIAS-chem, the system is applied to simulated artificial observations of both ash-contaminated and ash-free atmospheric columns using identical-twin experiments. Thus, in this idealized initial performance test the underlying meteorological uncertainty is neglected. The inversion system is applied to two notional sub-Plinian eruptions of the Eyjafjallajökull volcano, Iceland, with strong ash emission changes with time and injection heights. It demonstrates the ability of ESIAS-chem to retrieve the volcanic ash emission fluxes from the assimilation of column mass loading data only. However, the analyzed emission profiles strongly differ in their levels of accuracy depending of the strength of wind shear conditions. While the error is only 10 %–20 % for the estimated emission fluxes under strong wind conditions, it increases up to 60 % under weak wind shear conditions. In case of increasing wind shear, the performance of the analysis may benefit from extending the assimilation window, in which new observations potentially contribute valuable information to the analysis system. For our test cases using an artificial volcanic eruption, we found an assimilation window length of 18 h, i.e., 10 h after the eruption terminated, to be sufficient for analyzing the extent and location of the artificial ash cloud. In the performed test cases, the analysis ensemble predicts the location of high volcanic ash column mass loading in the atmosphere with a very high probability of > 95 %. Additionally, the analysis ensemble is able to provide a vertically resolved probability map of high volcanic ash concentrations to a high accuracy for both high and weak wind shear conditions.

Novel intelligent adjustment height method of Shearer drum based on adaptive fuzzy reasoning Petri net

The complexity of the coalface environment determines the non-linear and fuzzy characteristics of the drum adjustment height. To overcome this challenge, this study proposes an adaptive fuzzy reasoning Petri net (AFRPN) model based on fuzzy reasoning and fuzzy Petri net (FPN) and then applies it to the intelligent adjustment height of the shearer drum. This study constructs adaptive and reasoning algorithms. The former was used to optimize the AFRPN parameters, and the latter made the AFRPN model run. AFRPN could represent rules that had non-linear and attribute mapping relationships and could adjust the parameters adaptively to improve the accuracy of the output. Subsequently, the drum adjustment height model was established and compared to three models neural network (NN), classification and regression tree(CART) and gradient boosting decision tree (GBDT). The experimental results showed that this method is superior to other drum adjustment height methods and that AFRPN can achieve intelligent adjustment of the shearer drum height by constructing fuzzy inference rules.

P021 Vitamin D receptor-dependent protective effect of moderate hypoxia in a mouse colitis model Running title: Hypoxia prevents mouse colitis through VDR

Abstract Background Although physiological hypoxia is important for maintaining the intestinal barrier, its effect on the barrier during acute colitis and the underlying mechanisms are not fully understood. Here,we aim to explore the influence of hypoxia in dextran sulfate sodium (DSS)-induced colitis mice and the role of hypoxia-inducible factor (HIF) and vitamin D receptor (VDR) in the process Methods Colitis mice were subjected to hypoxia to detect colitis symptoms and intestinal barrier function changes. Finally, the mechanisms underlying the changes were explored in a colon cell line. Results First, compared with colitis mice without hypoxia stimulation, those with hypoxia stimulation showed significantly decreased pathological damage and improved permeability of the intestinal barrier. The expression of tight junction proteins (including occludin, claudin-1 and ZO-1), HIF-1α as well as VDR was up-regulated in colitis mice with hypoxia stimulation. However, in VDR gene knockout (VDR-KO) colitis mice, hypoxia treatment showed no protective effect, suggesting the VDR dependency of this effect. Second, we identified that although hypoxia stimulation could enhance the single-layer epithelial transmembrane electrical resistance in DLD-1 cells, these effects disappeared in VDR-knockdown cells. Furthermore, over-expression of HIF-1α in DLD-1 can increase the expression of VDR, then chromatin immunoprecipitation and luciferase assays confirmed that HIF-1α can bind to the promoter region of the VDR gene under hypoxia. Finally, compared with their wild-type siblings, VDR-KO mice showed reduced abundance of anaerobic bacteria and SCFA-producing bacteria. Conclusion Moderate hypoxia was protective against DSS-induced colitis, and VDR is instrumental in it. Furthermore, HIF-1α mediates the effect of hypoxia on the VDR signaling pathway. Moreover, intestinal flora may be an important link between hypoxia and VDR.

Box model trajectory studies of contrail formation using a particle-based cloud microphysics scheme

Abstract. We investigate the microphysics of contrail formation behind commercial aircraft by means of the particle-based LCM (Lagrangian Cloud Module) box model. We extend the original LCM to cover the basic pathway of contrail formation on soot particles being activated into liquid droplets that soon after freeze into ice crystals. In our particle-based microphysical approach, simulation particles are used to represent different particle types (soot, droplets, ice crystals) and properties (mass/radius, number). The box model is applied in two frameworks. In the classical framework, we prescribe the dilution along one average trajectory in a single box model run. In the second framework, we perform a large ensemble of box model runs using 25 000 different trajectories inside an expanding exhaust jet as simulated by the LES (large-eddy simulation) model FLUDILES. In the ensemble runs, we see a strong radial dependence of the temperature and relative humidity evolution. Droplet formation on soot particles happens first near the plume edge and a few tenths of a second later in the plume centre. Averaging over the ensemble runs, the number of formed droplets and ice crystals increases more smoothly over time than for the single box model run with the average dilution. Consistent with previous studies, contrail ice crystal number varies strongly with atmospheric parameters like temperature and relative humidity near the contrail formation threshold. Close to this threshold, the apparent ice number emission index (product of freezing fraction and soot number emission index) strongly depends on the geometric-mean dry core radius and the hygroscopicity parameter of soot particles. The freezing fraction of soot particles slightly decreases with increasing soot particle number, particularly for higher soot number emissions. This weakens the increase of the apparent ice number emission index with rising soot number emission index. Comparison with box model results of a recent contrail formation study by Lewellen (2020) (using similar microphysics) shows a later onset of our contrail formation due to a weaker prescribed plume dilution. If we use the same dilution data, our evolution and Lewellen's evolution in contrail ice nucleation show an excellent agreement cross-verifying both microphysics implementations. This means that differences in contrail properties mainly result from different representations of the plume mixing and not from the microphysical modelling. Using an ensemble mean framework instead of a single trajectory does not necessarily lead to an improved scientific outcome. Contrail ice crystal numbers tend to be overestimated since the interaction between the different trajectories is not considered. The presented aerosol and microphysics scheme describing contrail formation is of intermediate complexity and thus suited to be incorporated in an LES model for 3D contrail formation studies explicitly simulating the jet expansion. Our box model results will help interpret the upcoming, more complex 3D results.