Spatiotemporal Gradients Research Articles

Gradient-boosted spatiotemporal neural network for simulating underground hydrogen storage in aquifers

Underground hydrogen storage (UHS) in aquifers has emerged as a viable solution to address the seasonal mismatch between supply and demand in renewable energy. Numerical simulation of the UHS serves as a crucial foundation for optimizing storage operations and conducting system risk assessments. However, numerical simulation methods employed for these purposes often demand substantial data, making data collection challenging and computationally expensive, especially in scenarios involving the coupling of multiple physical fields. Deep learning serves as an effective tool in resolving this challenge. Here, we proposed a spatiotemporal neural network architecture with gradient enhancement, denoted as gradient-boosted spatiotemporal neural network (GSTNN) and its variant GSTNN-s. The GSTNN combines a convolutional neural network (CNN), a long short-term memory network (LSTM), and an autoencoder architecture. To incorporate physical constraints into the network, the spatiotemporal gradient operators from the gas-water seepage and gas convection-diffusion equations are introduced as regularization terms, imposing physics-informed constraints on the training process in both temporal and spatial dimensions. In predicting the multiphase flow of UHS in both homogeneous and heterogeneous formations, GSTNN outperforms CNN and CNN-LSTM in terms of the accuracy of pressure, saturation and H2 concentration fields. In terms of predicting UHS in formations with different permeabilities and porosities, GSTNN-s demonstrates improved performance as well. The proposed GSTNN architecture is promising in improving the efficiency of UHS numerical simulation, and has great potential to be applied for optimizing UHS operations in the future.

Pollinator functional group abundance and floral heterogeneity in an agroecological context affect mating patterns in a self-incompatible wild plant.

Restoration of seminatural field margins can elevate pollinator activity. However, how they support wild plant gene flow through interactions between pollinators and spatiotemporal gradients in floral resources remains largely unknown. Using a farm-scale experiment, we tested how mating outcomes (expected heterozygosity and paternity correlation) of the wild, self-incompatible plant Cyanus segetum transplanted into field margins (sown wildflower or grass-legume strips) were affected by the abundance of different pollinator functional groups (defined by species traits). We also investigated how the maternal plant attractiveness, conspecific pollen donor density, and heterospecific floral richness and density interacted with pollinator functional group abundance to modulate C. segetum mating outcomes. Multiple paternity increased (=lower paternity correlation) with greater local abundance of hoverflies (syrphids) and female medium-sized wild bees (albeit the latter's effect diminished with decreasing maternal plant attractiveness), and the presence of male bumblebees (Bombus) under low local floral richness. Cyanus segetum progeny genetic diversity increased with male Bombus presence but decreased with greater abundance of syrphids and honey bees (Apis mellifera). Overall, field margins supported plant-pollinator interactions ensuring multiple paternity and conservation of allelic diversity in C. segetum progeny. The contribution to plant mating outcomes of different pollinator functional groups was dictated by their local abundance or traits affecting pollen transfer efficiency. The local floral richness or maternal plant attractiveness sometimes modulated these relationships. This complex response of wild plant mating patterns to community interactions has implications for the use of field margins to restore functional pollination systems in farmed landscapes.

Long-term contamination by non-native fish assemblages in a Neotropical floodplain.

Biological invasions are a major threat to biodiversity in species-rich regions. Therefore, it is important to understand mechanisms behind the long-term establishment of non-native fish species in aquatic environments in the Neotropical region. Here, we associated fish biomass, species richness, and the proportion of non-native species (contamination and Kempton's indices) to quantify the non-native pressure over fish biodiversity in lakes and rivers of the Parana River floodplain, seasonally, from 2000 to 2017. We divided species into native and non-native assemblages sampled in spatio-temporal gradients. Temporal trends were examined using linear regressions and generalised additive models. Fish biomass in gillnets increased for both native and non-native fish species, but their Kempton indices were inversely correlated. Extinction of native species occurred locally with biotic differentiation of non-native species in lakes, rivers, and ecosystem contamination. A constant increase in fish biomass resulted in overwhelming biodiversity of non-natives at the end of the time series evaluated. Native biotic resistance to introductions was not detected in deterministic trends. The observed patterns were consistent with previous studies showing native biotic homogenisation and extinction of species in response to biological invasions, landscape fragmentation, and riverine impoundments. Increases in abundance and species richness of non-native fish were the biodiversity drivers that resulted in non-native species outweighing native species in the Parana floodplain.

Reconstructing MODIS normalized difference snow index product on Greenland ice sheet using spatiotemporal extreme gradient boosting model

The spatiotemporally continuous data of normalized difference snow index (NDSI) are key to understanding the mechanisms of snow occurrence and development as well as the patterns of snow distribution changes. However, the presence of clouds, particularly prevalent in polar regions such as the Greenland Ice Sheet (GrIS), introduces a significant number of missing pixels in the MODIS NDSI daily data. To address this issue, this study proposes the utilization of a spatiotemporal extreme gradient boosting (STXGBoost) model generate a comprehensive NDSI dataset. In the proposed model, various input variables are carefully selected, encompassing terrain features, geometry-related parameters, and surface property variables. Moreover, the model incorporates spatiotemporal variation information, enhancing its capacity for reconstructing the NDSI dataset. Verification results demonstrate the efficacy of the STXGBoost model, with a coefficient of determination of 0.962, root mean square error of 0.030, mean absolute error of 0.011, and negligible bias (0.0001). Furthermore, simulation comparisons involving missing data and cross-validation with Landsat NDSI data illustrate the model’s capability to accurately reconstruct the spatial distribution of NDSI data. Notably, the proposed model surpasses the performance of traditional machine learning models, showcasing superior NDSI predictive capabilities. This study highlights the potential of leveraging auxiliary data to reconstruct NDSI in GrIS, with implications for broader applications in other regions. The findings offer valuable insights for the reconstruction of NDSI remote sensing data, contributing to the further understanding of spatiotemporal dynamics in snow-covered regions.

BDNF Differentially Affects Low- and High-Frequency Neurons in a Primary Nucleus of the Chicken Auditory Brainstem.

Neurotrophins are proteins that mediate neuronal development using spatiotemporal signaling gradients. The chicken nucleus magnocellularis (NM), an analogous structure to the mammalian anteroventral cochlear nucleus, provides a model system in which signaling between the brain-derived neurotrophic factor (BDNF) and tyrosine receptor kinase B (TrkB) is temporally regulated. In the NM, TrkB expression is high early in development (embryonic [E] day 9) and is downregulated until maturity (E18-21). It is currently unknown how BDNF-TrkB signaling affects neuronal properties throughout development and across a spatial (i.e., frequency) axis. To investigate this, we exogenously applied BDNF onto NM neurons ex vivo and studied intrinsic properties using whole-cell patch clamp electrophysiology. Early in development (E13), when TrkB expression is detectable with immunohistochemistry, BDNF application slowed the firing of high-frequency NM neurons, resembling an immature phenotype. Current measurements and biophysical modeling revealed that this was mediated by a decreased conductance of the voltage-dependent potassium channels. Interestingly, this effect was seen only in high-frequency neurons and not in low-frequency neurons. BDNF-TrkB signaling induced minimal changes in late-developing NM neurons (E20-21) of high and low frequencies. Our results indicate that normal developmental downregulation of BDNF-TrkB signaling promotes neuronal maturation tonotopically in the auditory brainstem, encouraging the appropriate development of neuronal properties.

Effects of Colliding Laser Pulses Polarization on e^{-}e^{+} Cascade Development in Extreme Focusing.

The onset and development of electron-positron cascade in a standing wave formed by multiple colliding laser pulses requires tight focusing in order to achieve the maximum laser intensity. There, steep spatiotemporal gradients in the laser intensity expel seed particles from the high-intensity region and thus can prevent the onset of a cascade. We show that radially polarized laser pulses ensure that the seed electrons are present at the focal plane at the moment of the highest amplitude even in the case of extreme focusing. This feature reduces the required laser power for the onset of a cascade 100 times (80 times) compared to circularly (linearly) polarized laser pulses having the same focal spot radius and duration.

Diffusiophoresis promotes phase separation and transport of biomolecular condensates

The internal microenvironment of a living cell is heterogeneous and comprises a multitude of organelles with distinct biochemistry. Amongst them are biomolecular condensates, which are membrane-less, phase-separated compartments enriched in system-specific proteins and nucleic acids. The heterogeneity of the cell engenders the presence of multiple spatiotemporal gradients in chemistry, charge, concentration, temperature, and pressure. Such thermodynamic gradients can lead to non-equilibrium driving forces for the formation and transport of biomolecular condensates. Here, we report how ion gradients impact the transport processes of biomolecular condensates on the mesoscale and biomolecules on the microscale. Utilizing a microfluidic platform, we demonstrate that the presence of ion concentration gradients can accelerate the transport of biomolecules, including nucleic acids and proteins, via diffusiophoresis. This hydrodynamic transport process allows localized enrichment of biomolecules, thereby promoting the location-specific formation of biomolecular condensates via phase separation. The ion gradients further impart directional motility of condensates, allowing them to exhibit enhanced diffusion along the gradient. Coupled with a reentrant phase behavior, the gradient-induced enhanced motility leads to a dynamical redistribution of condensates that ultimately extends their lifetime. Together, our results demonstrate diffusiophoresis as a non-equilibrium thermodynamic force that governs the formation and transport of biomolecular condensates.

Food niche overlap between two stingless bee species along a spatio-temporal gradient: implications for keeping and conservation

Food niche overlap between two stingless bee species along a spatio-temporal gradient: implications for keeping and conservation

Automated Upscaling Via Symbolic Computing for Thermal Runaway Analysis in Li-Ion Battery Modules

Despite the availability of advanced numerical capabilities and computational power, complex multi-physical, multiscale systems continue to challenge modeling efforts due to their elusive behaviors and nontrivial couplings. High-fidelity simulations have paved the way for simulating such systems, but the associated computational costs prohibit such tools from being used in iterative routines for engineering design and optimization. While numerical strategies for accelerating complex at-scale multi-physical simulations exist, they often concede predictive accuracy and lack rigorous justification. Alternatively, methods for developing rigorous multiscale models (e.g., upscaling techniques) offer a priori modeling error guarantees, but are primarily used in academic settings due to their requirements in time and specialized expertise for handling analytically intractable derivations. In our previous work, we developed an automated upscaling engine, Symbolica, that uses symbolic computation to avoid these overheads and rapidly develop multiscale models via rigorous homogenization theory. In this work, we extend Symbolica’s model development capabilities to accommodate multi-domain structures and demonstrate Symbolica’s generality in model development by applying it to a real-world problem; namely, to analyze thermal runaway in Li-ion battery modules. We obtain homogenized models that capture the sharp spatiotemporal gradients associated with thermal runaway to within the guaranteed modeling error. Upon numerically validating the models, we demonstrate how Symbolica can be used to reduce the level of effort in rigorous multiscale modeling and model implementation for real-world applications. Figure 1

Scale Differences and Gradient Effects of Local Climate Zone Spatial Pattern on Urban Heat Island Impact—A Case in Guangzhou’s Core Area

With the continuous development of cities, the surface urban heat island intensity (SUHII) is increasing, leading to the deterioration of the urban thermal environment, increasing energy consumption, and endangering the health of urban residents. Understanding the spatio-temporal scale difference and gradient effect of urban spatial patterns on the impact of SUHII is crucial for improving the climate resilience of cities and promoting sustainable urban development. This paper investigated the characteristics of SUHII changes at different time periods based on local climate zones (LCZs) and downscaled land surface temperature (LST) data. Meanwhile, landscape pattern indicators and the multiscale geographically weighted regression (MGWR) model were utilized to analyze the impacts of urban spatial patterns on SUHII at multiple spatial–temporal scales. The results indicated that the SUHII of each LCZ type exhibited diverse patterns in different time periods. High SUHII occurred in summer daytime and autumn nighttime. Compact and high-rise buildings (LCZ1/2/4) showed markedly higher SUHII during the daytime or nighttime, except for heavy industry. The extent of influence and the dominant factors of LCZ spatial patterns on SUHII exhibit obvious scale differences and gradient effects. At the regional scale, highly regular and compacted built-up areas tended to increase SUHII, while single and continuously distributed built-up areas had a greater impact on increasing SUHII. At the local scale, the impact of the PLAND (1/2/4/5/10) on SUHII exhibited a trend of diminishing from urban to suburban areas. In urban areas, the PLAND of LCZ 1, LCZ 2, and LCZ4 was the major factor affecting the increase in SUHII, whereas, in suburban areas, the PLAND of LCZ 2 and LCZ 10 was the major influencing factor on SUHII. The results can provide a scientific reference for mitigating urban heat island effects and constructing an ecologically ‘designed’ city.

Redox-mediated Biomolecular information transfer in single electrogenetic biological cells

Electronic communication in natural systems makes use, inter alia, of molecular transmission, where electron transfer occurs within networks of redox reactions, which play a vital role in many physiological systems. In view of the limited understanding of redox signaling, we developed an approach and an electrochemical–optical lab-on-a-chip to observe cellular responses in localized redox environments. The developed fluidic micro-system uses electrogenetic bacteria in which a cellular response is activated to electrically and chemically induced stimulations. Specifically, controlled environments for the cells are created by using microelectrodes to generate spatiotemporal redox gradients. The in-situ cellular responses at both single-cell and population levels are monitored by optical microscopy. The elicited electrogenetic fluorescence intensities after 210 min in response to electrochemical and chemical activation were 1.3 × 108±0.30 × 108 arbitrary units (A.U.) and 1.2 × 108±0.30 × 108 A.U. per cell population, respectively, and 1.05 ± 0.01 A.U. and 1.05 ± 0.01 A.U. per-cell, respectively. We demonstrated that redox molecules' mass transfer between the electrode and cells – and not the applied electrical field – activated the electrogenetic cells. Specifically, we found an oriented amplified electrogenetic response on the charged electrodes' downstream side, which was determined by the location of the stimulating electrodes and the flow profile. We then focused on the cellular responses and observed distinct subpopulations that were attributed to electrochemical rather than chemical stimulation, with the distance between the cells and the stimulating electrode being the main determinant. These observations provide a comprehensive understanding of the mechanisms by which diffusible redox mediators serve as electron shuttles, imposing context and activating electrogenetic responses.

Functional connectome through the human life span.

The lifespan growth of the functional connectome remains unknown. Here, we assemble task-free functional and structural magnetic resonance imaging data from 33,250 individuals aged 32 postmenstrual weeks to 80 years from 132 global sites. We report critical inflection points in the nonlinear growth curves of the global mean and variance of the connectome, peaking in the late fourth and late third decades of life, respectively. After constructing a fine-grained, lifespan-wide suite of system-level brain atlases, we show distinct maturation timelines for functional segregation within different systems. Lifespan growth of regional connectivity is organized along a primary-to-association cortical axis. These connectome-based normative models reveal substantial individual heterogeneities in functional brain networks in patients with autism spectrum disorder, major depressive disorder, and Alzheimer's disease. These findings elucidate the lifespan evolution of the functional connectome and can serve as a normative reference for quantifying individual variation in development, aging, and neuropsychiatric disorders.

Developmental origin of oligodendrocytes determines their function in the adult brain

In the mouse embryonic forebrain, developmentally distinct oligodendrocyte progenitor cell populations and their progeny, oligodendrocytes, emerge from three distinct regions in a spatiotemporal gradient from ventral to dorsal. However, the functional importance of this oligodendrocyte developmental heterogeneity is unknown. Using a genetic strategy to ablate dorsally derived oligodendrocyte lineage cells (OLCs), we show here that the areas in which dorsally derived OLCs normally reside in the adult central nervous system become populated and myelinated by OLCs of ventral origin. These ectopic oligodendrocytes (eOLs) have a distinctive gene expression profile as well as subtle myelination abnormalities. The failure of eOLs to fully assume the role of the original dorsally derived cells results in locomotor and cognitive deficits in the adult animal. This study reveals the importance of developmental heterogeneity within the oligodendrocyte lineage and its importance for homeostatic brain function.

Development of the Fetal Brain Corticocortical Structural Network during the Second-to-Third Trimester Based on Diffusion MRI.

During the second-to-third trimester, the neuronal pathways of the fetal brain experience rapid development, resulting in the complex architecture of the interwired network at birth. While diffusion MRI-based tractography has been employed to study the prenatal development of structural connectivity network (SCN) in preterm neonatal and postmortem fetal brains, the in utero development of SCN in the normal fetal brain remains largely unknown. In this study, we utilized in utero dMRI data from human fetuses of both sexes between 26 and 38 gestational weeks to investigate the developmental trajectories of the fetal brain SCN, focusing on intrahemispheric connections. Our analysis revealed significant increases in global efficiency, mean local efficiency, and clustering coefficient, along with significant decrease in shortest path length, while small-worldness persisted during the studied period, revealing balanced network integration and segregation. Widespread short-ranged connectivity strengthened significantly. The nodal strength developed in a posterior-to-anterior and medial-to-lateral order, reflecting a spatiotemporal gradient in cortical network connectivity development. Moreover, we observed distinct lateralization patterns in the fetal brain SCN. Globally, there was a leftward lateralization in network efficiency, clustering coefficient, and small-worldness. The regional lateralization patterns in most language, motor, and visual-related areas were consistent with prior knowledge, except for Wernicke's area, indicating lateralized brain wiring is an innate property of the human brain starting from the fetal period. Our findings provided a comprehensive view of the development of the fetal brain SCN and its lateralization, as a normative template that may be used to characterize atypical development.

Corticogenesis and folding process of the neopallium in the South American plains vizcacha, Lagostomus maximus.

The plains vizcacha, Lagostomus maximus, is a precocial hystricomorph rodent with a gyrencephalic brain. This work aimed to perform a time-lapse analysis of the embryonic brain cortical development in the plains vizcacha to establish a species-specific temporal window for corticogenesis and the gyrencephaly onset. Additionally, a comparative examination with evolutionarily related rodents was conducted. Embryos from 40embryonic days (ED) until the end of pregnancy ( 154 ED) were evaluated. The neuroanatomical examination determined transverse sulci at 80ED and rostral lateral and caudal intraparietal sulci around 95ED. Histological examination of corticogenesis showed emergence of the subplate at 43ED and expansion of the subventricular zone (SVZ) and its division into inner and outer SVZs around 54ED. The neocortical layers formation followed an inside-to-outside spatiotemporal gradient beginning with the emergence of layers VI and V at 68ED and establishing the final six neocortical layers around 100ED. A progressive increment of gyrencephalization index (GI) from 1.005±0.003 around 70ED, which reflects a smooth cortex, up to 1.07±0.009 at the end of gestation, reflecting a gyrencephalic neuroanatomy, was determined. Contrarily, the minimum cortical thickness (MCT) progressively decreased from 61ED up to the end of gestation. These results show that the decrease in the cortical thickness, which enables the onset of neocortical invaginations, occurs together with the expansion and subdivision of the SVZ. The temporal comparison of corticogenesis in plains vizcacha with that in relative species reflects a prenatal long process compared with other rodents that may give an evolutionary advantage to L. maximus as a precocial species.

Intracortical recordings reveal vision-to-action cortical gradients driving human exogenous attention

Exogenous attention, the process that makes external salient stimuli pop-out of a visual scene, is essential for survival. How attention-capturing events modulate human brain processing remains unclear. Here we show how the psychological construct of exogenous attention gradually emerges over large-scale gradients in the human cortex, by analyzing activity from 1,403 intracortical contacts implanted in 28 individuals, while they performed an exogenous attention task. The timing, location and task-relevance of attentional events defined a spatiotemporal gradient of three neural clusters, which mapped onto cortical gradients and presented a hierarchy of timescales. Visual attributes modulated neural activity at one end of the gradient, while at the other end it reflected the upcoming response timing, with attentional effects occurring at the intersection of visual and response signals. These findings challenge multi-step models of attention, and suggest that frontoparietal networks, which process sequential stimuli as separate events sharing the same location, drive exogenous attention phenomena such as inhibition of return.

Using individual-based trait frequency distributions to forecast plant-pollinator network responses to environmental change.

Determining how and why organisms interact is fundamental to understanding ecosystem responses to future environmental change. To assess the impact on plant-pollinator interactions, recent studies have examined how the effects of environmental change on individual interactions accumulate to generate species-level responses. Here, we review recent developments in using plant-pollinator networks of interacting individuals along with their functional traits, where individuals are nested within species nodes. We highlight how these individual-level, trait-based networks connect intraspecific trait variation (as frequency distributions of multiple traits) with dynamic responses within plant-pollinator communities. This approach can better explain interaction plasticity, and changes to interaction probabilities and network structure over spatiotemporal or other environmental gradients. We argue that only through appreciating such trait-based interaction plasticity can we accurately forecast the potential vulnerability of interactions to future environmental change. We follow this with general guidance on how future studies can collect and analyse high-resolution interaction and trait data, with the hope of improving predictions of future plant-pollinator network responses for targeted and effective conservation.

Gambierdiscus species diversity and community structure in St. Thomas, USVI and the Florida Keys, USA

Ciguatera Poisoning (CP) is a widespread and complex poisoning syndrome caused by the consumption of fish or invertebrates contaminated with a suite of potent neurotoxins collectively known as ciguatoxins (CTXs), which are produced by certain benthic dinoflagellates species in the genera Gambierdiscus and Fukuyoa. Due to the complex nature of this HAB problem, along with a poor understanding of toxin production and entry in the coral reef food web, the development of monitoring, management, and forecasting approaches for CP has lagged behind those available for other HAB syndromes. Over the past two decades, renewed research on the taxonomy, physiology, and toxicology of CP-causing dinoflagellates has advanced our understanding of the species diversity that exists within these genera, including identification of highly toxic species (so called “superbugs”) that likely contribute disproportionately to ciguatoxins entering coral reef food webs. The recent development of approaches for molecular analysis of field samples now provide the means to investigate in situ community composition, enabling characterization of spatio-temporal species dynamics, linkages between toxic species abundance and toxin flux, and the risk of ciguatoxin prevalence in fish. In this study we used species-specific fluorescent in situ hybridization (FISH) probes to investigate Gambierdiscus species composition and dynamics in St. Thomas (USVI) and the Florida Keys (USA) over multiple years (2018–2020). Within each location, samples were collected seasonally from several sites comprising varying depths, habitats, and algal substrates to characterize community structure over small spatial scales and across different host macrophytes. This approach enabled the quantitative determination of communities over spatiotemporal gradients, as well as the selective enumeration of species known to exhibit high toxicity, such as Gambierdiscus silvae. The investigation found differing community structure between St. Thomas and Florida Keys sites, driven in part by differences in the distribution of toxin-producing species G. silvae and G. belizeanus, which were present throughout sampling sites in St. Thomas but scarce or absent in the Florida Keys. This finding is significant given the high toxicity of G. silvae, and may help explain differences in fish toxicity and CP incidence between St. Thomas and Florida. Intrasite comparisons along a depth gradient found higher concentrations of Gambierdiscus spp. at deeper locations. Among the macrophytes sampled, Dictyota may be a likely vector for toxin transfer based on their widespread distribution, apparent colonization by G. silvae, and palatability to at least some herbivore grazers. Given its ubiquity throughout both study regions and sites, this taxa may also serve as a refuge, accumulating high concentrations of Gambierdiscus and other benthic dinoflagellates, which in turn can serve as source populations for highly palatable and ephemeral habitats nearby, such as turf algae. These studies further demonstrate the successful application of FISH probes in examining biogeographic structuring of Gambierdiscus communities, targeting individual toxin-producing species, and characterizing species-level dynamics that are needed to describe and model ecological drivers of species abundance and toxicity.

Effect of environmental variables on mercury accumulation in sediments of an anthropogenically impacted tropical estuary (Buenaventura Bay, Colombian Pacific)

Estuaries are the main entry areas of mercury to the marine environment and are important to understand the effect of this contaminant on marine organisms, since it accumulates in the sediments becoming available to enter the food trophic chain. This study aims to determine the environmental variables that mainly influence the spatiotemporal dynamics of total mercury accumulation in sediments of tropical estuaries. Sediment samples were collected from interior and exterior areas of the estuary during the dry and rainy seasons, representing the spatiotemporal gradients of the estuary. The grain size, organic matter content (OM), and total mercury concentration (THg) of the sediment samples were determined. In addition, salinity, temperature, dissolved oxygen, and pH of the water column associated with each sediment sample were assessed. The variations in environmental conditions, OM and THg in sediment were in accordance with a gradient which goes from conditions influenced by fresh water in the inner estuary to conditions influenced by sea water in the outer part of the estuary. The OM and THg in sediments presented similar variation patterns; they were higher in the rainy season than in the dry season and in the interior area of the estuary than in the exterior area. Despite the complex dynamic observed in the distribution and accumulation processes of mercury in sediments, these processes could be modeled from OM and salinity parameters. Due to the correlations found, in the process of accumulation of mercury in sediments the OM could represents the pathway of transport and accumulation of THg, and salinity could represent the influence of the hydroclimatic variations and environmental gradients of the estuary.

Automated upscaling via symbolic computing for thermal runaway analysis in Li-ion battery modules

Despite the availability of advanced numerical capabilities and computational power, complex multi-physical, multiscale systems continue to challenge modeling efforts due to their elusive behaviors and nontrivial couplings. High-fidelity simulations have paved the way for simulating such systems, but the associated computational costs prohibit such tools from being used in iterative routines for engineering design and optimization. While numerical strategies for accelerating complex at-scale multi-physical simulations exist, they often concede predictive accuracy and lack rigorous justification. Alternatively, methods for developing rigorous multiscale models (e.g., upscaling techniques) offer a priori modeling error guarantees, but are primarily used in academic settings due to their requirements in time and specialized expertise for handling analytically intractable derivations. In our previous work, we developed an automated upscaling engine, Symbolica, that uses symbolic computation to avoid these overheads and rapidly develop multiscale models via rigorous homogenization theory. In this work, we extend Symbolica’s model development capabilities to accommodate multi-domain structures and demonstrate Symbolica’s generality in model development by applying it to a real-world problem; namely, to analyze thermal runaway in Li-ion battery modules. We obtain homogenized models that capture the sharp spatiotemporal gradients associated with thermal runaway to within the guaranteed modeling error. Upon numerically validating the models, we demonstrate how Symbolica can be used to reduce the level of effort in rigorous multiscale modeling and model implementation for real-world applications.