Hydrodynamic disconnectivity between surface water and groundwater is common in arid environments. It is also prone to affect shallow streams in wetter climate. Sediment layers with low permeability, owing to clogging, for instance, reduce hydrodynamic connectivity. The resistance of this clogging layer, results in unsaturated infiltration, which is characterized by the non-linear Richards equation. Either due to lack of field information, parsimony regarding computational resources or mere misunderstanding of the system, infiltration is often assumed saturated or drastically simplified in hydrogeological models. Here we show the existence of three simple generic asymptotic solutions to the unsaturated problem of vertical steady-state infiltration through a clogged profile, which we associate to three regimes: one dominated by the clogging layer, one by the underlying sediments and one depending on both layers. We also argue that infiltration rate roughly grows linearly with ponding depth. These observations motivate a refined definition of clogging potentially helpful in model selection as well as two novel approximate infiltration formulas. Our refined generic analytic framework is useful to better understand and formalize infiltration.

Satellite-based InSAR images have the potential to detect volcanic deformation prior to eruptions, but while a vast number of images are routinely acquired, only a small percentage contain volcanic deformation events. Manual inspection could miss these anomalies, and an automatic system modelled with supervised learning requires suitably labelled datasets. To tackle these issues, this paper explores the use of unsupervised deep learning on InSAR images for the purpose of identifying volcanic deformation as anomalies. We test three different state-of-the-art architectures, one convolutional neural network (PaDiM - Patch Distribution Modeling) and two generative models (GANomaly and DDPM). We propose a preprocessing approach to deal with noisy and incomplete data points. We further improve the performance of PaDiM by using a weighted distance, assigning greater importance to features from deeper layers. The final framework was tested with four different volcanoes, which have different characteristics and its performance was compared against an existing supervised learning method for volcanic deformation detection. The experiments show that our final anomaly detection outperforms the supervised learning, particularly where the characteristics of deformation are unknown. Our framework can thus be used to identify deformation at volcanoes without needing prior knowledge about the deformation patterns present there.

Transported mineral dust in a Saharan air layer (SAL) contains highly active ice-nucleating particles (INPs) that may be transported across the Atlantic Ocean and subsequently seed clouds in the Caribbean and the Americas. During an aircraft campaign around Houston and the western U.S. Gulf Coast, a widespread SAL advected into the sampling region allowing for measurement of the ice-nucleating ability of SD following long-range transport. Results showed that the mean INP concentrations were 3–4.5 times higher than non-Saharan dust (nSD), but only at temperatures < –21 ºC. Active surface site densities were also enhanced in the SD, exceeding the mean for nSD by over an order of magnitude at temperatures < –21 ºC. These INP measurements confirmed that SD remains a highly active INP even after > 8000 km westward transport across the Atlantic Ocean.

Valley width is controlled by lithology and upstream drainage area, but little work has focused on identifying the processes that control bedrock valley widening. Bedrock valleys widen by first laterally eroding bedrock valley walls, followed by the collapse of overlying bedrock material that must then be transported away from the valley wall before the valley can continue widening. We hypothesize that talus piles that cannot be transported by the river protect the valley wall and slow valley widening, while talus piles that are rapidly transported allow for uninterrupted valley widening. We use field measurements from 40 locations in both wide and narrow valleys along the Buffalo River, AR to test this hypothesis. Our data show that wide valleys tend to have fewer talus piles and smaller talus grain sizes, while talus in narrow valleys is larger in size and more continuous along valley walls. We calculated potential talus block entrainment at each site location and found that talus blocks in wide valleys are potentially entrained and moved away from valley walls during moderate and large flood events while talus in narrow valleys is very rarely moved. Our results show that the potential transport of talus piles protecting bedrock valley walls from widening is controlled by the block size of collapsed bedrock wall material relative to stream competency. Our results also suggest that persistence vs. mobility of collapsed talus piles is an important process in the development of wide bedrock valleys.

Estimating the probability of rare or yet unobserved rainfall extremes is essential for hazard quantification, especially in the context of an intensifying water cycle and of short observational records, common even in countries with long hydro-meteorological monitoring tradition. Data limitations can be mitigated by using regionalization techniques, which augment observational information, and by employing effective statistical models , such as the Metastatistical Extreme Value Distribution (MEVD), which maximizes the use of available observations. In this work, we develop the MEVD Regionalized framework (MEVD-R) to reduce uncertainty in estimating the probability of rare events that are not present in the training sample. Extensive testing using a global dataset of 40,000 rain gauges across Europe, North America, and Australia demonstrates that MEVD-R yields negligible systematic error and greatly reduces uncertainty compared to traditional approaches. MEVD-R can even provide estimates when available data are too sparse for conventional methodologies to offer reliable results.

Convection-permitting dynamical downscaling (CPDD) allows for an explicit representation of the storm-scale generators of tornadoes, hail, severe thunderstorm winds, and locally heavy precipitation. Possible changes in such hazardous convective weather (HCW) due to human-induced climate change are therefore projected with higher confidence using CPDD than with analyses of relatively coarse global climate models (GCM). However, the computational resources necessary for CPDD are significant and therefore CPDD-based future projections of HCW have tended to be based on a single experiment, and thus absent of uncertainty measures otherwise determined with an ensemble of experiments via an ensemble of GCMs. Herein we present “environment-informed” CPDD as a means to efficiently generate a CPDD ensemble driven by different GCMs. This variant of CPDD is applied only to a subset of days and geographical domains over which the meteorological conditions potentially favor HCW; unnecessary model integrations on meteorologically unfavorable days and domains are thereby eliminated. The selection procedure also accounts for GCM biases.The temporal and geospatial occurrence of historical HCW over the United States is demonstrated from the perspective of environment-informed CPDD as applied to eight different GCMs and ERA5 reanalysis. The overall geographical distributions in HCW vary considerably from downscaled GCM to GCM, thus demonstrating the value of an ensemble. The efficiency in which HCW is realized in favorable environments also varies considerably across the eight downscaled GCMs.

The detection and analysis of gravitational waves (GW) has opened new windows on the universe in recent years. Our team has developed the GWSkyNet and GWSkyNet-Multi models, which are designed to operate with the limited data publicly available from the LIGO-Virgo collaboration. We employ a combination of neural networks and explainable AI techniques to predict the nature of the detected signals. The XAI part helps us understand the strengths and biases of our models.

Using an automated novel approach we conduct a reproducible systematic survey of electromagnetic ion cyclotron wave activity detected by Van Allen Probe B during the time period 2013 January 1 – 2019 July 15. We identify approximately 500 hours of EMIC wave activity, an occurrence rate of ∼ 0.85%. Accounting for satellite dwell time, we find that EMIC waves preferentially occur on the dayside, between 9 and 15 magnetic local time. This is true for both the H+ and He+ wavebands. Higher amplitude waves are found at higher values of L shell, while weaker waves occur at low L. The highest amplitudes are concentrated at high L near dawn and dusk. It is also found that EMIC wave occurrence is enhanced during periods of strong geomagnetic activity, with an occurrence rate of 2.7%. During storm times, waves preferentially occur in the afternoon and early evening sectors. The full list of electromagnetic ion cyclotron wave detection times and their properties is made publicly available to the community. This provides a reference catalog for comparison with other magnetospheric phenomena and other wave databases.

To better understand the landscape dynamics and changes in habitat connectivity influenced by glacial and interglacial oscillations over the biodiversity-rich Hengduan Mountains (HM) region, high-resolution climate data for past periods are essential. We apply the non-hydrostatic limited-area model COSMO, with a resolution of 12 km over East Asia, to simulate the Last Glacial Maximum (LGM), a period characterized by a generally colder and drier climate compared to present-day conditions. We perform the downscaling with a novel approach for paleoclimate modelling, the Pseudo-Global Warming (PGW) method. The COSMO PGW simulation for the LGM shows that COSMO generally replicates the large-scale dynamics of the driving global climate model simulation in the colder climate. Both models suggest weaker Asian summer monsoon systems during this period. Consequently, regions such as the Bay of Bengal, and the South China Sea, which typically receive substantial monsoon rainfall, experience significantly reduced precipitation. However, despite these model similarities, the high-resolution COSMO simulation exhibits distinctive differences on a smaller scale—particularly over land. For instance, COSMO suggests a more pronounced southward shift of the jet stream during the LGM winter, with cooler conditions in southern China. Moreover, the COSMO simulation, despite the overall weaker summer monsoon circulation, features increased precipitation amounts for much of the HM. Additionally, COSMO suggests a more extensive increase in snowfall over the High Mountain Asia region. Our study suggests that the resource-saving PGW approach is a suitable method to bridge the gap between large-scale projections and regional climate impacts—also for past periods like the LGM.

This study applies a benchmarking framework to assess a 39-member ensemble of regional climate models which have dynamically downscaled CMIP6 models over the CORDEX Australasian region. Four modelling centres contributed regional climate models to this ensemble using three regional climate models (CCAM, WRF, and BARPA-R), and a total of five model configurations. Assessment is conducted over the Australian continent using a separation into four major climate zones over a 30 year historical climatological period (1985-2014). Rainfall and near-surface temperatures are compared against six benchmarks measuring mean state patterns, spatial and temporal variance, seasonal cycles, long-term trends and selected extreme indices. Benchmark thresholds are derived either from previous studies or comparison with the driving model ensemble. Major model biases vary between ensemble members and include dry biases in northern and southern Australia, winter wet biases and a persistent low bias in the winter diurnal temperature range across all the modelling centres. Daily variability at large length-scales is comparable in the driving global climate model and downscaled regional climate model length-scales, and long-term trends are largely determined by the driving global climate model. Overall, the ensemble was deemed to be fit for purpose for impact studies. Strengths and weaknesses of the systematic benchmarking framework used here are discussed.

Quasi-two-dimensional visualizations of microstructure in the thermocline are created by processing χpod signals in a non-standard way. As the moored instrument is pumped by surface waves, the fast thermistors at each end trace out vertically overlapping paths, from which we produce 2–3 m tall swaths. A swath capturing the unmistakable form of Kelvin–Helmholtz billows provides a proof of concept—albeit by relying on what is a rare event in our dataset. More commonly, swaths exhibit steppy temperature fields during weaker turbulence, an abundance of small overturns during stronger turbulence, or some combination thereof. To examine this continuum statistically, we take a 13-month dataset from an equatorial χpod at 120 m deep and divide it into 53 000 swaths, each 10 minutes long. Swaths are ordered based on their associated buoyancy Reynolds number (Reb) inferred from our standard χpod processing. Clear visual differences in swath characteristics arise when comparing across an order of magnitude or more (e.g., Reb < 10 vs Reb = 10–100 vs Reb > 100). With the help of a convolutional neural network to semi-objectively pick representative swaths, we present our best estimates of what microstructure looks like in two dimensions as a function of Reb.

The Solar Wind Magnetosphere Ionosphere Link Explorer (SMILE) is a future spacecraft mission supported by the European Space Agency and the Chinese Academy of Science. It is expected that the mission will be launched in 2025 by ESA. SMILE will target the magnetopause, cusp, and bow shock regions. To investigate magnetosphere-ionosphere coupling, the spacecraft will take auroral images using an ultraviolet imager (UVI). For scientific support and analysis of data collected by the UVI, a numerical model of the ionosphere and UV emission has been developed, which is presented in the poster presentation. The photoelectron transport module of the model is described in the poster. The SMILE UV-imager will take global auroral images that for calibration, require specification of the changing state of the ionosphere. We demonstrate progress in meeting this goal and showcase using an example how the model works. 

Clumped isotope thermometry (T(∆47)) of soil carbonates provides an estimate of soil temperature at the time of mineral formation. Historically, that temperature has been interpreted to represent a warm-season soil temperature based on modern calibration studies largely done in (very) coarse-grained soils. Additionally, T(∆47) gives us an estimate of the oxygen isotope composition of soil water (δ18Ow) in the past, but previous calibration work has not generated independent δ18Ow datasets with which to understand these archives. Here, we present a modern calibration study of pedogenic carbonate clumped isotope thermometry in three soils in Colorado and Nebraska, USA, that have a fine-medium grain size, contain clay, and are representative of many carbonate-bearing paleosols preserved in the rock record. At two of the three sites, Briggsdale, CO and Seibert, CO, T(∆47) overlaps with mean annual air temperature (MAAT), and the calculated δ18Ow overlaps within uncertainty with measured δ18Ow at carbonate bearing depths. At the third site in Oglala National Grassland, NE, mean T(∆47) is 9 – 10°C warmer than MAAT, and the calculated δ18Ow has a significantly higher isotope value than any observations of δ18Ow. At all three sites, even in the fall season, δ18Ow values at carbonate bearing depths indicate little to no evaporative enrichment of heavy isotopes. Moreover, δ18Ow values overlap with precipitation isotope values from spring precipitation. Altogether, these data indicate that soil grain size affects pedogenic carbonate formation, and highlight a need for continued research in modern systems that emulate key features of the geologic record.

The reliable operation of transformers is critical to power system stability, and early fault detection in transformers can significantly reduce the risk of catastrophic failures. Dissolved gas analysis (DGA) is a commonly used method to assess transformer health, as gas concentrations correlate with fault conditions inside the transformer. Methane (CH4), a gas indicative of thermal faults, serves as a key parameter in detecting early signs of fault. This study proposes a linear epsilon support vector regression (SVR) model fine-tuned with Grid Search for predicting methane concentration based on transformer temperature, oil temperature, winding resistance, load factor, service life, and transformer age using grey relational analysis. By predicting methane concentration in ppm, this model supports proactive fault management, optimizing maintenance schedules, and effectively reducing transformer downtime. Comparative results suggest that the proposed SVR model is an effective baseline for improving dissolved gas analysis accuracy, contributing to enhanced transformer diagnostics and preventive maintenance strategies.

• We propose a strategy combining geodetic reference frame definition and time series analysis to extract slow slip events from GNSS data. • We detect and characterize a shallow, 10-km-long, three-week-long SSE along the central section of the North Anatolian Fault. • We show that persistent aseismic slip and SSEs are spatially complementary and interact at shallow depth.

The full-hydrometeor four-dimensional variational (4D-Var) assimilation scheme in the Weather Research and Forecasting (WRF) model, based on the WRF single-moment 6-class microphysics scheme (WSM6), is utilized to assimilate precipitation data. The focus is on short-term convective precipitation forecasts influenced by the Northeast cold vortex (NCCV). Four assimilation experiments were designed to compare the warm rain scheme with the full-hydrometeor scheme, as well as to examine the differences between assimilating hourly surface rain gauge data and multi-source integrated precipitation products. Ten cases of intense convective precipitation related to NCCV were analyzed. The results demonstrate that the initial analysis of ice-phase hydrometeors was satisfactory across the three experiments utilizing the full-hydrometeor 4D-Var schemes. The assimilation of precipitation data full-hydrometeor scheme in WRF 4D-Var effectively adjusted atmospheric thermodynamic properties and decreased model spin-up time, leading to improved precipitation forecasts, especially for the 0-3 hour period. Furthermore, the assimilation of rain gauge data or multi-source integrated precipitation data has been demonstrated to be an effective approach for enhancing the accuracy of weather forecasts.

This paper presents a novel hybrid deterministic-stochastic river morphodynamics numerical modeling approach that integrates a two-dimensional (2D) hydrodynamic model (deterministic) with a bed morphodynamic model (deterministic) and a bank erosion model (Markovian stochastic). The model solves the 2D Shallow Water Equations and the standard k-epsilon turbulence model. Bedload transport is estimated using the Meyer-Peter and Muller formula, and bed evolution is solved using the Exner Equation. The Markovian stochastic bank erosion model uses a new method to evaluate bank erosion risk. The approach was applied to a meander bend cutoff event in the Maiqu River on the Tibetan Plateau. Sixteen different bank-material critical shear stress cases were considered, representing highly erodible banks to resistant banks. Ten statistical realizations were performed for each case with different bank-material erodibility to obtain ensemble-averaged results. Flow field and bed evolution in the cutoff channel suggest that the model can successfully simulate bank erosion processes during the cutoff channel evolution, and bank topographic irregularities are reasonably captured. A newly introduced calibration parameter, the ratio of mesh size to the coupling period between the bank erosion model and the hydrodynamic & morphodynamic model, is not as intuitive as the erosion resistance calibration parameters used by the traditional bank erosion model. The proposed approach necessitates estimating the size of a typical slump failure, or 10-30% of the cutbank height, to set the size of the computational mesh.

The Delaware Basin in West Texas and Southeast New Mexico has experienced a proliferation in seismic activity since 2016. The seismic activity is primarily due to subsurface injection of wastewater into both shallow and deep reservoirs. However, the precise mechanisms connecting pore-fluids to seismic activity is not well understood. To shed light on these processes, we measure rate-state friction and poromechanical properties of rocks sampled from the Delaware Mountain Group (DMG) at pressures and stresses representative of in-situ conditions. Experiments were conducted inside a pressure vessel and loaded in a true-triaxial stress state. The samples exhibit velocity-strengthening behavior and transition to a velocity-neutral behavior with increasing slip. We also measure frictional healing and demonstrate that the healing rates are consistent with those measured from quartz-feldspathic-rich rocks. Fault acceleration produces a transient increase in layer thickness (i.e, dilatancy), which in turn, reduces the local pore-pressure and causes dilatancy strengthening. Broadly speaking, the frictional and poromechanical data indicate that shallow faults within the DMG should favor aseismic creep as opposed to unstable slip. Hence, alternative mechanisms to an increase in pore-pressure being the direct causative agent to seismicity in the DMG need to be considered. We propose that seismicity in the DMG could be caused by a slip-weakening mechanism via a transition to velocity weakening behavior associated with shear localization at higher shear strains. Alternatively, seismic activity in the DMG could be a byproduct of aseismic creep as opposed to being triggered directly by the advancement of a pore-pressure front.