Over the last decade, the evolution of survey acquisition design composed of numerous receivers and shot points has exponentially increased the seismic data volume. As a result, the demand for efficient algorithms for detecting seismic first arrivals has been growing. Machine learning methods, especially convolutional neural networks, have shown great potential in previous studies. However, most existing methods lack generalization ability or be sensitive to noise. This paper proposes a new picking strategy based on the single-to multi-domain selftrained network to achieve automated picking for the first arrivals. Our self-trained network learns from a single domain to multiple domains while rejecting the wrong predicted picking and assigning the pseudo-first-arrivals-times to the unlabeled dataset. The model inherits the advantage from both single-and multi-domain picking strategies; it has high generalizability and noise robustness. In addition, the model is designed for a limited amount of manually annotated data to reduce human labor and total seismic processing time. In experiments on an open hard rock seismic multi-survey dataset, our picking method outperforms the benchmark on most evaluation metrics. Remarkably, our approach achieves approximately 10 % accuracy improvement against Lalor site data with a sampling rate that differs from training data. The experimental results demonstrate our method’s high generalizability and robustness for unseen datasets.

Accurately mapping the evolving bathymetry under energetic wave breaking is challenging, yet critical for improving our understanding of sandy beach morphodynamics. Though remote sensing is one of the most promising opportunities for reaching this goal, existing depth-inversion algorithms using linear approaches face major theoretical and/or technical issues in the surf zone, limiting their accuracy over this region. Here, we present a new depth-inversion approach relying on Boussinesq theory for quantifying nonlinear dispersion effects in nearshore waves. Using high-resolution datasets collected in the laboratory under diverse wave conditions and beach morphologies, we demonstrate that this approach results in enhanced levels of accuracy in the surf zone (errors typically within 10%) and presents a major improvement over linear methods. The new nonlinear depth-inversion approach provides significant prospects for future practical applications in the field using existing remote sensing technologies, including continuous lidar scanners and stereo imaging systems.

We conduct a series of eight 45-day experiments with a global storm-resolving model (GSRM) to test the sensitivity of relative humidity R in the tropics to changes in model resolution and parameterizations. These changes include changes in horizontal and vertical grid spacing as well as in the parameterizations of microphysics and turbulence, and are chosen to capture currently existing differences among GSRMs. To link the R distribution in the tropical free troposphere with processes in the deep convective regions, we adopt a trajectory-based assessment of the last-saturation paradigm. The perturbations we apply to the model result in tropical mean R changes ranging from 0.5% to 8% (absolute) in the mid troposphere. The generated R spread is similar to that in a multi-model ensemble of GSRMs and smaller than the spread across conventional general circulation models, supporting that an explicit representation of deep convection reduces the uncertainty in tropical R. The largest R changes result from changes in parameterizations, suggesting that model physics represent a major source of humidity spread across GSRMs. The R in the moist tropical regions is disproportionately sensitive to vertical mixing processes within the tropics, which impact R through their effect on the last-saturation temperature rather than their effect on the evolution of the humidity since last-saturation. In our analysis the R of the dry tropical regions strongly depends on the exchange with the extra-tropics. The interaction between tropics and extratropics could change with warming and presage changes in the radiatively sensitive dry regions.

There have been many extreme fire seasons in Maritime Southeast Asia (MSEA) over the last two decades, a trend which will likely continue, if not accelerate, due to climate change. Fires, in turn, are a major driver of atmospheric carbon monoxide (CO) variability, especially in the Southern Hemisphere. Previous studies have explored the relationship between climate variability and fire counts, burned area, and atmospheric CO through regression models that use climate mode indices as predictor variables. Here we model the connections between climate variability and atmospheric CO at a level of complexity not yet studied and make accurate predictions of atmospheric CO (a proxy for fire intensity) at useful lead times. To do this, we develop a regularization-based statistical modeling framework that can accommodate multiple lags of a single climate index, which we show to be an important feature in explaining CO. We use this framework to present advancements over previous modeling efforts, such as the inclusion of outgoing longwave radiation (OLR) anomalies, the use of weekly data, and a stability analysis that adds weight to the scientific interpretation of selected model terms. We find that the El Ni\ {n}o Southern Oscillation (ENSO), the Dipole Mode Index (DMI), and OLR (as a proxy for the Madden-Julian Oscillation) at various lead times are the most significant predictors of atmospheric CO in MSEA. We further show that the model gives accurate predictions of atmospheric CO at leads times of up to 6 months, making it a useful tool for fire season preparedness.

Earth angular momentum forecasts are naturally accompanied by forecast errors that typically grow with increasing forecast length. In contrast to this behavior, we have detected large quasi-periodic deviations between atmospheric angular momentum wind term forecasts and their subsequently available analysis. The respective errors are not random and have some hard to define yet clearly visible characteristics which may help to separate them from the true forecast information. These kinds of problems, which should be automated but involve some adaptation and decision-making in the process, are most suitable for machine learning methods. Consequently, we propose and apply a neural network to the task of removing the detected artificial forecast errors. We found, that a cascading forward neural network model performed best in this problem. A total error reduction with respect to the unaltered forecasts amounts to about 30% integrated over a 6 day forecast period. Integrated over the initial 3 day forecast period, in which the largest artificial errors are present, the improvements amount to about 50%. After the application of the neural network, the remaining error distribution shows the expected growth with forecast length. However, a 24 hourly modulation and an initial baseline error of 2*10-8 became evident that were hidden before under the larger forecast error.

This study’s objective was to evaluate the potential geospatial relationship between agricultural pesticide use and cancer incidences by using the concept of consolidated environmental burden indices. We conducted two sets of experiments, using pesticide data collected from the U.S. Geological Survey Pesticide National Synthesis Project database and using cancer data collected from the National Cancer Institute State Cancer Profiles and Cancer Data Registry of Idaho. In the first experiment, we analyzed the associations between several potentially carcinogenic pesticides and metals to the pediatric cancer incidence among the counties of Idaho. Principal component analysis was adopted to consolidate the environmental variables. This study identified that the principal components (PCs) were significantly associated with pediatric cancer at P-value < 0.01 and a t-statistic > 3. The PC’s improved model performance substantially, from NSE = 0.22 for the control case to NSE = 0.52 for the treatment case. In the second experiment, we evaluated the associations of pesticides to two cancer metrics (pediatric cancer incidence and total cancer incidence) across each of the 11 contiguous states in the Western U.S. at the state and county spatial scales. A multi-level model was developed using fumigant mass, fumigant mass tertiles as well as county and state boundaries, which strongly predicted total cancer incidence (R-squared = 0.95, NSE = 0.91, and SSR = 8.22). Further, this study also identified significant associations between total fumigant mass, total pesticide mass, medium and high fumigant class (tertiles > 0.33), and high pesticide class (tertile > 0.66), relative to total cancer incidence (P-value < 0.01) in the U.S. west. Fumigant application rate was shown to be significantly related to the incidence of total cancers and pediatric cancer. Moreover, this relationship was maintained regardless of the spatial resolution used in the analysis (county or state). Both studies reinforce inferences from previous studies that pesticides are related to an elevated risk for pediatric cancer. Similarly, among the pesticides, fumigants such as metam were shown to be important relative to the cancer incidences for both experiments, highlighting the significance of further analysis.

To improve our understanding of the impact of sea spray aerosols (SSA) on the Earth’s climate, it is critical to understand the physical mechanisms which determine the size-resolved sea spray aerosol source. Of the factors affecting SSA emissions, seawater salinity has perhaps received the least attention in the literature and previous studies have produced conflicting results. Here, we present a series of laboratory experiments designed to investigate the role of salinity on aerosol production from artificial seawater using a continuous plunging jet. During these experiments, the aerosol and surface bubble size distributions were monitored while the salinity was decreased from 35 to 0 g/kg. Three distinct salinity regimes were identified: 1) A high salinity regime, 10-35 g/kg, where decreasing salinity only resulted in minor reductions in particle number emissions but significant reductions in particle volume; 2) an intermediate salinity regime, 5-10 g/kg, with a local maximum in particle number; 3) a low salinity regime, < 5 g/kg, characterized by a rapid decrease in particle number as salinity decreased and dominated by small particles and larger bubbles. We discuss the implications of our results through comparison of the size-resolved aerosol flux and the surface bubble population at different salinities. Finally, by varying the seawater temperature at three specific salinities we have also generated a simple parameterization of the particle number concentration and effective radius as a function of seawater temperature and salinity that can be used to estimate the sea spray aerosol flux in low salinity regions like the Baltic Sea.

Magma chamber volume is critical for volcano monitoring and forecasting. Standard geodetic methods cannot constrain the total volume, only the change in volume. Here, we show that stress perturbations associated with trapdoor faulting allow bounds to be placed on the total chamber volume at Sierra Negra volcano, in the Galapagos. The deformation response of the magma chamber to faulting depends on both the absolute chamber volume and the compressibility of the magma. Bubble-free magma provides the lower limit on compressibility, thus an upper bound on the chamber volume of 13.6 to 20.6 km3, depending on fault dip. We estimate an upper limit on compressibility using a conduit model relating volatile content to lava fountain height, which is compared with observations from the 2005 eruption, constrained by volatile content of olivine melt inclusions. This yields a lower bound on chamber volume of 0.5 times the upper bound.

Quantifying the relative importance of the atmospheric drivers of surface melting on the Larsen C ice shelf is critical in the context of recent and future climate change. Here, we present analysis of a new multi-decadal, high-resolution model hindcast using the Met Office Unified Model (MetUM), described in part 1 of this study. We evaluate the contribution of various atmospheric conditions in order to identify the most significant causes of melting over the recent past. We find the primary driver of surface melting on Larsen C is solar radiation. Foehn events are the second most important contributor to surface melting, especially in non-summer seasons when relatively less solar radiation is received at the surface of the ice shelf. Thirdly, cloud influences surface melting via its impact on the surface energy balance (SEB); when the surface temperature is warm enough, cloud can initiate or prolong periods of melting. Lastly, large-scale circulation patterns such as the Southern Annular Mode (SAM), El Niño Southern Oscillation (ENSO) and Amundsen Sea Low (ASL) control surface melting on Larsen C by influencing the local meteorological conditions and SEB. These drivers of melting interact and overlap, for example, the SAM influences the frequency of foehn, which are commonly associated with leeside cloud clearances and sunnier conditions. Ultimately, these drivers matter because sustained surface melting on Larsen C could destabilise the ice shelf via hydrofracturing, which would have consequences for the fate of the ice shelf and sea levels worldwide.

Mega-fires have occurred in Australia during the 2019/2020 bushfire season, leading to enhanced concentrations of many tropospheric pollutants. Here we report on a fire plume with unusually high and persistent HONO levels that we could tracked during one day at free tropospheric levels over the Tasman Sea on 4 January 2020 using IASI and CrIS satellite observations. HONO concentrations up to about 8 ppb were retrieved during nighttime. Persistent HONO concentrations (>1ppb) were still observed at sunrise. Model simulations suggest a significant contribution of primary fire emissions and heterogeneous photo-induced reactions to explain the observed concentrations. However, many uncertainties and unknowns remain in the plume aerosol load and in the chemical processes which may explain the model inability to reproduce HONO concentrations at sunrise.

The initial list of Essential Water Variables (EWVs) evolved from wide meta-surveys of water data needs for research and applications that were carried out in 2010 to support GEO Societal Benefit Areas (SBAs). These EWVs were formalized in the Group on Earth Observations System of Systems (GEOSS) Water Strategy Report (WSR) “From Observations to Decisions”, released in 2014. Subsequently, discussions with additional user communities have augmented the list, for example with Surface Water Extent. Besides “primary” EWVs that identify key water variables, including precipitation, soil moisture, and water quality, a set of “supplementary” EWVs is also needed to complete the information that the formal list of primary EWVs should provide, such as Digital Elevation Models. It is clear that all available observing systems, employing both remote sensing and in situ observing instruments and networks are required to address the range of space/time resolutions, accuracies, and data latencies that the end-user applications require. In fact, there are still gaps in our ability to deliver all variables as required. In some cases this is a technical challenge, such as remote sensing capabilities for some water quality variables, while in many other cases it is a matter of administrative and resource challenges. This paper summarizes EWVs as currently defined and required by key end-user research and applications sectors. As a follow up to the WSR, we highlight the relevance of EWVs to the indicator monitoring objectives of the UN Sustainable Development Goals (SDGs), various international Conventions and Frameworks, and the GEO Global Water Sustainability (GEOGloWS) priority thematic communities.

A document by Asaf Inbal. Click on the document to view its contents.

Understanding marine soundscapes, including the biological, anthropogenic, and geological sounds, is essential to conserving protected species and their habitats. However, the marine resource managers often do not have a strong science background to interpret complicated soundscape data to facilitate them making decisions. The biological components of soundscapes can be useful to characterize biodiversity and monitor the distribution and behavior of individual species. Anthropogenic sound in the ocean is increasing and has been recognized as a threat to marine mammals for decades. To help the marine resource managers and the general public understand the impacts of ocean noise, we as nine undergraduate students from different majors of study at UC Berkeley’s Fung Fellowship Program utilized Human-Centered Design and created an interactive marine soundscape map (https://calsound.herokuapp.com), focusing on the California Current Ecosystem. Based on 14 interviews we conducted with researchers, policymakers, and environmental lobbyists, we decided to portray spectral soundscape metrics alongside the context of animal and human activities in a map format. We then created a digital hub to easily visualize, analyze, and synthesize marine-sourced soundscape data. Our website displays soundscape data over a range of spatial and temporal scales, acoustic detections of marine mammals, species habitat models, and anthropogenic sound source distributions as heat map layers and graphs. The platform not only displays ocean soundscape data, but also provides an overview of marine soundscape technology, as well as related articles and websites. The website is designed so that users who are not familiar with marine soundscape data, such as coastal managers and the public, can guide themselves through a tutorial and explore on their own to gain a better understanding of oceanographic sound. In the future, we will add more features to the website, such as allowing users to upload their own data to the website to visualize them online. The website will be self-sustainable and continue to serve more people. Our website will facilitate people to visualize and understand marine soundscapes, their impacts and our solutions.

Xavier University of Louisiana, in collaboration with the Institute of Earth Science Research and Education (IESRE) run the ECoSTEM project. This project aims to improve the quality of undergraduate STEM education by introducing a computational, community-oriented component into STEM programs at Xavier University. Specifically, undergraduate students develop microcontroller-based systems for collecting airborne particulate data. They work with public school teachers (and their students) and Louisiana Department of Environmental Quality (LDEQ) officers to deploy the sensors at locations around New Orleans and to analyze the resulting data. Airborne particulates are often the pollutant of greatest concern in minority and other underserved communities. This project provide undergraduates at Xavier and the local teachers (and their students) with learning experiences that are relevant to them and their communities. ECoSTEM uses a tiered mentoring system in which the project team members mentor undergraduates, who in turn work with public school teachers and mentor their students. Research has shown positive effects from student involvement in such real-world, personally relevant STEM projects, including increases in student interest in STEM-related careers after graduation. As a result, the project has the potential to broaden student interest in STEM and increase pursuit of STEM-related majors.

Airborne LiDAR has become an essential data source for large-scale, high-resolution modeling of forest biomass and carbon stocks, enabling predictions with much higher resolution and accuracy than can be achieved using optical imagery alone. Ground noise filtering – that is, excluding returns from LiDAR point clouds based on simple height thresholds – is a common practice meant to improve the ‘signal’ content of LiDAR returns by preventing ground returns from masking useful information about tree size and condition contained within canopy returns. Although this procedure originated in LiDAR-based estimation of mean tree and canopy height, ground noise filtering has remained prevalent in LiDAR pre-processing, even as modelers have shifted focus to forest aboveground biomass (AGB) and related characteristics for which ground returns may actually contain useful information about stand density and openness. In particular, ground returns may be helpful for making accurate biomass predictions in heterogeneous landscapes that include a patchy mosaic of vegetation heights and land cover types. We applied several ground noise filtering thresholds while mapping two regions within New York State, one a forest-dominated area and the other a mixed-use landscape. We observed that removing ground noise via any height threshold systematically biases many of the LiDAR-derived variables used in AGB modeling. By fitting random forest models to each of these predictor sets, we found that that ground noise filtering yields models of forest AGB with lower accuracy than models trained using predictors derived from unfiltered point clouds. The relative inferiority of AGB models based on filtered LiDAR returns was much greater for the mixed land-cover study area than for the contiguously forested study area. Our results suggest that ground filtering should be avoided when mapping biomass, particularly when mapping heterogeneous and highly patchy landscapes, as ground returns are more likely to represent useful ‘signal’ than extraneous ‘noise’ in these cases.

Lahars, or volcanic mudflows, are one of the most devastating natural, volcanic hazards. We present an analysis of several rain-triggered lahar events at Volcán Fuego in Guatemala with an infrasound array and time-lapse camera network deployed along the Las Lajas river channel. While infrasound detects the passage of low frequency sound waves associated with these mudflows, the time-lapse cameras can be used to verify lahar passages and channel responses as these flows move downstream. Twenty-three infrasound microphones and five time-lapse cameras were deployed during the 2021 rainy season (May-October) in the Las Lajas drainage on the southeasterly side of Volcán de Fuego. With the data collected over this field campaign, we hope to quantify flow parameters such as volumes, velocities, and the frequency of these rain-triggered mudflows, as well as characterize flow behaviors in our infrasound signals by correlating data to the time-lapse imagery. This study allows us to identify the occurrence of several lahars, quantify important characteristics, and use a multi-faceted approach to verify and delineate important flow behaviors related to these volcanic hazards.

Slow slip events have previously been observed along the Hikurangi subduction zone beneath the North Island of New Zealand. These slow slip episodes occur both on the shallow plate interface (< 15km depth) and at the deeper end of the seismogenic zone (> 30km depth). We present the first catalog of low-frequency earthquakes (LFEs) in the Hikurangi subduction zone, located beneath the Kaimanawa Range on the central Hikurangi margin, downdip of a region that regularly (every 4-5 years) hosts M7 slow slip events. To systematically detect LFEs using continuous seismic data recorded by GeoNet, we developed a matched-filter technique with template waveforms derived from previous observations of tectonic tremor. The workflow presented in this work is composed of two iterations of a matched-filter search. In each iteration, the detections were gathered into families and their common waveforms postprocessed with machine-learning methods to extract high-quality waveforms, allowing us to pick seismic phase arrivals with which to locate the LFEs. We found that LFEs occur in episodes of intense activity during the neighboring updip M7 slow slip events. We also observe a recurrence time of 2 years between other large bursts of LFEs, suggestive of a shorter cycle of slow slip. We hypothesize that these and other frequent LFE episodes highlight smaller slow transients that have not yet been geodetically observed.

The North Polar Layered Deposits (NPLD) of Mars are ice-rich sedimentary layers that formed under the influence of Mars’ modern climate and thus record the recent climatic history of Mars, analogous to terrestrial ice sheets. The 2013-2023 Planetary Science Decadal Survey recommends a lander mission to sample the NPLD for climatic records; however, linking the geologic record to the climatic history will require quantitative dating of the NPLD. In this study we use orbital reflectance spectroscopy to show for the first time that dateable mafic lithics are present throughout the NPLD. We find significant glass as well as diverse crystalline minerals, which suggests that surface processes like impacts and volcanism were active during the late Amazonian and transported sediments from across the planet to the north pole. In situ investigation of the NPLD will thus provide critical quantitative constraints on both the recent geologic and climatic histories of Mars.