The southeast Atlantic Ocean provides an excellent natural laboratory to study smoke-cloud interactions, a large driver of uncertainty in climate projections. The value of studying this in particular region is largely attributable to two factors---the expansive, bright, semi-permanent stratocumulus cloud deck and the fact that southern Africa is the largest source of biomass-burning aerosols in the world. We study this region using the WRF-Chem model with CAM5 aerosols and in situ observations from the ORACLES, LASIC, and CLARIFY field campaigns, all of which overlapped in August 2017. Across these campaigns, we compare aerosol, cloud, and thermodynamic variables to quantify model performance and expand upon observational findings of aerosol-cloud effects. Specifically, our approach is to analyze aerosol and cloud properties along flight tracks, picking out uniform legs within tropospheric smoke plumes and in the boundary layer. This unique approach allows us to sample the high spatiotemporal variability that can get lost to large-scale averaging. It also allows process-level comparison of local cloud responses to aerosol conditions, and measure model performance in those same processes. Along with better quantifying model predictive power, we find and justify updates to model parameters and processes to better emulate observations, notably aerosol size parameters. Preliminary results suggest that WRF-CAM5 is activating a smaller percentage of aerosols into cloud droplets than shown in observations, which could lead to biased modeling of aerosol indirect radiative effects on a larger scale. We explore this effect further with CCN activation tendency, updraft, particle sizing, and composition analysis, as well as broader dynamics like entrainment and removal rates. Comparing the model with similar instrument suites across multiple colocated campaigns also allows us to quantify instrument uncertainty in ways that a focus on a single campaign cannot and gives further context to the model performance.

Southeastern Greenland (SEG) provides a complex habitat area consisting of dozens of deep fjords that connect land-based ice with the open ocean. Within these fjords, glacial ice mixes with sea ice and intricate topography can create niche local conditions. This area is important for a number of species, including polar bears, and better understanding of both land ice and floating glacial ice as biological habitat motivates the need to study the physical environment itself. Studying SEG, however, posed a variety of challenges, including difficult access for in-situ work or instrument deployment, cloudy conditions that can obscure optical satellite instruments, and steep, complex terrain that can complicate identification of surface conditions. Here, we discuss our work to leverage several remote sensing products to determine the geospatial patterns of landfast sea ice and solid glacial ice during 2015-2019 in five SEG fjords with high polar bear use. We further connect these data with measurements of solid glacial ice discharge and glacial and terrestrial freshwater flux across SEG. The landfast sea ice season in our focus fjords is quite short, extending on average only ~2-4 months, and including substantial variability. Because of the fjord connections to marine-terminating glaciers, however, solid glacial ice creates a potential alternative ice platform on the fjord surfaces that can complement the short sea ice season. Challenges remain in automating this type of surface classification, and we discuss how our manual digitization work can be leveraged to support other ongoing efforts to create machine learning surface identification algorithms.

We give a preliminary report on results of detecting low-frequency earthquakes (LFs) occurring at Mt. Fuji, Japan, using the matched filter method (MF method: e.g., Peng & Zhao, 2009). LFs have been observed in the depth 10-25km beneath Mt. Fuji (Hamada, 1981; Ukawa et al, 2005). These LFs seem to occur at an almost constant rate at all times, but it may become remarkably active as in the fall of 2000 (Yoshida et al., 2006). Because it is considered that the activity of LFs is associated with behavior of magmatic fluid at depth (e.g., Nakamichi et al., 2003), an investigation into the relationship between LFs and volcanic activity (e.g., Harada et al., 2010) is important. Understanding of the details of LFs activity is the first step for this investigation. Here, a system using the MF method for detecting LFs at Hakone volcano, Japan (Yukutake, 2017; Yukutake et al., 2019), was modified to be applicable to the detection of LFs at Mt. Fuji. Then, this was applied to continuous seismic record at seismic stations around Mt. Fuji during the period of 2012-2020. Next, the template waveforms of LFs were prepared on the basis of the earthquake catalog maintained by the Japan Meteorological Agency (JMA). Then, the cross-correlation analysis was conducted between the template waveforms and the seismic records. Finally, a catalog of LFs, obtained by using the MF method, was created. Using this catalog, we confirmed that LFs in 2012-2020 occurred at an almost constant rate, and that this is also true for LFs included in the JMA catalog. However, our case shows that LFs occurred at a rate of about 1,250 per year, which is about 10 times higher than that shown for the JMA case (a rate of about 125 per year). It was also confirmed that the larger LFs tend to have fewer numbers and smaller LFs tend to have more numbers, again a feature found by using the JMA catalog. Our research is underway, and tackling challenges such as selection of appropriate template waveforms of LFs, correction of magnitude estimate, and extension of analysis period will improve our results, which will be reported in the presentation.

One of the prominent effects of space weather is the variation of electric currents in the magnetosphere and ionosphere, which can cause localized, high amplitude Geomagnetic Disturbances (GMDs) that disrupt ground conducting systems. Because the source of localized GMDs is unresolved, we are prompted to model these effects, identify the physical drivers through examination of the model we use, and improve our prediction of these phenomena. We run a high-resolution configuration of the Space Weather Modeling Framework (SWMF) to model the September 7, 2017 event, combining three physical models: Block Adaptive Tree Solar wind Roe Upwind Scheme (BATS-R-US), an ideal magnetohydrodynamic model of the magnetosphere; the Ridley Ionosphere Model (RIM), a shell ionosphere calculated by solving 2-D Ohm’s Law; and the Rice Convection Model (RCM), a kinetic drift model of the inner magnetosphere. The configuration mirrors that which is used in Space Weather Prediction Center (SWPC) operations; however, the higher grid resolution can reproduce mesoscale structure in the tail and ionosphere. We use two metrics to quantify the success of the model against observation. Regional Station Difference (RSD) is a metric that uses dB/dt or geoelectric field to pinpoint when a single magnetometer station records a significantly different value than others within a given radius, indicating a localized GMD. Regional Tail Difference (RTD) performs the same calculation using relevant variables in the magnetosphere at points that map down along field lines to the magnetometer station locations on the ground. We theorize two distinct causes of RSD, the first being small-scale structure in the tail and the second being station field lines mapping to spatially separated locations in the tail. We examine the differences between RSD spikes that we can reproduce in the model and those that we cannot. We categorize spikes by cause of localized GMDs to examine model capability for each theorized cause. We investigate the improvements in our model when we switch from empirical specification of ionosphere conductance to a physics-based one, MAGNetosphere-Ionosphere-Thermosphere (MAGNIT) Auroral Conductance Model. For small-scale effects we cannot reproduce, we explore the deficiencies in our model.

Geophysical surveys are efficient ways to obtain information on areas that are promising for geothermal energy. One of the geophysical techniques commonly used is the magnetic method, which is useful to detect shallow structures and changes in magnetization due to processes related to geothermal activity, such as faulting and hydrothermal alterations. Despite the richness of available geothermal resources in South America and Ecuador, the use of these resources for electricity production is very limited. Chachimbiro, in northern Ecuador, is one of the potential sites for developing a geothermal power plant. Our objective is to provide complementary magnetometry data to improve the existing model of the geothermal area. We performed high resolution ground magnetometry survey of ~30 m spacing around the prospective drilling area in order to better understand the shallow structures above the reservoir. We also performed two additional survey lines with ~5 m spacing across possible fault locations. After necessary data reductions the magnetic anomaly map was compared with a digital elevation model and a geological map of the area. This helped to understand the distribution of the anomalies and their relation with the presence of high magnetic susceptibility materials, hydrothermal alterations and topography. Major anomalies observed in the magnetic profiles were compared with forward fault models, allowing us to distinguish topographic from fault effects. We then compared our new magnetometry results with previous geophysical models of the Chachimbiro geothermal system. The large long-wavelength negative anomaly on the Northeast side of the survey area seems to coincide with the suggested location of the clay cap, and can therefore be used to improve the existing models. The new magnetic exploration of Chachimbiro therefore shows the usefulness of this method to locate magnetic anomalies related to faulting and hydrothermal alterations.

Solar-Induced Chlorophyll Fluorescence (SIF) is a powerful proxy for gross primary productivity (GPP) in Boreal ecosystems. However, SIF and GPP are fundamentally different quantities that describe distinct, but related, physiological processes. Recent work has highlighted non-linearities between SIF and GPP at finer spatial (leaf- to canopy- level) and temporal (half-hourly) scales. Therefore, questions have arisen about when, where, and why SIF is a good proxy for GPP and what the potential sources for divergence between the two are. The goal of this study is to answer two specific questions: 1) At what temporal scale is SIF a good proxy for GPP and 2) What are the predominant physical and ecophysiological drivers of nonlinearity between SIF and GPP in boreal ecosystems? We collected tower-based measurements of SIF (and other common vegetation indices) with PhotoSpec (a custom spectrometer system) and eddy-covariance GPP data at a 30-minute resolution at the Southern Old Black Spruce Site (SOBS) in Saskatchewan, CA. We applied a combination of statistical and machine learning approaches to disentangle the influence of structural/illumination effects and ecophysiological variations on the SIF signal. Our results show that at a high temporal resolution (half-hourly), SIF and GPP are predominantly dependent on photosynthetically active radiation (PAR). Therefore, the non-linear light response of GPP drives non-linearity between SIF and GPP. Additionally, canopy structure and illumination effects become important to the SIF signal at high temporal resolutions. At the seasonal timescale, SIF and GPP exhibit co-varying responses to PAR, even when accounting for changes in canopy structure. We attribute changes in the light responses of SIF and GPP to sustained photoprotection over winter which co-varies with changes in temperature. Finally, we show that the relationship between SIF and GPP has a seasonal dependence caused by small differences between the light use efficiencies of fluorescence and photosynthesis. Accounting for this seasonally variable relationship will improve the use of SIF as a proxy for GPP.

Though Venus’s atmospheric conditions and composition have been directly measured, the composition of the Venus lower atmosphere near the surface is generally still poorly known. It was extrapolated from observational data at other altitudes by assuming the constancy of elemental composition without condensation (Krasnopolsky 2007). Both in-situ measurement and remote-sensing observations reveals the most abundant components that exceed the mixing ratio of 10-4 to be CO2, N2, and SO2 (Bezard & de Bergh 2007, JGR 112, E04S07). Water and formation of photochemical H2SO4 — and condensation of cloud-forming H2SO4 — is only important at higher altitudes (Krasnopolsky 2012, Icarus 191, 25). In this work, the balancing of chemical-gravitational-thermal diffusive potentials for the ternary mixture of CO2, N2, and SO2, which represent the neutral Venusian lower atmosphere near the surface, is addressed to obtain the composition grading and to evaluate the tendency toward supercritical density-driven separation of CO2 and N2 (Lebonnois & Schubert 2017, Nat. Geosci. 10, 473). Even though dynamic atmospheric systems, including advective mixing, are more realistic, the static cases evaluated in this work provide stationary states where every dynamic process would eventually proceed to. Hence, our modeling is of a limiting case of the systems of interest, which could help explain some indications of compositional grading. The CRYOCHEM equation of state, which has been successfully applied in describing phase equilibria of Titan’s atmosphere and the surface liquid (Tan & Kargel 2018, Fluid Phase Equilib. 458, 153), as well as that involving solid phases on Pluto’s surface (Tan & Kargel 2018, MNRAS, 474, 4254), is used in this work on the supercritical Venus’s lower atmosphere. In the absence of direct measurement of composition of the lower atmosphere, as well as no lab evidence of CO2 and N2 separation under Venusian surface conditions (Lebonnois et al. 2020, Icarus 338, 113550), the results from this study may at least introduce some new concepts that would entail some tendency for molecular fractionation.

Uturuncu volcano in southern Bolivia is a member of a distinctive class of volcanoes – systems that show unrest despite not having erupted in the Holocene. Uturuncu has not erupted in 250 kyr, but has been deforming (uplift with a moat of subsidence) for several decades, along with seismic swarms and active, sulfur-encrusted fumaroles. Our work builds on previous geophysical imaging at Uturuncu by jointly analyzing multidisciplinary datasets, focusing on imaging the shallow (<15 km depth below surface) structure of the system with geophysical and geochemical data. Whereas previous research pointed to andesite melt at depths >15 km depth, results were ambiguous as to what proportions of melts vs. brines are present at shallower depths. Identifying fluids (melt, brine, etc.) and structures at shallow depths is key for evaluating the hazard potential of the volcano and understanding the source of the unrest. We present new results from gravimetry, seismology (hypocenter relocation, seismic velocity and attenuation tomography), gas geochemistry, and InSAR observations. The results point to an extensive and active hydrothermal system extending ~20 km laterally and ~10 km vertically from Uturuncu, with possible connections at depth to the deeper magmatic system. A combined view of the new density, seismic velocity and attenuation models, and the existing resistivity model is crucial for revealing key features of the hydrothermal system: a vapour-rich conduit beneath Uturuncu (low resistivity/high attenuation column extending from 1.5 to 12.5 km depth), an extensive alteration zone surrounding Uturuncu (complex zone of annular shaped anomalies surrounding Uturuncu from 1.5 to 12.5 km depth), and a possible zone of sulfide deposition just below the western flank of Uturuncu at 1.5 km depth (high density/low resistivity/high attenuation). High fluxes of diffuse CO2 degassing at sub-magmatic temperatures and a small area directly above a low resistivity anomaly subsiding from 2014 to 2017 show that the hydrothermal system is currently active. Analyzed jointly, this multidisciplinary data set suggests that current activity within the shallow structure at Uturuncu is dominated by hydrothermal, rather than magmatic processes.

Oceanic basalts provide an invaluable window into evolutionary processes governing mantle spatial and temporal chemical heterogeneity. Ocean island basalts (OIBs) and enriched mid-ocean ridge basalts (E-MORBs) are powerful tracers of mantle melting and crust-mantle recycling processes. Whether the elemental and isotopic variations observed in both E-MORBs and OIBs are derived from similar mechanisms, however, remains under debate. Investigating compositional differences between E-MORBs and OIBs is a simple approach to constrain their origins, a technique for which machine learning classification algorithms are optimal. Here we implemented a novel machine learning approach complemented by mantle component mixing models to highlight compositional differences between E-MORBs and OIBs and further investigate their petrogenesis (data sourced from GEOROC database and Gale et al., 2013). Considering Random Forest-based Gini indexes, elements sensitive to pressure and degree of melting (FeO, TiO2, Lu, and Sr) were identified as the best discriminators between E-MORBs and OIBs. Our Gaussian process classification algorithm successfully classified OIBs and E-MORBs better than 97% of the time when considering 1) Sr & FeO and 2) TiO2 & Lu. The probabilistic nature of Gaussian process modeling permitted calculation of new quantitative discriminant diagrams rooted in probability (Sr vs. FeO and TiO2 vs. Lu). Complementary trace element modeling yielded compositionally similar E-MORB and OIB sources with moderately incompatible element enrichments in the OIB source due to the influence of recycled oceanic crust (Prytulak & Elliott, 2007). Our source compositions are consistent with a simple, joint model for E-MORB and OIB petrogenesis after Donnelley et al. (2014): low-degree partial melts of subducted slabs metasomatize the depleted mantle producing a re-fertilized mantle (RM). RM is randomly sampled at mid-ocean ridges to produce E-MORB, while upwelling plumes sample both RM and recycled oceanic crust, yielding OIB. References: Donnelly et al. (2004). Earth and Planet. Sci. Lett., 226(3–4), 347–366. Gale et al. (2013). Geochem., Geophys., Geosyst., 14(3), 489–518. Prytulak & Elliott (2007). Earth and Planet. Sci. Lett., 263(3–4), 388–403.

Efficient use of available water resources is key to sustainable viticulture management in California (CA) and other regions with limited water availability in the western US and abroad. This requires remote and frequent field-scale information on vineyard water status. Though the Sentinel-2 sensors offer good spatial (10-60m) and temporal (~5 days) coverages, their utility in monitoring vineyard evapotranspiration (ET) has not been considered viable primarily due to the lack of a thermal band. However, recently, a new spectral-based Shuttleworth Wallace (SW) ET model, which uses a contextual framework to determine dry and wet extremes from the Sentinel-2 (SW-S2) surface reflectance data, has shown promise when tested over a single GRAPEX (Grape Remote-sensing Atmospheric Profile and Evapotranspiration eXperiment) site in CA. However, current knowledge on its applicability across a climate gradient in CA with different topography, soils, trellis design and vine variety is lacking. Moreover, how the selection of modeling domain and meteorological forcing data influence model output is limited. Consequently, this presentation expands the evaluation of the SW-S2 model across multiple domains and meteorological inputs covering all three GRAPEX vineyard sites spanning a north to south climate gradient over three recent growing seasons (2018-2020). In comparison with flux tower observations, the size of the modeling domain and the source and quality of meteorological forcing data on the performance of the SW-S2 model as well as application to the three different vineyard study sites will be presented. Future research on merging output from more-frequent spectral and less-frequent thermal-based ET models to reduce latency in ET monitoring of California vineyards will also be discussed.

The Southwest Watershed Research Center (SWRC) of the United States Department of Agriculture-Agricultural Research Service has been conducting arid and semiarid (dryland) watershed research since 1953. This included establishment and continuous operation of the Walnut Gulch Experimental Watershed (WGEW) in southeast Arizona. The 149 km2 ephemeral watershed is one of the most densely instrumented dryland research catchments in the world with a drainage area greater than 10 km2. This instrumentation captures many aspects of the hydrological cycle including how precipitation is partitioned into soil moisture, runoff and evapotranspiration and its subsequent effects on sediment transport, vegetation productivity, carbon sequestration, and groundwater recharge. The long-term, high-resolution record of observations on the WGEW enables understanding of the mean and variability of the hydrological processes, not well characterized with shorter term records, that fail to capture the large variability common to dryland regions. This presentation will highlight trends in temperature, precipitation, and runoff over the WGEW observation period. Additional research findings made by the SWRC and collaborators on erosion; plant productivity and carbon sequestration; the facilitation of soil water redistribution by plant roots; and groundwater recharge will also be presented.

Enhanced water management systems depend on accurate estimation of hydraulic properties of subsurface formations. This is while hydraulic conductivity of geologic formations could vary significantly. Herein, we studied an intensively managed area located in the Upper Sangamon Watershed in Central Illinois, U.S.A., and generated 2D maps of hydraulic conductivity over a large-scale region with quantified uncertainties in different depth layers. In doing so, we made use of low cost, small-scale measurements obtained from the Electrical Earth Resistivity together with more accurate, more expensive pumping tests in a calibration framework based on Kriging. We offered a cost-effective approach to reliably characterize the hydraulic conductivity properties in under-sampled sites and can be particularly used in obtaining large-scale parameter maps for a region using small-scale measurements in an efficient way. This work also includes optimal sensor placement, where the best locations for future data collection are selected by considering the current confidence levels estimated by the Kriging model, which is related to the expected value of information from future sensor data. Our approach is based on the Bayesian experimental design, which selects the best locations, out of a set of candidate locations, based on the value of information that each location is expected to offer.

Deep earthquakes in the lower continental lithosphere - the lower crust and uppermost mantle - are frequently too poorly located in depth to be definitively labelled as having occurred above or below the Mohorovičić discontinuity (Moho; base of the crust). Our Sn/Lg methodology utilizes two regional seismic waves to determine the depth of an earthquake relative to the Moho: Sn and Lg waves, which are phases that propagate through reflections in the mantle lid and the crust, respectively. Therefore, an analysis of Sn and Lg waves can provide a robust understanding of an earthquake’s depth relative to the Moho. We present our Sn/Lg analysis through reduced-velocity record sections, which show Sn and Lg energy in the waveform, allowing measurements of RMS amplitudes, and maps of ray-paths and Sn/Lg amplitude ratios which allow us to visualize the propagation of Sn and Lg in all directions. We demonstrate the efficacy of our approach by applying it to a well-known upper-mantle earthquake in Wyoming and a shallow earthquake with a similar epicenter. We then use our method to study other deep-crustal/upper-mantle earthquakes in North America. A cluster of earthquakes with reported depths from 10–50 km ± ~10 km, spans the border between Alberta, Canada and Montana, U.S.A. where the crustal thickness increases from ~30 km in the SW to ~45km in the NE. For an earthquake occurring in the crust-mantle transition zone, the dipping Moho should tend to block Sn (decrease the Sn/Lg ratio) in the direction of crustal thickening, and tend to block Lg (increase the Sn/Lg ratio) in the direction of crustal thinning. We studied seven earthquakes of magnitude>2.5. A m=2.7 earthquake, previously reported (USGS PDE) to be at 50±10 km where CRUST 1.0 shows a 49 km Moho depth, and a m=3.5 reported at 38 km depth above a nominal 43 km Moho, both show much stronger Sn/Lg ratios than earthquakes with nearby epicenters at nominal dots of 15 and 21 km. Hence the “50-km” and the “38-km” earthquakes must occur in the upper mantle, or so close to the upper mantle as to preferentially excite Sn. The better-recorded of these earthquakes also shows clear evidence of Sn enhancement along azimuths into regions of thinner crust.

Drought is a recurring and extreme hydroclimatic hazard with serious impacts on agriculture and overall society. Delineation and forecasting of agricultural and meteorological drought are essential for water resource management and sustainable crop production. Agricultural drought assessment is defined as the deficit of root-zone soil moisture (RZSM) during active crop growing season, whereas meteorological drought is defined as subnormal precipitation over months to years. Several indices have been used to characterize droughts, however, there is a lack of study focusing on comprehensive comparison among different agricultural and meteorological drought indices for their ability to delineate and forecast drought across major climate regimes and land cover types. This study evaluates the role of RZSM from Soil Moisture Active Passive (SMAP) mission along with two other soil moisture (SM) based indices (e.g., Palmer Z and SWDI) for agricultural and meteorological drought monitoring in comparison with two popular meteorological drought indices (e.g., SPEI and SPI) and a hybrid (Comprehensive Drought Index, CDI) drought index. Results demonstrate that SM-based indices (e.g., Palmer Z, SMAP, SWDI) delineated agricultural drought events better than meteorological (e.g., SPI, SPEI) and hybrid (CDI) drought indices, whereas the latter three performed better in delineating meteorological drought across the contiguous USA during 2015–2019. SM-based indices showed skills for forecasting agricultural drought (represented by end-of-growing season gross primary productivity) in the early growing seasons. The results further confirm the key role of SM on ecosystem dryness and corroborate the SM-memory in land-atmosphere coupling.

The Gandalf Staff is a mobile tool designed to be a flexible device supporting crewed and uncrewed operations on the lunar surface. The core of the device is a 24v battery with communications and data storage systems. Initial optional components supporting crewed Extra-Vehicular Activity (EVA) include a LiDAR and 360˚ camera. These provide 3D mapping of the traverse for documentation, and to aid future planning. The mapping also creates outreach opportunities for the public to “stand beside” the astronaut in Virtual Reality (VR). The staff provides external lighting for field site illumination in the south polar region low sun angle environment. Navigation instruments for crew position determination with Lunar Search and Rescue (LunaSAR) are also included. The staff itself can be used as a walking aid or as a splint for Incapacitated Crew Rescue (ICR). As a stand-alone device, the staff operates as a long duration untended science platform collecting environmental data and sending it to a lunar base station. The stand-alone mode requires connection to an auxiliary power source (e.g. solar array) and energy storage system (e.g. battery), so it could become an electrical recharging station. To make rapid progress in the 1st year, and also to demonstrate innovative project management techniques, NASA guided a private industry partner, T STAR, in leading Capstone Engineering student teams at Texas A&M University (TAMU) for proof-of-concept development and testing. These teams developed the power system and demonstrated successful integration of LiDAR, WiFi communications, and external lighting subsystems. Another industry team at Jacobs Technology prototyped a tripod to hold the staff upright. For the 2nd year (FY’22), NASA will again collaborate with partners to prototype enhanced power and lighting concepts. Year 2 will also add new capability for LunaSAR and geophysical science instrumentation using a heat probe. The heat probe is based upon Apollo heritage but modified to measure subsurface volatile ice regimes at the Artemis landing site. Components of the Gandalf Staff can be developed, tested, and deployed independently, or on the integrated staff, rovers, or utility trailers. The project supports crew safety, lunar sample curation, mission science, and public outreach goals of NASA.

Bangladesh is an extremely flood-prone country due to its geographical location at the downstream end of the Ganges, Brahmaputra and Meghna (GBM) river basin. Flood destroys agricultural products of large areas and causes loss of lives and damage to infrastructures. Heavy rainfall during the monsoon season is the major cause of flooding in this region which occurs almost every year. However, the lack of observations of rainfall in the upper catchment areas outside Bangladesh makes flood forecasting challenging in this region. In addition, errors in rainfall forecasts and lack of high-resolution bathymetry and topographic data put major constraints to flood forecasting in Bangladesh through hydrologic and hydrodynamic models. Currently Flood Forecasting and Warning Centre (FFWC) of Bangladesh Water Development Board (BWDB) is producing short-range flood forecasts with a lead time of up to three days. However, medium-range (3 to 5 days) forecasts are crucial for reducing flood-related losses as they provide more time for decision making and preparation compared to short-range forecasts. In this study, a flood forecast model based on Artificial Neural Network (ANN) has been developed for the Kushiyara river which is one of the major rivers of the northeastern region of Bangladesh. Rainfall data from the fifth generation European Centre for Medium-Range Weather Forecasts Reanalysis (ERA5), daily Terrestrial Water Storage (TWS) from the Global Land Data Assimilation System with the Gravity Recovery and Climate Experiment Data Assimilation (GRACE-DA) and daily Surface Soil Moisture data from Soil Moisture Active Passive (SMAP) have been used as input to the model. The model shows reasonable accuracy in forecasting the water level of the Kushiyara river at Sheola station with a lead time of up to seven days. For 1-day lead time, the correlation coefficient (R) between the observed and simulated water levels is 0.97. The performance of the model is also promising for a medium-range forecast (R=0.93 for 7-day lead time). This study indicates that the release of daily GRACE gravity field solutions in near-real-time may enable us to forecast and monitor high volume flood events in this region.

Geologists need help classifying microscope rock images of sigma clasts; a type of mantled porphyroclasts widely used as kinematic indicators in rocks. Knowledge about the shear sense of sigma clast during formation (either CCW or CW shearing) gives insights into rock formation history. This work reports on early investigation of machine learning techniques for automatic detection and classification of sigma clasts and their rotation from photomicrographs. Convolutional Neural Networks (CNNs) are used to extract and leverage defining features of sigma clasts, such as shape, color, texture, and tail direction to improve accuracy. We leverage existing models that are pre-trained on very large collections of images, and use transfer learning techniques to apply them to microstructure images. We used YOLOv3 to identify different sigma clasts in a given image. We also experimented with other large pre-trained models such as ResNet50, VGG19, InceptionV3 with two additional layers trained specifically on our dataset. In order to facilitate exploration of different models with different settings, we are developing a computational experimentation environment to visualize different CNN network layers, classification heatmaps, and comparative metrics. Finally, since models perform better when more data are available, we are developing a web application to collect additional data from geoscientists and incentivize their participation in open science. The website allows researchers to upload images of rock microstructures, showing them the classification of the images based on the best models available, and allows them to correct any errors which can be used to improve the models.

Regional-scale, continuously-operating Global Navigation Satellite System (GNSS) networks are a powerful tool to monitor plate motion and surface deformation. Since their inception, their size, density, and length of observation record have steadily increased throughout the world. Simultaneously, researchers have had to write accompanying software to enable the analysis (especially the decomposition) of the ever-increasing amount of available timeseries in an efficient way. These codes and respective studies have individually set standards for different subsets of the following desirable qualities: portability (between locations), speed (code runtime), automation (avoiding or simplifying manual inspection of each station), use of spatial correlation (exploiting the fact that stations experience common signals), availability (open source), and documentation (of the usage and underlying methods). In this study, we present the DISSTANS Python package, which aims to combine the aforementioned achievements in a single software by offering generic (therefore portable), parallelizable (fast) methods that can exploit the spatial structure of the observation records in a user-assisted, semi-automated framework, including uncertainty propagation. The code is open source, includes an application interface documentation as well as usage tutorials, is easily extendable, and is based on the previously published and validated method of Riel et al. (2014). We also present two case studies to validate our code, one using a synthetic dataset and one using real GNSS network timeseries.