Precipitation flag (precipitating or not; stratiform or convective) is a key parameter for us to make betterretrieval of precipitation characteristics as well as to understand the cloud-precipitation physicalprocesses. The Global Precipitation Measurement (GPM) Core Observatory’s Microwave Imager (GMI)and Dual-Frequency Precipitation Radar (DPR) together provide ample information on globalprecipitation characteristics. As an active sensor in particular, DPR provides an accurate precipitationflag assignment, while passive sensors like GMI were traditionally believed not to be able to tell apartprecipitation types. Using collocated precipitation flag assignment from DPR as the “truth”, this project employs machinelearning models to train and test the predictability and accuracy of using passive GMI-only observationstogether with ancillary atmosphere information from reanalysis. Precipitation types are classified intothe following classes: convective, stratiform, convective-stratiform mixed, no precipitation, and otherprecipitation. Sub-sampling with different probabilities is employed to construct a balanced trainingdataset. A variety of classification algorithms are tested, including Support Vector Machines, NaiveBayes, Random Forests, Gradient Boosting, and Neural Networks (Multilayer Perceptron Network), andtheir results are evaluated and compared. The trained model has ~ 85% of prediction accuracy for everytype of precipitation. High-frequency channels (166 GHz and 183 GHz channels) and 166 GHzpolarization difference are found among the most important factors that contribute to the modelperformance, which shed light on future instrument channel selection.

Accurate descriptions of natural fault surfaces and associated fault rocks are important for understanding fault zone processes and properties. Slickensides–grooved polished surfaces that record displacement and wear along faults– develop measurable roughness and characteristic microstructures during fault slip. We quantify the roughness of natural slickensides from three different fault surfaces by calculating the surfaces power spectra and height distributions and analyze the microstructures formed above and below the slickensides. Slickenside surfaces exhibit anisotropic self-affine roughness with corresponding mean Hurst exponents in directions parallel– 0.53±0.07– and perpendicular –0.6±0.1– to slip, consistent with reports from other fault surfaces. Additionally, surfaces exhibit non-Gaussian height distributions, with their skewness and kurtosis roughness parameters having noticeable dependence on the scale of observation. Below the surface, microstructural analyses reveal that S-C-C’ fabrics develop adjacent to a C-plane-parallel principal slip zone characterized by a sharp decrease in clast size and a thin (≤100 µm) nanoparticulate-rich principal slip surface (PSS). These microstructures are present in most analyzed samples suggesting they commonly form during slickenside development regardless of lithology or tectonic setting. Our results suggest that 1) PSS likely arise by progressive localization along weaker oriented fabrics 2) deformation along PSS’s is energetic enough to comminute the rocks into nanometric grains, and 3) fault geometry can be further characterized by studying the height distributions of fault surfaces, which are likely to impact stress distributions and frictional responses along faults.

Aerosol-cloud interactions are both uncertain and important in global and regional climate models, and especially in the southeast Atlantic Ocean. This uncertainty in the region is largely due to two correlated 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 and LASIC field campaigns in August-October of 2016 through 2018. We compare aerosol and cloud properties to measure and improve model performance and expand upon observational findings of aerosol-cloud effects. Relevant comparison variables include aerosol number concentration, mean particle diameter and spread, CCN activation tendency, hygroscopicity, and cloud droplet number concentrations. Specifically, our approach is to analyze colocated model data along flight tracks to resolve aerosol-cloud interactions. Within and between single-day flights, there is high spatiotemporal variability that can get lost to large-scale averaging analyses. We have found that CCN is substantially under-represented in the model compared to observations. For a given aerosol number concentration, size, supersaturation and hygroscopicity, the model will consider fewer particles as CCN than observations indicate. We plan to explore this result further, diagnosing the model-observation differences more consistently and updating the model with more physically accurate values of aerosol size, concentration, or hygroscopicity based on observations. We will also intercompare multiple instrument platforms involved with the ORACLES and LASIC campaigns. With improved small-scale aerosol-cloud interactions, this work also shows promise to substantially improve that representation in climate models.

Trans–ionospheric high frequency (HF) signals experience a strong attenuation following a solar flare, commonly referred to as Short–Wave Fadeout (SWF). Although solar flare-driven HF absorption is a well-known impact of SWF, the occurrence of a frequency shift on radio wave signal traversing the lower ionosphere in the early stages of SWF, also known as “Doppler Flash”, is newly reported and not well understood. Some prior investigations have suggested two possible sources that might contribute to the manifestation of Doppler Flash: first, enhancements of plasma density in the D and lower E regions; second, the lowering of the reflection point in the F region. Observations and modeling evidence regarding the manifestation and evolution of Doppler Flash in the ionosphere are limited. This study seeks to advance our understanding of the initial impacts of solar flare-driven SWF. We use WACCM-X to estimate flare-driven enhanced ionization in D, E, and F-regions and a ray-tracing code (Pharlap) to simulate a 1-hop HF communication through the modified ionosphere. Once the ray traveling path has been identified, the model estimates the Doppler frequency shift along the ray path. Finally, the outputs are validated against observations of SWF made with SuperDARN HF radars. We find that changes in refractive index in the D and lower E regions due to plasma density enhancement are the primary cause of Doppler Flash.

In this study, we forecast hourly relativistic (>2 MeV) electron fluxes at geostationary orbit for the next 72 hours using a deep learning model. For this we consider three deep learning methods, such as multilayer perceptron (MLP), LSTM, and sequence-to-sequence based on LSTM. The input data of the model are solar wind parameters (temperature, density and speed), interplanetary magnetic field (|B| and Bz), geomagnetic indices (Kp and Dst), and electron fluxes themselves. All input data are hourly averaged ones for the preceding 72 consecutive hours. We use electron flux data from GOES-15 and -16, and perform cross-calibration to match the two data. Total period of the data is from 2011 January to 2021 March (GOES-15 data for 2011-2017 and GOES-16 data for 2018-2021). We divide the data into training set (January-August), validation set (September), and test set (October-December) to consider the solar cycle effect. Our main results are as follows. First, the MLP model, which is the best, successfully predicts hourly electron fluxes for the next 72 hours. Second, root-mean-square error (RMSE) of our model is from 0.18 (for 1h prediction) to 0.68 (for 72h prediction), and prediction efficiency (PE) is from 0.97 to 0.53, which are much better than those of the previous studies. Third, our model well predicts both diurnal variation and sudden increases of electron fluxes associated with fast solar winds and interplanetary magnetic fields. Our study implies that the deep learning model can be applied to forecasting long-term sequential space weather events.

The new dual-frequency radar satellite ASAR-ISRO provide simultaneously and for the first time a wide wavelength spectrum which is critical to discriminate surface roughness based on different backscattering characteristics. Here we use such dual L-band and S-band airborne SAR system to characterize and map various volcanic areas in the Northern Cascades through their backscattering properties. Mapping volcanic flows (lava flows, pyroclastic currents, lahars) is vastly improved by using backscattering as a metric of surface roughness. Various types of volcanic flow deposits and surface textures are distinguished by their roughness measured with radar systems. The ability of radar systems to distinguish volcanic flow textures, represented by roughness, is a key factor in understanding the processes and timescales of flow emplacement. For instance, transition from pahoehoe to aa lava flows is associated with change in flux and steepness of topographic gradient. Therefore, lava textures are essential data for calibrating and improving lava flow emplacement codes. Similarly, the textures of high velocity, and more deadly, pyroclastic currents and lahars change along their flow paths, also revealing critical data about the mechanisms of flow emplacement. As with lava flows, we use the ASAR-ISRO L+S SAR system to characterize the run-out distances of pyroclastic currents and lahars where large numbers of blocks accumulate in such deposits (places where flow momentum was lost, whether due to friction or break-in-slope and vast quantities of blocks accumulate in a relatively small area compared to the total area inundated by the flow). Mapping volcanic flow textures in a variety of volcanic terrains will provide clues about modes and rates of emplacement, and change in these through time, in a way that is simply unavailable by traditional geologic mapping. The next generation of volcanic flow maps, used for hazard assessment, will rely on radar data to delineate these textures.

My earth science focused PhD research included the analysis of annual fluorescent laminae in cave stalagmites, under the supervision of karst hydrologist Peter Smart (Bristol, UK). At the time, the source of this fluorescence was uncertain, opening new research opportunities characterizing what is now understood to be soil-derived, water-soluble fluorescent dissolved and colloidal organic matter that is transported to the cave by vadose zone percolation waters. Aquatic organic matter fluorescence research benefitted at this time from significant laboratory analytical advances that were commercially driven, including Sony’s Blu-ray technology and the use of fluorescent labelling in the biomedical sciences. Faster analyses at increasingly higher energy excitation energies opened opportunities for novel fingerprinting of organic matter in hydrological science, including landfill leachates and sewage contamination. Today, hand-held fluorescence sensors can instantaneously determine microbial water quality. And back in the field of speleothem (cave deposit) science, annual geochemical laminae are now recognized to be widely preserved in regions where there is a seasonality in recharge, providing a precise chronology for the stalagmite paleoenvironmental archive. Currently, we are utilizing this precise chronology to generate high-resolution fire history records, where the fire proxy is water-soluble ash-derived elements transported from soil to cave by vadose zone percolation waters.

The solution to high-speed-railway seismic source function is of great theoretical significance for the inversion of high-speed-railway focal mechanism, propagation law and energy evolution, which may provide a reference for the seismic exploration of rock mass structure in shallow ground. In this paper, the simulation signal characteristic is studied through MATLAB, by working out the Green function of high-speed ailway seismic source in near-field and proceeding forward calculation. Based on the comparative analysis of simulation signal and measured data, the Green function is verified. In addition, wth further analysis of vibration curves, frequency-spectrum diagrams and energy distribution maps, the results show that in the near-field of high-speed railway seismic source: The seismic signal is broadband, of which the vibration differences at each observation position in the horizontal direction indicate the propagation law of high-speed railway seismic wave, while the energy attenuation characteristic can be reflected by that in the vertical direction; There are clear discrete spectral lines in high-speed railway seismic wave, of which the feature frequency varies from 60Hz to 370Hz, affected by the propagation distance; The spacial energy distribution of high-speed railway seismic wave radiation in near field depends on the frequency response property at the seismic source, and the frequency response property shows significant difference with the frequency ranging from 40Hz to 500Hz.

We present probabilistic centroid-moment tensor solutions inferred from the combination of Hamiltonian Monte Carlo sampling and a 3-D full-waveform inversion Earth model of the Japanese archipelago. While the former provides complete posterior probability densities, the latter allows us to exploit waveform data with periods as low as 15 s. For the computation of Green’s functions, we employ spectral-element simulations through the radially anisotropic and visco-elastic model, leading to substantial improvements of data fit compared to layered models. Focusing on Mw 4.8 - Mw 5.3 offshore earthquakes with a significant non-double-couple component, we simultaneously infer the centroid location, time and moment tensor without any a priori constraints on the faulting mechanism. Furthermore, we perform the inversions across several period bands, varying the minimum period between 15 s and 50 s. Accounting for 3-D Earth structure at shorter periods can increase the double-couple component of an event, compared to the GCMT solution, by tens of percent. This suggests that at least some of the non-double-couple events in the GCMT catalog might result from unmodeled Earth structure. We also observe that significant changes in source parameters, and the double-couple component in particular, may be related to only small waveform changes, thereby accentuating the importance of a reliable Earth model. Posterior probability density distributions become increasingly multimodal for shorter-period data that provide tighter constraints on source parameters. This implies, in our specific case, that stochastic approaches to the source inversion problem are required for periods below ~ 20 s to avoid trapping in local minima.

When organic peat soils are sufficiently dry, they become flammable. In Southeast Asian peatlands, widespread deforestation and associated drainage create dry conditions that, when coupled with El Niño-driven drought, result in catastrophic fire events that release large amounts of carbon and deadly smoke to the atmosphere. While the effects of anthropogenic degradation on peat moisture and fire risk have been extensively demonstrated, climate change impacts to peat flammability are poorly understood. These impacts are likely to be mediated primarily through changes in soil moisture. Here, we used neural networks (trained on data from the NASA SMAP satellite) to model soil moisture as a function of climate, degradation, and location. The neural networks were forced with regional climate model projections for 1985-2005 and 2040-2060 climate under RCP8.5 forcing to predict changes in soil moisture. We find that reduced precipitation and increased evaporative demand will lead to median soil moisture decreases about half as strong as those observed during recent El Niño droughts. Such reductions may be expected to accelerate peat emissions. Our results also suggest that soil moisture in degraded areas with less tree cover may be more sensitive to climate change than in other land use types, motivating urgent peatland restoration. Climate change may play an important role in future soil moisture regimes and by extension, future peat fire in Southeast Asian peatlands.

We utilize the Planet Model Code [1] to model the probability of survival of Earth-like life on habitable worlds (to 3 billion years post-abiogenesis). The Planet Model Code [1] was originally created by Tyrrell [2] to investigate the chances of intelligent life. This necessitates stability in a rather narrow temperature range. Here we expand the focus to the survival of any life, including extremophiles, with greater tolerance for variations in temperature. Specifically, we investigate the relationship between the long-term survival probability of life and larger temperature variations. Keywords: Exoplanet; Habitable; Prokaryote; Matlab; Planet Model Code References: 1. T. Tyrrell, Planets model code, (Oct. 2020) https://doi.org/10.5281/zenodo.4081451. 2. T. Tyrrell, “Chance played a role in determining whether earth stayed habitable”, Commu- nications Earth & Environment 1, 1–10 (2020).

Estimating surface heat fluxes via direct covariance measurements or bulk formulae is observation-intensive and costly. We present a methodology whereby we estimate net surface heat fluxes as the difference between the depth-integrated heat tendencies and the depth-integrated horizontal heat exchanges in a hydrodynamic model. We calibrate the model to achieve a good representation of mixing and advection and then assimilate satellite sea-surface-temperature (SST) observations into the model at an eight-day scale. The SST data assimilation forces a good representation of observed temperatures and heat tendencies both at the surface and throughout the water column. We estimate the horizontal heat exchange directly from the model output and then infer the surface fluxes required to close the budget. When we apply this methodology to a model with prescribed surface heat fluxes and without data assimilation, we can recover the prescribed fluxes with an RMS error of ±10 Wm−2 and an r2 of 0.998. When we compare our results to those estimated using COARE bulk formulae with observations in western Long Island Sound, we find similarly good agreement.

Surface heterogeneities below the spatial resolution of thermal infrared (TIR) instruments result in anisothermality and produce emissivity spectra with negative slopes at longer wavelengths. Sloped spectra arise from an incorrect assumption of either a uniform surface temperature or a maximum emissivity during the temperature-emissivity separation of radiance data. Surface roughness and lateral mixing of differing sub-pixel surface units result in spectral slopes that are distinct, with magnitudes proportional to the degree of temperature mixing. Routine Off-nadir Targeted Observations (ROTO) of the Thermal Emission Imaging Spectrometer (THEMIS) are used here for the first time to investigate anisothermality below the spatial resolution of THEMIS. The southern flank of Apollinaris Mons and regions within the Medusae Fossae Formation are studied using THEMIS ROTO data acquired just after local sunset. At higher emission angles, differing relative proportions of rocky and unconsolidated surface units are observed. This produces a range of sloped TIR emission spectra dependent on the magnitude of temperature differences within a THEMIS pixel. Spectral slopes and wavelength-dependent brightness temperature differences are forward-modeled for a series of two-component surfaces of varying thermal inertia values. This creates a thermophysical model suggesting a local rock abundance 6 times greater than currently published results and four orders of magnitude more sensitive than those relying on nadir data High-resolution visible images of these regions indicate a mixture of surface units from boulders to dunes, providing credence to the model.

Oceanic lithosphere created at mid-ocean ridges (MOR) is shaped by a complex interplay of magmatic, tectonic and hydrothermal processes. Where melt budget decreases at (ultra)slow spreading rates, the tectonic mode changes from normal to detachment faulting. We investigate an endmember type of detachment faulting, the so-called “flip-flop” detachment mode, observed exclusively at almost amagmatic sections of ultraslow-spreading MORs (e.g. Southwest Indian Ridge 62°E to 65°E): Active detachments migrate across the ridge axis towards the hanging wall side until they are superseded by a new on-axis fault of opposite polarity. Using a steady-state temperature field, numerical experiments by Bickert et al. (EPSL, 2020, 10.1016/j.epsl.2019.116048) show that an axial temperature maximum is essential to trigger flip-flop faults by focusing flexural strain in the footwall of the active fault. However, ridge segments without a significant melt budget are more likely to be in a transient thermal state controlled, at least in part, by the faulting dynamics themselves. In our study we investigate (1) which processes have first order control on the thermal structure of the lithosphere, (2) their respective feedbacks on the mechanical evolution of the lithosphere, and (3) how they facilitate flip-flop detachment faulting. We present results of 2-D thermo-mechanical numerical modelling including serpentinization reactions and dynamic grain size evolution. The model features a novel form of parametrized hydrothermal cooling along fault zones as well as the thermal and rheological effects of periodic sill intrusions. We find that both hydrothermal cooling of the active fault zone and periodic sill intrusions in the footwall are essential for entering the flip-flop detachment mode. Hydrothermal cooling of the fault zone pushes the temperature maximum into the footwall, while intrusions near the temperature maximum further weaken the rock and facilitate the opening of new faults with opposite polarity. We conclude that this interplay of hydrothermal cooling and magmatic intrusions is required to produce the observed flip-flop faulting. Our model allows us to put constraints on the magnitude of both processes, and we obtain reasonable melt budgets and hydrothermal heat fluxes only if both are considered.

Dry deposition is an important ozone sink that impacts ecosystem carbon and water cycling. Ozone dry deposition in forests is regulated by vertical transport, stomatal uptake, and non-stomatal processes including chemical removal. However, accurate descriptions of these processes in deposition parameterizations are hindered by sparse observational constraints on individual sink terms. Here we quantify the contribution of canopy-atmosphere turbulent exchange and chemical ozone removal by soil-emitted nitric oxide (NO) to ozone deposition in a North-Italian broadleaf deciduous forest. We apply a multi-layer canopy exchange model to interpret campaign observations of nitrogen oxides (NOx=NO+NO2) and ozone exchange above and inside the forest canopy. Two state-of-science parameterizations of in-canopy vertical diffusivity, based on above-canopy wind speed or stability, do not reproduce the observed exchange suppressed by canopy-top radiative heating, resulting in overestimated dry deposition velocities of 10-19\% during daytime. Applying observation-derived vertical diffusivities in our simulations largely resolves this overestimation. Soil emissions are an important NOx source despite the observed high background NOx levels. Soil NOx emissions decrease the gradient between canopy and surface layer NOx mixing ratios, which suppresses simulated NOx deposition by 80% compared to a sensitivity simulation without soil emissions. However, a sensitivity analysis shows that the enhanced chemical ozone sink by reaction with soil-emitted NO is offset by increased vertical ozone transport from aloft and suppressed dry deposition. Our results highlight the need for targeted observations of non-stomatal ozone removal and turbulence-resolving deposition simulations to improve quantification and model representation of forest ozone deposition.

We develop a data assimilation scheme with the Icosahedral Non-hydrostatic Earth System Model (ICON-ESM) for operational decadal and seasonal climate predictions at the German weather service. For this purpose, we implement an Ensemble Kalman Filter to the ocean component as a first step towards a weakly coupled data assimilation. We performed an assimilation experiment over the period 1960-2014. This ocean-only assimilation experiment serves to initialize 10-year long retrospective predictions (hindcasts) started each year on 1 November. On multi-annual time scales, we find predictability of sea surface temperature and salinity as well as oceanic heat and salt contents especially in the North Atlantic. The mean Atlantic Meridional Overturning Circulation is realistic and the variability is stable during the assimilation. On seasonal time scales, we find high predictive skill in the tropics with highest values in variables related to the El Niño/Southern Oscillation phenomenon. In the Arctic, the hindcasts correctly represent the decreasing sea ice trend in winter and, to a lesser degree, also in summer, although sea ice concentration is generally much too low in both hemispheres in summer. However, compared to other prediction systems, prediction skill is relatively low in regions apart from the tropical Pacific due to the missing atmospheric assimilation. In addition, we expect a better fine-tuning of the sea ice and the oceanic circulation in the Southern Ocean in ICON-ESM to improve the predictive skill. In general, we demonstrate that our data assimilation method is successfully initializing the oceanic component of the climate system.

Origins of the interdecadal fluctuations of the ENSO-Indian summer monsoon (ISM) teleconnection strength measured by running window correlation coefficients have been much debated. The main question is whether it’s due mostly (i) to sampling errors, or, (ii) under internal variability decadal time scale drivers modulate it, and/or, (iii) it has undergone some forced change recently. A new statistical test does not detect any of (ii-iii) in observations. However, it does not mean that there could not be such effects, just that the data is insufficient. Large ensemble data sets are well suited to investigate these questions. Previously, (iii) was found to be small (even if possibly important) in the MPI-GE data set, and the situation is similar in the new CESM2-LE data set. This time we investigate (ii), too, and find that the decadal variability of e.g. the Dipole Mode Index (DMI*, decorrelated from Nino3) of the Indian Ocean “explains” 4% of the apparent variability of the ENSO-ISM teleconnection. If the CESM2-LE were faithful to reality, then this signal would take two millenia to have a 50% chance to detect. We also find that this apparent DMI* influence is due to heteroscedasticity, namely, to the apparent DMI* influence on ENSO variance. However, we also find that this influence is mostly apparent, indeed, as it is the ENSO variance that dominantly influences the DMI*.

High-conductivity anomalies of 0.1–1 S/m are widely distributed in the mid-lower crust of the Tibetan Plateau. Dehydration of amphibole-bearing rocks may play an important role in explaining these anomalies. To survey the anomalies’ origin, therefore, the electrical conductivities of amphibole-bearing samples, containing varying amphibole content, are measured at 1.5 GPa and 600–1300 K. Our experiments show that dehydration melting occurs at about 1100 K. Proton conduction and ionic conduction dominate the conduction mechanisms before and after dehydration melting, respectively. The dehydration melting of felsic rocks, containing 25 vol% of amphibole, is unable to account for the high-conductivity anomalies of 0.1–1 S/m. In contrast, the dehydration melting of garnet amphibolite, with an amphibole content higher than 60 wt%, can enhance the bulk conductivity to higher than 0.1 S/m under the lower-crust conditions beneath the Tibetan Plateau. The melt fraction of the garnet-amphibolite is estimated to be 3.8–36 vol% in the partial molten region based on a cube-model simulation.