The SuperCam instrument onboard Perseverance rover has remote imaging (RMI), VISIR, LIBS, Raman and Time-Resolved Luminescence (TRL) capabilities. RMI images of the rocks at the Octavia Butler landing site have revealed important granular texture diversities. VISIR raster point observations have revealed important differences in the 2.10-2.50 µm infrared range (metal-hydroxides); many include water features at 1.40±0.04 and 1.92±0.02 µm [1]. LIBS observations on the same points analyzed by VISIR revealed important differences in the concentrations of major elements, suggesting mineral grain sizes larger than the laser beam (300-500 µm). LIBS and VISIR show coherent results in some rock surfaces that are consistent with an oxy-hydroxide (e.g., ferrihydrite) [1]. LIBS elemental compositions are consistent with pyroxenes, feldspars, and more often feldspar-like glass, often enriched in silica. Olivine compositions [1, 2] have been observed so far in LIBS data (up to Sol 140) exclusively in rounded regolith pebbles. They have not yet been observed in the rocks themselves, which are MgO-poor compared to regolith and are consistent with FeO bearing pyroxenes (e.g., hedenbergite, ferrosilite). A 3x3 LIBS and VISIR raster (9x9 mm) acquired on a low-standing rock on sol 90 exemplifies these finding. A dark L-shaped filled void sampled by points 1 and 2 with possible ferrihydrite (H seen in LIBS and VISIR spectra). Point 5 contains abundant silica and alkali elements but is Al-depleted relative to feldspars, consistent with dacitic glass composition. Point 7 has TiO2 content consistent with ilmenite. Comparisons to (igneous) Martian meteorites are potentially useful, e.g. [3], to explain the presence of several minerals, although most Martian meteorites are olivine-rich, e.g., more mafic than the rocks at the landing site. In summary, the bedrock at Octavia Butler landing site can be interpreted as showing evidence for relatively coarse-grained weathered pyroxenes, iron and titanium oxides and feldspars, while the local soil contains pebbles from a different source (richer in MgO) incorporating olivine grains. References: [1] Mandon et al. 2021 Fall AGU, New Orleans, LA, 13-17 Dec. ; [2] Beyssac et al. 2021 Fall AGU, New Orleans, LA, 13-17 Dec. ; [3] Garcia-Florentino et al.(2021), Talanta, 224, 121863.

Agriculture is a significant source of carbon dioxide (CO2) and methane (CH4) and is the dominant source of anthropogenic nitrous oxide (N2O) emissions. Changes in agricultural land management practices that reduce overall greenhouse gas (GHG) emissions have been suggested to help mitigate climate change, but a better understanding of the timing, magnitude, and drivers of GHG fluxes is needed. Alfalfa agroecosystems may be significant sources of N2O given their ability to increase N inputs through symbiotic N2 fixation and frequent irrigation events that create conditions for hot moments of N2O production. However, few studies have explored long-term N2O emissions and their associated drivers in alfalfa ecosystems. We collected over 108,000 CO2, CH4 and N2O soil flux measurements over four years using cavity ring-down spectroscopy from a conventional flood-irrigated alfalfa field in California, USA. This ecosystem was a consistent source of N2O (annual mean: 624.4 ± 27.8 mg N2O m-2 yr-1, range: 263.6. ± 5.6 to 901.9 ± 74.5 mg N2O m-2 yr-1) and a small net sink of CH4 (annual mean: -53.5 ± 2.5 mg CH4 m-2 yr-1, range: -78.2 ± 8.8 to -31.6 ± 2.5 mg CH4 m-2 yr-1). Soil CO2 fluxes averaged 4925.9 ± 13.5 g CO2 m-2 yr-1 and were greater than other alfalfa ecosystem estimates, likely driven by elevated temperatures and plant productivity throughout the growing season. Hot moments of N2O emissions represented only 0.2% to 1.1% of annual measurements but were 31.6% to 56.8% of the annual flux. We found that both the magnitude and the contribution of N2O hot moments to annual emissions decreased over time. Normalized difference vegetation index (NDVI), soil temperature, moisture, and O2 were all significantly correlated with soil CO2, N2O, and CH4 fluxes, although associations varied across both soil depth and timescales. Our results suggest that flood-irrigated alfalfa is a significant source of agricultural N2O emissions, and that plant productivity and soil moisture effects on O2 availability may modulate the net GHG budget of alfalfa agroecosystems.

Coastal flooding associated with hurricanes and other major storm events along the U.S. Coast results from complex interactions between freshwater flows, tides, storm surge, and wave effects. We have developed a two-way coupled model consisting of the National Water Model (NWM), the Advanced Circulation Ocean Model (ADCIRC), and WAVEWATCH III (WWIII) to quantify these interactions and compute total water levels in the coastal zone after significant riverine and coastal flooding events. This coupled continental coastal model covers the US Gulf and Atlantic Coasts, extending from the US-Canada border to the US-Mexico border. The Delft3D FM, D-Flow Flexible Mesh (D-Flow FM) model simulates coastal flooding on a 2D unstructured mesh within the National Water Model (NWM)/ADCIRC/WWIII coupled system. We developed a high quality 2D unstructured mesh using a sizing function that assigns element sizes based on proximities of coastal features at given spatial locations. Data sources used to identify relevant coastal features included NWM streamlines, the National Hydrography Dataset (NHD), and United States Army Corps of Engineers (USACE) data, allowing integration of D-Flow FM with the NWM and optimization of the number of computational points. The system obtains freshwater inflow boundary conditions to D-Flow FM from the NWM channel network. Offshore water levels boundary conditions for D-Flow FM come from the coupled ADCIRC-WWIII model. Domain sub-setting keeps runtimes within reasonable limits, as it does execution of the detailed hydrodynamic model within a user-defined area enclosing the storm landfall site. The advantage of this approach comes from the fact that the same coupled model setup allows simulation of coastal flooding for different storm events; only the sub-setting enclosure and the atmospheric forcing require updating from case to case. Model validation, consisting of water level comparisons against observations from simulations using the coupled system for historical storm events. The model simulations satisfactorily reproduced observed spatial and temporal variations of total water levels. In conclusion, this study presents performance of the sub-setting approach in reducing runtime considerably without compromising the accuracy of the coupled modeling system solution.

Relative sea-level rise is a major coastal hazard affecting about half the population of the United States. The Chesapeake Bay is characterized by the fastest rate of sea-level rise along the Atlantic coast of North America, in part because of land subsidence. Previous studies have quantified a range of land subsidence rates in the Chesapeake Bay (~1-4 mm/yr) from various measurement techniques that contribute to high rates of relative sea-level rise. In this study, we present progress towards developing a new vertical land motion map for the Chesapeake Bay region to provide more robust constraints on estimates of relative sea-level rise. We are using a combination of GNSS observations and InSAR interferograms. Available continuous GNSS data in the region that span November 2014 - September 2020 are processed with GAMIT-GLOBK to align temporally with available Sentinel-1 InSAR satellite data. We are using an approach that combines the two geodetic observations to provide a new solution of vertical land motions for the Chesapeake Bay. Additionally, this project is collecting new campaign GNSS observations across the Chesapeake Bay each fall for 5 years, beginning in 2019. We will also present about the 2020 and planned 2021 campaign GNSS observations, which will ultimately be incorporated into our new map of vertical land motions for the region. The impacts of this work will be improved flooding and inundation hazard maps, as well as updated projections for municipal flood mitigation planning that will be created using the new dataset.

The Surface Biology and Geology (SBG) mission is one of the core missions of NASA’s Earth System Observatory (ESO). SBG will acquire high resolution solar-reflected spectroscopy and thermal infrared observations at a data rate of ~10 TB/day and generate products at ~75 TB/day. As the per-day volume is greater than NASA’s total extant airborne hyperspectral data collection, collecting, processing/re- processing, disseminating, and exploiting the SBG data presents new challenges. To address these challenges, we are developing a prototype science pipeline and a full-volume global hyperspectral synthetic data set to help prepare for SBG’s flight. Our science pipeline is based on the science processing operations technology developed for the Kepler and TESS planet-hunting missions. The pipeline infrastructure, Ziggy, provides a scalable architecture for robust, repeatable, and replicable science and application products that can be run on a range of systems from a laptop to the cloud or an on-site supercomputer. Our effort began by ingesting data and applying workflows from the EO- 1/Hyperion 17-year mission archive that provides globally sampled visible through shortwave infrared spectra that are representative of SBG data types and volumes. We have fully implemented the first stage of processing, from the raw data (Level 0) to top-of-the-atmosphere radiance (Level 1R). We plan to begin reprocessing the entire 55 TB Hyperion data set by the end of 2021. Work to implement an atmospheric correction module to convert the L1R data to surface reflectance (Level 2) is also underway. Additionally, an effort to develop a hybrid High Performance Computing (HPC)/cloud processing framework has been started to help optimize the cost, processing throughput and overall system resiliency for SBG’s science data system (SDS). Separately, we have developed a method for generating full-volume synthetic data sets for SBG based on MODIS data and have made the first version of this data set available to the community on the data portal of NASA’s Advanced Supercomputing Division at NASA Ames Research Center. The synthetic data will make it possible to test parts of the pipeline infrastructure and other software to be applied for product generation.

The most simple representation of the dynamics of the global energy budget is the 0-dimensional energy balance model (EBM) introduced by Budyko (1965). Budyko’s EBM assumes a linear relationship between the Earth’s radiative response and the global surface temperature such that the dynamics of the global energy budget reads CdTs/dt = N = F - λ Ts, where Ts is the global surface temperature, N is the Earth Energy Imbalance, C is the ocean heat capacity and λ is the constant climate feedback parameter. Such simple conceptual model depicts reasonably well the centennial time scale response of the steady state preindustrial global energy budget under an anomalous forcing such as the increase of atmospheric greenhouse gases concentrations. For this reason it has served as the basis for the definition of the effective climate sensitivity to atmospheric CO2 concentrations. However recent studies identified limitations to Budyko’s EBM. Indeed climate model simulations show that the radiative response of the Earth not only depends on the global surface temperature but also on its geographical pattern: the so-called “pattern effect”. It arises from changes in the mix of radiative forcings, lag-dependent responses to forcings, or unforced variability and it leads to an apparent time variation in λ. This time variation must be accounted for in Budyko’s EBM to represent the longer term response of the global energy budget under increased CO2 concentrations. Here, a simple theory is developed to account for the time dependency of λ in the global energy budget. The resulting differential equation accurately reproduces the long term response (i.e. >200 years) of climate under abrupt changes in CO2 concentrations as simulated in the longrunmip experiment. Analysis of the asymptotic form of the differential equation yields a new expression of the climate sensitivity which not only depends on the climate feedback parameter but also on its temporal change. We evaluate this new climate sensitivity for all runs of the longrunmip experiment and show how it relates with the classical effective climate sensitivity from Gregory et al. (2004). We find that the spread in climate sensitivity among climate models of the longrunmip experiment is essentially due to different temporal changes in λ (and thus different pattern effect) among models.

Understanding the dynamical behavior of plasma and energetic particles in Earth’s inner magnetosphere requires carefully designed and calibrated instrumentation. The Van Allen Probes Mission included two instruments capable of measuring the proton distribution function in-situ. The Energetic Particle Composition and Thermal Plasma Suite (ECT) – Helium Oxygen, Proton, and Electron (HOPE) spectrometer (Spence et al., 2013; Funsten et al., 2013) used a top-hat detector designed to measure protons from the SC potential through 50 KeV in logarithmic energy steps. The Radiation Belt Storm Probes Ion Composition Detector (RBSPICE) instrument (Mitchell, 2013) used a time of flight and SSD detector design to measure protons from approximately 7 KeV through 650 KeV in logarithmic energy steps. Using the overlap of energy channels between the two instruments, the two instrument teams have worked diligently during the final Phase F of the mission to calibrate the observations so that a continuous distribution function can be resolved on nearly a spin-by-spin basis. During the life of these two instruments calibration changes have been required both on-board the spacecraft as well as within the final production datasets. Manweiler (2018) provided an early report on the intercalibration factors between HOPE and RBSPICE with a nominal factor of two difference between the proton data sets in the energy range between 7 and 50 KeV. With the final production of each of these data sets occurring in Fall 2021, both teams have been worked together to provide for an understanding of the required intercalibration factors to be used so that a full distribution function is available on a spin-by-spin basis. In this poster we report on the final efforts to provide this calibrated set of data products between the two instruments. Details of the intercalibration calculations are presented as well as year by year L by MLT maps of the factors required to match both datasets. Finally, we report on a supplementary data set that is to be made available which contains the spin-by-spin factors required to match the ECT/HOPE and RBSPICE/TOFxPH proton datasets. Funsten, H.O., et al. Space Sci Rev 179, 2013 Manweiler, J. W., et al., 2018 GEM Summer Workshop. Mitchell, D.G., et al., Space Sci. Rev., 179, 2013 Spence, H.E., et al. Space Sci Rev 179, 2013

Sea ice leads are produced from deformational forces, which break apart the ice surface and expose open water areas or leads. Leads are the primary regulators of heat in Arctic sea ice during the polar night. Partially-frozen and re-frozen leads produce smaller heat exchanges than open leads due to the absence of a warm water surface. Thus detecting leads is important because they may be used as an indirect way of estimating air-sea heat fluxes. To quantify winter-time leads, we utilize Sentinel-1 C-Band Synthetic Aperture Radar (SAR) data to examine sea ice images through heavy cloud cover. We employ a support vector machine learning technique in a cloud computation environment (Google Earth Engine) to detect and quantify lead areas. With the use of dual-polarization data, we improve the separation of leads from other elongated features (e.g., ridges) in the Sentinel-1 dataset by adding altimetry information from ICESat-2. In addition to typical texture analysis to assess surface roughness, the ICESat-2 ATL-10 data allows us to train the algorithm by discretizing leads and ridges by their freeboard values. Performing this method in a cloud environment allows processing of a large volume of satellite data and converting it into a time series of leads properties. Overall, our method improves lead detection with dual-polarization SAR data while simultaneously providing a big data solution for SAR image processing. The interannual variability of leads and newly formed ice fractions were found for the winters of 2017-2020. Finally, we compare the results from previous studies to validate our cloud-derived sea ice lead detection maps.

Floating photovoltaic solar energy (FPV) are solar photovoltaic systems that float on bodies of water. They are a rapidly expanding renewable energy source emerging as an alternative to land-intensive ground-mounted solar arrays. The applications of this technology are commonly explored through technical potential assessments; a vital step in the development of renewable energy resources that allow for the identification of feasible installation sites and provide an estimation of costs, power generation, and capacity (Lee and Roberts 2018). These assessments are carried out to aid planners, policymakers, and other decision-makers in predicting and achieving goals related to the development of renewable energy; however, some considerations may be overlooked. Assessing the technical potential of solar energy without a standardized methodological framework for site selection may lead to an inconsistent range of generation outcomes, adding to the confusion and lack of confidence that has been a significant barrier to the growth of renewable energy in recent years (Seetharaman 2019) This study systematically reviewed criteria in the published literature emphasizing assessment of FPV technical potential and related siting studies, especially those using geographic technologies. The systematic review was performed in alignment with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. Preliminary findings of this systematic review suggest that FPV site selection criteria can be categorized into economic, social, environmental, and technical considerations. We aim to elucidate which criteria within each classification are important to include in an FPV siting study based on the current literature. Results from this analysis will inform the standardization of a site-selection framework used in future technical potential studies, which may, in turn, improve the accuracy of generation estimates and offer a meaningful and realistic idea of how FPV installations could transform the direction of renewable energy science and development.

As the sea level rises, it is alarming that the threat from flooding induced by tropical cyclones would cause more severe damages to coastal regions worldwide. In order to address this threat, optimizing coastal protective or mitigation strategies is necessary, given limited resources. The optimization methodology must incorporate feedback from stakeholders for practical use. Multiple interviews were conducted by engineering model developers and social scientists with stakeholders who have first-hand knowledge and varied backgrounds in New York. The protective strategies have been tuned to the critical infrastructure's particular and contextual risks due to flood hazards by engaging and integrating stakeholders' knowledge on the interdependency of the infrastructures and other aspects after the first interview. The second interview was conducted for further model improvement.

The temporal dynamics of aboveground woody biomass across the landscape is one of the most uncertain aspects of the global carbon cycle. Previous studies suggest that the forests and savannas of Africa act as a sink of biomass carbon, while others indicate that it might already be turning into a source. In this study we analysed continental-wide biomass dynamics using a time-series of aboveground biomass maps for the 2007 to 2017 period. We developed these maps at a spatial resolution of 100m using LiDAR, SAR and multispectral spaceborne sensors within a machine-learning framework. Our results show that the aboveground biomass stocks in Africa were 120.5 Pg. At the same time, we can observe a continuous increase in the annual rate of biomass loss in the continent due to the increasing deforestation rates occurring in the Congo Basin.

As part of the (AC)3 Arctic Amplification project, we are studying the influence of specific sea ice conditions like the presence of leads or polynyas on micro- and macrophysical cloud properties such as cloud fraction, altitude, thickness, thermodynamic phase, and their coupling state with respect to the underlying surface during the MOSAiC expedition’s legs 1 to 3. Micro- and macrophysical properties of surface-coupled clouds are analyzed as a function of sea ice concentration (SIC) in the vicinity of the ground-based atmospheric remote-sensing observations onboard the RV Polarstern. Only situations are analyzed where wind favored the transportation of air from location where open sea ice is detected. Cloud microphysical properties are obtained from the CloudNet cloud target classification algorithm which uses the atmospheric remote-sensing instrumentation suite on board of RV Polarstern provided by the US Atmospheric Radiation Measurement (ARM) mobile facility, the TROPOS ship-borne Atmosphere observation suite (OCEANET) and liquid water path retrievals by the University of Cologne. Primarily, the classical Matlab-based CloudNet classifications retrieved by TROPOS are used. Furthermore, the recently released ARM “evaluation” Active Remote Sensing Clouds (ARSCL) data product for the KA-band cloud radar is also evaluated by the new Python CloudNet version developed at the Finish Meteorological Institute. Discrepancies between those two CloudNet versions will be evaluated and reported as feedback for the ARM evaluation data set. High resolution (1-km) merged AMSR2-MODIS satellite retrievals of Sea Ice Concentration by the University of Bremen are used as information for sea ice monitoring. The present contribution only exploits SIC data, however future studies will focus on MOSAiC specific products for the classification of leads. Statistics for the cloud properties as a function of SIC will be presented as first approach to investigate the influence of sea ice conditions to central Arctic clouds.

Regional crop production estimates are important in both public and private sectors to ensure the adequacy of a food supply and aid policymakers and farmers in managing harvest, storage, import/export, transportation, and anticipate market fluctuations. Food security will be progressively challenged by population growth and climate change. Thus, the prediction of accurate regional crop yield is essential for national food security and the sustainable development of the Indian agriculture sector. In this study, we have selected Punjab, the highest wheat yielding state in India. The district-wise wheat yield data were available for the year 2000 – 2019. We have used several covariates for crop health viz. normalized difference vegetation index (NDVI), leaf area index (LAI), fraction of absorbed photosynthetically active radiation (fAPAR); meteorological indicators viz. land surface temperature (LST), and evapotranspiration (ET); and surface characteristics viz. protrusion coefficient (PC). These indicators were generated at 250 m spatial resolution from the MODIS data using Google Earth Engine. The whole data was divided into two groups for training (2000 – 2009, 2011, 2013, 2014, 2016 - 2019) and testing (2010, 2012, 2015), which were randomly selected. This study uses the random forest (RF) regression method to create a wheat yield prediction model. We created several combinations of covariates and found that fAPAR and ET are highly correlated with NDVI and do not have much influence on the model’s prediction accuracy. Hence, only four out of six covariates were selected for final training. The coefficient of determination between district-level yield vs. (NDVI/LAI/PC/LST) was 0.37/0.31/0.15/0.13 respectively. We used randomized search cross-validation as well as grid search cross-validation for hyper-parameter tuning. Furthermore, we used mean absolute error (MAE) and accuracy as quality metrics. The MAE for training was 0.1870 t/Ha with 95.81% accuracy, whereas the MAE on test data was obtained as 0.4293 t/Ha with 90.02% accuracy. The results of this study are within acceptable error limits of the published research articles. Overall, this study demonstrates that covariates derived from coarse resolution satellite data can predict district-level crop yield with reasonable accuracy.

On the 22nd of May 2021, although no alarming precursory unrest had been reported, Nyiragongo volcano erupted and lava flows threatened about 1 million of inhabitants living in the cities of Goma (Democratic Republic of Congo) and Giseny (Rwanda). After January 1977 and January 2002, it was the beginning of the third historically known flank eruption of Nyiragongo volcano and the first ever to be recorded by dense measurements both on the ground and from space. In the following days, seismic and geodetic data as well as fracture mapping revealed the gradual southward propagation of a shallow dike from the Nyiragongo edifice underlying below Goma airport on May 23-24, then Goma and Gisenyi city centers on May 25-26 and finally below the northern part of Lake Kivu on May 27. Southward migration of the associated seismic swarm slowed down between May 27 and June 02. Micro seismicity became more diffuse, progressively activating transverse tectonic structures previously identified in the whole Lake Kivu basin. Here we exploit ground based and remote sensing data as well as inversion and physics-based models to fully characterize the dike sized, the dynamics of dike propagation and its arrest against a structural lineament known as the Nyabihu Fault. This work highlights the shallow origin of the dike, the segmented dike propagation controlled by the interaction with pre-existing fracture networks and the incremental crater collapse associated with drainage which led to the disappearance of the world’s largest long-living lava lake on top of Nyiragongo.

Evaluation of spatially distributed crop coefficient (Kc) for estimating evapotranspiration (ETc) based on remotely sensed imagery has become an essential topic in managing the demand for agricultural water. Currently, satellite (MODIS, Landsat, etc.) imageries are not insufficient to detect variability within the small agricultural field due to its lack of desired spatial and temporal resolutions. Unmanned Aerial Vehicle (UAV) equipped with various sensors like Multispectral (MS), Thermal, and Hyperspectral cameras is becoming an emerging technology to overcome these limitations over small agricultural fields. A field experiment is carried out in the Agricultural and Food Engineering (AGFE) Department, IIT Kharagpur, to estimate Kc over the small Agri. Field using UAV-based MS cameras during Kharif (monsoon) 2019-2020 season. Lysimeters are used for estimating daily ETc for conventionally irrigated paddy crops. Reference evapotranspiration (ET0) is also calculated using the weather data of the study area. High-resolution multispectral imageries are acquired using a quad-copter UAV. The imageries are pre-processed using Pix4Dmapper software, and various vegetation indices (such as NDVI, TNDVI, NDRE, RVI, GNDVI, and LCI) are evaluated. The vegetation indices (VIs) are correlated with ground truth Kc values and spatially distributed Kc maps for the whole study area are generated based upon the excellent correlation between the VIs and ground Kc. The spatial Kc maps clearly show the variation in Kc within the plots and will be helpful for the calculation of Kc for any field without a lysimeter experiment. Generated Kc maps describe the crop water demand by visual color variations within the field. This approach may be helpful in understanding the variability in crop water requirements within the field Keywords: UAV, Crop Coefficient (Kc), Crop Evapotranspiration (ETc), Vegetation Indices, Remote Sensing.

The frequency, size, and intensity of wildfires in California have increased substantially in recent years, leading to widespread mandatory evacuations affecting millions of residents. However, because evacuation orders are implemented by local agencies, there is limited quantitative evidence on the scope of evacuations statewide. In order to improve the understanding of wildfire evacuations, we assembled information on historical evacuation orders for two distinct wildfire-prone regions --- Fresno and Sonoma county --- in California. This data was used to understand how the frequency and extent of evacuations have changed over time. We then combined this information with census data to characterize which populations have been most affected by evacuation orders. Ultimately, our work aims to quantify this important element of wildfire impacts in key regions around California. Collectively, it provides a starting point for a public database of evacuation orders that could be used by researchers and policymakers to better understand dynamics and improve decision-making around wildfire evacuations.

Microbes play a critical role in regulating the size, composition, and turnover of dissolved organic matter (DOM), which is one of the largest pools of carbon in aquatic ecosystems. Global change may alter DOM-microbe associations with implications for biogeochemical cycles, although disentangling these complex interactions remains a major challenge. Here we develop a framework called Energy-Diversity-Trait integrative Analysis (EDTiA) to examine the associations between DOM and bacteria along temperature and nutrient gradients in a manipulative field experiment on mountainsides in contrasting subarctic and subtropical climates. In both study regions, the chemical composition of DOM correlated with bacterial communities, and was primarily controlled by nutrients and to a lesser degree by temperature. At a molecular-level, DOM-bacteria associations depended strongly on the molecular traits of DOM, with negative associations indicative of decomposition as molecules are more biolabile. Using bipartite networks, we further demonstrated that negative associations were more specialized than positive associations indicative of DOM production. Nutrient enrichment promoted specialization of positive associations, but decreased specialization of negative associations particularly at warmer temperatures in subtropical climate. These global change drivers influenced specialization of negative associations most strongly via molecular traits, while both molecular traits and bacterial diversity similarly affected positive associations. Together, our framework provides a quantitative approach to understand DOM-microbe associations and wider carbon cycling across scales under global change.

During an explosive eruption, large volumes of ash and gases are ejected into the atmosphere, forming a volcanic plume which is transported by the wind. The dispersion of volcanic ash in atmosphere represents a threat for aviation safety, whereas the tephra fallout, together with gas emission, may strongly affect population health and damage to environment and infrastructure as well. Volcanic monitoring from space offers now a powerful tool to quantify hazards on both population and air traffic and gain insight into processes and mechanism of violent explosive eruptions. Here we propose a machine learning (ML) algorithm that exploits the Thermal Infrared (TIR) bands of the images acquired by the sensor Spinning Enhanced Visible and InfraRed Imager (SEVIRI), on board Meteosat Second Generation (MSG) geostationary satellite, to identify the components of ash and SO2 gas in a volcanic plume. The detection and assessment of volcanic ash clouds has been performed applying the brightness temperature difference (BTD) approach, between bands at 10.8 μm and 12.0 μm, which highlights the presence of thin volcanic ash, while the algorithm for the SO2 retrieval is based on the contributions given by the bands at 10.8 μm and 8.7 μm. Combining the latter two bands with the 10.8 μm band in the RGB channels, it is possible to create an Ash RGB image, used both day and night for the detection and monitoring of volcanic ash and sulphur dioxide gas. The advantage of the machine learning algorithm is to detect and extract automatically these features from an Ash RGB image. As test cases, we considered the sequence of explosive eruptions occurred at Etna volcano (Italy) in early 2021, which produced very long and high plume columns. Thanks to the high temporal resolution of SEVIRI (one image every 15 minutes), it was possible to visualize and to follow the plumes, from their formation to their complete dispersion in the atmosphere. The comparison of our ML algorithm with the consolidated procedure based on a RGB channels combination in the visible (VIS) spectral range showed a good agreement.