UMBC GESTAR II

Permanent URI for this collectionhttp://hdl.handle.net/11603/24116

n December 2021, GESTAR II partnered with NASA Goddard Space Flight Center’s Earth Science Division to advance Earth science and Goddard’s leadership by providing a competitive environment to hire and retain high-quality scientists who are on track to be leaders at NASA, in academia and in industry. GESTAR II exemplifies the power of mentorship, embracing a career development strategy that only a university research center can provide. In GESTAR II, early-career researchers and students can build outstanding resumes, launching them to become the Earth science leaders of tomorrow.

Browse

Now showing 1 - 20 of 1152

Tutorial on Causal Inference with Spatiotemporal Data
(ACM, 2024-11-04) Ali, Sahara; Wang, Jianwu
Spatiotemporal data, which captures how variables evolve across space and time, is ubiquitous in fields such as environmental science, epidemiology, and urban planning. However, identifying causal relationships in these datasets is challenging due to the presence of spatial dependencies, temporal autocorrelation, and confounding factors. This tutorial provides a comprehensive introduction to spatiotemporal causal inference, offering both theoretical foundations and practical guidance for researchers and practitioners. We explore key concepts such as causal inference frameworks, the impact of confounding in spatiotemporal settings, and the challenges posed by spatial and temporal dependencies. The paper covers synthetic spatiotemporal benchmark data generation, widely used spatiotemporal causal inference techniques, including regression-based, propensity score-based, and deep learning-based methods, and demonstrates their application using synthetic datasets. Through step-by-step examples, readers will gain a clear understanding of how to address common challenges and apply causal inference techniques to spatiotemporal data. This tutorial serves as a valuable resource for those looking to improve the rigor and reliability of their causal analyses in spatiotemporal contexts.
Evolution of Reactive Organic Compounds and Their Potential Health Risk in Wildfire Smoke
(ACS, 2024-10-22) Pye, Havala O. T.; Xu, Lu; Henderson, Barron H.; Pagonis, Demetrios; Campuzano-Jost, Pedro; Guo, Hongyu; Jimenez, Jose L.; Allen, Christine; Skipper, T. Nash; Halliday, Hannah S.; Murphy, Benjamin N.; D’Ambro, Emma L.; Wennberg, Paul O.; Place, Bryan K.; Wiser, Forwood C.; McNeill, V. Faye; Apel, Eric C.; Blake, Donald R.; Coggon, Matthew M.; Crounse, John D.; Gilman, Jessica B.; Gkatzelis, Georgios I.; Hanisco, Thomas F.; Huey, L. Gregory; Katich, Joseph M.; Lamplugh, Aaron; Lindaas, Jakob; Peischl, Jeff; St Clair, Jason; Warneke, Carsten; Wolfe, Glenn; Womack, Caroline
Wildfires are an increasing source of emissions into the air, with health effects modulated by the abundance and toxicity of individual species. In this work, we estimate reactive organic compounds (ROC) in western U.S. wildland forest fire smoke using a combination of observations from the 2019 Fire Influence on Regional to Global Environments and Air Quality (FIREX-AQ) field campaign and predictions from the Community Multiscale Air Quality (CMAQ) model. Standard emission inventory methods capture 40–45% of the estimated ROC mass emitted, with estimates of primary organic aerosol particularly low (5–8×). Downwind, gas-phase species abundances in molar units reflect the production of fragmentation products such as formaldehyde and methanol. Mass-based units emphasize larger compounds, which tend to be unidentified at an individual species level, are less volatile, and are typically not measured in the gas phase. Fire emissions are estimated to total 1250 ± 60 g·C of ROC per kg·C of CO, implying as much carbon is emitted as ROC as is emitted as CO. Particulate ROC has the potential to dominate the cancer and noncancer risk of long-term exposure to inhaled smoke, and better constraining these estimates will require information on the toxicity of particulate ROC from forest fires.
Accelerating Subglacial Bed Topography Prediction in Greenland: A Performance Evaluation of Spark-Optimized Machine Learning Models
(2024) Cham, Mostafa; Tabassum, Tartela; Shakeri, Ehsan; Wang, Jianwu
Landslide Hazard Is Projected to Increase Across High Mountain Asia
(AGU, 2024-10-03) Stanley, Thomas; Soobitsky, Rachel B.; Amatya, Pukar; Kirschbaum, Dalia B.
High Mountain Asia has long been known as a hotspot for landslide risk, and studies have suggested that landslide hazard is likely to increase in this region over the coming decades. Extreme precipitation may become more frequent, with a nonlinear response relative to increasing global temperatures. However, these changes are geographically varied. This article maps probable changes to landslide hazard, as shown by a landslide hazard indicator (LHI) derived from downscaled precipitation and temperature. In order to capture the nonlinear response of slopes to extreme precipitation, a simple machine-learning model was trained on a database of landslides across High Mountain Asia to develop a regional LHI. This model was applied to statistically downscaled data from the 30 members of the Seamless System for Prediction and Earth System Research large ensembles to produce a range of possible outcomes under the Shared Socioeconomic Pathways 2-4.5 and 5-8.5. The LHI reveals that landslide hazard will increase in most parts of High Mountain Asia. Absolute increases will be highest in already hazardous areas such as the Central Himalaya, but relative change is greatest on the Tibetan Plateau. Even in regions where landslide hazard declines by year 2100, it will increase prior to the mid-century mark. However, the seasonal cycle of landslide occurrence will not change greatly across High Mountain Asia. Although substantial uncertainty remains in these projections, the overall direction of change seems reliable. These findings highlight the importance of continued analysis to inform disaster risk reduction strategies for stakeholders across High Mountain Asia.
AI to the rescue: how to enhance disaster early warnings with tech tools
(Nature, 2024-10-01) Kuglitsch, Monique M.; Cox, Jon; Luterbacher, Jürg; Jamoussi, Bilel; Xoplaki, Elena; Thummarukudy, Muralee; Radwan, Golestan Sally; Yasukawa, Soichiro; McClain, Shanna N.; Albayrak, Arif; Oehmen, David; Ward, Thomas
Artificial intelligence can help to reduce the impacts of natural hazards, but robust international standards are needed to ensure best practice.
Spectral correlation in MODIS water-leaving reflectance retrieval uncertainty
(Optica, 2024-01-15) Zhang, Minwei; Ibrahim, Amir; Franz, Bryan A.; Sayer, Andrew; Werdell, P. Jeremy; McKinna, Lachlan I.
Spectral remote sensing reflectance, Rᵣₛ(λ) (sr⁻¹), is the fundamental quantity used to derive a host of bio-optical and biogeochemical properties of the water column from satellite ocean color measurements. Estimation of uncertainty in those derived geophysical products is therefore dependent on knowledge of the uncertainty in satellite-retrieved Rᵣₛ. Furthermore, since the associated algorithms require Rᵣₛ at multiple spectral bands, the spectral (i.e., band-to-band) error covariance in Rᵣₛ is needed to accurately estimate the uncertainty in those derived properties. This study establishes a derivative-based approach for propagating instrument random noise, instrument systematic uncertainty, and forward model uncertainty into Rᵣₛ, as retrieved using NASA’s multiple-scattering epsilon (MSEPS) atmospheric correction algorithm, to generate pixel-level error covariance in Rᵣₛ. The approach is applied to measurements from Moderate Resolution Imaging Spectroradiometer (MODIS) on the Aqua satellite and verified using Monte Carlo (MC) analysis. We also make use of this full spectral error covariance in Rᵣₛ to calculate uncertainty in phytoplankton pigment chlorophyll-a concentration (chlₐ, mg/m³) and diffuse attenuation coefficient of downwelling irradiance at 490 nm (Kₔ(490), m⁻¹). Accounting for the error covariance in Rᵣₛ generally reduces the estimated relative uncertainty in chlₐ by ~1-2% (absolute value) in waters with chlₐ < 0.25 mg/m³ where the color index (CI) algorithm is used. The reduction is ~5-10% in waters with chlₐ > 0.35 mg/m³ where the blue-green ratio (OCX) algorithm is used. Such reduction can be higher than 30% in some regions. For Kₔ(490), the reduction by error covariance is generally ~2%, but can be higher than 20% in some regions. The error covariance in Rᵣₛ is further verified through forward-calculating chlₐ from MODIS-retrieved and in situ Rᵣₛ and comparing estimated uncertainty with observed differences. An 8-day global composite of propagated uncertainty shows that the goal of 35% uncertainty in chlₐ can be achieved over deep ocean waters (chlₐ ≤ 0.1 mg/m³). While the derivative-based approach generates reasonable error covariance in Rᵣₛ, some assumptions should be updated as our knowledge improves. These include the inter-band error correlation in top-of-atmosphere reflectance, and uncertainties in the calibration of MODIS 869 nm band, in ancillary data, and in the in situ data used for system vicarious calibration.
VIIRS Version 2 Deep Blue Aerosol Products
(AGU, 2024) Lee, Jaehwa; Hsu, N. Christina; Kim, Woogyung V.; Sayer, Andrew; Tsay, Si-Chee
NASA's Deep Blue aerosol project has developed global aerosol data records using consistent retrieval algorithms applied to various satellite sensors. The primary components of these data records are derived from the series of Visible Infrared Imaging Radiometer Suite (VIIRS) aboard the Suomi National Polar-orbiting Partnership (SNPP) and the National Oceanic and Atmospheric Administration or NOAA-20+ satellites as well as the Moderate Resolution Imaging Spectroradiometer (MODIS), among others. These instruments provide over 23 years of measurements with similar radiometric characteristics for aerosol retrievals. The algorithms used for the initial Version 1 SNPP VIIRS data set were based on the MODIS Collection 6.1 Deep Blue algorithm over land and Satellite Ocean Aerosol Retrieval (SOAR) algorithm over water. For VIIRS Version 2 data reprocessing, major updates have been made to the algorithm suite, including better accounting for effects of surface pressure, improved determination of surface reflectance, and the inclusion of fine-mode aerosol optical models to better represent anthropogenic aerosols over land. Cross-calibration gain factors are derived for the NOAA-20 VIIRS measurements to be consistent with the SNPP VIIRS, which allows the use of a unified algorithm package for both instruments. Comparisons against AERONET observations indicate that the Version 2 AOD data from SNPP VIIRS are significantly better than the Version 1 counterpart over land and slightly degraded over water in exchange for better spatial coverage. The AOD data from SNPP and NOAA-20 VIIRS are comparable, indicating that cross-calibration enables the creation of consistent aerosol data records using the series of VIIRS sensors.
GESTAR II Center Awarded $47 Million Extension On Cooperative Agree With NASA Goddard Space Flight Center
(UMBC News, 2024-09-26) Fraser, Adriana; Demond, Marlayna
The UMBC-led Goddard Earth Science Technology and Research (GESTAR II) center has been awarded a two-year, $47 million extension to continue its cooperative agreement with the NASA Goddard Space Flight Center.
Relationship between the sub-micron fraction (SMF) and fine-mode fraction (FMF) in the context of AERONET retrievals
(EGU, 2023-03-03) O'Neill, Norman T.; Ranjbar, Keyvan; Ivănescu, Liviu; Eck, Thomas; Reid, Jeffrey S.; Giles, David M.; Pérez-Ramírez, Daniel; Chaubey, Jai Prakash
The sub-micron (SM) aerosol optical depth (AOD) is an optical separation based on the fraction of particles below a specified cutoff radius of the particle size distribution (PSD) at a given particle radius. It is fundamentally different from spectrally separated FM (fine-mode) AOD. We present a simple (AOD-normalized) SM fraction versus FM fraction (SMF vs. FMF) linear equation that explains the well-recognized empirical result of SMF generally being greater than the FMF. The AERONET inversion (AERinv) products (combined inputs of spectral AOD and sky radiance) and the spectral deconvolution algorithm (SDA) products (input of AOD spectra) enable, respectively, an empirical SMF vs. FMF comparison at similar (columnar) remote sensing scales across a variety of aerosol types. SMF (AERinv-derived) vs. FMF (SDA-derived) behavior is primarily dependent on the relative truncated portion (ε꜀) of the coarse-mode (CM) AOD associated with the cutoff portion of the CM PSD and, to a second order, the cutoff FM PSD and FM AOD (εբ). The SMF vs. FMF equation largely explains the SMF vs. FMF behavior of the AERinv vs. SDA products as a function of PSD cutoff radius (“inflection point”) across an ensemble of AERONET sites and aerosol types (urban-industrial, biomass burning, dust, maritime and a mixed class of Arctic aerosols). The overarching dynamic was that the linear SMF vs. FMF relation pivots clockwise about the approximate (SMF, FMF) singularity of (1, 1) in a “linearly inverse” fashion (slope and intercept of approximately 1-ε꜀ and ε꜀) with increasing cutoff radius. SMF vs. FMF slopes and intercepts derived from AERinv and SDA retrievals confirmed the general domination of ε꜀ over εբ in controlling that dynamic. A more general conclusion is the apparent confirmation that the optical impact of truncating modal (whole) PSD features can be detected by an SMF vs. FMF analysis.
Intercomparison of aerosol optical depths from four reanalyses and their multi-reanalysis consensus
(EGU, 2024-05-31) Xian, Peng; Reid, Jeffrey S.; Ades, Melanie; Benedetti, Angela; Colarco, Peter R.; da Silva, Arlindo; Eck, Thomas; Flemming, Johannes; Hyer, Edward J.; Kipling, Zak; Rémy, Samuel; Sekiyama, Tsuyoshi Thomas; Tanaka, Taichu; Yumimoto, Keiya; Zhang, Jianglong
The emergence of aerosol reanalyses in recent years has facilitated a comprehensive and systematic evaluation of aerosol optical depth (AOD) trends and attribution over multi-decadal timescales. Notable multi-year aerosol reanalyses currently available include NAAPS-RA from the US Naval Research Laboratory, the NASA MERRA-2, JRAero from the Japan Meteorological Agency (JMA), and CAMSRA from Copernicus/ECMWF. These aerosol reanalyses are based on differing underlying meteorology models, representations of aerosol processes, as well as data assimilation methods and treatment of AOD observations. This study presents the basic verification characteristics of these four reanalyses versus both AERONET and MODIS retrievals in monthly AOD properties and identifies the strength of each reanalysis and the regions where divergence and challenges are prominent. Regions with high pollution and often mixed fine-mode and coarse-mode aerosol environments, such as South Asia, East Asia, Southeast Asia, and the Maritime Continent, pose significant challenges, as indicated by higher monthly AOD root mean square error. Moreover, regions that are distant from major aerosol source areas, including the polar regions and remote oceans, exhibit large relative differences in speciated AODs and fine-mode versus coarse-mode AODs among the four reanalyses. To ensure consistency across the globe, a multi-reanalysis consensus (MRC, i.e., ensemble mean) approach was developed similarly to the International Cooperative for Aerosol Prediction Multi-Model Ensemble (ICAP-MME). Like the ICAP-MME, while the MRC does not consistently rank first among the reanalyses for individual regions, it performs well by ranking first or second globally in AOD correlation and RMSE, making it a suitable candidate for climate studies that require robust and consistent assessments.
Letter to the Editor regarding Chappell et al., 2023, “Satellites reveal Earth's seasonally shifting dust emission sources”
(Elsevier, 2024-07-20) Mahowald, Natalie; Ginoux, Paul; Okin, Gregory S.; Kok, Jasper; Albani, Samuel; Balkanski, Yves; Chin, Mian; Bergametti, Gilles; Eck, Thomas; Pérez García-Pando, Carlos; Gkikas, Antonis; Gonçalves Ageitos, María; Kim, Dongchul; Klose, Martina; LeGrand, Sandra; Li, Longlei; Marticorena, Beatrice; Miller, Ronald; Ryder, Claire; Zender, Charles; Yu, Yan
Investigating Causal Cues: Strengthening Spoofed Audio Detection with Human-Discernible Linguistic Features
(2024-09-09) Khanjani, Zahra; Ale, Tolulope; Wang, Jianwu; Davis, Lavon; Mallinson, Christine; Janeja, Vandana
Several types of spoofed audio, such as mimicry, replay attacks, and deepfakes, have created societal challenges to information integrity. Recently, researchers have worked with sociolinguistics experts to label spoofed audio samples with Expert Defined Linguistic Features (EDLFs) that can be discerned by the human ear: pitch, pause, word-initial and word-final release bursts of consonant stops, audible intake or outtake of breath, and overall audio quality. It is established that there is an improvement in several deepfake detection algorithms when they augmented the traditional and common features of audio data with these EDLFs. In this paper, using a hybrid dataset comprised of multiple types of spoofed audio augmented with sociolinguistic annotations, we investigate causal discovery and inferences between the discernible linguistic features and the label in the audio clips, comparing the findings of the causal models with the expert ground truth validation labeling process. Our findings suggest that the causal models indicate the utility of incorporating linguistic features to help discern spoofed audio, as well as the overall need and opportunity to incorporate human knowledge into models and techniques for strengthening AI models. The causal discovery and inference can be used as a foundation of training humans to discern spoofed audio as well as automating EDLFs labeling for the purpose of performance improvement of the common AI-based spoofed audio detectors.
Developing Small Satellite Ground Support Software for Orbit Tracking and Target Acquisition of the HARP Cubesat
(Frontiers, 2024-10-10) Sienkiewicz, Noah; Martins, J. Vanderlei; Xu, Xiaoguang; McBride, Brent; Remer, Lorraine
Small satellites are efficient at performing Earth science from space due to their limited cost and size. Small satellites (cubesats) achieve much with limited power production/storage, heat dissipation, data storage, and ground contact points/bandwidth. As such it is beneficial to offload as much as possible to ground support systems. Consider the HyperAngular Rainbow Polarimeter (HARP) Cubesat. Its goals were to serve as a technical demonstration prior to the development of HARP2 aboard the NASA Plankton Aerosol Cloud and ocean Ecosystem (PACE) mission and to serve as an Earth viewing remote sensing platform which measured the characteristics of clouds and aerosols. HARP cubesat was limited to taking 5-minute capture sequences once every 24 h. It took approximately 10 such captures before it needed to perform data downlink and have its memory cleared for continued use. A ground station at NASA Wallops supported HARP with approximately three points of contact each day. To maximize the value of each capture, ground support software was developed leveraging public data to inform the schedule of each capture. In this paper, we review the algorithms and data sources that allowed us to: 1; predict the HARP orbital track a week in advance, 2; predict also the location of other remote sensing satellites and ground stations relative to HARP, 3; predict the ground view geometry of the instrument along its orbital track, 4; compare global climatological data products of clouds and aerosols along the predicted orbital tracks, and 5; identify and integrate important ground target locations based on remote sensing literature and ongoing natural phenomena. This HARP Orbital Prediction System (HOPS) made HARP into a successful technical demonstration which also offered significant science value. The HOPS system presents a valuable methodology for small satellites to operate efficiently despite their limited capabilities. HOPS is also a useful testbed for studying the sensitivity of scene geometry. Using HOPS, we show that for a wide field-of-view (FOV) instrument, like HARP, latitude/longitude geolocation varies by approximately 0.1° at a height of 8–10 km. Scattering angles vary less than 0.01° at similar heights, with the worst performance near direct backscatter (180°).
Atmospheric Gravity Wave Detection Using Transfer Learning Techniques
(IEEE, 2022-12) González, Jorge López; Chapman, Theodore; Chen, Kathryn; Nguyen, Hannah; Chambers, Logan; Mostafa, Seraj Al Mahmud; Wang, Jianwu; Purushotham, Sanjay; Wang, Chenxi; Yue, Jia
Atmospheric gravity waves are produced when gravity attempts to restore disturbances through stable layers in the atmosphere. They have a visible effect on many atmospheric phenomena such as global circulation and air turbulence. Despite their importance, however, little research has been conducted on how to detect gravity waves using machine learning algorithms. We faced two major challenges in our research: our raw data had a lot of noise and the labeled dataset was extremely small. In this study, we explored various methods of preprocessing and transfer learning in order to address those challenges. We pre-trained an autoencoder on unlabeled data before training it to classify labeled data. We also created a custom CNN by combining certain pre-trained layers from the InceptionV3 Model trained on ImageNet with custom layers and a custom learning rate scheduler. Experiments show that our best model outperformed the best performing baseline model by 6.36% in terms of test accuracy.
Effect of Dust Morphology on Aerosol Optics in the GEOS-Chem Chemical Transport Model, on UV-Vis Trace Gas Retrievals, and on Surface Area Available for Reactive Uptake
(AGU, 2024-09-26) Singh, Inderjeet; Martin, Randall V.; Bindle, Liam; Chatterjee, Deepangsu; Li, Chi; Oxford, Christopher; Xu, Xiaoguang; Wang, Jun
Many chemical transport models treat mineral dust as spherical. Solar backscatter retrievals of trace gases (e.g., OMI and TROPOMI) implicitly treat mineral dust as spherical. The impact of the morphology of mineral dust particles is studied to assess its implications for global chemical transport model (GEOS-Chem) simulations and solar backscatter trace gas retrievals at ultraviolet and visible (UV-Vis) wavelengths. We investigate how the morphology of mineral dust particles affects the simulated dust aerosol optical depth; surface area, reaction, and diffusion parameters for heterogeneous chemistry; phase function, and scattering weights for air mass factor (AMF) calculations used in solar backscatter retrievals. We use a mixture of various aspect ratios of spheroids to model the dust optical properties and a combination of shape and porosity to model the surface area, reaction, and diffusion parameters. We find that assuming spherical particles can introduce size-dependent and wavelength-dependent errors of up to 14% in simulated dust extinction efficiency with corresponding error in simulated dust optical depth typically within 5%. We find that use of spheroids rather than spheres increases forward scattered radiance and decreases backward scattering that in turn decrease the sensitivity of solar backscatter retrievals of NO₂ to aerosols by factors of 2.0–2.5. We develop and apply a theoretical framework based on porosity and surface fractal dimension with corresponding increase in the reactive uptake coefficient driven by increased surface area and species reactivity. Differences are large enough to warrant consideration of dust non-sphericity for chemical transport models and UV-Vis trace gas retrievals.
Integrating Fourier Transform and Residual Learning for Arctic Sea Ice Forecasting
(IEEE, 2023-12) Lapp, Louis; Ali, Sahara; Wang, Jianwu
Arctic sea ice plays integral roles in both polar and global environmental systems, notably ecosystems, commu-nities, and economies. As sea ice continues to decline due to climate change, it has become imperative to accurately predict the future of sea ice extent (SIE). Using datasets of Arctic meteorological and SIE variables spanning 1979 to 2021, we propose architectures capable of processing multivariate time series and spatiotemporal data. Our proposed framework consists of ensembled stacked Fourier Transform signals (FFTstack) and Gradient Boosting models. In FFTstack, grid search iteratively detects the optimal combination of representative FFT signals, a process that improves upon current FFT implementations and deseasonalizers. An optimized Gradient Boosting Regressor is then trained on the residual of the FFTstack output. Through ex-periment, we found that the models trained on both multivariate and spatiotemporal time series data performed either similar to or better than models in existing research. In addition, we found that integration of FFTstack improves the performance of current multivariate time series deep learning models. We conclude that the high flexibility and performance of this methodology have promising applications in guiding future adaptation, resilience, and mitigation efforts in response to Arctic sea ice retreat.
Large-Scale Climate Features Control Fire Emissions and Transport in Africa
(Wiley, 2024-09-15) Dezfuli, Amin; Ichoku, Charles; Bosilovich, Michael G.
Recent increase in extreme wildfire events has led to major health and environmental consequences across the globe. These adverse impacts underlined the need for better understanding of this phenomenon and to formulate mitigating actions. While previous research has focused on local weather drivers of wildfires, our knowledge about their large-scale climatic controls remains limited, especially in tropical Africa, which stands out as a global hotspot for fire emissions. Here, we show that interannual variability of carbon emission due to fires in the southern Congo Basin is strongly linked to low-level winds that are controlled by the Indian Ocean subtropical high. The interhemispheric transport of these emissions to West Africa relies on the intensity and position of both Indian and South Atlantic subtropical highs. Combined effects of this transport mechanism and carbon production in the source region explain a majority of the interannual variability of black carbon in West Africa.
Hybrid Ensemble Deep Graph Temporal Clustering for Spatiotemporal Data
(2024-09-19) Nji, Francis Ndikum; Faruque, Omar; Cham, Mostafa; Janeja, Vandana; Wang, Jianwu
Classifying subsets based on spatial and temporal features is crucial to the analysis of spatiotemporal data given the inherent spatial and temporal variability. Since no single clustering algorithm ensures optimal results, researchers have increasingly explored the effectiveness of ensemble approaches. Ensemble clustering has attracted much attention due to increased diversity, better generalization, and overall improved clustering performance. While ensemble clustering may yield promising results on simple datasets, it has not been fully explored on complex multivariate spatiotemporal data. For our contribution to this field, we propose a novel hybrid ensemble deep graph temporal clustering (HEDGTC) method for multivariate spatiotemporal data. HEDGTC integrates homogeneous and heterogeneous ensemble methods and adopts a dual consensus approach to address noise and misclassification from traditional clustering. It further applies a graph attention autoencoder network to improve clustering performance and stability. When evaluated on three real-world multivariate spatiotemporal data, HEDGTC outperforms state-of-the-art ensemble clustering models by showing improved performance and stability with consistent results. This indicates that HEDGTC can effectively capture implicit temporal patterns in complex spatiotemporal data.
Extreme Rainfall Anomalies Based on IMERG Remote Sensing Data in CONUS: A Multi-Decade Case Study via the IPE Web Application
(2024-09-23) Ekpetere, Kenneth Okechukwu; Mehta, Amita V.; Coll, James Matthew; Liang, Chen; Onochie, Sandra Ogugua; Ekpetere, Michael Chinedu
A web application - IMERG Precipitation Extractor (IPE) was developed that relies on the Integrated Multi-satellite Retrievals for Global Precipitation Measurement (IMERG-GPM) data available at a global coverage. The IPE allows users to query, visualize, and download time series satellite precipitation data for various locations, including points, watersheds, country extents, and digitized areas of interest. It supports different temporal resolutions ranging from 30 minutes to 1 week. Additionally, the IPE facilitates advanced analyses such as storm tracking and anomaly detection, which can be used to monitor climate change through variations in precipitation frequency and intensity. To validate the IMERG precipitation data for anomaly estimation over a 22-year period (2001 to 2022), the Rainfall Anomaly Index (RAI) was calculated and compared with RAI data from 2,360 NOAA stations across the conterminous United States (CONUS), considering both dry and wet climate regions. In the dry region (e.g., Nevada), the results showed an average correlation coefficient (CC) of 0.94, a percentage relative bias (PRB) of -22.32%, a root mean square error (RMSE) of 0.96, a mean bias ratio (MBR) of 0.74, a Nash-Sutcliffe Efficiency (NSE) of 0.80, and a Kling-Gupta Efficiency (KGE) of 0.52. In the wet region (e.g., Louisiana), the average CC was 0.93, the PRB was 24.82%, the RMSE was 0.96, the MBR was 0.79, the NSE was 0.80, and the KGE was 0.18. Median RAI indices from both IMERG and NOAA indicated an increase in rainfall intensity and frequency since 2010, highlighting growing concerns about climate change. The study suggests that IMERG data can serve as a valuable alternative for modeling extreme rainfall anomalies in data-scarce areas, noting its possibilities, limitations, and uncertainties. The IPE web application also offers a platform for extending research beyond CONUS, advocating for further global climate change studies.
The 2021-2024 Winter Precipitation Ground Validation Field Campaign at The University of Connecticut
(IEEE, 2024-09-05) Cerrai, Diego; Filipiak, Brian; Spaulding, Aaron; Wolff, David B.; Tokay, Ali; Helms, Charles N.; Loftus, Adrian M.; Chibisov, Alexey V.; Schirtzinger, Carl; Bliven, Larry; Pabla, Charanjit S.; Boulanger, Mick J.; Kim, EunYeol; Thant, Hein; Junyent, Francesc; Chandrasekar, Venkatachalam; Notaros, Branislav; De Azevedo, Gustavo B. H.
During three consecutive winter seasons, between December 2021 and April 2024, several ground-based wintry precipitation measurement instruments were deployed at the University of Connecticut’s main campus. The instruments included an assortment of K-band and W-band profiling radars and Ka-Ku band scanning radars, weighing, and tipping bucket pluviometers, laser disdrometers, high-speed and high-resolution cameras for quantitative precipitation measurement, weather stations, and an unmanned aircraft system for environmental variables. The goal of this field campaign is to provide a dataset for validating NASA Global Precipitation Measurement (GPM) products, and to examine the error characteristics of co-located ground-based instruments. In this manuscript, we present the instrument suite and discuss possible uses of this unique set of measurements for remote sensing applications.

Browse

Recent Submissions