“A perfect forecast on its own cannot meet the mission of the weather service.” It is the interpretation and communication of the forecast that will provide decision makers with the context they need to protect life, livelihoods, and property. One of the main long-term strategic goals of the National Weather Service (NWS) is to  assess and communicate probabilistic forecast information to partners to assist in such critical decisions. This social science research explores how forecasters in various  offices and regions of the NWS currently convey the uncertainty of the forecast and the array of potential forecast outcomes. Semi-structured interviews with 17 NWS  forecasters representing each region of the NWS and several National Forecast Centers were conducted and analyzed. The interviews revealed layers of organizational processes that contribute to the agency’s contemporary probabilistic communication strategies, including tactics for identifying valuable probability  information to partners, navigation of practical and operational challenges, and considerations for dissemination. The presentation will go into these key themes in  detail. The study reflects the organizational processes undertaken by a primarily physical science-based agency to accomplish a multidisciplinary goal. This analysis  provides a foundation for evaluating these practices as an integrated system and highlights areas needing further development. 

This review explores the extensive efforts undertaken by NOAA to study and address urban air pollution, emphasizing the critical role of the Climate Program Office's  Atmospheric Chemistry, Carbon Cycle, and Climate (AC4) program. Urban air pollution poses significant health risks, contributing to a substantial global disease burden  as highlighted by recent reports attributing 8.1 million deaths worldwide in 2021 to air pollution. NOAA's multidisciplinary approach, encompassing various offices and  labs, focuses on understanding the complex interactions between air pollution and urban environments. The AC4 program complements NOAA's initiatives by funding  innovative research projects that address key areas such as wildfire impacts on urban air quality, emissions and heat in urban areas, and the effects of reduced human  activity during the COVID-19 pandemic on air quality. These projects provide crucial data and insights that inform policy, improve air quality forecasting models, and  support public health initiatives. Furthermore, NOAA's integration of advanced satellite monitoring systems, ground-based measurements, and modeling techniques  enhances our ability to predict and mitigate the impacts of urban air pollution. This comprehensive review underscores the importance of continued research and  collaboration in tackling the challenges posed by urban air pollution, particularly in the context of a changing climate and increasing urbanization.

Feedback from NWS forecasters suggests that vertical sounding data is currently difficult to access in AWIPS, and is often disregarded. As a potential solution, data  retrieved from GOES-16, GOES-18, and NUCAPS satellites, GFS and NAM models, and radiosondes is stored in the AWS cloud to produce a fast integrated visualization  tool. Utilizing VUGraF, a Python framework, the webpage VuSkew was created directly from forecaster feedback as a suggested operational replacement to increase  spatial and temporal access to sounding data. The user can select individual soundings to view on a traditional Skew-T Log-P diagram, including major parameters,  heights, and wind barbs. However, VuSkew also grants users the ability to compare multiple soundings on the same diagram or average soundings via individual  selection or a polygon. To further equip forecasters, a spatial visualization tool, VuSkew Visualizer, allows users to spatially distinguish temperature, dew point, dew  point depression, and relative humidity values at different pressure levels. For simple temporal accessibility, a timeline displays relative amounts of soundings along  with each source, with an animation feature exhibiting the spatial arrangement and changing values of selected soundings over time. For operational purposes, the data  is automatically updated and preferences can be set for a WFO, archive and forecast times, color bars, and more. Future updates to VuSkew include incorporating  ACARS and HRRR data and continually collaborating with forecasters for practicality.

Feedback from NWS forecasters suggests that vertical sounding data is currently difficult to access in AWIPS, and is often disregarded. As a potential solution, data  retrieved from GOES-16, GOES-18, and NUCAPS satellites, GFS and NAM models, and radiosondes is stored in the AWS cloud to produce a fast integrated visualization  tool. Utilizing VUGraF, a Python framework, the webpage VuSkew was created directly from forecaster feedback as a suggested operational replacement to increase  spatial and temporal access to sounding data. The user can select individual soundings to view on a traditional Skew-T Log-P diagram, including major parameters,  heights, and wind barbs. However, VuSkew also grants users the ability to compare multiple soundings on the same diagram or average soundings via individual  selection or a polygon. To further equip forecasters, a spatial visualization tool, VuSkew Visualizer, allows users to spatially distinguish temperature, dew point, dew  point depression, and relative humidity values at different pressure levels. For simple temporal accessibility, a timeline displays relative amounts of soundings along  with each source, with an animation feature exhibiting the spatial arrangement and changing values of selected soundings over time. For operational purposes, the data  is automatically updated and preferences can be set for a WFO, archive and forecast times, color bars, and more. Future updates to VuSkew include incorporating  ACARS and HRRR data and continually collaborating with forecasters for practicality.

Feedback from NWS forecasters suggests that vertical sounding data is currently difficult to access in AWIPS, and is often disregarded. As a potential solution, data  retrieved from GOES-16, GOES-18, and NUCAPS satellites, GFS and NAM models, and radiosondes is stored in the AWS cloud to produce a fast integrated visualization  tool. Utilizing VUGraF, a Python framework, the webpage VuSkew was created directly from forecaster feedback as a suggested operational replacement to increase  spatial and temporal access to sounding data. The user can select individual soundings to view on a traditional Skew-T Log-P diagram, including major parameters,  heights, and wind barbs. However, VuSkew also grants users the ability to compare multiple soundings on the same diagram or average soundings via individual  selection or a polygon. To further equip forecasters, a spatial visualization tool, VuSkew Visualizer, allows users to spatially distinguish temperature, dew point, dew  point depression, and relative humidity values at different pressure levels. For simple temporal accessibility, a timeline displays relative amounts of soundings along  with each source, with an animation feature exhibiting the spatial arrangement and changing values of selected soundings over time. For operational purposes, the data  is automatically updated and preferences can be set for a WFO, archive and forecast times, color bars, and more. Future updates to VuSkew include incorporating  ACARS and HRRR data and continually collaborating with forecasters for practicality.

Heat and momentum transfer between the Earth’s surface and the atmosphere is important for characterizing the behavior of the atmospheric boundary layer. In field  campaigns surface fluxes are difficult to measure directly, so the fluxes are often modeled using bulk aerodynamic formulas and mean flow quantities at a set distance  from the surface. The use of mean flow quantities, approximated by time averages, can lead to a question of sample convergence and its influence on the calculated  surface fluxes. For this study the influence of sample size on flux calculations is explored using data from large eddy simulations (LES) of a neutral and highly convective  boundary layer. Fluxes are calculated directly from the LES data as well as with the bulk aerodynamic formulas built into the COARE algorithm. The impact of sample size  on accuracy of the calculated values is assessed with averages over varied temporal and spatial domains. The convergence of the statistics is compared to the expected  behavior with increasing sample size in the literature to observe the sensitivity of result quality on sample size.

The Hurricane Analysis and Forecast System (HAFS) is the next generation of hurricane computer modeling. Evaluating the HAFS on past instances can allow for  improvements to both hurricane forecasting and the hurricane model itself. One way to evaluate the model is to run it repeatedly and tweak it slightly with each run,  creating an ensemble set that can help expose periods of enhanced uncertainty and provide a realistic range of possibilities. This allows for a more robust analysis of the  sensitivity of relevant synoptic-scale and vortex-scale features. In this case, HAFS was run at an early forecast period for Hurricane Ian. It was run with 30 ensemble  members twice, each with a different convective scheme, which essentially determines how sub grid-scale convective processes are resolved. The goal of this project is  to analyze these convective scheme and ensemble member differences using methods such as visual analysis, clustering, and member comparison. Indeed, notable track  dispersion occurs in just the short-term forecast, especially when comparing between convective schemes. The primary causes of this short-term track dispersion are  differences in the mid-level ridging and vortex-scale convective processes. This initial dispersion increases with time, leading to a wide range of landfalls along the  Cuban and eventually Floridian coastlines. Expanding this analysis to different forecast hours and different storms could allow for stronger conclusions to be made regarding the benefits and drawbacks of the different convective schemes.

This presentation highlights my team and I's efforts to enhance oceanographic data processing and visualization tools through the development of three key projects:  the netCDF to GRIB2 Converter, the Climate Stripes (For Ocean Level Rise) visualization tool, and the Ocean Maps Variable analysis tool. The netCDF to GRIB2 Converter  improves data compatibility and usability in weather and climate models, while the Climate Stripes tool creates visual representations of ocean level rise using climate  stripe charts to illustrate historical trends and patterns. The Ocean Maps Variable Checker ensures the accuracy and consistency of variables in oceanographic maps,  thereby maintaining data integrity for ocean models and research projects. Together, these projects address critical needs in the field of oceanography, enhancing data  compatibility, clarity, and overall precision.

Over the last three years, NOAA Fisheries Alaska Fisheries Science Center (AFSC) Resource Assessment and Conservation Engineering (RACE) Division’s Groundfish  Assessment Program has worked to develop new public-facing data platforms and products. These initiatives aim to enhance transparency, facilitate reproducible  research, and promote accountability by openly sharing survey catch and environmental findings with the broader community. Key products include near real-time survey progression and temperature map webpages, automated data reports and outreach materials, and various data access and exploration platforms. These resources are openly accessible without cost or restrictions, fostering widespread distribution among the general public, private entities, and NOAA partners. This  presentation introduces a new R Shiny application in support of the AFSC’s open-science and data transparency efforts. This innovative tool will enable the dynamic  exploration of over four decades of bottom survey temperature data which spans the Bering Sea, Aleutian Islands, and Gulf of Alaska. By enabling interactive  engagement with a comprehensive historical dataset, this application will help to foster collaborative discussions and insights essential for advancing our understanding  of global ecosystems and organisms.

The Arctic is experiencing accelerated climate-driven environmental changes, including extreme heat, permafrost thaw, and ice loss. These changes affect marine and  coastal ecosystems and impact Arctic communities’ food security, infrastructure, and health. The U.S. Arctic Observing Network (AON) - a collaborative network of  agencies, organizations, and communities - seeks to improve data collection and observing to support resilience in the region. To identify problematic observing and data  gaps, U.S. AON has developed a methodology to systematically link the observing system and data products to applications that provide societal benefits, like climate  resilient communities, revealing strengths and gaps in their effectiveness. In 2023, U.S. AON held 40+ dialogs with federal, state, and academic partners to identify urgent and shared topics for assessment case studies focused on risks and hazards. Dialogs were analyzed using a standardized coding system in MAXQDA, a qualitative coding  software. Using this approach, coastal inundation, along with its impacts on subsistence food security and infrastructure, emerged as an important case study scope for  assessment in the Alaskan Arctic. This presentation will cover the key information gathered from the assessment, including organizations engaged in coastal inundation  work, their primary services and products, and the most critical consequences of coastal inundation. Visualizations of coded qualitative data from the dialogs reveal gaps  in and potential improvements of the observing system. The gaps assessment will help U.S. AON identify specific investments in data collection and observing to address coastal inundation hazards, improving climate resilience in the Alaskan Arctic.

The Weather Prediction Center is responsible for generating the Day 3-7 Hazards Outlook, which illustrates the location and duration of weather-related threats to  operations, life, and property across the United States. Prevalent among these threats is Hazardous Heat, defined as a probability greater than 40% of exceeding  widespread Heat Advisory Criteria. The criteria referenced are specific to the climatology of the region at the county level. First Guess Fields for the spatial and time  scales of hazards are a vital tool that increase forecaster efficiency by summarizing data visually in a probabilistic format to highlight regions of interest and concentrate  the efforts of skilled labor in operations. To create such fields, this project gathers data from the experimental WPC HeatRisk product, the EFI (Extreme Forecast Index)  from ECMWF, and WPC Heat Index Forecasts. These datasets are analyzed probabilistically to produce a union where EFI exceeds 0.8, HeatRisk is Major or Extreme, or  the Heat Index shows greater than a 40% chance of exceeding Heat Advisory Criteria. Information is synthesized in ArcGIS Pro and ArcMAP layers designed to be  interactive for forecasters, and using color to emphasize the most relevant information while blending with the convention for existing products.

The Tropical Analysis and Forecast Branch of the National Hurricane Center provides multiple products that aim to protect mariners from hazards on the high seas.  These products often contain information about non-tropical gale- and storm-force wind conditions in the Eastern Pacific and Atlantic basins. However, all information  on the history of these warnings are contained within text products, making it extremely difficult to get a large overview of their occurrences, leaving mariners and  residents of affected countries underinformed about their patterns. This project aimed to visualize data pertaining to such events to draw connections and identify  patterns regarding their occurrences. We created separate climatologiesfor both gale- and storm-warnings in the Eastern Pacific and Atlantic basins, as well as  individual climatologiesfor the Caribbean, Gulf of Mexico, and larger areas of the Atlantic basin to provide more localized data. The results from these tables provide  insight into correlations in the number of warnings issued and changes in instrumentation, data quality, techniques, location, and climate variations. These findings will  help aid forecasters in their work and could be an educational tool to help keep mariners in the Eastern Pacific and Atlantic basins informed about these events. By  creating educational tools, mariners and countries affected by such events will be better protected from them, thus upholding NOAA’s mission to protect life and  property. In this presentation we will discuss the background, methodology, and results of the project, as well as possible next steps for using this data in the future.

The Marine Biodiversity Observation Network (MBON) supports advances to better understand spatial and temporal changes in sea ecosystems using environmental  DNA (eDNA) analysis. Marine eDNA is any extractable DNA from a volume of seawater left behind from living cells, shed tissue, or other dissolved molecules by  organisms residing in that ecosystem. The large number of derived inferences available from the raw data collected from marine eDNA together with metadata  observations have the potential to inform marine conservation and management by creating an improved information management system to track ecosystem statuses.  eDNA samples were collected every two months from 2015 through 2024 from 50 inshore and offshore stations in South Florida waters. This project characterizes these  samples’ spatial and temporal microbial community distributions and biodiversity using 18S ribosomal RNA metabarcoding. Sequencing data and metadata are run  through the Tourmaline amplicon processing workflow using QIIME2 and Snakemake, then analyzed in Jupyter notebooks using the Python packages pandas and  Seaborn. These data will be used to inform various state and federal agencies of the spatial and temporal shifts in the microbial diversity of South Florida waters.

Machine learning (ML) has become increasingly popular in geosciences due to its ability to skillfully model complex processes. One commonly used ML method is  eXtreme Gradient Boost (XGBoost) which uses a decision tree-based method to process tabular data. An XGBoost classification model creates decision trees such that  each new tree attempts to improve on the errors of the previous tree, creating a more accurate model. The outputs of these trees are then combined to create the  model prediction. Previous work within MDL has explored the use of ML methods such as Random Forests (RF) and Convolutional Neural Networks (CNNs) to predict  lightning probability across the Contiguous United States using observed total lightning from Earth Networks and Vaisala and High Resolution Rapid Refresh (HRRR)  output. XGBoost was used to attempt to improve on these previous ML methods. When using XGBoost with inputs from a single location, the model produced results  similar to the RF model. When using inputs from the area around a point, the model produced significant improvements over the single-point model with results similar  to the spatially-aware CNN. In addition to model development, explainable artificial intelligence (xAI) methods which break down predictions into the contributions  from each input were applied to the XGBoost model. These methods showed that the observed lightning input was the largest contributor for short-term forecasts  while thermodynamic variables were the largest contributors for longer forecasts. In this presentation, the techniques used will be described and the results  summarized.

Traditional genesis potential indices incorporate both thermodynamic and dynamic variables, but recent research suggests that dynamic environmental factors may be  more reliable indicators of TC genesis (TCG) in a warming climate. This study investigates the importance of Dynamic Genesis Potential Index (DGPI) variables in  distinguishing between genesis and non-genesis systems in the Central Pacific during ENSO-neutral years from 2001 to 2019. IBTrACS storm track data was combined  with Climate Data Store ERA-5 reanalysis data used to calculate the four key DGPI variables: vertical wind shear (vs), vertical pressure velocity at 500 hPa (omega500),  meridional gradient of zonal wind at 500 hPa (uy), and absolute vorticity at 850 hPa (vort850_abs). Statistical analyses, including the Mann-Whitney U-Test and  Hotelling's T-squared Test, revealed significant differences in the distributions of vs, omega500, and vort850_abs between genesis and non-genesis cases. A Random  Forest model demonstrated 72% accuracy in distinguishing between genesis and non-genesis events, with vertical wind shear reported as the most important  contributor (29.6%), followed by omega500 (25.4%), uy (22.9%), and vort850_abs (22.1%). This balanced importance suggests that a combination of dynamic factors is  crucial for TCG prediction. Findings support the usage of dynamics in TCG indices, especially for future climate projections where thermodynamic factors may be less  reliable, providing insights into predicting genesis events in the Central Pacific region during ENSO-neutral seasons. While this study focused on four dynamic variables,  future research could investigate how other factors may modify these results.

Determining and quantifying the sources of methane (CH 4), which is a potent greenhouse gas, is crucial for ensuring effective science-based policy development to  reduce and regulate these emissions. An investigation of emissions of greenhouse gases and air pollutants was conducted in the Denver-Julesburg Oil and Gas Basin  (DJB). Methane sources in this region include Oil and Gas (O&G) operations (thermogenic methane sources emitting methane and ethane) or Concentrated  Animal Feeding Operations (CAFOs) and landfills (biogenic sources that emits only methane with little ethane). We compared mobile measurements of methane and  ethane at the surface the oil with and aircraft measurements of the ethane and methane in the air to calculate the dilution factor of the ethane to methane ratio. Using  this dilution factor we estimate the relative contributions to the total methane emissions from these two categories of methane sources in this area. We compared the  results from this study with a previous study conducted in fall 2023. Results from this study will help make targeted policiesto reduce methane emissions from this  area.

There is increasing demand for probabilistic guidance to inform maritime weather forecasts and warnings issued by the Ocean Prediction Center (OPC). Probabilistic  forecasts of winds exceeding marine warning thresholds (Gale, Storm, Hurricane-Force) can be computed from ensemble systems like the Global Ensemble Forecast  System (GEFS). This project assessed the usefulness of available oceanic probabilistic wind guidance and explored different techniques for generating and interpreting  such forecasts using GEFS data. Experimental probabilistic guidance for wind and wave hazards during the Week 2 time frame was developed in prior work at OPC. This  experimental product selects the tail of a distribution function and computes probabilities of winds exceeding warning thresholds within constant spatial  neighborhoods. These Week 2 forecasts were compared against weekly composites from the Global Data Assimilation System (GDAS), and the utility of using the GDAS as a verification tool was analyzed. Another method of computing probabilities utilizing ensemble agreement scales (EAS) was investigated. In this method, the  spatial neighborhoods used to compute probabilities vary based on the similarity of the ensemble members’ forecasts. The EAS method was demonstrated to have  sharper and smaller regions of high probability associated with Greenland tip jets. The EAS method may better highlight the regional extent of orographically forced  wind events.

The NOAA’s Ocean Acidification Program (OAP) runs annual ocean acidification research cruises on the East Coast, on the West Coast, and in the Gulf of Mexico. On each  cruise, oxygen is sampled in two ways: through titration, and through CTD sensors. Titrations are more accurate but also consume a lot of time. This statistical analysis  aims to estimate the minimum number of oxygen titrations that are needed to accurately calibrate the CTD data on ocean acidification research cruises around the USA.  Using R and RStudio, titration oxygen profiles were compared to sensor oxygen profiles graphically and numerically to determine their relationship at different depths,  locations, temperatures, and salinities. The central analysis calculates the amount of oxygen titration values that can be replaced with uncalibrated CTD data before the  calibrated CTD data deviates by 1% or more. This calculation was then plotted using various graphs and maps in order to assist cruise personnel in determining how  many titrations should be done for any given site’s location and depth, as well as to provide some suggestions based on the data. Future project development will  consist of developing a Shiny App that will make this project’s findings more accessible and interpretable.

The use of satellites for weather has made great strides in research and development of public safety. To maintain a “birds eye view” of weather events, satellites must  have consistent communication with ground stations. There are many factors that lead to the use of satellite communications. Satellite communications are dependent  on the location of the satellite in the atmosphere, the inclination of the orbit, the frequency that the satellite is transmitting on, and the ground stations that hold  dynamic links with those corresponding satellites. For other parties to launch satellites, they must go through a process where their satellite communication is analyzed with the variables that they provide to see if there is interference with one of ours. This interference is referred to as “mutual visibility” and frequency coordination. If a  satellite has mutual visibility with one of our own, it does not mean that anything needs to change unless further analysis shows that the interfering satellite surpasses  a threshold put in place by the victim satellite. If the threshold is met, the other party cannot proceed with their interfering and must employ mitigation techniques to  their satellites.

The Environment Modeling Center (EMC) produces dozens of numerical weather prediction (NWP) models that are deployed for operation across the other National  Centers for Environmental Prediction (NCEP), as well as the rest of the National Weather Service (NWS). In recent years, machine learning weather prediction (MLWP)  models have become increasingly popular and competitive with traditional NWP models. One breakthrough model that was recently developed was GraphCast, a global  weather model that utilizes graph neural networks (GNNs) by Google that outperformed a comparable NWP model 90% of the time. From this model, a group of  researchers are developing a similar model based on GraphCast with the key difference of making regional forecasts instead of global forecasts called Neural-LAM. The  EMC believes that this state-of-the-art model can be utilized to enhance their current numerical ozone prediction model. This project involves utilizing the recently released AQMv7.1 dataset, which contains atmospheric and meteorological variables that all facilitate air pollution, including average O 3 , average NO 2 , u and v  components of wind, specific humidity, and temperature. This is planned to be a multi-year effort with constant communication between NOAA and the John Hopkins  University’s Applied Physics Laboratory (JHU APL), so this project is only in its preliminary phase. Hence, the data being experimented on now is only a subset of what  will be used in the final product and will be expanded on in the future. The responsibility of this project will be transferred to the JHU APL, who will continue with it  after the summer concludes.

The Storm Events Database documents the occurrence of thunderstorms or other severe weather phenomena that can cause damage and disruption to lives and  property in the United States, dating back to 1950. The National Weather Service enters the data to the Storm Events Database from a multitude of sources, but as of  October 2006, uses comma separated text files (CSV) to import data to the Database. The SED provides users two ways of accessing the Database: a narrative text search  option and a state and area option. Both options lead to the same Event Details page for a said event. An improvement to the Storm Events Database is by integrating  radar loops from the Next-Generation Radar (NEXRAD) Archive and satellite imagery from the University of Wisconsin-Madison’s Space Science and Engineering Center’s (SSEC) GOES Online Geostationary Archive into the Event Details page of the Storm Events Database to help users get a better perspective of the size and scale of  thunderstorms (whether if a thunderstorm was isolated or part of a larger system, and if the system was synoptic-scale or mesoscale). This project was done using the  Open Information Stewardship Service (OISS), a holistic metadata and workflow organization layer developed by NCEI contractor Ryan Berkheimer. OISS was used to  create a process to store data from NEXRAD and SSEC that matched SED terms entered by a user. A further improvement to the SED could include an image-uploaded  search in conjunction with the text search to improve chatbot meteorological phenomena identification.

Tropopause folding is a vital mechanism influencing the stratosphere-troposphere exchange (STE) and has significant effects for both atmospheric composition and  dynamics. This mechanism is pivotal in transporting ozone-rich, dry stratospheric air, into the troposphere, contributing to thetropospheric ozone budget. Building on  previous studies, our research integrates components of the tropopause folding model (Pan et al. 2014). In this model, there is emphasis on exchanges from synoptic  scale weather systems to refine the complexities through which these atmospheric interactions occur. Our study investigates cases of stratospheric air intrusion in  Idabel, Oklahoma during April 13th–April 20th, 2011. This particular event was associated with a notable severe tornado outbreak in the southeastern United States  characterized by an enhanced upper-level jet stream and a negatively tilted trough. This study uses high-resolution meteorological data, ozonesonde measurements,  and the HYSPLIT trajectory model to better characterize the evolution and impacts of this event and provides an excellent, in situ data set aligned with the Pan et al. model.

As wildfire intensity and frequency escalate in times of unprecedented climate change and severe droughts, it is important to enhance forecasting models that accurately and efficiently predict wildfire smoke movement. NOAA’s experimental Rapid-Refresh Forecasting System- Smoke/Dust (RRFS-SD) aims to improve upon the  High-Resolution Rapid Refresh- Smoke (HRRR-S), by extending the domain to include all of North America and incorporating a broader range of variables, including  smoke particulate concentrations. Our work explores the process of adding satellite Aerosol Optical Depth (AOD) data as an additional method of verification of the  RRFS-SD experimental model. Using a case study on early-summer wildfires in Alaska and Canada, we analyzed AERONET and AIRNOW data alongside the experimental  RRFS-SD model, applying the Real-Time Verification System to assess and explain discrepancies in AOD and PM 2.5 concentrations. Our findings provide insights into the  strengths and limitations of RRFS-SD, contributing to the refinement of wildfire smoke forecasting.

The Environmental Modeling Center's Coupled Modeling System provides the essential foundation for operational forecast systems at NOAA. WaveWatch III (WW3) is a  third generation spectral wave model which is used in several center’s applications to forecast wave conditions globally and regionally. WW3’s capabilities have recently  expanded to include running on unstructured meshes. Several packages can generate these meshes, including OceanMesh2D, Gmsh, and JIGSAW. This project  investigates the reproducibility of various mesh generation techniques for wave modeling by employing a series of multiple regression models on a high-performance  computing platform and post-processing the results to evaluate the performance of each modification in mesh refinement. A key aspect of this research is the creation  of meshes specifically designed for regional modeling, which require high-fidelity mesh nearshore to accurately capture complex physical processes. These criteria  include the proximity to the coastline, the gradient of the coastal bathymetry, and the number of waves that need to be accurately resolved. Furthermore, this project  explores the potential advantages of renumbering nodes within the generated mesh, a technique aimed at enhancing computational efficiency and possibly reducing  the computation times.Through a comprehensive series of experiments and comparative analyses, this approach aspires to establish the best practices for mesh  generation in regional wave modeling. By doing so, it aims to substantially improve the operational capability and accuracy of wave forecasting models.

The purpose of this project is to provide the scientific arguments for a Hyperspectral Microwave Sensor (HyMS) to support weather applications. This project was done  using simulated HyMS and operational ATMS data. Simulated data was generated using the Community Radiative Transfer Model (CRTM) and Global Forecasting  System (GFS) model data. The simulated data accounts for instrument noise characteristics. The ATMS sensor provides microwave observations at 22 channels spanning  from 23 GHz to 183 GHz. On the other hand, simulated HyMS data provides simulated microwave observations at 1266 channels between 12 GHz and 220 GHz. The  simulated HyMS and operational ATMS data are run through the Multi-Instrument Inversion and Data Assimilation Preprocessing System-Artificial Intelligence  (MIIDAPS-AI) to retrieve atmospheric temperature, water vapor, and cloud parameters for each hurricane case. In this work, the quality of the retrieved geophysical  parameters was assessed with the GFS model data. The assessment of the retrieved data involves comparisons of HyMS and ATMS to the GFS model output, it involves  statistical analysis of this data with the GFS output. The performance of HyMS is expected to represent an enhancement of the ATMS performance due to the increase in  the information content associated with an improvement in the spectral resolution sampling in the microwave region.

The ongoing development of the community-based Unified Forecast System (UFS) brings together the Weather, Water, and Climate Enterprise to innovate and improve NOAA’s operational numerical Earth system prediction applications. The Weather, Water, and Climate Enterprise is composed of academia, industry, private sector, and federal government. The UFS Student Ambassador role builds on work completed by previous William M. Lapenta interns, focusing on outreach and communication  activities with academia. Outreach includes surveying students and academia attending the Unifying Innovations in Forecasting Capabilities Workshop (UIFCW). This presentation will highlight outreach activities accomplished by the UFS Student Ambassador throughout her William M. Lapenta internship. She will also highlight recommendations to the community on how to better engage with and involve students working with the UFS. Additionally, the UFS Student Ambassador will provide  updates to the UFS Student Engagement Plan to encourage the next generation of students to join the UFS community. The Student Ambassador performed student  outreach through a survey, directed towards students and academia that attend UIFCW 2024. These survey responses influenced the recommendations given by the  ambassador in updates to the Student Engagement Plan. The work completed by the ambassador is significant to improving community engagement within the UFS.

Data assimilation (DA) is a critical component of modern weather forecasting and earth system modeling, it enables the integration of atmospheric observations into models to increase forecast accuracy. This project aimed to create an automated DA diagnostic visualization and monitoring tool with Python. The tool will automatically read GSI and JEDI diagnostic files and create a comprehensive set of statistics, plots, and maps of key assimilation metrics like OmB (Observation minus  Background) and OmA (Observation minus Analysis). The motivation for the tool is to facilitate and speed up analysis of DA performance in both real-time and  retrospective scenarios, allowing for the timely identification of errors and comparison of models The tool can be run with a bash script and configuration file for  integration into real-time workflows while also capable of being used in an interactive mode on Jupyter Notebooks. To demonstrate the capabilities of the tool, a case  study comparing data assimilation of temperature observations in Colorado between the Rapid Refresh Forecast System (RRFS) model and the Real-Time Mesoscale  Analysis (RTMA) model was done. RTMA tunes the data assimilation of surface temperature observations but not upper air temperature observations, so it is important  to confirm this difference was present. The tool was used to create time series, maps, plots, and statistics, which identified these differences and confirmed proper  functioning of the data assimilation tuning in RTMA.

Rapid warming of the Arctic is causing widespread melting of permafrost on Alaska’s North Slope. A new NOAA atmospheric observatory building in Barrow, Alaska,  built-in 2020 on piles anchored into the permafrost, could one-day face structural risks from this melt. In an effort to minimize these risks, NOAA/GML installed an  instrument system to monitor ground thaw and provide advanced warning on melt conditions that may impact the building infrastructure. This system was installed in  September 2023 to measure permafrost temperature at various depths and locations around the observatory site. This study begins by describing the temperature monitoring system hardware, its installation and examples of the data collected during the first 9 months of its operation, along with an initial assessment  of the data quality. Furthermore, sources of uncertainty were identified to address the next steps that should be taken to verify the system's performance. Finally,  this temperature data may potentially introduce future scientific research and monitoring benefits providing an example to enhance larger permafrost thawing  monitoring programs in the Arctic regions or even globally. This data, including derived permafrost thawing rates could also potentially provide scientific foundations for better understanding greenhouse gases contributions to Earth’s total energy budget.

The National Data Buoy Center (NDBC) engineers aim to develop a method for retrieving an experimental concept mooring design to be deployed in 2025. I researched  and evaluated different ways in which the NDBC could retrieve their moorings. The motivation for recovery are environmental benefits, reusability of components, and  post-deployment study. The new concept mooring design will be tested on 2-3 buoy stations and if successful could improve reliability by potentially replacing many  existing Weather and DART buoy moorings. NDBC’s current buoy types are Weather, DART (Deep Ocean Assessment and Reporting Tsunami), and TAO (Tropical  Atmospheric Ocean). The buoys are positioned with a mooring line that could potentially reach up to 6000 meters depth and these moorings are divided into upper,  middle, and lower sections. Some of the researched methods that could address these difficulties are an airbag system for the anchor, AUV’s and ROV’s, but these are  currently too expensive or not suitable for deep water depths. The recommended retrieval method for the NDBC consists of a three tiered method. The primary  retrieval method includes an acoustic release, which is presently in use. Should this method prove ineffective, a line cutter will be deployed. As a final contingency, an 11/16” nylon rope section could be installed, to serve as a weak link when recovering the mooring.

The accurate prediction of Southern Ocean clouds poses considerable challenges in many numerical models and results in large discrepancies in short wave radiation  and the resulting global energy budget. The Common Community Physics Package Single Column Model (CCPP SCM), produced by the Developmental Testbed Center  (DTC), is used to extract vertical columns from the Unified Forecast System (UFS) and uses horizontal advection and prescribed vertical velocity to mimic how the  surrounding environment impacts the column state, effectively replacing the forcing from dycore. Physics suites enabled with aerosol aware cloud microphysical  parameterizations are used to explore the representation of clouds over the Southern Ocean in the UFS. The comparison of UFS data with flight and radiosonde data  from the SOCRATES field campaign facilitates the evaluation of UFS prediction skill. Furthermore, this research aims to assess if shortwave radiation biases in coupled  global climate models are also present in the UFS. Improvements to the physical representation of clouds in the SCM are utilized to decrease biases with the goal of  proposing solutions for improving cloud simulations and reducing shortwave radiation biases in numerical models.

The Visible Infrared Imager Radiometer Suite (VIIRS) Vegetation Index (VI) is an operational product that NOAA/NESDIS developed to produce Top of Atmosphere (TOA),  Normalized Difference Vegetation Index (NDVI), Top of Canopy (TOC), NDVI, and Top of Canopy Enhanced Vegetation Index (TOC EVI). The VIIRS sensor is attached to  the Suomi National Polar-orbiting Partnership (S-NPP), NOAA-20, and NOAA-21 satellites and collects data as input to the VIIRS VI operational product system. The  motivation of this project is to validate the algorithm through agricultural monitoring of Knox County, Illinois by producing high-resolution vegetation data. The data  used are Moderate Resolution Imaging Spectroradiometer (MODIS) and Sentinel 2(S2) data to downscale lower resolution to higher resolution imaging. As an additional  step, we show subset data during the downscaling process. The output of the modeling system is thus compared to SNPP and NOAA20. However, the product’s output  data creates grounds for a deeper investigation of discrepancies or limitations of the application (Carter et al., 2019).

Each year, the Weather Program Office (WPO) hosts funding competitions to support weather research, inviting academic and private sector stakeholders to submit  proposals aimed at advancing weather forecasting, developing innovative weather-related products and services, and improving the delivery of life-saving information  to the public. To gain insight into applicants’ experiences applying for these funding opportunities, WPO first launched the Applicant Customer Experience and  Satisfaction (ACES) survey at the close of the Fiscal Year (FY) 2023 competition. This feedback was then used to refine the FY2024 funding competition application processes. Notable changes from the FY23 to FY24 competition included releasing Letter of Intent (LOI) feedback two weeks earlier, standardizing the quality of LOI  feedback across competitions, and promoting additional resources located on the WPO website. The present research sought to evaluate the effectiveness of these  changes by comparing the 2023 and 2024 ACES data. The results showed applicants were more satisfied with the timeliness of receiving LOI feedback. However,  discrepancies in the constructiveness and clarity of reviewer feedback between the two competitions persisted. Additionally, applicant awareness of the resources on the WPO website remained low, with 41.18% of applicants unaware of at least one of the website resources. This analysis also identified key gaps in applicant  experience data. To address these gaps, additional questions will be proposed for inclusion in the 2025 ACES survey and a focus group protocol to solicit more in-depth  applicant feedback specifically on the proposal review, selection, and award processes has been developed.

The National Weather Service (NWS) provides near real-time Flood Inundation Mapping (FIM) services depicting flood impacts for neighborhoods and civil engineering  infrastructure. Baseline mapping capabilities have been established using the Height Above Nearest Drainage (HAND) terrain model derived from 10-meter resolution  digital elevation model (DEM) data. The NWS is interested in understanding how high-resolution 1-meter DEMs increase the skill of FIM capabilities to better facilitate  the provisioning of impact-based decision support services (IDSS). However, the computational demand of using 1-meter data for both the HAND model and FIM is  substantial. This study investigates the cost efficiencies of employing a subgrid modeling approach to generate high-resolution, 1-meter relative elevation model HAND  data for producing inundation maps. 1-meter resolution inundation results are compared to those generated using 10-meter resolution data. Inundation accuracy  evaluation is focused on an infrastructure impact-based skill assessment rather than a pixel coverage FIM forecast benchmark comparison. This method allows for  meaningful assessment of how model enhancements improve IDSS. Initial results indicate that storage requirements for the subgrid HAND approach are 290 times  greater than those for 10-m HAND for a single HUC8 watershed. These findings will guide future FIM model development, ensuring more accurate and reliable flood  impact predictions that also balance computational efficiency.

The United States state of California is vulnerable to extreme precipitation events and has experienced widespread associated hazards including flooding, landslides, infrastructure damage and economic losses. Atmospheric rivers (ARs) are responsible for much of the extreme precipitation acrossthe state of California, and despite  their associated hazards, can provide well-needed precipitation during periods of drought. In order to properly prepare for AR storms - both by mitigating flood risk and  improving water storage management - we must improve the forecasting of such events. Certain regions present gaps in observational data that, if filled, could improve  forecast modeling through data assimilation. In an attempt to fill these data gaps and improve AR forecasting, the AR Reconnaissance (AR Recon) project uses aircraft to  release dropsondes over the North Pacific within impactful landfalling ARs. Here we evaluate the influence of such dropsonde data on forecast accuracy during a  particularly influential sequence of ARs from January 6-18, 2023. AR Recon efforts during this event consisted of 13 operations, 23 flights, and 625 dropsondes. Upper-and-lower-level dynamics, moisture transport, and precipitation forecasts from control (CTRL; includes dropsonde data) and denial (DENY; excludes dropsonde  data) runs of the NCEP Global Forecast System (GFS) were compared against the Global Data Assimilation System (GDAS). Dropsonde inclusion leads to an improvement  in the GFS model initial conditions as well as forecasts of large-scale circulation patterns, moisture transport, and precipitation, indicating the importance of data  collection during influential precipitation events in California.

Climate modes are typically defined as large scale patterns of coupled atmospheric and oceanic variability occurring over interannual to multi-decadal time scales.  Identifying ocean regions of covariability allows for understanding these climate modes and the underlying dynamical processes. Here we used hierarchical clustering on  sea level anomalies across observational datasets and dynamical model output. These clusters can be used to assess the fidelity of climate models, specifically with  regards to large scale ocean dynamics. Using ~25 years of monthly sea level anomalies from satellite altimetry data and NOAA GFDL’s coupled (CM4) and ocean only  (OM4) models we implement our hierarchical clustering algorithm and explore model-observational similarities. With this tool, volume distribution patterns become more clear, and we can relate inner-cluster volume redistribution to climate mode indexes to gain physical understanding of our hierarchical clusters. These results  point towards some spatial discrepancies of climate modes in coupled models, and similarities in ocean only models. Further work could be done to cluster subsurface  ocean regions utilizing climate models, to visualize areas unavailable when using altimetry data. These methods can also be applied to non-sea level related data and  can be used to find other regions of co-variability. This assessment and associated tool can inform model development and ultimately lead to improvements in climate simulations.

Measured and modeled lunar irradiances have long been used for vicarious satellite calibration. This project advances a novel approach using lunar radiance, defined as  the irradiance per unit solid angle. Three areas of interest (AOIs) were selected for their widespread visibility and differing radiances on GOES-16 satellite images.  Coordinates of each AOI were manually obtained from a reference GOES-16 image. A key challenge is projecting these coordinates onto a G-16 sample image when  phase angle and libration differ. To address this, the detection and matching of multiple control points across images is essential. Previous methods employed classical  computer vision techniques like SIFT. Here, a new algorithm was developed, leveraging robust AI techniques such as ConvolutionalNeural Networks (CNNs) and Graph Neural Networks (GNNs). SuperPoint, a self-supervised CNN architecture, was applied on the reference and sample images to extract control points and their  descriptors. Unlike traditional methods which require extensive labeled data for training, SuperPoint is trained on synthetic data generated from homographic  transformations, allowing it to effectively learn and describe interest points in various conditions. Next, the SuperGlue GNN architecture was applied to match control  points. By leveraging the spatial relationships and descriptor similarities from SuperPoint, SuperGlue is able to optimize the matching process. A homography transformation was then calculated from these matches to project AOI coordinates from the reference to the sample image, enabling lunar radiance extraction.  Prediction models were constructed using time and solar/view zenith angles as proxies for lunar radiance.

Rip currents - dangerous currents of water moving away from the coastline - continue to be a leading cause of coastal drownings and rescues, with an annual average of  100 fatalities per year. Despite advances in technology and the accessibility of the National Weather Service (NWS) rip current Beach Hazard Statements, rip current  deaths have not been prevented. Local Weather Forecast Offices issue these statements during a moderate to high likelihood of rip currents forming at the shore. As of  July 10, 2024, 21 fatalities have been reported due to rip currents, highlighting the urgent need for more accurate rip current modeling, forecasting and improved risk  communication. This project aimed to improve this forecast resolution by refining the experimental National Ocean Service (NOS) probabilistic rip current forecast  model using machine learning techniques. The model analyzed several datasets, such as water level data, wave data and qualitative rip currents lifeguard reports. To  verify the accuracy of the model, lifeguards reports for Wilmington, NC were compared with the model predictions. Additionally, this project aimed to further improve  NWS rip current risk communication and hazard statements by conducting a focus group for various stakeholders (emergency managers, lifeguards, media, among others). This feedback will be used to inform model development and improve probabilistic rip current forecast models as well as improve mechanisms to better  communicate risk and help beachgoers to avoid or get out of dangerous conditions.

Providing reliable and timely access to environmental information from satellites is a critical function and need at the National Oceanographic and Atmospheric  Administration (NOAA), especially for numerical weather predictions. Innovations in this sector are experimenting with new data collection methods, enhancing data  quality by reducing Radio Frequency Interference (RFI), and developing new methods for measuring environmental variables. Ground-based bistatic radars using  Software Defined Radios (SDRs) are being developed to detect and utilize “Signals of Opportunity'' to measure variables such as sea level rise, ocean surface wind  speeds or soil moisture, or to reduce the noise of a signal, to validate satellite products. For instance, GNSS Reflectometry is a method used in both satellite and ground based observations to monitor the Earth. In this study, a rotating GNSS antenna is used to separate the reflected signal from the direct GNSS signal, to support the  validation of satellite measurements over water and land. Open source software, such as RTKLIB and GNSS-SDRLIB, were used to receive and process the raw I/Q data.  This new and active research area has a great future for the improvement of satellite data collection, processing and validation, especially for the purposes of enhancing  our understanding and predictions of our changing environment. This study will help validate products that retrieve and analyze data from satellites in new and  innovative ways, supporting current and future satellite missions like CYGNSS and SNOOPI.

Fire weather is a fast-growing field of research due to recent severe events causing extensive damage and loss of life. Scientists at the Storm Prediction Center (SPC) and  throughout the National Weather Service (NWS) are developing tools to better forecast fire weather to provide those in high-risk areas adequate preparation for  wildfires. Understanding climatology trends is essential for forecasting fire risk areas in the future. Data from the Fire Program Analysis Fire Occurrence Database spans  the years 1992 to 2020, creating a 29-year collection of data near to the 30-year standard for examining climatology. This data allows for a study of climatological trends in wildfire location, size, number, and occurrence time of year. Another useful tool is examining the probability of a fire occurring at any location for a given day,  month, or season. Statistical analysis of the trends shows a significant increase in the mean size of fires throughout the country. Regional trends reveal variations from  the overall observations, including a relative increase in fire size in the Western U.S. Additionally, only wildfires in the contiguous United States (CONUS) have been  used to create a mapped climatology in the past, leaving out Alaska and Hawaii. Mapping these risk areas helps expand wildfire prediction and forecasting for these  locations, which are also being adversely affected by climate change. By understanding which areas are becoming more prone to fires, in particular large fires, more  time and resources can be allocated to the locations in preparation for damaging events.

The impact-based decision support services (IDSS) approach first emerged alongside the Weather-Ready Nation concept in 2011 and has transformed National Weather  Service (NWS) operations. The aim of IDSS is to better inform and prepare partners for hazardous weather events, ultimately influencing actions and decisions that can  prevent injury, loss of life, and damage to property. Impacts of hazards are not uniform across the nation, however. The types of impacts and their magnitudes can vary  dramatically depending on characteristics of the affected community. Currently, there is a notable lack of vulnerability information available with prognostic NWS products that could aid partner decision making. This project explores a potential method of intertwining vulnerability and risk factor data with existing Weather  Prediction Center (WPC) and Storm Prediction Center (SPC) products to facilitate IDSS and situational awareness in NWS Western Region. Social vulnerability data was  acquired from the 2020 U.S. Census and the Centers for Disease Control and Prevention’s Social Vulnerability Index and PLACES project. Social vulnerability data was  then spatially joined with WPC’s HeatRisk and Excessive Rainfall Outlook and SPC’s Convective and Fire Outlooks. Using bivariate analyses, forecast hazards and vulnerabilities can be displayed simultaneously, emphasizing areas that are most likely to see significant impacts from the hazard.

Sea ice monitoring is a well-established application of NOAA Joint Polar Satellite System (JPSS) observations, one that is essential in the Arctic cryosphere, where rapidly  warming temperatures have increased sea-ice melt. Since this condition can have both positive and negative impacts on Arctic nations, NESDIS data and derived services  help decision-makers make short and long-term forecasts to improve livelihoods and promote sustainable economic growth. This ongoing study of the socioeconomic  impact of sea ice products is centered on a value-tree analysis, as depicted in our Sankey diagram. Based on conversations with end-users and data assessed in the  NOAA TPIO database, the diagram demonstrates the relative impact of Lower Earth Orbit (LEO) sensors on user sectors, through the usage of sea ice products. Three user sectors were further selected from the group based on case studies assessing the benefits of the products: 1) maritime transportation and services, 2) oil and gas  industry, and 3) fisheries and aquaculture. The Sankey diagram reveals which observing systems contribute to which products, and the importance of such  contributions. To begin its construction information from the NOAA TPIO database (foundational data source) was sorted under five categories that were linked, from  left to right, in the following order: Sensor —> Product Family (KPS) —> Outcome —>Mission Service Area (MSA) —> User Sector. Next, the following percentage  relationships were calculated : weight of sensor to KPS, weight of KPS to Outcome, weight of Outcome to MSA, and weight of MSA to User Sector. We observe that the  relationships between categories and weight metrics flow from one to the next; the wider, curved bands connecting each of the categories are the flows (or links) which  are sized proportionally to the quantitative values they represent.

Ocean Exploration reveals new species inhabiting the water column each year, many of which are sensitive to marine heat events. While satellites provide thorough sea  surface temperature coverage, generating subsurface ocean temperature data necessitates in-situ observation platforms such as the Argo array. This project aimed to  identify marine heat anomalies in 2023 and 2024 using data collected from Argo floats in the Gulf of Mexico. Climatology data from 2005-2017 available from the World  Ocean Atlas was compared to Argo data from 2023 and 2024 to establish baseline temperatures. The Argo data broadly exhibited warmer temperatures than the  climatology at all depths measured for both years. 2024 data displayed much warmer summer temperatures than the climatology in comparison to 2023. The temporal  resolution of the Argo data was determined to be too coarse to identify marine heatwaves by the accepted definition. The results from this project support the  continuation of Argo float deployment in the Gulf of Mexico to obtain better spatial and temporal data resolution to observe future extreme heat events. This greater  data coverage could improve sensitive species conservation efforts in the Gulf by identifying refuges and hotspots vulnerable to heat anomalies.

Accurate wintertime precipitation forecasts and estimates enhance public safety and are criticalfor water supply forecasts. Complex terrain is known to make forecasts  and estimates of precipitation challenging for the High Resolution Rapid Refresh (HRRR) model and Multi-Radar Multi-Sensor (MRMS) product. We hypothesize the  quality of gridded precipitation products is related to the availability of radar observations. To assess this, cool-season precipitation from the HRRR 1-hour forecast, the MRMS pass 2 product, and lidar-derived Snow Water Equivalent (SWE) data were used. The difference in seasonal accumulation from HRRR and MRMS  was compared to radar quality index (RQI) values within five Colorado watersheds. Since not all cool-season precipitation falls as snow, and lidar-derived SWE  measurements observe the state of the snowpack which may melt before it is observed, it is expected that the HRRR and MRMS precipitation will be greater than what  is observed by the lidar-derived SWE. Where adequate radar data existed (RQI >= 0.6), the HRRR and MRMS precipitation totals were greater than the lidar-derived  SWE. Where inadequate radar data existed (RQI <= 0.3), the HRRR and MRMS precipitation was less than the lidar-derived SWE. This means that HRRR and MRMS were  both underestimating wintertime precipitation. Where RQI was between 0.3 and 0.6, HRRR and MRMS precipitation were substantially lower than the lidar-derived  SWE. In this case, the models were substantially underestimating. These results suggest that patterns of seasonal precipitation bias exist within HRRR 1-hour forecasts  and the MRMS pass 2 product in relation to RQI.

Microplastics are a ubiquitous environmental contaminant that has been documented in marine animals. To account for this growing marine pollutant, the National  Center for Environmental Information (NCEI) is analyzing and adding submitted animal data to its marine microplastics geodatabase. This analysis includes both ArcGIS  spatial analysis and statistical analysis using R Studio. Out of the 21 species analyzed, S. maderensis and S. aurita showed positive relationships between microplastics  concentrations and both sample length and weight . This relationship between sample size and microplastics concentrations can be attributed to individual microplastic  exposure accumulating with age, as well as larger individuals experiencing trophic transfer from their prey. D. angolensis and M. merlangus also exhibited a similar but  weaker relationship due to the differing environmental characteristics of their respective habitats. For the species that did not demonstrate an obvious correlation with  microplastics concentrations, further research is needed to better characterize their individual exposure and ingestion of microplastics. The spatial analysis done on  each species did not turn out to be as valuable as expected, due to the lack of spatial diversity in some of the samples collected. To better comprehend the potential  spatial relationships that marine animals may have with their microplastics concentrations, additional sampling is needed in areas with better spatial diversity.

Marine mammals are highly vocal and depend on sound for essential activities such as finding food, communicating, reproducing, and detecting predators. Because they  rely heavily on sound, they can be negatively impacted by anthropogenic disturbances like ship noise and seismic airgun surveys, both of which dominate the Gulf of  Mexico (GOM) soundscape. The LISTEN GoMex (Long-term Investigations into Soundscapes, Trends, Ecosystems, and Noise in the Gulf of Mexico) project collected  acoustic recordings using High-frequency Recording Packages (HARPs) in the western GOM from July 2022 to 2023. Automated detectors classified various cetacean  clicks to species, including sperm whales, which are endangered in the GOM. Using Triton, a MATLAB-based software program, acoustic data were processed into Long Term Spectral Averages (LTSAs) to manually identify acoustic events with spermwhale vocalizations. Then, the toolkit Where’s Whaledo, which was developed to assist researchers in reconstructing the behavior of these animals using arrays of acoustic recording devices, was used to create 3D diving tracks of individual sperm whales by  analyzing the time difference of arrival (TDOA) from two hydrophone arrays, with the ultimate goal of determining the exact location of the whale. This research  underscores the importance of localizing animal sounds, providing valuable insights into whale behavior and population dynamics. The findings emphasize the  importance of understanding the impact of anthropogenic noise on sperm whales and to develop strategies to mitigate these effects.

'Candidatus Aquarickettsia,' a potentially parasitic bacterium, plays a pivotal yet mysterious role in the microbiome of Acropora cervicornis, an endangered coral species  in the Caribbean protected under the Endangered Species Act (ESA). The abundance of this bacteria has been linked to the health and resilience of A. cervicornis,  particularly during thermal stress events. In the face of increasing threats to coral reefs, the decline of A. cervicornis has become a cause for concern. In 2023  unprecedented ocean temperatures were recorded in Florida, which led to massive coral die-offs. In this study, we aimed to understand ‘Candidatus Aquarickettsia’  response to a natural thermal stress event in A. cervicornis across multiple genotypes used for coral restoration. We investigated the role of the ‘Ca. Aquarickettsia’  using qPCR and 16S rRNA analysis before and during the 2023 Florida heatwave. We screened two replicates of 38 genotypes of A. cervicornis at the University of  Miami’s Key Biscayne offshore nursery during low and high temperatures. Tissue biopsies were DNA extracted and used to quantify with qPCR the tlc1 and CAM genes from Aquarickettsia rohweri and A. cervicornis, respectively. DNA was also used to sequence the 16S rRNA for microbiome analysis. The qPCR ratios showed that there  was genotype-specific response to heat stress, with most genotypes showing an overall decline in Aquarickettsia abundance, however, 5 genotypes showed an increase  in Aquarickettsia abundance. The microbiome analysis showed that samples that declined in Aquarickettsia had relative increases in potential putative coral pathogens.  Our data suggests that increases in Aquarickettsia during heat stress may have beneficial roles by reducing opportunistic pathogen increases.

Tsunamis are no-notice, fast-onset ocean hazards that can cause catastrophic humanitarian, social, economic, and physical impacts. To mitigate these risks, after  piloting for 11 years in the Caribbean and other ocean basins, the UNESCO-IOC Assembly approved the Tsunami Ready Recognition Programme in 2022 with guidelines  and indicators to minimize the loss of life, livelihoods, and property to these hazards. This project evaluates the effectiveness of the program in facilitating tsunami  preparedness and response among Tsunami Ready communities within the Intergovernmental Coordination Group for the Tsunami and Other Coastal Hazards Warning  System for the Caribbean and Adjacent Regions (ICG/CARIBE-EWS). It specifically targeted 19 communities recognized since 2019. The survey solicited feedback from  National and Regional Tsunami Ready Boards and Local Tsunami Ready Committees, revealing the program’s pivotal role in bolstering community readiness through  education, outreach, and risk assessment efforts. However, it also identified gaps needing to be addressed, including communication barriers, data accuracy and  collection issues, and resource allocation, especially pertinent for Small Island Developing States. The project also highlighted the necessity for improving and  monitoring program efficacy through the establishment of a national Tsunami Ready Contact, the implementation of the survey upon recognition, and the  implementation of an annual reporting mechanism on Tsunami Ready indicators to enable continual progress monitoring. Ultimately, this project contributes to one of  the two overarching goals of the UNESCO-IOC Ocean Decade Tsunami Programme: "Ensure 100 percent of communities at risk of tsunami are prepared for and resilient  to tsunamis by 2030.”

Arctic sea ice decline has accelerated in recent decades which has enhanced efforts to create accurate models and further the understanding of climate teleconnection  patterns’ influence on the sea ice extent. A recent study utilized hindcast regression models to predict September sea ice extent using teleconnection indices from 1948  to 2000. In this work, we build off the previous study to extend it to 2020 in order to improve prediction accuracy. The monthly teleconnection indices of the Arctic  Oscillation, the Central Arctic Index (a feasible version of the Arctic Dipole Anomaly), the North Atlantic Oscillation, the El Niño-Southern Oscillation, the Atlantic  Multidecadal Oscillation, and the Pacific Decadal Oscillation were used to create hindcast regression models of September sea ice extent. Using the same hindcast  regression models from the previous study which were based on the idea of increasing statistical significance by including the contribution of each teleconnection and  their interactions into a single model. The models were trained using different time periods of 1948-1984, 1948-2000, 1984-2020, and 1948-2020, they were then  compared to the original equation from the previous study. The extended time frame of the current study (1948 – 2020) improved model performance as compared to  the original study (1948 – 2000) while maintaining the 95% significance level.

Wildfires are a preeminent threat to life and property on the Southern Plains. In the aftermath of destructive wildfire outbreaks in Texas and Oklahoma early this  century, policymakers invested in a body of research to support predictive services, firefighting response, and public safety. The resultant Southern Great Plains Wildfire  Outbreak Working Group (SGPWO WG) is a collaborative multi-agency online research-to-operations community focused on science-based support to state forestry  agencies in Texas, Oklahoma, and Kansas. Since 2011, the group has produced probabilistic significant wildfire potential outlooks which have informed strategic pre staging of firefighting equipment and personnel before dangerous wildfire outbreaks. This study quantifies economic impacts of these activities during the 2022-2024  fire seasons in Oklahoma and Texas. We leverage records from Texas A&amp;M Forest Service and Oklahoma Forestry Services corroborated by census data to estimate  the value, scale, and efficiency of decisions influenced by collaborative National Weather Service (NWS) meteorologist and fire analyst predictions. For the 2022-2024  fire seasons, strategic preparatory mobilizations informed by such outlooks enabled crews to save a reported 10,955 structures (80% of homes within fire perimeters in some incidents) worth more than $1 billion at a cost of $13.5 million. Net economic impacts, determined by the value of saved structures minus values of structures  lost, agricultural losses, and expenses of pre-fire resource deployments totaled over $803 million. While inherent limitations in available financial data exist, this work  provides insights into the economic value of the NWS’s impact-based decision support services approach and core partnerships.

The Great Lakes Water Quality Agreement (GLWQA) is a binational agreement between Canada and the U.S. designed to restore and protect the Great Lakes while  recognizing individual national actions that affect this shared resource. The Agreement includes ten focused annexes, including Lakewide Action and Management Plans  (LAMPs) for each lake. A binational agreement with many subcommittees can lead to challenges with communication and collaboration between many different  organizations. Additionally, understanding an organization’s participation, specifically NOAA, can be complex. A series of interviews with each Annex and LAMP NOAA  representative was conducted to gauge what projects each Annex or LAMP are working on, how often each Annex or LAMP meets, and how active a role NOAA plays within the Annex or LAMP itself. These interviews revealed that, depending on the Annex or LAMP, there is a general trend of inactivity or lack of collaboration  within NOAA or within the Annex or LAMP committees themselves. Possible reasons for these ebbs is due to the lack of funding, unorganized communication and  collaboration within the Annex or LAMP committee itself, or simply because NOAAs mission does not align with the needs of the project. Based upon the results of the  surveys, to strengthen NOAA’s involvement with the GLWQA, NOAA should continue conducting projects that align with the Annexes or LAMPs objectives despite the lack of guidance from the GLWQA committee itself, being a more active voice in Annex or LAMP meetings, and collaborating and communicating between the  science and policy supporting the Great Lakes.

Today, the most reliable and accurate total lightning detection network is the Lightning Mapping Array (LMA). However, in a recent study about one-third of lightning  was not detected by the LMA during the DC-3 aircraft field campaign (Brune 2021). The missing lightning was termed “sub-visible lightning”, and was identified by  chemical constituents sampled by the campaign that act as lightning indicators. This study investigates whether sub-visible flashes are present in thunderstorms. To get  more sensitive measurements than the LMA, we used the LongWavelengthArray (LWA), which is a VHF radio telescope located in New Mexico. The LWA is capable of  observing signals as low as -53 dBW/MHz, which is 20 dB more sensitive than the most sensitive LMA deployments. It has 256 VHF antennas with 40 MHz of observing bandwidth at two locations, LWA-SV and LWA-1. All 256 LWA antennas were combined into a single beam to measure radio waves in 10 ms increments continuously for  12-hour periods. The beam is steered to observe directly overhead, and the large number of antennas enables us to detect signals below the galactic background noise  level. We will present a detailed comparison between LMA and LWA observations to identify how much lightning the LMA is seeing. Most LMA flashes aligned well with  LWA signals, but there are emissions the LMA missed. In addition, the LWA observed up to minute-long VHF glows with no corresponding LMA sources, many which  correspond to solar radio bursts identified by LWA Orville imagery.

The Argo program, a decades-long international collaboration to establish and maintain a comprehensive ocean observation network, has revolutionized understanding  of temperature structure in the world ocean. Argo’s core floats allow the structure of the world ocean’s temperature field to be spatiotemporally compared, providing a  valuable resource for analyzing Earth’s climate system. Here we use a 1° × 1° × 1-month map of ocean temperatures constructed using machine learning techniques that  use satellite data as predictors and Argo float temperature profiles as training data. The year 2023 was the warmest in the modern record, with mean ocean  temperatures over 1°C higher than pre-industrial temperatures for the first time. Higher ocean temperatures are an important ingredient for severe weather formation,  as well as increasing ocean stratification, raising sea-levels, and impacting marine ecosystems. In this project, we analyze and compare 2023’s global and regional ocean  temperature fields with those of the previous 30 years, identifying key features of and contributors to 2023’s unusual temperature field. We fit local means, trends, and  seasonal cycles to the 1993-2022 ocean temperature data. We then examine the impact of the mean, trend, and anomalies on the changing vertical temperature structure of the ocean, both globally and regionally. We show that the linear trend, the buildup of El Niño, and unprecedented marine heatwaves all played roles in  2023’s exceptional temperatures.

Forecasting of tropical cyclones presents an unmet challenge, especially for their rapid changes in intensity. Improvements in intensity forecasting have lagged that of  track forecasting in Atlantic and East Pacific basin tropical cyclones. This is partially related to the unknown accuracy of air-sea energy exchanges, such as latent and  sensible heat fluxes as major energy sources for tropical cyclone evolution. Forecast models struggle to resolve these processes due to sparse in-situ observations near  the ocean surface. With the introduction of saildrone uncrewed ocean surface vehicles, surface observations within tropical cyclones have been taken in real-time for  the first time. During the 2023 Hurricane Season, 6 tropical cyclones were intercepted 19 times across 10 different saildrones in the Atlantic. Model validation with  surface observations is crucial in improving understanding of surface processes and forecasting. In this presentation, we will be validating ocean latent and sensible heat  fluxes from the European Centre for Medium-Range Weather Forecasts (ECMWF) ensemble forecast of 50 members across 11 Saildrone observations in the Atlantic  basin from August – October 2023. This study will focus on evaluating forecasts across lead times up to 15 days against observation. Forecast accuracies and errors are  quantified by correlation, error spreads and distributions. This analysis will help shed light on the degree to which the ensemble forecasts resolve sea-air processes and  on how to improve forecasts for tropical cyclone evolution.

The biological carbon pump describes the flow of carbon from marine organisms to the ocean depths, typically through sinking of an organism’s waste or body parts.  This process is an important driver of ocean carbon cycling and impacts other nutrient cycles and influences the productivity of fisheries. Understanding how ocean  particulates affect the biological carbon pump is critical to our understanding of ocean function and identifying changes over time. Learning more about the types of  organisms associated with deep ocean particulates that make up the biological carbon pump provides additional insights into how this process is driven. One approach  to examining the composition of the organisms that make up this flux is through environmental DNA (eDNA), or DNA that is cast off from a marine organism through  microbes, organism parts of the body, or waste products. A sediment trap in the northern Gulf of Mexico at a depth of 550 m is collecting sediments containing eDNA.  Seasonal trends in taxonomy of eukaryotes were assessed using 18S rRNA gene metabarcoding data collected approximately every two weeks for a 6-month period  from December 2021 to May 2022 from the sediment trap. A bioinformatics workflow called Tourmaline (utilizing QIIME2 and Snakemake) was used to process and  analyze the genetic data. The results were exported into a Jupyter Notebook to match eDNA data with other metadata and conduct further analysis. Results indicated  high genetic reads for dinoflagellates, apicomplexa, and opisthokonta metazoa across the full 6-month time series. Comparison to sediment geochemistry indicated a  relationship between unique genetic read count with time and total mass flux. Future research opportunities include additional investigation of the relationship  between unique genetic read count and carbonate or silicate.

The L-band (1-2 GHz) radio frequency is vital for NOAA operations, such as satellite downlink for Earth observation data of hurricanes, and satellite missions like SMAP providing soil moisture data used in weather prediction models. This data also informs irrigation decisions, and crop yield predictions. Radio frequency  interference (RFI) increases the uncertainty of these measurements, making corrupted data unusable. This study investigates the 1420 MHz forbidden hydrogen  transition with software defined radio, identifies RFI, and will experiment with the latest machine learning (ML) methods to distinguish and potentially remove RFI. The  1420 MHz signal is extremely weak, which makes it a useful target to study and distinguish from RFI, due to its similarity tothe weak soil moisture and ocean salinity  signals at a similar frequency (~1419 MHz). Using a ML method has the advantage of using digital signal processing to clean up corrupted data with AI, which can be  done with large amounts of data quickly once trained. Better digital RFI cleanup methods mean less hardware having to be launched on satellites, which leads to  reduced mission costs and better quality data. This method could also be applied to data from previous missions, giving us better long-term prediction and forecasting.

The National Tsunami Warning Center (NTWC), located in Palmer, Alaska, maintains and operates sea level gauges in the Barry Arm fjord, Prince William Sound, to  detect tsunamis from a potentially catastrophic landslide. While this low probability, high-impact event has yet to occur, numerous smaller events have created  measurable wave disturbances, including from ship wakes and glacier calving events. While ship wakes are high-frequency direct arrivals, the calving events generate  seiches in the narrow fjord whose frequencies are independent of the size of the initial displacement. Identifying these events and determining their characteristics can be used to verify numerical models that are used to model a large landslide-induced tsunami that could have a devastating impact on communities in Prince William  Sound. Detecting seiche resonances is challenging because they are low-amplitude events contaminated by additive sensor noise and structured interference, such as  wind-driven waves. Fortunately, their frequency content is repeatable, which allows them to be isolated from the background interference. We identify these events by  decomposing the time series from a known calving event over the frequency band of interest to create a target function prototype. We then perform a hypothesis test  to determine if a test signal is statistically the same as the target signal. Sea level data recorded from 2022 to 2024 are used to identify several events that can be used  for more in-depth oceanographic analysis. This analysis may benefit real-time operations by helping to discriminate between potential wave sources.

Communication is key in all things. Traditionally coined to offer stability within a relationship, The National Weather Service and The National Hurricane Center look to offer people constancy in the evolving digital age. When faced with natural disasters ranging from storm surge to hurricane force winds, the NHC looks to provide  information through channels mostly dominated by influencers and “How To” videos. The change in online communication and the need for reachability and  engagement is new to the agency but not something left untouched. The importance of social media usage is made apparent and this presentation shows its growth and  how the NHC utilizes these tools to inform more of the public over time. By reaching into social media demographics it opens the lines of communication not just  through a Facebook like or a Twitter retweet but by accessing tools to branch into podcasting, livestreams and more. The National Hurricane Center is making sure the  professionals tracking the oceans are the ones giving you the direct information all through an app on your phone.

Rapid onset droughts are associated with quickly developing drought conditions driven by low precipitation and increased evaporative demand which increase heat and  water stress thereby negatively impacting ecological health and agricultural productivity. Flash droughts occur within the subseasonaltimescale (2-8 weeks) and are  difficult to predict within forecast models due complex feedbacks between land, atmosphere, and ocean as well as simulations with lower resolution models which  cannot explicitly solve momentum and mass fluxes. The overall goal of this study is to develop and produce real-time SESR-based drought monitoring and forecasts to  support the National Oceanic and Atmospheric Administration Climate Prediction Center (NOAA CPC) flash drought monitoring and forecasts. We retrieved  evapotranspiration (ET), potential evapotranspiration (PET), and reference ET (refET) from North American Land Data Assimilation System (NLDAS) Noah land surface  model for years 1981-2020. We conducted a thorough investigation of ET, PET, refET, and the Standardized Evaporative Stress Ratio (SESR) and compared results across  two climatological periods, a.) 1981-2020, and b.) 2000-2019. Flash drought accuracy was assessed with the Heidke skill score for different drought intensification  categories. We applied the SESR methodology on Global Ensemble Forecast System version 12 (GEFSv12) five-week ensemble retrospective forecasts over the period  2000-2019 and evaluated deterministic and probabilistic skill across both climatological periods. Forecast accuracy was assessedseasonally and during enhanced modes of subseasonal predictability including the MJO and ENSO. Final deliverables will include the SESR implemented into the CPC operational pipeline to increase  early-warning predictability and economic resilience against rapid onset flash droughts.

The El Niño-Southern Oscillation (ENSO) and the Madden-Julian Oscillation (MJO) are drivers of seasonal and intraseasonal climate variability, but there has been  limited research on their extratropical impacts within U.S. domestic airspace. Prior research found that Convective SIGMETs (CSIGs) followed typical convective patterns  over the United States (Slemmer, 2005). Understanding convection patterns during ENSO and MJO phases can help aviation forecasters provide impact-based decision  support services to the aviation community. To explore these topics, a searchable CSIGs dataset (May 2012 - May 2024) was analyzed based on ENSO and MJO phases.  Time periods for each ENSO phase (El Niño Advisory, La Niña Advisory, and periods with no active watches/advisories) were based on the Climate Prediction Center’s ENSO Diagnostic Discussions and time periods for MJO phases were obtained from the Australian Bureau of Meteorology. As a first step to explore the potential  influences of MJO, this study focused on known extratropical impacts of MJO including extreme rainfall events in western North America (Higgins et al. 2000), the North  American Monsoon (Higgins and Shi, 2001), enhanced subtropical jet streams (Gottschalck, CPC, 2022, personal communication), and tropical cyclones along the U.S.  Gulf Coast (Maloney and Hartman 2000). Extensive analysis revealed distinct monthly climatological patterns of CSIGs during different ENSO phases, but we did not see similarly strong patterns while exploring specific extratropical impacts during the associated MJO phase. Additional research is necessary to validate the initial  results regarding the spatiotemporal distribution of CSIGs related to ENSO phases and to further investigate the distribution of CSIGs during MJO phases.

One of the key tasks, amongst many, of operational hydrologists is to monitor and interpret river gauges which are then used to assess water flows, inform analysis and  flood forecasting. To assist the National Weather Service Hydro Program management, an arcgis online application was created to visualize river gauge data using the  National Water Prediction Systems Application Programming Interface. Development of the application used open source libraries within python, Survey123 Connect,  Web Experience Builder and ArcGIS Online. This hydrological dashboard application aids in visualization of data, a filtration system (that helps assess gauge details such  as location/equipment history and visitation dates), and an embedded survey to update river gauge site visitation history. The hydrological dashboard application is  expected to be expanded to a national scale to service other forecast/hydrological offices after further development. This presentation will provide details on the  dashboard's development, how it is being used and iteratively improved by service hydrologists within the FW and HGX offices, and how it is envisioned to be a key  graphics-based tool for improving services for water decision makers.

Climate literacy is an understanding of the relationship, interaction and influence between humans and climate, including processes, causes, and effects of climate  change. The importance of climate literacy cannot be understated - being climate literate allows individuals to understand and evaluate information about,  communicate about, act in ways that acknowledge the dimensions of, and make informed decisions regarding climate change. The U.S. Global Change Research Program  (USGCRP) brought together multiple Agencies and Departments to create a 3rd edition of The Climate Literacy Guide, which explains Essential Principles to understanding and addressing climate change, and will be shared through a PDF and website publication. This project focused on the visual elements of the Climate  Literacy Guide, including selection of images, graphics, and artworks to be included in the guide, layout of the PDF version of the guide, and production of a video to  promote the guide and methods to increase climate literacy. The resulting visual elements support the complexity, intersectionality, and urgency surrounding climate  change, taking into account the importance and implications of various aspects of representation and visual qualities to create and promote a cohesive and impactful  Guide with visual elements that connect and enhance written topics.

With recent upticks in the rate and severity of hazardous weather, it is now more important than ever to ensure that life-saving weather information is reaching all  members of our communities. This includes those with limited English proficiency, individuals with disabilities, persons with limited internet access, families who are  living in low-income housing, and our unhoused neighbors. Communities like these who have been historically underserved and underrepresented must be at the  forefront of these initiatives. The Indianapolis Weather Forecast Office has approached this call to action by inviting a diverse group of community representatives to a  full day conference to discuss how we can better serve their organizationsand their constituents. Together we will have ongoing conversations about how to get the  word out about weather safety and preparation to those who may be especially at risk. By offering up a seat at the table and walking in with open minds, we are  allowing trusted community advocates the space to express their concerns and propose solutions to the gaps in communication we have seen in the past. We are here  to prove that equitable weather messaging saves lives and that the National Weather Service can be an active member in each community that we serve across the country.

With an emerging focus on conditional peak hazard intensity forecasting at the Storm PredictionCenter(SPC), there is a need for accurate, reliable guidance for severe  hail size classification. Numerous composite indices, including Large Hail Parameter (LHP); Significant Hail Parameter(SHIP); and SARS Hail Size, seek to distinguish severe (>=1” diameter) hail from significant-severe (>=2” diameter) hail using a combination of environmental parameters. These indices are incorporated into SPC  Mesoanalysis, which provides hourly assessments of the environment from the gridded 40-km RAP-based surface objective analysis (sfcOA) scheme and are utilized by  forecasters for mesoscale discussion and watch issuance. However, no work exists that directly compares and verifies these indices. Therefore, this project  comparatively evaluates composite indices and environmental parameters from the SPC’s sfcOA scheme for conditional hail size classification, with sfcOA data gathered  for every initialization hour and nearest grid containing Storm Data hail reports from January 2013 to December 2023. While all indices demonstrate some statistically  significant skill in distinguishing hail intensity, SHIP has the greatest separation between each hail size bin. Still, these differences are minimal, and the predictability of  conditional hail size by existing indices remains limited. Environmental parameters not used in existing indices, such as effective bulk shear and entrainment-adjusted CAPE, demonstrate greater individual skill in size classification than the components of existing indices. Tree-based machine learning models incorporating these  parameters were also developed during this project and evaluated according to the aforementioned methods, demonstrating a substantial improvement in  performance over existing indices.

The National Weather Service (NWS) Office of Central Processing is automating the translation of key NWS text products into multiple languages using Artificial  Intelligence (AI) language models. Linguistic experts refine and train the AI through the use of real time feedback to improve the translation accuracy and performance.  Validating these translations is essential to ensure that accurate weather reports reach communities with limited English proficiency. To assess translation accuracy and  identify areas needing more training, "back translations" are performed using AI, by translating English text into the target language and then back into English. The  original English text is compared to the re-translated version using a correlation coefficient. If the correlation coefficient indicates that an inaccurate translation has  occurred, the project translators can be notified to edit the sentences in the weather products where the AI model needs significant translation training corrections by  human translators. To trace exactly which data needs to be re-evaluated, the data is split into segments, given an ID, and a URL to guide the translator to this exact  segment is generated. The aim of this project is to ensure that people across all communities and cultures are provided with accurate information about potential  weather threats.

As global climate change increasingly impacts ocean ecosystems, understanding the effects on oceanographic conditions and species becomes critical. The Climate, Ecosystems, and Fisheries Initiative (CEFI) is a collaborative effort across NOAA to develop national oceanic models and decision support tools. The purpose of this study was to validate hydrodynamic models in the Gulf of Mexico (GOM) by comparing bottom temperature and salinity data with outputs from the MOM6 NWA12 model, in  support of CEFI initiatives. Observational data were collected as part of the SEAMAP Trawl Surveys throughout the northern region of the GOM from 1979 to 2019.  These data were used to compare to model outputs to identify correlations and discrepancies between datasets. Key findings indicate the model generally aligns well  with observed bottom temperature data (RMSE: 2.12), with no significant bias over time but slight warming bias in cooler months. Salinity validation revealed highly  variable differences (RMSE: 12.21), with season trends and significant variability across different depths. These results highlight the model’s strength in capturing  seasonal patterns bit demonstrate a possible need for further model refinement to improve accuracy of model predictions at depth. Future steps of this project include  comparing additional observational datasets and assessing the performance of other hydrodynamic models to determine which model best supports ecosystem-based  fisheries management.

Lake effect snow (LES) is a mesoscale weather phenomenon that occurs when cold, dry airflows over a warm body of water, becoming warm and saturated. As the air  continues to cross the lake, the resulting clouds can grow into bands that exhibit vigorous convection and intense 
snowfall. One of the challenges associated with monitoring LES is the limited availability of nowcasting tools. LES features are often very shallow (cloud top heights  typically below 3 km in the northern Great Lakes) . Radar beams quickly overshoot the shallow clouds, creating “blind spots” in the radar mosaic that limit the  information available to forecasters. A possible supplement to radar-derived snowfall is quantitative precipitation estimates (QPE) using geostationary satellite data.  This study demonstrates the use of an experimental satellite-derived QPE product for shallow LES events and investigates the potential relationships between snowfall  intensity and other satellite-derived variables to improve satellite snowfall estimates. 
This study utilized Next Generation Weather Radar (NEXRAD), Multi-Radar/Multi-Sensor (MRMS), and Clouds from AVHRR Extended System (CLAVR-x) – a processing  system that generates quantitative cloud products for all NOAA operational satellites – data. Four LES events were studied, with three producing moderate to heavy LES  to most of Michigan’s Upper Peninsula and the other producing intense snowfall to La Porte County, IN as a result of embedded mesovortices. While LES structures for  all of these events were effectively monitored by NEXRAD observations near specific radar sites, significant radar observational blind spots are noted. Snowfall intensity  was also likely underestimated, especially far from radar locations. CLAVR-x cloud water path and cloud effective radius variables, however, effectively identified LES  structures for these events and offer a way to augment NEXRAD observations to improve situational awareness. CLAVR-x variables were also compared to radar-derived  snowfall rates through nearest neighbor pixel matching to find potential relationships that could improve satellite-derived QPE and offer a path forward to further satellite-derived QPE algorithm development.