This page lists past seminars and presentations by STAR
scientists and visiting scientists. These seminars include the STAR
Science Forum and similar events. Presentation materials for
seminars will be provided when available.
Title: (Talk 1): Dynamics in phytoplankton size fractionated primary production in the East Bering Sea(Talk 2): Assessing Water Clarity in the Chesapeake Bay using Satellite-derived Secchi Disk DepthCoastWatch Seminar Series
Abstract: (Abstract 1): Phytoplankton net primary production (NPP) forms the base of the global marine food web, where the size structure of the phytoplankton community strongly influences trophic transfer efficiency through the food web. The East Bering Sea, a critical fishery supporting~50% of the total U.S. harvest volume, is undergoing rapid change including warming temperatures and loss of sea ice and it remains unclear how these changes impact the total and size fractionated NPP and subsequent trophic levels. Here, we present initial results from a Phytoplankton Size Class Absorption based Production Model (PSC-AbPM) which derives NPP contributions from pico-, nano-, and microphytoplankton size classes. The PSC-AbPM was applied to 20 years of ocean color data from OC-CCI over the East Bering Sea, and compared to other widely used NPP models and in situ observations. The PSC-AbPM is able to capture the seasonal succession of phytoplankton size classes, with a massive spring bloom dominated by microphytoplankton, followed by summertime low production dominated by the smaller picophytoplankton class. Additionally ,the model captures the tight seasonal relationship between the spring bloom timing and ice extent over the winter. Lastly, we show the dynamical approach used to derive each size class photosynthetic efficiency parameter throughout the year. (Abstract 2): Secchi disk depth (SDD) is a conventional in situ optical method of measuring water clarity, which is an important characteristic of water quality and ecosystem health. Assessments of SDD from remotely-sensed ocean color data benefit from wide spatial and temporal coverage but are subject to low accuracy resulting from unstable relationship between water transparency and the color of water surface. A semi-analytical SDD algorithm (here after called CB-SDD algorithm) proposed for turbid coastal waters was extensively evaluated in the Chesapeake Bay. Satellite SDD maps from MODIS-Aqua and VIIRS-SNPP generated using CB-SDD algorithm showed lower bias and higher accuracy as compared to other methods and found that SDD are shallower in the northern bay and upstream in its tributaries, while SDD are deeper along the main stem int he middle and lower parts of the bay. Expected seasonal and interannual variation in SDD in the Bay are evident in the satellite results in relation to seasonal and interannual variations river discharge patterns. Our results show that satellite data can be fit-for-purpose for water quality management across the Chesapeake Bay and this satellite method can be considered to extend water quality observations where and when in situ observations are lacking in the Bay.
Bio(s): Jonathan Sherman received his PhD in Oceanography from Rutgers University in 2021 where he focused on control mechanisms on photosynthetic energy conversion efficiency in the global ocean. He completed his postdoc at City University of New York working on remote sensing of water quality and carbon cycling in the coastal ocean. He joined NOAA NESDIS/STAR CoastWatch Applications and Research Team in March 2023 through Global Science & Technology.Seunghyun Son received both his BS and MS from Pusan National University, Department of Marine Sciences, in Pusan, South Korea. He received his PhD from University of New Hampshire in 2004 and was a postdoctoral researcher at University of Maine. He joined NOAA/NESDIS/STAR& Univ. Maryland/ESSIC/CISESS, College Park, MD as a Senior Research Associate in 2007.
Slides, Recordings, Other Materials: NOAA CoastWatch Seminars are recurring monthly contributions to the STAR and NOAA Science Seminar Series and are not recorded, but slide decks are made available here after the presentation.
Abstract: The Advanced SCATterometer (ASCAT) is a vertically polarized C-band ocean wind radar sensor carried on the Metop series of three polar-orbiting satellites launched between October 2006 and November 2018. The NOAA produces two global ocean wind products with the resolution of 12.5 km and 25.0 km up to 15 km of the coast for its operational users. While the ASCATs provide invaluable data in the open ocean, due to land contamination of the signal, most inner coastal zones are left void of the data. Most coastal marine activity occurs within a few kilometers of the coast, coastal observations are also needed for ocean forcing for upwelling affected areas. In order to retrieve winds closer to the coast, a coastal wind retrieval algorithm that utilizes enhancement resolution technique and the land contamination removal was developed and applied to the ASCAT measurements. This allowed us to retrieve winds within 20 km inner coastal zone. The enhanced resolution can be achieved by utilizing overlapping measurements of the ASCAT antenna gain. For each near coastal measurement amount of the land signal contamination is determined by computing land contamination ratio (LCR). The normalized radar cross section (NRCS) measurements over near by land mass are used to calculate a mean and a standard deviation of the land brightness for each coastal observation. By using the LCR and the mean and the standard deviation of the land brightness we have developed the land contamination correction for each coastal NRCS slice is determined within a few iterations. However in the vicinity of strong land brightness, the proposed NRCS corrections alone cannot completely remove land contamination. A post wind retrieval processing is developed and applied before final coastal wind product is produced. This post wind retrieval processing involves processing of the corrected NRCS using varying LCR threshold. First Pilot coastal wind and ice ASCAT product in US coastal regions is being produced in NRT for operational validation. New product will be presented and discussed.
Bio(s): Seubson Golf Soisuvarn joined the NOAA/NESDIS OceanSurface Winds Science Team in 2006 as a UCAR visiting scientist and iscurrently a UCAR Project Scientist. His research focuses on the development ofactive and passive microwave remote sensing techniques for the ocean surface,with an emphasis on retrieving ocean surface wind fields. His work includes improvingwind retrieval algorithms and developing new products. Seubson has a backgroundin electrical engineering, earning a B.Eng. from Kasetsart University inBangkok, Thailand, in 1998. He later earned an M.S.E.E. and a Ph.D. from theUniversity of Central Florida in 2001 and 2006, respectively.
Slides, Recordings, Other Materials: NOAA CoastWatch Seminars are recurring monthly contributions to the STAR and NOAA Science Seminar Series and are not recorded, but slide decks are made available here after the presentation.
Abstract: Satellite-derived chlorophyll-a (Chl-a) data are crucial for monitoring and understanding aquatic ecosystems. However, existing satellite Chl-a products are typically sensor-specific, requiring separate development and calibration for each sensor, which introduces inconsistencies and complicates multi-sensor data merging. This study proposes a novel machine-learning approach that unifies spectral information from diverse sensors by leveraging a transformer-based model. This method compels the model to learn latent representations of both the spectral response functions and top-of-atmosphere (TOA) reflectance, respectively, allowing a single sensor-agnostic model to ingest spectral inputs from multiple sensors.By training on a combined dataset comprising field-measured Chl-a data matched up with coincident MODIS, MERIS, VIIRS, and OLCI observations, the model is able to learn and generalize effectively across diverse band configurations. Results show that the sensor-agnostic model performs comparably to, and often surpasses, four sensor-specific machine-learning models trained separately for each individual sensor, while reducing inter-sensor biases and offering a more unified product.Beyond immediate improvements in accuracy, this transformer architecture demonstrates the potential to build a foundation model in ocean color remote sensing using similar frameworks, enabling more efficient fine-tuning for new sensors or new regions without requiring extensive retraining.
Bio(s): Guangming Zheng is an Associate Research Scientist at the CISESS/ESSIC at the University of Maryland, College Park. He received his Ph.D. from Scripps Institution of Oceanography in 2013. Dr. Zheng's research interest focuses on training and applying artificial-intelligence models to monitor and forecast coastal and inland water quality using satellite remote-sensing data.
Slides, Recordings, Other Materials: Supply links here for slides, recordings, and any other supporting materials. You may also note here that materials will be linked here after the seminar.
Recordings: Supply information about availability of seminar recordings, if they are going to be posted.
