Applying Deep Learning to Clear-Sky Radiance Simulation for VIIRS with Community Radiative Transfer Model—Part 1: Develop AI-Based Clear-Sky Mask

نویسندگان

چکیده

A fully connected deep neural network (FCDN) clear-sky mask (CSM) algorithm (FCDN_CSM) was developed to assist the FCDN-based Community Radiative Transfer Model (FCDN_CRTM) reproduce Visible Infrared Imaging Radiometer Suite (VIIRS) radiances in five thermal emission M (TEB/M) bands. The model design referenced and enhanced from its earlier version (version 1), trained tested global ocean domain using six dispersion days’ data 2019 2020 as inputs a modified NOAA Advanced Clear-Sky Processor over Ocean (ACSPO) CSM product reference labels. improved FCDN_CSM 2) further by including daytime data, which not collected 1. then employed predict VIIRS multiple days an accuracy stability check. results were validated against biases between sensor observations CRTM calculations (O-M). objectives (1) enhance performance include analysis, improve stability, accuracy, efficiency; (2) understand based on combination of statistics physical interpretation. According analyses F-score, prediction result showed ~96% ~97% for day night, respectively. type Cloud most accurate, followed Clear-Sky. O-M mean are comparable ACSPO all bands, both night. standard deviations (STD) slightly degraded long wave IRs (M14, M15, M16), mainly due contamination 3% misclassification Cloud, may require be fine-tuned future. However, consistent means STDs persist throughout period, suggesting that 2 is robust does have significant overfitting. Given high F-scores, spatial long-term efficiency, acceptable STDs, deemed ready use FCDN_CRTM.

برای دانلود باید عضویت طلایی داشته باشید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Comparison of clear-sky surface radiative fluxes simulated with radiative transfer models

The surface fluxes of several important radiatively active gases, including H2O, CO2, CH4, N2O, O3, and the chlorofluorocarbons CFC11 and CFC12, were simulated with the radiation band models from the National Center for Atmospheric Research (NCAR) community climate model 3 (CCM3), the single-column community atmospheric model (SCAM), and the Canadian global climate model 3 (GCM3). These results...

متن کامل

Clear sky direct radiative effects of aerosols over Southeast Asia based on satellite observations and radiative transfer calculations

a r t i c l e i n f o Keywords: Aerosol Radiative Effects Southeast Asia MODIS CERES Radiative transfer model Using the Moderate Resolution Imaging Spectroradiometer (MODIS), Clouds and the Earth's Radiant Energy System (CERES) instrument, and a radiative transfer model (RTM), we provide a quantitative assessment of regional cloud-free diurnally averaged shortwave Aerosol Radiative Effects (ARE...

متن کامل

Preliminary Inter-Comparison between AHI, VIIRS and MODIS Clear-Sky Ocean Radiances for Accurate SST Retrievals

Clear-sky brightness temperatures (BT) in five bands of the Advanced Himawari Imager (AHI; flown onboard Himawari-8 satellite) centered at 3.9, 8.6, 10.4, 11.2, and 12.3 μm (denoted by IR37, IR86, IR10, IR11, and IR12, respectively) are used in the NOAA Advanced Clear-Sky Processor for Oceans (ACSPO) sea surface temperature (SST) retrieval system. Here, AHI BTs are preliminarily evaluated for s...

متن کامل

Digital imaging of clear-sky polarization.

If digital images of clear daytime or twilight skies are acquired through a linear polarizing filter, they can be combined to produce high-resolution maps of skylight polarization. Here polarization P and normalized Stokes parameter Q are measured near sunset at one inland and two coastal sites. Maps that include the principal plane consistently show that the familiar Arago and Babinet neutral ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

ژورنال

عنوان ژورنال: Remote Sensing

سال: 2021

ISSN: ['2315-4632', '2315-4675']

DOI: https://doi.org/10.3390/rs13020222