Dissertations and Theses
Date of Award
2021
Document Type
Thesis
Department
Computer Science
First Advisor
Michael Mandel
Keywords
Machine Learning, Artificial Intelligence, GOES-R, Predictive Calibration, Loop-Heat-Pipe, Data Quality Flag
Abstract
The GOES-R series is a product line of four satellite, with two currently on-orbit (GOES-16 “East” and GOES-17 “West”). GOES-17 is susceptible to a Loop-Heat-Pipe (LHP) phenomenon where during Fall and Spring seasons, there are times of day where some of the infrared bands records inaccurate readings from the Advanced Baseline Imager (ABI). This occurs from joint astronomical behavior and position of the GOES-17. This calibration issue occurs when the LHP instrument fails to radiate the heat of the sun out of ABI. Predictive Calibration (pCal) is an algorithm developed by instrument vendors for the National Oceanic Atmospheric Agency (NOAA) to correct the readings of GOES-17. NOAA implemented the algorithm July 26, 2019. pCal is a regression that corrects for the average temperature in a region of interest where a threshold of points may be susceptible to LHP. pCal has two components: an equation where the rapidly changing calibration parameters are linearly interpolated in time, the second is more frequent calibration events, such as looking at the internal calibration target. There are sixteen channels per satellite, specialized to take measurements of various properties. In this project we explore a multi-layer perceptron (MLP), neural network, to train a model using various sample size. We compare our R-square scores, mean square error (MSE), and mean absolute (MAE) between pCal and MLP models. In addition, we explore different artificial intelligence, machine learning algorithms to detect image anomalies which has broader flagging applications than just correcting for temperatures.
Recommended Citation
Adomako, Ronald, "GOES-R Supervised Machine Learning" (2021). CUNY Academic Works.
https://academicworks.cuny.edu/cc_etds_theses/927