The Use of Neural Networks in Identifying Error Sources in Satellite-Derived Tropical SST Estimates 英文参考文献.docVIP

Sensors 2011, 11, 7530-7544; doi:10.3390/s110807530 OPEN ACCESS sensors ISSN 1424-8220 /journal/sensors Article The Use of Neural Networks in Identifying Error Sources in Satellite-Derived Tropical SST Estimates Yung-Hsiang Lee 1, Chung-Ru Ho 1,*, Feng-Chun Su 2, Nan-Jung Kuo 1 and Yu-Hsin Cheng 1 1 Department of Marine Environmental Informatics, National Taiwan Ocean University, Keelung 20224, Taiwan; E-Mails: .tw (Y.-H.L.); c0021@.tw (N.-J.K.); .tw (Y.-H.C.) 2 Department of Marine Biotechnology and Resources, National Sun Yat-sen University, Kaohsiung 80424, Taiwan; E-Mail: fengchunsu@ * Author to whom correspondence should be addressed; E-Mail: b0211@.tw; Tel.: +886-2-2462-2192; Fax: +886-2-2462-0912. Received: 17 June 2011; in revised form: 20 July 2011 / Accepted: 20 July 2011 / Published: 29 July 2011 Abstract: An neural network model of data mining is used to identify error sources in satellite-derived tropical sea surface temperature (SST) estimates from thermal infrared sensors onboard the Geostationary Operational Environmental Satellite (GOES). By using the Back Propagation Network (BPN) algorithm, it is found that air temperature, relative humidity, and wind speed variation are the major factors causing the errors of GOES SST products in the tropical Pacific. The accuracy of SST estimates is also improved by the model. The root mean square error (RMSE) for the daily SST estimate is reduced from 0.58 K to 0.38 K and mean absolute percentage error (MAPE) is 1.03%. For the hourly mean SST estimate, its RMSE is also reduced from 0.66 K to 0.44 K and the MAPE is 1.3%. Keywords: infrared sensor; data mining; neural network; sea surface temperature; tropical pacific 1. Introduction Sea Surface Temperature (SST) is an important factor in ocean processes with a major impact on weather and climate [1]. Since the 1980s, infrared and microwave sensors onboard satellites have been  Sensors 20