A Multivariate Regression Approach to Adjust AATSR Sea Surface Temperature to In Situ Measurements

Abstract

Abstract-The Advanced Along-Track Scanning Radiometer (AATSR) onboard Envisat is designed to provide very accurate measurements of sea surface temperature (SST). Using colocated in situ drifting buoys, a dynamical matchup database (MDB) is used to assess the AATSR-derived SST products more precisely. SST biases are then computed. Currently, Medspiration AATSR SST biases are discrete values and can introduce artificial discontinuities in AATSR level-2 SST fields. The new AATSR SST biases presented in this letter are continuous. They are computed, for nighttime and best proximity confidence data, by linear regression with different MDB covariables (wind speed, latitude, aerosol optical depth, etc.). As found, the difference between dual-view and nadir-only SST products explains most of the variability (26%).

Linear Gaussian state-space model with irregular sampling: application to sea surface temperature

“…We also plan to investigate more elaborated measurement equations and include covariates to model the changing biases and variances of the different satellites (see e.g. [25]). Then, the Markovian structure of the model leads to efficient methods for the statistical inference.…”

“…The observation equation (1) could be modified to take into account these fluctuations in the accuracy of the data. In the same way, we could include the various covariates which alter the quality of the satellite measurements (see [25]) or assume that the parameters H and R depend on the satellite if the observed time series was obtained by mixing data from different satellites. Then we assume that the latent process {X t } is a simple Ornstein-Uhlenbeck process, that is a stationary solution of the following stochastic differential equation:…”

Land and Sea Surface Temperature Estimation by Support Vector Regression

“…In the case of SST, compositing methods have been proposed that combine SST estimates obtained from multiple swaths in order to increase the spatial coverage and to reduce the impact of possible cloud cover (Haines et al 2007;Li et al 2001) and multivariate regression have been used in Tandeo et al (2009) to compensate for the effect of the solar zenith angle (and of other possible variables, such as wind speed or aerosol optical depth) on the SST estimation accuracy.…”

Analysis and Modeling Methodologies for Heat Exchanges of Deep-Sea In Situ Spectroscopy Detection System Based on ROV

“…In this process, the measurement system on ROV needs to ensure not only perennially cold (274–276 K) in seafloor [ 20 , 21 ], but also temperature gradient in seawater profile. For the in situ measurement system, it has to ensure temperature changes from 303 K to 275 K during the deep diving process (with depth of 2000 m) [ 22 ], which may cause frequency shift and calculation error [ 23 , 24 ] of spectrum data. So, it will be useful to build a straightforward model to predict the system cabin condition development trend during profile detection [ 25 ].…”