A Neural Network Multiparameter Algorithm for SSM/I Ocean Retrievals: Comparisons and Validations☆
Introduction
We evaluate an empirical neural network (NN) algorithm that simultaneously retrieves several geophysical parameters over the ocean from brightness temperature (BT) measurements acquired by the Special Sensor Microwave Imager (SSM/I) (Hollinger et al., 1990) with primary emphasis on retrieving wind speed. This recently developed NN-based algorithm (OMBNN3), which retrieves wind speed, columnar water vapor, and columnar liquid water (see http://polar.wwb.noaa.gov/winds/), is described elsewhere Krasnopolsky et al. 1996, Krasnopolsky et al. 1997, Krasnopolsky et al. 1999. Here we present a comparison of this algorithm with other global SSM/I wind speed, W (m/s), columnar water vapor, V (mm), and cloud liquid water, L (mm), retrieval algorithms. We also examine different BT retrieval flags and their relation to different forms of atmospheric moisture, an important topic that is usually ignored. We also present various wind speed retrieval error analyses.
We describe global SSM/I retrieval algorithms that were considered in this study. Our new multiparameter empirical approach and the NN-based SSM/I retrieval algorithm, OMBNN3 (the corresponding FORTRAN code is available upon request at: [email protected]), are briefly described. We compare OMBNN3 with existing algorithms and perform error analyses using different retrieval flags. Finally, we discuss some operational applications of the geophysical fields that are retrieved by the new algorithm and present a recent example. OMBNN3 has been running as the operational SSM/I retrieval algorithm in the data assimilation system at the National Centers for Environmental Prediction since April 1998.
Section snippets
Algorithms selected for comparison
The SSM/I generates brightness temperatures in seven channels at four frequencies (19 GHz, 22 GHz, 37 GHz, and 85 GHz), each with vertical and horizontal polarization (the 22-GHz channel provides only vertical polarization). The spatial resolution is about 50 km at 19 GHz and 22 GHz, about 30 km at 37 GHz, and 15 km at 85 GHz. The SSM/I is a passive instrument that infers brightness temperatures (BTs) from the ocean surface, receiving microwave radiation emitted both by the ocean surface and
A new multiparameter retrieval algorithm
The NN algorithm presented here is an empirical multiparameter NN algorithm called OMBNN3 Krasnopolsky et al. 1996, Krasnopolsky et al. 1997. The OMBNN3 algorithm uses five SSM/I channels: 19 GHz and 37 GHz (horizontal and vertical polarizations) and 22 GHz (vertical polarization). It contains three new elements. First, it is a multiparameter retrieval algorithm. The covariability of related atmospheric and surface parameters that can be extracted from the same set of brightness temperatures is
Wind Speed
SSM/I wind speed retrieval algorithms are usually used together with so-called retrieval flags that are based on various BT criteria. These retrieval flags serve as delimiters for BT over the ranges of applicability for a given algorithm. They are also used to indicate reliability and accuracy of retrieved wind speeds. These flags are usually derived a posteriori based on analyses of retrieval error statistics. Because retrievals are performed with a particular algorithm, each set of flags is
Conclusions
A new empirical, multiparameter SSM/I retrieval algorithm based on the neural network approach (OMBNN3) is compared with other global SSM/I retrieval algorithms. This algorithm simultaneously retrieves wind speed, columnar water vapor, columnar liquid water, and SST using only SSM/I brightness temperatures. The accuracy of the wind speed retrievals from the new OMBNN3 algorithm (algorithm RMS error 1.0 m/s under clear conditions and 1.3 m/s under clear plus cloudy conditions) is consistently
Acknowledgements
We thank D. B. Rao for his thorough and critical review of this paper. Also, we want to thank those who provided us with expanded collocated SSM/I-buoy data sets; Gene Poe of the Naval Research Laboratory for providing a preliminary raw version of the new NRL matchup database; David Kilham of Bristol University for providing us with additional matchup data for high latitudes; and Michael McPhaden and Linda Magnum for providing information concerning TOGA-TAO buoys. Without this comprehensive
References (16)
- et al.
Determination of oceanic total precipitable water from the SSM/I
IEEE Trans. Geosci. Remote Sens.
(1990) - et al.
Approximation capability to functions of several variables, nonlinear functionals and operators by radial basis function neural networks
Neural Networks
(1995) - et al.
The use of SSM/I data in operational marine analysis
Weather and Forecasting
(1999) - et al.
Remote sensing of ocean surface winds with the special sensor microwave imager
JGR
(1989) - et al.
SSM/I instrument evaluation
IEEE Trans. Geosci. Remote Sens.
(1990) - Krasnopolsky, V. (1997), Neural networks for standard and variational satellite retrievals. Technical Note, OMB...
- et al.
A neural network as a nonlinear transfer function model for retrieving surface wind speeds from the Special Sensor Microwave/Imager
J. Geophys Res.
(1995) - Krasnopolsky, V., Gemmill, W. H., and Breaker, L. C. (1996), A new transfer function for ssm/I based on an expanded...
Cited by (43)
Determination of inactive water quality variables by MODIS data: A case study in the Kızılırmak Delta-Balik Lake, Turkey
2021, Estuarine, Coastal and Shelf ScienceCitation Excerpt :In the case of very few neurons, the self-learning ability and network accuracy would be reduced. At the same time, if too many neurons are chosen, then the training time would increase while the generalization ability of the network would be reduced (Krasnopolsky et al., 2000, 2005; Liu et al., 2015). In this study, neuron numbers from 1 to 20 are considered experimentally, and finally 10 neuron use in the hidden layer is determined to be the optimum for the analysis.
Sea surface temperature retrieval based on simulated space-borne one-dimensional multifrequency synthetic aperture microwave radiometry
2022, Frontiers in Environmental ScienceRetrieval of sea surface wind speed by one-dimensional synthetic aperture microwave radiometer
2021, Journal of Tropical Meteorology
- ☆
This paper is OMB Contribution No. 159.