Improved neural network scatterometer forward models

Dan Cornford, Ian T. Nabney, Guillaume Ramage

Research output: Contribution to journalArticle

Abstract

Current methods for retrieving near-surface winds from scatterometer observations over the ocean surface require a forward sensor model which maps the wind vector to the measured backscatter. This paper develops a hybrid neural network forward model, which retains the physical understanding embodied in CMOD4, but incorporates greater flexibility, allowing a better fit to the observations. By introducing a separate model for the midbeam and using a common model for the fore and aft beams, we show a significant improvement in local wind vector retrieval. The hybrid model also fits the scatterometer observations more closely. The model is trained in a Bayesian framework, accounting for the noise on the wind vector inputs. We show that adding more high wind speed observations in the training set improves wind vector retrieval at high wind speeds without compromising performance at medium or low wind speeds. Copyright 2001 by the American Geophysical Union.

Original languageEnglish
Pages (from-to)22331-22338
Number of pages8
JournalJournal of Geophysical Research
Volume106
Issue numberC10
DOIs
Publication statusPublished - 15 Oct 2001

Fingerprint

scatterometers
scatterometer
neural networks
Neural networks
wind speed
wind velocity
retrieval
surface wind
backscatter
ocean surface
sea surface
oceans
sensor
flexibility
education
sensors
Sensors

Bibliographical note

Not subject to U.S. copyright. An edited version of this paper was published by AGU. Copyright 2001 American Geophysical Union. Cornford, Dan; Nabney, Ian T. and Ramage, Guillaume, (2001), Improved neural network scatterometer forward models, Journal of Geophysical Research - Oceans, 106, C10, 10.1029/2000JC000417. To view the published open abstract, go to http://dx.doi.org/10.1029/2000JC000417

Keywords

  • near‐surface winds
  • scatterometer observations
  • ocean surface
  • forward sensor model
  • mapping
  • wind vector
  • measured backscatter
  • hybrid neural network forward model
  • local wind vector retrieval
  • Bayesian framework
  • wind vector inputs

Cite this

Cornford, Dan ; Nabney, Ian T. ; Ramage, Guillaume. / Improved neural network scatterometer forward models. In: Journal of Geophysical Research. 2001 ; Vol. 106, No. C10. pp. 22331-22338.
@article{af0281e253e34de88444d7985e2969b0,
title = "Improved neural network scatterometer forward models",
abstract = "Current methods for retrieving near-surface winds from scatterometer observations over the ocean surface require a forward sensor model which maps the wind vector to the measured backscatter. This paper develops a hybrid neural network forward model, which retains the physical understanding embodied in CMOD4, but incorporates greater flexibility, allowing a better fit to the observations. By introducing a separate model for the midbeam and using a common model for the fore and aft beams, we show a significant improvement in local wind vector retrieval. The hybrid model also fits the scatterometer observations more closely. The model is trained in a Bayesian framework, accounting for the noise on the wind vector inputs. We show that adding more high wind speed observations in the training set improves wind vector retrieval at high wind speeds without compromising performance at medium or low wind speeds. Copyright 2001 by the American Geophysical Union.",
keywords = "near‐surface winds, scatterometer observations, ocean surface, forward sensor model, mapping, wind vector, measured backscatter, hybrid neural network forward model, local wind vector retrieval, Bayesian framework, wind vector inputs",
author = "Dan Cornford and Nabney, {Ian T.} and Guillaume Ramage",
note = "Not subject to U.S. copyright. An edited version of this paper was published by AGU. Copyright 2001 American Geophysical Union. Cornford, Dan; Nabney, Ian T. and Ramage, Guillaume, (2001), Improved neural network scatterometer forward models, Journal of Geophysical Research - Oceans, 106, C10, 10.1029/2000JC000417. To view the published open abstract, go to http://dx.doi.org/10.1029/2000JC000417",
year = "2001",
month = "10",
day = "15",
doi = "10.1029/2000JC000417",
language = "English",
volume = "106",
pages = "22331--22338",
journal = "Journal of Geophysical Research",
issn = "0148-0227",
publisher = "Wiley-Blackwell",
number = "C10",

}

Cornford, D, Nabney, IT & Ramage, G 2001, 'Improved neural network scatterometer forward models', Journal of Geophysical Research, vol. 106, no. C10, pp. 22331-22338. https://doi.org/10.1029/2000JC000417

Improved neural network scatterometer forward models. / Cornford, Dan; Nabney, Ian T.; Ramage, Guillaume.

In: Journal of Geophysical Research, Vol. 106, No. C10, 15.10.2001, p. 22331-22338.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Improved neural network scatterometer forward models

AU - Cornford, Dan

AU - Nabney, Ian T.

AU - Ramage, Guillaume

N1 - Not subject to U.S. copyright. An edited version of this paper was published by AGU. Copyright 2001 American Geophysical Union. Cornford, Dan; Nabney, Ian T. and Ramage, Guillaume, (2001), Improved neural network scatterometer forward models, Journal of Geophysical Research - Oceans, 106, C10, 10.1029/2000JC000417. To view the published open abstract, go to http://dx.doi.org/10.1029/2000JC000417

PY - 2001/10/15

Y1 - 2001/10/15

N2 - Current methods for retrieving near-surface winds from scatterometer observations over the ocean surface require a forward sensor model which maps the wind vector to the measured backscatter. This paper develops a hybrid neural network forward model, which retains the physical understanding embodied in CMOD4, but incorporates greater flexibility, allowing a better fit to the observations. By introducing a separate model for the midbeam and using a common model for the fore and aft beams, we show a significant improvement in local wind vector retrieval. The hybrid model also fits the scatterometer observations more closely. The model is trained in a Bayesian framework, accounting for the noise on the wind vector inputs. We show that adding more high wind speed observations in the training set improves wind vector retrieval at high wind speeds without compromising performance at medium or low wind speeds. Copyright 2001 by the American Geophysical Union.

AB - Current methods for retrieving near-surface winds from scatterometer observations over the ocean surface require a forward sensor model which maps the wind vector to the measured backscatter. This paper develops a hybrid neural network forward model, which retains the physical understanding embodied in CMOD4, but incorporates greater flexibility, allowing a better fit to the observations. By introducing a separate model for the midbeam and using a common model for the fore and aft beams, we show a significant improvement in local wind vector retrieval. The hybrid model also fits the scatterometer observations more closely. The model is trained in a Bayesian framework, accounting for the noise on the wind vector inputs. We show that adding more high wind speed observations in the training set improves wind vector retrieval at high wind speeds without compromising performance at medium or low wind speeds. Copyright 2001 by the American Geophysical Union.

KW - near‐surface winds

KW - scatterometer observations

KW - ocean surface

KW - forward sensor model

KW - mapping

KW - wind vector

KW - measured backscatter

KW - hybrid neural network forward model

KW - local wind vector retrieval

KW - Bayesian framework

KW - wind vector inputs

UR - http://www.scopus.com/inward/record.url?scp=0035166557&partnerID=8YFLogxK

UR - https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2000JC000417

U2 - 10.1029/2000JC000417

DO - 10.1029/2000JC000417

M3 - Article

VL - 106

SP - 22331

EP - 22338

JO - Journal of Geophysical Research

JF - Journal of Geophysical Research

SN - 0148-0227

IS - C10

ER -