Assimilation of image sequences in numerical models

Abstract

Understanding and forecasting the evolution of geophysical fluids is a major scientific and societal challenge. Forecasting algorithms should take into account all the available information on the considered dynamic system. The variational data assimilation (VDA) technique combines all these informations in an optimality system (O.S.) in a consistent way to reconstruct the model inputs. VDA is currently used by the major meteorological centres. During the last two decades about 30 satellites were launched to improve the knowledge of the atmosphere and of the oceans. They continuously provide a huge amount of data that are still underused by numerical forecast systems. In particular, the dynamic evolution of certain meteorological or oceanic features (such as eddies, fronts, etc.) that the human vision may easily detect is not optimally taken into account in realistic applications of VDA. Image Assimilation in VDA framework can be performed using ‘pseudo‐observation’ techniques: they provide apparent velocity fields, which are assimilated as classical observations. These measurements are obtained by certain external procedures, which are decoupled with the considered dynamic system. In this paper, we suggest a more consistent approach, which directly incorporates image sequences into the O.S.

On the use of Finite-Time Lyapunov Exponents and Vectors for direct assimilation of tracer images into ocean models

“…In this paper we exploit two properties of backward FTLE-V to construct adapted observation operators for direct ocean tracer image assimilation (Titaud et al, 2010).…”

“…4. In this study, we adopt the direct image assimilation approach described in Titaud et al (2010): first, we will consider FTLE field as it were a tracer field which in turn can be considered as an image. This makes sense because of the almost-Lagrangian nature of FTLE (Shadden et al, 2005):…”

“…We emphasize that we are not interested in the assimilation of such measurements as in the classical cost function (1), but in the assimilation of the structured data that can be extracted from the image. With this in mind Titaud et al (2010) suggest adding an image term to the classical cost function. This takes the general form…”

“…This approach takes into account little physical information about the underlying processes that are observed in the image. The second approach, referred to as Direct Image Assimilation (DIA) in this paper, directly assimilates the image data into the dynamic model: in Corpetti et al (2009) the image pixel is assimilated as a tracer concentration measurement; in Titaud et al (2010) the structured information extracted from the image are assimilated using an adapted observation operator. The latter does not use the pixel basis of the image but higher levels of interpretation (e.g.…”

“…We suggest using sophisticated observation operators based on the computation of FTLE and FTLV. Their construction does not require the numerical advection of a synthetic passive tracer, as was the case in Titaud et al (2010).…”