Multifractal analysis of oceanic chlorophyll maps remotely sensed from space

Montera, Louis de; Jouini, Manel; Verrier, Sébastien; Thiria, Sylvie; Crépon, Michel

doi:10.5194/os-7-219-2011

Cited by 17 publications

(25 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To resolve 85% of turbulent eddy induced 1,500‐km SST subgrid variability, one needs an even smaller scale than 10 km. The SST turbulent eddy scaling exponents H were estimated from the universal multifractal exponents (see supporting information S1: Table S1) and the theory of Schmitt et al () and De Montera et al (, ). They were found to be within the 0.33–0.4 interval, and this places the passive tracer turbulent SST ℓ N around the 1‐km scale.…”

Section: Resultsmentioning

confidence: 99%

“…(3) The scale of ℓ N ≈1 km is clearly indicative of the fact that turbulent eddies drive the small‐scale SST variability. The exact value of ℓ N is however subject to some relatively large uncertainties: First, the model of Schmitt et al () and De Montera et al (, ) is based on several assumptions, such as the closeness between the energy and tracer concentration variance flux multifractal parameters (De Montera et al, ). Second, in the SO eddies move with a velocity of roughly 22 km per week (Frenger et al, ); in a 5‐day model output, they will as a consequence be smeared through a larger volume.…”

Section: Resultsmentioning

confidence: 99%

“…Above the largest eddy scale, there is no eddy‐induced variability, so the scaling slope flattens ( H = 0). We confirmed that the 0.3–0.45 turbulent passive scalar range of H values is relevant for this study by calculating universal multifractal SST parameters for the EO data (see supporting information S1: Table S1) and using the approach of Schmitt et al () and De Montera et al (, ). Using the EO SST data and the universal multifractal theory has advantage over analyzing model outputs, as we avoid some of the limitation imposed on eddies by the NEMO model resolution.…”

Section: Data Theory and Methodsmentioning

confidence: 99%

“…Ocean is a highly stratified fluid with horizontal dimensions much larger than the vertical dimension, so it is highly desirable to use universal multifractal model for anisotropic (i.e., different scaling in the vertical than in the horizontal direction) inhomogeneous turbulence (e.g., Lovejoy & Schertzer, 2010;Schertzer & Lovejoy, 2011). The anisotropic universal multifractal approach typically predicts the passive scalar turbulent H in horizontal to have values between 0.3 and 0.45 (De Montera et al, 2011;Lovejoy & Schertzer, 2010), consistent with SST universal multifractal spectra from Schmitt et al (1996), , , Lovejoy et al (2001), and De Montera et al (2010). This scaling is expected to be valid up to the scale of the largest eddies (∼150 km, Frenger et al, 2015).…”

Section: Model Input Drivers For Sst Dynamicsmentioning

confidence: 99%

See 3 more Smart Citations

SST Dynamics at Different Scales: Evaluating the Oceanographic Model Resolution Skill to Represent SST Processes in the Southern Ocean

et al. 2019

View full text Add to dashboard Cite

In this study we demonstrate the many strengths of scale analysis: we use it to evaluate the Nucleus for European Modelling of the Ocean model skill in representing sea surface temperature (SST) in the Southern Ocean by comparing three model resolutions: 1/12°, 1/4°, and 1°. We show that while 4–5 times resolution scale is sufficient for each model resolution to reproduce the magnitude of satellite Earth Observation (EO) SST spatial variability to within ±10%, the representation of ∼100‐km SST variability patterns is substantially (e.g., ∼50% at 750 km) improved by increasing model resolution from 1° to 1/12°. We also analyzed the dominant scales of the SST model input drivers (short‐wave radiation, air‐sea heat fluxes, wind stress components, wind stress curl, and bathymetry) variability with the purpose of determining the optimal SST model input driver resolution. The SST magnitude of variability is shown to scale with two power law regimes separated by a scaling break at ∼200‐km scale. The analysis of the spatial and temporal scales of dominant SST driver impact helps to interpret this scaling break as a separation between two different dynamical regimes: the (relatively) fast SST dynamics below ∼200 km governed by eddies, fronts, Ekman upwelling, and air‐sea heat exchange, while above ∼200 km the SST variability is dominated by long‐term (seasonal and supraseasonal) modes and the SST geography.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Data Theory and Methodsmentioning

confidence: 99%

Section: Model Input Drivers For Sst Dynamicsmentioning

confidence: 99%

See 2 more Smart Citations

SST Dynamics at Different Scales: Evaluating the Oceanographic Model Resolution Skill to Represent SST Processes in the Southern Ocean

et al. 2019

View full text Add to dashboard Cite

show abstract

“…From the known models the most successful are the stable attractive cascades termed universal multifractals [ Schertzer and Lovejoy , ]. The statistical model of universal multifractals has been widely used to model the Earth atmosphere [ Lovejoy et al ., ; Lovejoy and Schertzer , ; Schertzer and Lovejoy , ; Schmitt et al ., ], the Earth topography [ Gagnon et al ., ; Lavalee et al ., ], climate [ Lovejoy and Schertzer , ; Lovejoy , ], and also the oceanic fields [ Lovejoy et al ., ; de Montera et al ., ; Seuront et al ., ; Seuront and Lagadeuc , ; Seuront et al ., ; Seuront and Schmitt , ; Skakala and Smyth , ]. Universal multifractals lead to models exhaustively described by the values of three parameters.…”

Section: Introductionmentioning

confidence: 99%

Using multifractals to evaluate oceanographic model skill

et al. 2016

View full text Add to dashboard Cite

We are in an era of unprecedented data volumes generated from observations and model simulations. This is particularly true from satellite Earth Observations (EO) and global scale oceanographic models. This presents us with an opportunity to evaluate large‐scale oceanographic model outputs using EO data. Previous work on model skill evaluation has led to a plethora of metrics. The paper defines two new model skill evaluation metrics. The metrics are based on the theory of universal multifractals and their purpose is to measure the structural similarity between the model predictions and the EO data. The two metrics have the following advantages over the standard techniques: (a) they are scale‐free and (b) they carry important part of information about how model represents different oceanographic drivers. Those two metrics are then used in the paper to evaluate the performance of the FVCOM model in the shelf seas around the south‐west coast of the UK.

show abstract

Scaling and stochastic cascade properties of NEMO oceanic simulations and their potential value for GCM evaluation and downscaling

2014

View full text Add to dashboard Cite

Spectral scaling properties have already been evidenced on oceanic numerical simulations and have been subject to several interpretations. They can be used to evaluate classical turbulence theories that predict scaling with specific exponents and to evaluate the quality of GCM outputs from a statistical and multiscale point of view. However, a more complete framework based on multifractal cascades is able to generalize the classical but restrictive second-order spectral framework to other moment orders, providing an accurate description of probability distributions of the fields at multiple scales. The predictions of this formalism still needed systematic verification in oceanic GCM while they have been confirmed recently for their atmospheric counterparts by several papers. The present paper is devoted to a systematic analysis of several oceanic fields produced by the NEMO oceanic GCM. Attention is focused to regional, idealized configurations that permit to evaluate the NEMO engine core from a scaling point of view regardless of limitations involved by land masks. Based on classical multifractal analysis tools, multifractal properties were evidenced for several oceanic state variables (sea surface temperature and salinity, velocity components, etc.). While first-order structure functions estimated a different nonconservativity parameter H in two scaling ranges, the multiorder statistics of turbulent fluxes were scaling over almost the whole available scaling range. This multifractal scaling was then parameterized with the help of the universal multifractal framework, providing parameters that are coherent with existing empirical literature. Finally, we argue that the knowledge of these properties may be useful for oceanographers. The framework seems very well suited for the statistical evaluation of OGCM outputs. Moreover, it also provides practical solutions to simulate subpixel variability stochastically for GCM downscaling purposes. As an independent perspective, the existence of multifractal properties in oceanic flows seems also interesting for investigating scale dependencies in remote sensing inversion algorithms.

show abstract

Multifractal analysis of oceanic chlorophyll maps remotely sensed from space

Cited by 17 publications

References 31 publications

SST Dynamics at Different Scales: Evaluating the Oceanographic Model Resolution Skill to Represent SST Processes in the Southern Ocean

SST Dynamics at Different Scales: Evaluating the Oceanographic Model Resolution Skill to Represent SST Processes in the Southern Ocean

Using multifractals to evaluate oceanographic model skill

Scaling and stochastic cascade properties of NEMO oceanic simulations and their potential value for GCM evaluation and downscaling

Contact Info

Product

Resources

About