PML Thermal Ocean Fronts Maps

Summary

Provider: Plymouth Marine Laboratory Prospect Place Plymouth PL1 3DH UK

Contact person: Peter Miller, Send email

Contact for production: NEODAAS Helpdesk Send mail

Algorithm development: PML

Frontal analyses provide estimates of the location of thermal fronts in the open ocean, based on SST measurements from the Aqua-MODIS sensor (see related section).The front maps may be used as a visual aid, or metrics can be provided to indicate the gradient strength (in units of K km-1) and persistence (in units of probability) of each line. Ocean fronts are often productive zones which influence the distribution of certain pelagic fish, and hence may be of use in managing and predicting fisheries. Long term analysis of fronts may be of use for fisheries management, e.g. identifying hotspots for assisting the designation of marine protected areas.

Data Portal: PML MultiView website.

Click the arrow below to show additional information about this product.
Click the arrow below to hide the additional information about this product.

Platform & processing

Aqua MODIS SST data, at 1 km resolution.

Algorithm

Fronts are detected on every SST scene and the combined using the 'composite front map' software, proprietary to PML. Cloud masking is applied to the SST scenes prior to front detection.

Parameters include: the size of median filter used to smooth the SST field prior to front detection; the window size for local spatial statistics; the minimum difference in mean SST on either side of a front.

Algorithms developed by PML enable fronts to be located accurately and objectively. The composite front map technique combines the location, gradient, persistence and proximity of all fronts observed over a given period (e.g. 7 days) into a single map (Miller, 2009). Fronts are detected on every individual SST scene, and then combined statistically. Detection is based on local window statistics specific to frontal structures, not simply on horizontal SST gradients. The output is in the form of maps of frontal contours, coloured according to one of several metrics: Fmean - mean gradient across the front; Pfront - persistence of the front; Fcomp - combined gradient and persistence, and is the default metric provided. The algorithm is generic and has been applied to many other European and worldwide areas. The algorithm was published in: Miller, P.I. (2009).

Product uncertainties

Satellites only observe surface fronts, though strong and persistent surface fronts usually indicate a front through the whole surface layer. Therefore when considering MCZ designation, the major fronts may offer some indication of mobile pelagic species, but not benthic fauna. Thermal infrared EO data are limited by cloud cover, though the SST processing and composite front maps techniques optimise the visualisation of fronts by combining all observations derived from sequences of partially cloudy scenes. Several observations of the same feature are preferable in order to assess its persistence. Cloud cover may lead to biases in data analysis, for instance if features such as upwelling or stratification fronts are correlated with clear skies. Several factors prevent front detection within a few kilometres of the coast.

Ocean fronts may be interpreted as a proxy for enhanced productivity, though it is not currently possible to quantify the enhancement or to predict at what times the effect will be greatest. The maps may indicate many frontal lines; the most significant fronts are often seen as a cluster of dark parallel lines representing the spatial advection of a persistent structure. Most mesoscale fronts have some spatial variability, e.g. a few kms over 7 days due to advection and changes in tidal currents; hence, the detected fronts will indicate the range of movement over this period. The fronts should be a reasonable estimate of locations for at least 3 days following the end date of the composite.

Product dissemination

Within 1 hour of acquisition of the source MODIS SST data. Highly operational, robust and automated. See Data Portal above.

Product validation

The composite front map algorithm has so far been used within six peer-reviewed papers in addition to the methodology paper (Miller, 2009). The methodology paper includes validation of the detected front locations against those provided by a trained image analyst, for 10 sample periods for the Iberian upwelling front (Miller, 2009).

Data characteristics

SST maps ideally at 1 km resolution though also possible at 4 km resolution, after cloud masking. Sequence of instantaneous EO scenes required (ideally several per day), not composites which blur fronts and introduce artefacts. MODIS SST are obtained and processed in near-real time from NASA OceanColor FTP site.Notes

Front detection is used exactly as described in Miller (2009). DevCoCast will be trialling modifications to improve front detection near to coast and cloud.

Useful references:

Miller, P.I. (2009) Composite front maps for improved visibility of dynamic sea-surface features on cloudy SeaWiFS and AVHRR data. Journal of Marine Systems, 78(3), 327-336. doi:10.1016/j.jmarsys.2008.11.019.