Sea Surface Temperature maps - what's shown

We have three categories of maps that you choose between at left of screen, covering three general scales: 'local', 'state' and the whole Australian region. They all show various combinations of satellite-derived and in-situ observations, in ways appropriate to the scale of the map. The multi-year archive is updated daily with data that is available in near-real-time (within the last few weeks) and re-generated or extended back in time as it becomes possible to include other data from the IMOS in-situ arrays. All maps include keys that tell you details of what data are shown. Here we explain how to interpret those keys.

Local scale maps

These maps of SST and other data are designed for being viewed as animations. For every overpass of a NOAA satellite, we generate a new image frame (from a single IMOS L3U data file) which is the previous one updated only where there are data of good quality in clear-sky regions. If those clear-sky regions are large and last for several satellite overpasses, animation of the imagery shows the complex patterns of movement clearly. Where the clear-sky regions are small and transient, however, this process leads to a messy patchwork of little value. View the animation several times and you will see the varying quality and usefullness of the images as the weather swings between clear and cloudy. Animations in fli format from the date selector page. Day-time satellite passes, especially in warmer regions, are often affected by a 'warm skin' effect that occurs if there is very little wind, allowing the uppermost few centimeters of the ocean to become several degrees warmer than immediately below. A warm-skin-affected image shows an accurate measurement of sea surface temperature, but it is not very useful for visualizing the ocean currents. The colourbar of the image, apart from showing the temperature scale, lists the NOAA satellite number and the fraction of the scene with new data of acceptable quality. The leading cause of data gaps is cloud cover, and poor data quality is mostly due to contamination by thin or patchy clouds. Bear this in mind as you step through images, which is most quickly done using the animations.

Overlain on the temperature imagery are several other data, including the sea level (contours at 0.1m; see description below), the geostrophic current velocity field derived from gradients of sea level, and the more localised data from HF radars, Argo floats, gliders, current meters and surface drifters, as detailed below. [FFI].

Velocity estimates from the various sources are shown in one of three ways. Straight arrows are used for current meters and gliders. Curved arrows are used for HF radar and altimetry. The scaling is the same for both arrow types, the difference is that the curved arrows use the 2-dimensional information of the gridded velocity types, by integrating Lagrangian trajectories. For both arrow types, the scaling has dimensions of time, which can be interpreted as how far a particle will drift. For zoomed-in maps of high-current regions we use short (e.g. 3 or 6h) times to avoid over-crowding while for weak current regions we use longer times (e.g. 12 or 24h) as shown in the velocity key.

The intervals between map times is determined by the SST imagery and is not uniform, or synchronised with the other data. The sea level information is a few days older than the latest temperature image, because it takes many days for the altimeters to sample the globe. The maps are replotted daily until these two data types become contemporaneous. other data are available at ranges af times as indicated.

Regional maps SST Maps GO

The colour-coded field is a map of tidal-residual, isostatically-adjusted sea level anomaly, valid for the analysis date T_a shown, which is normally five days ago. By 'tidal residual', we mean that the (relatively) rapid oscillations of sea level associated with tides are excluded. By 'anomaly', we mean the difference from the long-term average. The atmospheric pressure map (blue contours are lows, white contours are high) used for making the isostatic adjustment is shown because features of the circulation (eg near the coast, or under a tropical cyclone) can sometimes be explained by the winds.

How do we know that the ocean has areas where the water is raised or lowered by half a meter or so, for 100's of km?. If you look closely, you will see lines of little white, magenta or black dots. These lines show where satellites carrying radar altimeters have flown over, measuring the distance from the satellite down to the water. That distance is a little shorter where the sea level is raised a bit. The fact that that difference can be accurately measured is a great triumph of engineering, and is one of the key breakthroughs responsible for the present revolution in ocean observation. The colour of the dot indicates when the satellite flew over. White means more recently than T_a, magenta means the three days previous, black means longer ago.

The four satellites presently (December 2011) flying altimetry missions are called Jason-1, Jason-2, EnviSat and CryoSat. The bar plot shows the history, from 7 days before, to 3 days after T_a, of the daily number of observations made by each of these, within the region shown. (More precisely, each sea level 'observation' is a 2km-wide average along 25km of the flight path). The satellites can't sample the whole world every day because they can only measure directly beneath them. To make a complete 'quasi-synoptic' map, we must therefore use data that is up to 10 days old. The ocean changes more slowly than the atmosphere, so that is OK. The older data points are down-weighted compared to the newer ones in making the map. Where there is only old data or no data at all, the estimated anomaly relaxes to zero and the map is obviously least useful.

The other (and much older) way of measuring sea level is by tide gauge. Australia has many of these in ports all around the country [FFI]. We include these data in our maps by averaging-out the tides and making the same atmospheric pressure correction as with the altimeter estimates, then interpolating the results at many points along the coastline between the gauges. Both the observed and interpolated coastal observations are shown on the map. Coastal sea level changes more rapidly than deep-ocean sealevel, so it is just as well that the coastal observations are made much more frequently than those by satellite over the deep sea.

Regional maps

The right panel shows the sea level map again, but this time as total sea level, ie, the anomaly plus an CAST2008 estimate of the mean surface dynamic height with respect to 2000m included. Being overlain on SST, it is shown just as white contour lines. The result is the oceanographer's weather-map, with sea level taking the place of air pressure, and ocean current taking the place of the wind. The physics associated with the earth's rotation is analogous to weather systems: geostrophic currents run with the high on their left in the southern hemisphere. Black arrow heads depict the direction and strength of the ocean currents. Reference should be made to the distribution (left panel) of the available data in order to judge the reliability of these estimates. Where no recent data are available, this map will show our estimate of the time-mean sealevel slopes and currents (the reliability of which also varies regionally depending on numerous factors).

The colour-coded field is a map of Sea Surface Temperature (SST) which is formed by compositing single images over a three-day period, in order to obtain a relatively cloud-free map without averaging-out useful detail.

The SST map is best for precisely locating where the ocean currents are, while the sea level map is good for resolving ambiguities of which way the currents are generally flowing, and whether they are weak or strong. The accuracy of both maps is instructively judged by comparing them with the trajectories and speeds of Surface Velocity Program (SVP) drifters [FFI], which are shown in magenta. Drifters can sometimes travel at up to twice the speed that we estimate from the sea level maps because they measure the velocity at a point, while the map shows the average over many kilometers and a period of days.

Please note: neither the satellite nor the coastal sealevel near-real-time data streams are fully quality-controlled, and errors do occur.

Large Area Australasian maps SST Maps GO

When you enter this series of maps of the Australasian region, you are first shown a map of the Sea Surface Temperature seasonal anomaly. This is the difference of SST (composited over 6 days) from the CARS2009 atlas estimate for the time of year. Also shown are contours of isostatically-adjusted sea level, geostrophic current velocities and drifters as described above.

[sst_s] and [sst_n]: These links take you to plots of the 6-day composite of SST, with appropriate (for the time of year) temperature scales for the southern and northern waters, respectively.

[ht] shows the sea level height anomaly map overlaid with the data used in its estimation, as well as atmospheric pressure.

[uv] shows the sea level with the CAST2008 estimate of the mean surface dynamic height with respect to 2000m included, and the geostrophic surface velocity field.