May 13, 2011 12:56 PM
by Erick Geiger
Predicting Sea Surface Salinity from Space
The simplest definition of salinity is how salty the ocean is. Easy enough, right? Why is this basic property of the ocean so important to oceanographers? Well, along with the temperature of the water, the salinity determines how dense it is. The density of the water factors into how it circulates and mixes…or doesn't mix. Mixing distributes nutrients allowing phytoplankton (and the rest of the food web) to thrive. Mixing also distributes heat between the poles of the earth's poles and the equator which can affect climate. Climate change is important to oceanographers; therefore, salinity is important to oceanographers.
Spring Salinity Climatology for the Chesapeake
Salinity doesn't vary that much in the open ocean, but it has a wide range in the coastal ocean since the coast is where fresh water from rivers and salt water in the ocean mix. Our measurements of salinity along the coast provide us with a picture of how this complex mixing between fresh and salt water occurs and also affects the local biology, physics, and chemistry of the seawater. Our picture, however, is very small. Salinity data is collected by instruments on ships, moorings, and more recently underwater vehicles such as gliders. While these measurements are trusted to be very accurate, their spatial and temporal resolution leaves much to be desired when compared to say near-daily sea surface temperature estimated from a satellite in space.
Why can't we do the same thing we do with salinity then? Let's just measure salinity from a satellite and get the big picture. Well, it's not as simple, but it is possible. NASA's Aquarius mission http://aquarius.nasa.gov/ scheduled to launch this year is taking advantage of a set of three advanced radiometers that are sensitive to salinity (1.413 GHz; L-band) and a scatterometer that corrects for the ocean's surface roughness. With this they plan on measuring global salinity with a relative accuracy of 0.2 psu and a resolution of 150 km. This will provide a tremendous amount of insight on global ocean circulation, the water cycle, and climate change. Awesome, global salinity problem solved. What about coastal salinity? What if I wanted to know the salinity in the Chesapeake Bay? That's much smaller than 150 km wide.
That's where my project comes in. It involves NASA's MODIS-Aqua satellite (conveniently already in orbit: http://modis.gsfc.nasa.gov/), ocean color, and a basic understanding of the hydrography of the coastal Mid-Atlantic Ocean. Here's how it works. We already know a few things about the color of the ocean, that is, the sunlight reflecting back from the ocean measured by the MODIS-Aqua satellite. We know enough that we can estimate the concentration of the photosynthetic pigment chlorophyll-a. So not only can we see temperature from space, but we can estimate phytoplankton biomass too! Anyway, there are other things in the water that absorb light besides phytoplankton and alter the colors we measure from the satellite.
Spring Salinity Climatology for the Mid-Atlantic
We group these other things into a category we call colored dissolved organic material or CDOM. CDOM is non-living detritus in the water that either washes off from land or is generated biologically. It absorbs light in the ultraviolet and blue wavelengths, so it's noticeable from satellites. In coastal areas especially, its main source of production is runoff from land. So, CDOM originates from land and we can see a signal of it from satellites that measure color. What's that have to do with salinity?
You may have already guessed it, but water from land is fresh. So, water in the coastal ocean that is high in CDOM should be fresher than surrounding low CDOM water. That's our basic understanding of the hydrography of the coastal Mid-Atlantic Ocean, how it relates to ocean color, and why we need the MODIS-Aqua satellite to measure it. With this information, I compiled a lot of historic salinity data from ships (over 2 million data points) in the Mid-Atlantic coastal region (Chesapeake, Delaware, and Hudson estuaries) and matched it with historic satellite data from the MODIS-Aqua satellite. Now I have a dataset that contains ocean color and salinity. Using a non-linear fitting technique, I produced an algorithm that can predict what the salinity of the water should be given a certain spectral reflectance. I made a few of these algorithms in the Mid-Atlantic, one specifically for the Chesapeake Bay. It has an error of ±1.72 psu and a resolution of 1 km. Not too bad, but there's always room for improvement. Even so, this is an important first step for coastal remote sensing of salinity. An algorithm like this can be used to get salinity data as often as sea surface temperature. That's pretty useful. Now we just need to apply it and make it better.
Climatologies of salinity can be downloaded here: http://modata.ceoe.udel.edu/dev/egeiger/salinity_climatologies/
Dr. Art Trembanis
No comments:
Post a Comment