Monitoring Sea Surface Salinity from Space; Awesomeness of Remote Sensing

Few days ago, I met some scientist from Geology department. While talking, I have found Geologist/hydrologist/ ecologist ,…, are very much interested to find out a way to estimate water salinity without going to the field and measuring salinity by  electrical conductivity. The first solution what came to my mind is very simple… I asked them “why don’t you monitor salinity from space?”. Then they asked, “how?”. I didn’t knew the answer then. That is how I get involved in these sort of topic and have started to look for a proper solution.

Sea Surface Salinity (SSS) dynamic studies are in the heart of several major scientific domains which play a vital role in sea water circulation patterns, influences the spatial distribution of many marine organisms, and affects seawater density in both coastal systems and open oceans.

Satellite remote sensing offers the potential to estimate salinity across entire water bodies at the frequency of satellite overpass, dramatically enhancing our monitoring capabilities relative to in situ observation networks. To date, however, satellite missions targeting salinity have focused on open ocean rather than coastal applications. NASA’s Aquarius mission, launched in June 2011, and the European Space Agency’s Soil Moisture and Ocean Salinity (SMOS) mission launched in November 2009, are capable of measuring SSS from space across the world’s oceans, but the 150 km spatial and 7-day temporal resolution of Aquarius and the 250 km spatial and 10–30 day average temporal resolution of SMOS are too coarse for coastal and Earnestine environments.


Attempts to successfully map SSS in the coastal region via remote sensing have ranged from Skylab photography to microwave radiometer measurements, decametric wave ranges, ESTAR measurements, and Landsat TM data. The use of satellite imagery to map SSS in an estuary was first performed in the San Francisco Bay. This pioneer study found correlations between Landsat TM color bands and SSS in an estuarine environment. Other studies explore the empirical relationships between colored dissolved organic matter (CDOM) and salinity using remotely sensed ocean color in a coastal setting. These sort of studies showed that a straight-line relationship between CDOM and salinity is expected dependent on the ratio of the flushing time of an estuary and the timescale of the source variation.

Surface Salinity of Chesapeake Bay; derived from MODIS image (the salinity observation network (left); overlay of salinity estimation and observation (middle);  overlay of temperature estimation and observation (right) )

Recently one study has successfully  predicted SSS in the Chesapeake Bay at 1 km resolution using MODIS-Aqua ocean color bands using multiple statistical models and an artificial neural network (ANN) algorithm applied to MODIS-Aqua data. Besides, some scientist modeled SSS from MODIS satellite data in the east coast of Malaysia, using oleast square model as compared to 2nd polynomial order algorithm .Others have also used satellite remote sensing to map low-salinity water plume in the East China Sea. There are also a similar kind of exercise to predict salinity from remote sensing for the Columbia River plume.

salinity and CDOM relationship

all these studies tried to predict salinity in the coastal ocean for an specific region using proper regional algorithms from large in-situ measurements of apparent and inherent optical properties. these studies have one common drawback and that is none of these algorithms and method can successfully predict the salinity of other region then their own. then again these studies have given a proper signal of the ability of measuring salinity from space but suffers from data and methodical limitations.

PROPOSAL: A good research proposal could be, an attempt or to establish a generic predictive algorithm which can successfully predict salinity of most of the coastal regions with a tolerable accuracy. To assure satellite data consistency with higher spatial and temporal resolution, the study may involve testing several method of atmospheric correction and interpolation techniques. the study also can be aided by several point process models from probability wing and/or some spatial autoregressive models (lattice approach) which can be devised to run auto regression in space and time simultaneously. As those spatio-temporal models are data dependent so the study may also involve extensive in-situ salinity measurement with higher temporal and spatial resolution from several/different physical and biological costal settings.

Say thanks for this article (0)
#Capella Space #ICEYE #Maxar #Optical #Planet #SAR #Transportation
Earth Observation Plays a Vital Role in Global Supply Chain Monitoring
Hassan Safieddine 06.27.2022
Space ESG. How might the evolution of space industry cause us to reflect on our own mistakes?
Aleks Buczkowski 04.17.2022
#Atmospheric #Copernicus #Environmental Protection #ESG #Landsat #Multispectral #Optical
How Satellite Data Helps Us Understand Climate Change
Justyna Więcławska 04.22.2022
Next article

The frozen continent without ice: New Topography of Antarctica

One of the earth’s most alien landmasses, the Antarctic’s rocky surface, concealed by a vertical ice crust of up to more than 3km (average thickness 1,6km), was unveiled, at least digitally, at an unprecedented detail in spring 2013. Scientists have created a significantly improved map of the continent’s bedrock in the frame of the bedmap2 project led by the British Antarctic Survey. While the previous version ‘bedmap’, produced in 2001, gave a general overview of the Antarctic’s topography, the high spatial resolution of bedmap2 shows a landscape of mountain ranges, hills and vast plateaus dissected by canyons and gorges. Thus, it conveys the cartography of the bedrock terrain much closer to reality than ever portrayed. The project was prompted by the motivation to calculate the total extension and mass of ice that occurs in the Antarctic, which is essential to estimate the sea level rise induced by the melting of the polar ice. Therefore, the shape and structure of bedrock beneath the ice sheet was needed to know in detail. Bedmap2 relies on information from NASA’s IceSAT1 satellite (Cloud and Land Elevation Satellite) gathered between 2003 and 2009 and ice thickness data from the airborne based Operation IceBridge (radio echo sounding measurements, seismic techniques as well as satellite and cartographic data).

The new level of measurements delivers important information on variables that are of interest for research, particularly climate related study due to the key role of the Antarctic in this matter. Some new facts are:

  • The volume of Antartic ice is 4,6% higher than previously assumed (27 million km³)
  • Similar to previous estimates the melting of Antarctic ice shield would cause a 58m rise of sea level
  • The mean depth of the bed beneath the grounded ice sheet is 72 m deeper than thought to date
  • The ice sheet grounded on bed below sea level is increased by 10 % (which means that a considerable greater extent of bedrock extends below sea surface in bedmap2)
  • The Bedmap 2 survey returned more accurate data on the lowest point on the Earth’s crust, which extends below the Byrd Glacier in Victoria Land at around 2,870 metres below sea level (around 400 metres lower than was previously thought).

The full survey of Antarctica via Bedmap2 is available for free via The Cryosphere, an open-access journal of the European Geosciences Union.

British Antarctic Survey – bedmap2

NASA – IceSAT 1&2

Operation IceBridge

Image showing the Antarctic' surface with the ice sheet as seen from space

Image showing the Antarctic’ surface with the ice sheet as seen from space

The image shows the differences between the most recent map of Antarctic's rocky surface bedmap2 (left) and its previous version bedmap (right) from 2001. The topography in bedmap2 reveals much more detail than in bedmap.

The image shows the differences between the most recent map of Antarctic’s rocky surface bedmap2 (left) and its previous version bedmap (right) from 2001. The topography in bedmap2 reveals much more detail than in bedmap.

Read on