Calibrating and validating satellite data
Earth observation satellites are very useful for taking images of things – for example, lakes, ice sheets or penguin colonies – on a regular basis. Satellite images are taken hundreds of kilometres above the Earth’s surface so scientists take on-the-ground measurements to calibrate and validate the data to ensure the satellite imagery is accurate.
Jargon alert:
Calibrate: to determine the accuracy of a measuring instrument.
Validate: to prove something is accurate. With remote sensing, it is a comparison of data obtained via satellite with data collected by on-the-ground observations, sampling, etc.
Satellite signal: what the satellite ‘sees’ – through the air column, on the ground and/or through the water column.
Water column: the vertical section of water between the bottom of the lake and the lake’s surface.
Questions for discussion:
Why does Moritz use colour swatches when he’s trying to determine the colour of the lake water?
How can the atmosphere interfere with satellite signals?
Why is it important for scientists to calibrate and validate satellite data?
Why do you think Moritz and other scientists share the data that they have collected?
Transcript
Dr Moritz Lehmann
Aquatic remote sensing scientist Senior Scientist, Xerra Earth Observation Institute Adjunct Senior Research Fellow, University of Waikato
When we plan a space mission, we do very thorough measurements and calculations to tell us what the signal that we’re trying to measure looks like on the ground and what it looks like up in space at the level of the satellite.
Some of the uncertainty we’re facing in looking at an image and being able to say how much cyanobacteria is in that water is that other things look very similar to cyanobacteria. So we’re collecting water samples and taking them to the lab to count cyanobacteria. We measure chlorophyll – the main green pigment of photosynthesis. We also measure water clarity, and then we measure suspended particulate matter. So we look at suspended sediments that are mixed in the water column and the brown colour that comes in from say leaf litter – those principal components of the water make up the colour overall.
And for us, it's difficult to discern which portion of that colour comes from the cyanobacteria. So one thing I like to do when I go to a lake and take the electronic and the scientific measurements is to actually look at the lake and see what colour it has to the naked eye. And in order to make that a little more objective, I have a series of colour plates here or colour swatches, and I’m trying to match the colour that I see in a body of water to one of those colour fields.
One of the issues we have to account for in looking at a satellite image is that much of the satellite signal that the satellite sees actually comes from scattering in the atmosphere. So light that comes from the Sun bounces off an air molecule, for example, and goes into the camera of the satellite. So you look at a mountain in the distance, and it’s kind of hazy and milky, and that’s the stuff we need to get rid of in order to see what the mountain really looks like. In some cases, 90% of the data that the satellite sees actually comes from the atmosphere and not from the thing that you’re interested in, so we have to account for that. So I have collected a lot of data from New Zealand lakes and shared it with an international community of scientists to develop algorithms to remove this atmospheric signal to get a clearer view from the ground.
Acknowledgements Dr Moritz Lehmann, Xerra Earth Observation Institute and University of Waikato Te Whare Wānanga o Waikato Dinghy on lake, Dr Moritz Lehmann using viewer and swatches to determine lake colour, Dilshan DaSilva Dr Moritz Lehmann lowering light sensor into water from dinghy and taking a water sample and still of Ian Hawes lowering a Secchi disc, Kelsi Doscher Light sensor on Lake Forsyth, Secchi disc, Secchi disc data visualisation and lake images before and after atmospheric noise removal, Dr Moritz Lehmann and Xerra Earth Observation Institute Satellite image of lakes, Copernicus Sentinelhub, European Space Agency (ESA) Drone sample comparison images, David Schmale Images of Mount Pirongia, Angela Schipper
