Remote sensing reflectance from Sentinel 2 in SNAP, ArcGIS or QGIS

Unlocking Earth’s Secrets: A Human’s Guide to Sentinel-2 Reflectance with SNAP, ArcGIS, and QGIS

Ever gazed at a satellite image and wondered what all those colors really mean? A big part of that answer lies in something called remote sensing reflectance. Think of it as the Earth’s way of showing off what it’s made of – literally reflecting sunlight back into space, where satellites like Sentinel-2 can capture it. Sentinel-2, part of the European Union’s Copernicus Programme, is a real workhorse, constantly snapping high-resolution pictures that help us monitor our environment, manage land use, and respond to disasters. But to truly understand these images, we need to get down to brass tacks: processing that raw data into something usable. That’s where reflectance comes in, and where software like SNAP, ArcGIS, and QGIS become our trusty sidekicks.

Sentinel-2 and Reflectance: The Dynamic Duo

So, what’s the big deal with Sentinel-2? Well, imagine having not one, but two eyes in the sky! That’s essentially what Sentinel-2A and Sentinel-2B are. These twin satellites work in tandem, zipping around the Earth and giving us a fresh look at the same spot every few days. This rapid revisit time is a game-changer, especially when tracking fast-changing events like floods or wildfires. Each satellite is armed with a MultiSpectral Instrument (MSI), a fancy camera that sees the world in 13 different “colors,” or spectral bands. These bands span from the colors we see (visible light) to infrared wavelengths we can’t, each revealing unique information about the Earth’s surface. And with resolutions ranging from a super-sharp 10 meters to a still-pretty-good 60 meters, we can really zoom in on the details.

Now, Sentinel-2 data comes in a couple of flavors: Level-1C (L1C) and Level-2A (L2A). L1C is like a raw photo, straight from the camera – it’s been tweaked for clarity, but still has atmospheric haze. This is TOA, or Top-Of-Atmosphere reflectance. Level-2A, on the other hand, is like a professionally edited shot. It’s been atmospherically corrected to remove that haze, giving us a much clearer picture of what’s actually on the ground. This is BOA, or Bottom-Of-Atmosphere reflectance, also known as surface reflectance. For most serious work, L2A is the way to go. Trust me, skipping the extra atmospheric correction saves a ton of time and lets you directly compare reflectance values from different images.

Getting Your Hands on Sentinel-2 Data

The best part? Sentinel-2 data is totally free! You can grab it from the Copernicus Open Access Hub. Just punch in your area of interest, the dates you need, how much cloud cover you can tolerate, and which product type you want (L1C or L2A). Another great resource is Digital Earth Africa, which provides pre-processed Sentinel-2 Level-2A Collection 1 data – a real time-saver!

Processing with SNAP: ESA’s Gift to Earth Observers

SNAP (Sentinel Application Platform) is like the Swiss Army knife of Sentinel data processing. Developed by the European Space Agency (ESA), it’s free, open-source, and packed with tools for everything from atmospheric correction to data analysis.

Tackling Atmospheric Correction:

Got L1C data? No sweat! SNAP’s built-in Sen2Cor processor will handle the atmospheric correction for you. It’s the official tool for Sentinel-2, so you know it’s legit.

Decoding Reflectance Values:

Sentinel-2 L2A data stores reflectance as digital numbers (DNs). To turn these into actual reflectance values (between 0 and 1), you’ll need to do a little math: divide the DNs by 10,000. This scaling factor is hiding in the metadata file that comes with the Sentinel-2 product. SNAP’s Band Maths tool makes this conversion a breeze.

ArcGIS: The Industry Standard

ArcGIS, the commercial powerhouse of the GIS world, also plays nicely with Sentinel-2 data.

From DNs to Reflectance:

Image-Based Atmospheric Correction:

ArcGIS also lets you get creative with image-based atmospheric correction. This involves estimating atmospheric scattering and subtracting it from the TOA reflectance values to get an approximation of surface reflectance.

QGIS: The Open-Source Hero

QGIS, the free and open-source alternative to ArcGIS, is another great option for processing Sentinel-2 data.

Reflectance Conversion:

Semi-Automatic Classification Plugin (SCP):

The Semi-Automatic Classification Plugin (SCP) is a real gem for QGIS users. It simplifies Sentinel-2 processing, converting L1C data to TOA reflectance and applying a basic atmospheric correction using the DOS1 method. It also handles the DN-to-reflectance conversion for L2A data.

Diving into Water: Remote Sensing Reflectance (Rrs)

If you’re interested in studying water bodies, you’ll want to calculate remote sensing reflectance (Rrs). This is basically the amount of light bouncing off the water, divided by the amount of light hitting it. While L2A data gives you a good starting point, specialized software like ACOLITE is often used for more accurate atmospheric correction and Rrs retrieval, especially in complex coastal and inland waters. ACOLITE takes into account factors like adjacency effects that can mess with reflectance measurements over water.

The Takeaway

Turning Sentinel-2 data into accurate reflectance values is a crucial step for unlocking its full potential. Whether you’re using SNAP, ArcGIS, or QGIS, understanding the different data levels, atmospheric correction techniques, and reflectance conversion methods is key. So, go forth, process your data, and start uncovering the secrets hidden in those satellite images!