Poor Landsat 5 coverage in Google Earth Engine

The Patchy Legacy: Digging into Landsat 5’s Coverage Quirks in Google Earth Engine

Landsat 5. What a legend! This satellite wasn’t just hanging around up there; it broke records, earning a spot in the Guinness Book for its incredible run as an Earth observer. Launched way back on March 1, 1984, it kept snapping photos of our planet for over 28 years before finally retiring on June 5, 2013. But here’s the thing: if you’ve ever tried using Landsat 5 data in Google Earth Engine (GEE), you’ve probably run into some frustrating gaps. So, what’s the deal with these missing pieces? Let’s dive in and see why this amazing satellite’s data isn’t always complete.

Landsat 5: A Quick Look Back

This mission, a team effort by NASA, NOAA, and USGS, had two main cameras: the Multispectral Scanner System (MSS) and the Thematic Mapper (TM). These instruments were workhorses, providing images that helped us understand everything from environmental changes to crop health and disaster impacts. Orbiting at 705 km, Landsat 5 circled the Earth every 16 days, capturing the whole planet at a resolution of about 30 meters. Pretty impressive, right?

Why the Holes in the Data?

Okay, so why can’t we just pull up a complete, seamless Landsat 5 view of the world in GEE? Well, there are a few key reasons:

• Old School Tech Limitations: Back in the day, satellites weren’t exactly swimming in storage space or high-speed internet. Landsat 5 simply couldn’t store and beam down all the data it collected, especially when it was out of range of US ground stations. A primary TDRSS transmitter failed in 1988, and a redundant one failed in 1992, making things even harder. Imagine taking photos 24/7 but not being able to save them all – that’s basically what happened! I remember working on a project years ago where we desperately needed data from a specific region in Africa, only to find huge chunks missing due to this downlink issue. Talk about frustrating!

• Gremlins in the System: Landsat 5 wasn’t immune to glitches. Things like the “Sun Glint Anomaly,” where sunlight messed with the signal, caused data loss. Then there were the “Christmas Tree” anomalies (weird telemetry data popping up in the images) and dropped scans (blips with the main mirror or transmission). It was like the satellite had a mind of its own sometimes!

• Wear and Tear: Let’s face it, after 28 years, things start to wear out. The MSS instrument was actually turned off for a while in 1995, then brought back to life in 2012. The TM instrument finally gave up the ghost in November 2011, and eventually, a gyroscope failure led to the whole mission being shut down. So, the Landsat 5 collections in Google Earth Engine end around May 2012.

• The Commercialization Era: There was a period when the Landsat program was run more like a business. If no one was willing to pay for the data right away, it often wasn’t even saved. This led to even more gaps, especially between 1984 and 1999.

Filling in the Blanks: What You Can Do

So, what can you do when you hit those annoying data gaps? Don’t despair! Here are a few tricks:

• Bring in the Newer Models: Landsat 7 (launched in 1999) and Landsat 8 (launched in 2013) generally have better coverage. Landsat 7 has its own issues (the infamous scan line corrector failure), but you can often combine its data with other sources to get a more complete picture.

• Become a Gap-Filling Ninja: There are ways to estimate the missing data. You can use the average of the images before and after the gap, or try more sophisticated methods to fill in the blanks.

• Expand Your Horizons: Don’t be afraid to look at other satellites. Sentinel-2, for example, gives you high-resolution images with lots of different bands. The Harmonized Landsat Sentinel-2 (HLS) data product is also super useful, as it combines data from both Landsat and Sentinel-2.

• Be Strategic About Location: Data coverage isn’t uniform across the globe. Some regions, like the Continental US, have more complete Landsat 5 archives.

• Stick to Collection 2: Make sure you’re using the latest Landsat Collection 2 data in Google Earth Engine. The old Collection 1 is no longer supported.

• Mask Those Clouds! Clouds can obscure the view, so use cloud masking techniques to clean up your images.

Final Thoughts

Sure, Landsat 5 has its data gaps, but its long history still makes it an invaluable resource for studying long-term changes on our planet. By understanding why these gaps exist and using the right strategies, you can still unlock a wealth of information from this legendary satellite. So, don’t let those missing pieces discourage you – get creative and keep exploring!