Free satellite imagery has never been easier to get — but the number of sources is now the problem. Different archives suit different jobs, and picking the wrong one wastes hours on data you cannot use.

Start with the big three free systems. The Copernicus Data Space Ecosystem is the home of Sentinel data: Sentinel-2 covers the whole planet at 10 metres roughly every five days. USGS EarthExplorer is the front door to the Landsat archive, reaching back to 1972 at 30 metres. NASA Earthdata Search indexes hundreds of datasets beyond plain imagery.

If you want to look before you download, use a viewer first. NASA Worldview shows the whole planet within hours of the overpass. EOS LandViewer lets you search and preview Sentinel and Landsat data with built-in band math. Google Earth Engine lets you process petabytes of imagery in the cloud instead of downloading it.

Match resolution to the job, not to the maximum available. Thirty-metre Landsat data suits regional analysis; 10-metre Sentinel-2 suits national and regional mapping; only sub-metre commercial imagery suits building-level work.

Understand the licenses before you build anything on free data. Sentinel and Landsat data are genuinely open, but some portals add their own terms for derived products, and commercial previews are not free data.

A sensible first workflow: define your question, pick the coarsest resolution that can answer it, preview in a viewer, then download only the scenes you need, filtered by date and cloud cover.

Once you are comfortable downloading scenes, learn about cloud-native catalogs, which are changing how satellite data is accessed. The SpatioTemporal Asset Catalog, or STAC, is an open standard that lets you search many archives through one interface, and cloud-optimized formats let you stream just the part of an image you need instead of downloading whole scenes. For large projects this is a major saving in time and storage: instead of filling a hard drive with data you may never use, you query the catalog for your area and dates and process the results in the cloud or stream them into QGIS. The traditional download portals are not going away, but the catalog approach is where the field is heading, and it is worth learning early.