EXPLORER FIELD GUIDE / EARTH
Earth Data: source reliability, freshness and request status
Read the Earth Data observatory correctly: HTTP success, record age, comparisons and unavailable providers are separate kinds of evidence.
Open Earth ↗Opens the interactive application in a new tab. Your existing explorer stays open.
A source registry you can inspect
In Earth, open Sources & requests to inspect the Data Observatory. The panel groups providers by subject and exposes browser requests as well as server-side observations when the backend is available. Its search and category filters help find one provider without combining all sources into one misleading reliability score.
What a percentage actually measures
The HTTP success percentage uses completed successful and failed requests for the selected source and scope. Cancelled requests and unknown outcomes are displayed separately. Browser resource loading, browser API traffic and server-to-provider requests are not interchangeable measurements. A 100% success rate can describe an old response; it does not certify scientific correctness, spatial coverage or the number of active users.
Freshness is a property of the data
The observatory records evidence timestamps and distinguishes historical data where supported. Compare the observation or valid time with the question you are asking. A historical wildlife record can be legitimate while being unsuitable for a claim about animals currently present on one street. A cached orbital catalogue can drive a prediction while its age increases uncertainty. Missing timestamps should remain unknown, not silently become today’s observations.
Sampling is an explicit action
The panel offers source sampling and shows unavailable or failed results rather than making all providers look verified. Loading this reading page performs no provider sample and publishes no endpoint responses. Source licences and attribution differ across mapping, imagery and scientific datasets; follow the original provider and per-record terms before reuse. Request counters are operational diagnostics, not a published scientific dataset or a guarantee of service uptime.
Editorial review: · Not a live data feed.