- Home
- Air-Quality Forecast Valid Time vs Update Time
Forecast timestamp guide
Air-Quality Forecast Valid Time vs Update Time
Air-quality forecast valid time vs update time begins with one rule: identify what each timestamp actually means. Valid time is the date and hour a forecast point describes. “Update time” can refer to a model run, provider publication, application fetch, cache age, or page refresh, so the label must name the clock. OpenWeather’s Air Pollution API documents dt for each current or forecast point but does not document a separate model-run timestamp for this feed. AirAgora therefore preserves point times, records fetch and cache state in technical provenance, and surfaces an explicit stale label without inventing a provider publication time. A browser refresh does not create a new model run, and every timestamp remains modeled provenance rather than a street-level observation.

The short definition: represented hour versus a named freshness clock
Valid time answers “when are these modeled conditions supposed to apply?” An update or freshness time answers a different question only after the interface defines it: it might mean model production, provider publication, application fetch, cache creation, or page display. Those clocks are not interchangeable, and a generic “updated” label should never be treated as a model-run time without documentation.
Use valid time to align a forecast point with an activity, alert threshold, or later observation. AirAgora’s fetch and cache times belong to technical provenance; the planner currently exposes the provider point time and any stale warning rather than an exact fetch-age field. Use a provider run or publication time only when the provider actually supplies and defines one. None of these timestamps measures personal exposure or guarantees local conditions.
OpenWeather forecast points carry a Unix UTC time
OpenWeather documents the dt field as date and time in Unix UTC. That value identifies the represented forecast point. AirAgora converts the absolute instant to the selected location’s IANA timezone for display. The conversion changes the clock label and possibly the local calendar date, but it does not change which instant the forecast point represents.
This distinction matters near midnight and daylight-saving changes. Two locations can display different local clock times for the same absolute instant. A repeated local hour can refer to two different instants, while a skipped clock label should not be invented. Store the absolute timestamp and timezone when an exact later comparison matters.
OpenWeather does not document a separate run time for this feed
The Air Pollution API documentation defines dt for each current or forecast point. It does not define a separate model-run, issue, or publication timestamp in the documented response. That means a point valid tomorrow cannot be paired with an invented “run time” merely because the application fetched the response today.
AirAgora does not promise how often OpenWeather publishes a new air-pollution dataset. A newly fetched response may contain changed points or the same points, but fetch time alone cannot prove a new provider run. The honest public evidence is the documented point time and any stale warning; AirAgora’s technical provenance separately records fetch and cache state.
Fetch time and view time are application clocks
Fetch time records when AirAgora asked the provider or read a valid cache entry. View time is when the browser displayed the page. Those operational clocks help explain performance and reuse, but they do not redefine the environmental hour. They belong to technical provenance rather than an exact fetch-age field in the current planner. A user can open a page at 10:05, receive a cached response fetched at 09:55, and inspect a forecast point valid at 14:00 local.
The example needs no invented pollution value: the timestamps alone show the roles. The 14:00 point should be compared with other values valid at 14:00. The 09:55 fetch can help explain cache age. Unless a provider separately documents a run time, neither the fetch nor the 10:05 page view identifies when a model cycle was produced.
Stale data preserves point times and the original fetch context
AirAgora treats the forecast endpoint cache as fresh for 20 minutes and the current endpoint cache as fresh for 10 minutes. After a provider error, a valid retained entry may be served for at most two hours with an explicit stale label. The original point timestamps and fetch context remain attached; the failed refresh time does not become a model-run time or a current observation.
Once the retention limit expires, or when no valid cached response exists, AirAgora reports unavailable. It does not rewrite old timestamps, fill gaps, switch to scraping, or silently use another provider. The timestamp contract is therefore also a truthfulness contract: older evidence remains older evidence even when it is temporarily the only evidence available.
Use the right timestamp for the question
For ordinary planning, align the activity with valid hours in local time and use the displayed provider point time plus any stale warning. For an operational comparison that has access to technical provenance, align the same valid hour and retain each response’s fetch and cache timestamps. View time can document when a person saw the page. None of those application clocks is a provider model run unless the provider documents it.
When comparing another screen, also align location, pollutant, unit, averaging context, provider scale, and modeled-versus-observed provenance. A timestamp match alone does not make two values equivalent. This page is about clock roles, not a declaration that one source is more accurate than another.
A compact timestamp checklist
Before acting, ask: What local hour does the value represent? Which absolute timestamp and timezone produced that label? Does the provider document a separate run time, or only the point time? Is the planner marked stale? When did I view it? If technical provenance is available, when did AirAgora fetch or reuse the response? Recording those answers prevents a future-hour forecast from being called a current measurement and prevents a page refresh from being mistaken for a model update.
Then apply the normal boundaries. The value remains modeled for a grid, the OpenWeather 1–5 provider scale remains distinct from regulatory AQIs, and a matching timestamp does not establish street-level exposure or universal safety. Official local guidance remains primary for urgent events.
Direct answers
Frequently asked questions
What is valid time in an air-quality forecast?
Valid time is the date and hour represented by a forecast value. AirAgora formats the absolute provider timestamp in the selected location’s timezone so the period can be matched to a real plan.
Does OpenWeather publish a separate update or model-run time?
Not in the documented Air Pollution API response. OpenWeather defines dt for each point, while AirAgora separately records fetch and cache state in technical provenance and surfaces any stale warning rather than inventing a provider run timestamp.
What does AirAgora’s “Forecast time” represent?
It is the timestamp attached to the provider’s current modeled point, displayed separately from future hourly valid times and alongside any stale warning. AirAgora does not market it as an exact model-publication time.
Does refreshing the page create a new update time?
No. The request can reuse a valid cache or fetch a response with unchanged point times. A page view or network request does not prove that the provider published a new model run.
Which timestamp stays on stale data?
The original point timestamps stay visible, technical fetch context remains attached to the response, and an explicit stale warning is shown. The failed refresh attempt does not make the retained forecast current or newly observed.
Which time should I save for later verification?
Save the valid time, absolute timestamp, location timezone, displayed provider point time, any stale warning, place, pollutant, unit, provider scale, and the forecast values the user actually saw. If technical provenance is available, save fetch and cache timestamps separately. Save a provider run time only if the provider supplies one.
Evidence register
Sources and further reading
The links below identify the primary references used on this page. Selection, summary, and presentation are AirAgora modifications.
Keep exploring
AirAgora provides informational planning guidance, not medical advice. Local conditions can differ from modeled forecasts; follow official guidance during severe events.