NOAA Space Weather Prediction Center (SWPC): Intelligence Source Guide

SWPC publishes the operational space weather picture as open JSON and text: solar wind at the L1 point, geomagnetic indices, X-ray flare flux, aurora forecasts and formal alerts. It is the source that explains why HF went dead, GNSS drifted and a satellite decayed faster than modelled.

At a glance

Source	NOAA Space Weather Prediction Center (SWPC)
Category	Aviation & Space › Space Weather & Astronomy
Homepage	https://www.swpc.noaa.gov/
Machine interface	`https://services.swpc.noaa.gov/`
Format	`JSON`
Access	Open — no account required
Disciplines	Meteorological Intelligence
Mission domains	Space & Satellite Intel, RF & Signals Intel, Energy Security, Aviation Security

Solar wind, geomagnetic (Kp/Dst), aurora, X-ray flares, planetary indices and alerts as open JSON/text products. — as catalogued in the platform’s own source registry.

The Space Weather Prediction Center is a US National Weather Service centre in Boulder, Colorado, and it is one of very few organisations anywhere issuing operational, round-the-clock space weather forecasts and warnings. Its public data service exposes the underlying products as machine-readable files rather than as a query API: directories of JSON and plain text served over HTTPS, refreshed on schedules ranging from one minute to daily, with no authentication and no key. The content falls into several families. Real-time solar wind from spacecraft at the first Lagrange point – speed, density, temperature and the interplanetary magnetic field components, of which the north-south component is the operationally decisive one. Geomagnetic activity indices, principally the planetary K index and its estimated real-time variant, plus derived storm classifications. Solar X-ray flux from the GOES satellites, which is what defines a flare's class and therefore the radio blackout scale. Energetic particle flux, relevant to satellite anomalies and to radiation exposure on polar flight routes. Aurora forecasts from a model producing hemispheric maps of expected auroral power. Human-authored forecast products including a three-day forecast, a forecast discussion and a longer-range outlook. And a formal alerts, watches and warnings stream, which is the product with operational weight because it is what downstream industries actually act on.

The analytical job this source does, and no other source in this catalogue does, is discriminate between natural and deliberate degradation of the electromagnetic environment. When HF communications fail across a region, when GNSS positions wander or a receiver loses lock, when a satellite link degrades, when a low-orbit spacecraft's tracking solution stops fitting – each of those has both a hostile explanation and a solar one, and the difference matters enormously. An analyst who reports GPS interference over a region without first checking whether an X-class flare had just produced a radio blackout, or whether a geomagnetic storm was driving ionospheric scintillation, is producing an attribution claim that will not survive contact with a physicist. Running SWPC data as a standing background layer converts that from a special check into a default one. Beyond the deconfliction role, the data supports timing and planning: HF propagation conditions for maritime and aviation communications, polar route viability during solar radiation storms, satellite drag conditions that determine how fast orbital element sets go stale, and geomagnetically induced current risk to power grids and pipelines. It is METOCINT for the space and radio-frequency domains, and it belongs in the baseline of any operation that depends on radio, satellites or the grid.

Who publishes it, and why that matters

SWPC is a US federal government centre operating as part of the National Weather Service under NOAA, funded by appropriation, with a formal operational mandate rather than a research one. That mandate matters: it staffs forecasters continuously, it issues products on defined schedules with defined lead times, and its alerts are consumed under service agreements by airlines, power utilities, satellite operators and national authorities. The data is published as a public good with no charge, no registration and no commercial tier, which removes the usual questions about incentive distortion. The genuine risks are institutional rather than commercial. Its observing capability depends on spacecraft that are ageing and on replacements that depend on continued funding, and the loss of a solar wind monitor at L1 would degrade warning lead times globally with no immediate substitute. Product paths, file formats and directory structures do change as systems are modernised, occasionally without much notice, so an integration should fail loudly rather than silently when a file stops appearing. And as a national service it produces a US-centric product framing – the scales, the alert thresholds and the impact statements are written for US infrastructure operators – even though the physics is global.

Provenance is the first question to ask of any dataset and the one most often skipped. Who collects it, what their incentive is, whether they publish a methodology, and whether they correct the record when they get something wrong all bear directly on how much weight a finding drawn from it can carry.

What a record actually contains

The fields you will be working with, what each one means, and whether it is something you can pivot on. Read the meanings carefully — more analysis is wrecked by misreading a field than by failing to find one, and a field that looks like an observation is often an inference.

Field	Type	What it means	Pivot value
`time_tag`	timestamp	The observation or model time in UTC. Real-time solar wind products carry the time the measurement was made at the spacecraft, not the time the resulting disturbance reaches Earth, and the difference is roughly half an hour to an hour.	Correlation with any timestamped RF, GNSS, satellite or grid event; the anchor for every space weather timeline.
`bz_gsm`	int	North-south component of the interplanetary magnetic field in nanotesla. Sustained negative values allow efficient coupling between the solar wind and Earth's magnetosphere and are the single best short-lead predictor of geomagnetic storming.	Geomagnetic storm onset prediction, correlation with the K index a few hours later, aurora extent.
`bt`	int	Total interplanetary magnetic field strength in nanotesla. High total field with a strongly southward component is the storm-producing combination; high total field alone is not.	Coronal mass ejection arrival identification, shock detection when combined with a density and speed jump.
`speed`	int	Solar wind bulk speed in kilometres per second. A sharp step increase marks the arrival of a shock or a stream interaction region; sustained high speed indicates a coronal hole high speed stream rather than an ejection.	Event classification, transit time reconstruction back to a solar source, recurrence analysis on the solar rotation period.
`density`	int	Solar wind proton density in particles per cubic centimetre. A simultaneous jump in density, speed and field strength is the classic signature of an interplanetary shock arriving at the monitoring spacecraft.	Shock arrival timing, magnetopause compression estimates, correlation with sudden storm commencements.
`kp_index`	int	The planetary geomagnetic activity index on a quasi-logarithmic scale from zero to nine, nominally over three-hour intervals. Real-time values are estimates; the definitive index is produced later by the international custodian and can differ.	Storm classification on the G scale, aurora visibility latitude, historical geomagnetic activity comparison.
`estimated_kp`	int	A one-minute estimate of planetary activity derived from a subset of ground magnetometers. It exists because operators cannot wait three hours, and it should never be cited as the K index without saying it is the estimate.	Near-real-time storm monitoring, alert triggering, correlation with the definitive index once published.
`xray_flux`	int	Solar soft X-ray irradiance from the GOES sensors in watts per square metre. Flare class is defined directly from this number, and the radio blackout scale is defined directly from flare class.	Flare classification, D-region absorption and HF blackout onset, correlation with reported communications failures.
`proton_flux`	int	Integral flux of energetic protons above a threshold energy, in particle flux units. Elevated levels define a solar radiation storm and drive polar route diversions, satellite single-event upsets and radiation exposure concerns.	Solar radiation storm classification, aviation polar route impact, satellite anomaly correlation.
`noaa_scale`	enum	The G, S and R scale values for geomagnetic storms, solar radiation storms and radio blackouts, each running from one to five. These are the operational language of space weather and the level at which non-specialist decision makers act.	Impact statements, alert thresholds, communication with operational stakeholders who will not read flux values.
`message_code`	enum	The identifier of an alert, watch or warning product in the alerts stream. Watches are issued in advance of expected conditions, warnings when conditions are imminent or occurring, and alerts when a threshold has been crossed.	Event timeline construction, lead time analysis, reconstruction of what was known and when.
`issue_datetime`	timestamp	When the alert product was issued, distinct from the time the condition began. The gap between the two is the operational lead time, and it is the number that matters when reconstructing whether a warning was actionable.	Warning lead time analysis, after-action reconstruction of decision timelines.
`aurora_power`	int	Modelled hemispheric auroral power and the derived probability grid. This is model output driven by solar wind input, not an observation, and it forecasts roughly the next half hour to hour.	Auroral oval extent, HF absorption in the auroral zone, geomagnetically induced current risk regions.
`observed_flag`	enum	Whether a value in a forecast product is observed, estimated or predicted. Mixing the three within a single time series is the most common way this data is misread, and the flag is what prevents it.	Quality gating; separate observed history from forecast in any analysis that will be reviewed.

Coverage — and what is not in it

The physical coverage is global by nature – geomagnetic storms and radio blackouts are planetary phenomena and the solar drivers are observed from a single vantage point that serves the whole Earth. The practical coverage varies sharply by product. Solar wind measurement is a single-point sample at L1, roughly a million and a half kilometres sunward, which gives warning lead times of roughly thirty to sixty minutes depending on solar wind speed; it tells you what is about to arrive, not what is arriving everywhere. X-ray flux is continuous and immediate, since the photons arrive at light speed and the effect on the sunlit hemisphere is simultaneous with the observation. Geomagnetic indices are ground-derived from a global network of magnetometers, so they are planetary averages that can badly understate or overstate local conditions at any specific location. Temporal coverage of the real-time files is short by design – typically the last few hours to a few days, with older data rolling off – because these are operational products, not an archive. Deeper history is held by the national geophysical data archives rather than by the real-time service, which is a distinction that catches out anyone trying to build a multi-year study from the live endpoints. Cadence ranges from one-minute solar wind and magnetometer products through three-hour indices to daily and three-day forecast text.

Known blind spots

Absence of evidence here is not evidence of absence. These are the conditions under which NOAA Space Weather Prediction Center (SWPC) will not show you something that is nevertheless real:

Real-time indices are estimates that get revised. The definitive planetary K index is produced later by the international custodian and the definitive ring current index by a different institution, and both can differ from the real-time value used to trigger an alert.
Solar wind is measured at a single point upstream. Structures narrower than the spacecraft's sampling geometry, or arriving at an angle that misses it, produce effects at Earth with no corresponding upstream signature and therefore no warning.
Geomagnetic indices are global averages and hide enormous local variation. A moderate planetary index is entirely compatible with severe local geomagnetic disturbance at a specific high-latitude site, which is exactly where the infrastructure impacts occur.
Real-time product files hold only a short rolling window. Anything more than a few days old has to come from the separate archival services, and an integration that assumes the live files are an archive will lose data silently.
Coronal mass ejection arrival time predictions carry uncertainty measured in many hours, so a forecast arrival window is not a schedule and should never be presented as one to an operational decision maker.
Instrument outages and spacecraft anomalies produce gaps and bad values in the real-time stream, and some products propagate obviously wrong numbers rather than dropping them, so range validation is mandatory rather than optional.
The impact statements attached to the storm scales are written for United States infrastructure and mid-latitude conditions, and translating them directly to other regions – particularly high-latitude or equatorial ones – overstates or understates the real effect.
Ionospheric effects on GNSS and HF are highly local and frequency-dependent, and no product here tells you what happened at a specific receiver on a specific band; the data explains conditions, it does not measure your link.
The service depends on ageing observing spacecraft, and a gap in solar wind monitoring would remove the primary warning lead time with no substitute available at short notice.

Write the blind spot into the product. A statement that something “was not observed in NOAA Space Weather Prediction Center (SWPC)” is defensible; a statement that it “did not happen” is not, and the difference is what survives cross-examination.

Access, licensing and what you may do with it

Access model: Open — no account required

There is no key, no registration and no rate limit gate. You fetch files over HTTPS from the services host, which serves directories of JSON and text products organised by family. The practical work is discovering which files exist and what their update cadence is, and the correct method for that is to browse the service directories directly rather than to trust any hardcoded list, including one from documentation that may be older than the current product set. Build your collector to fetch by explicit path, validate the schema of what comes back, and fail loudly when a file disappears or changes shape, because these are operational systems that get modernised and a silently empty product is worse than a visible error. Use conditional requests where the server supports them so you are not re-downloading unchanged files, and match your polling interval to the product cadence – polling a one-minute product every ten seconds is pure waste, and polling a daily forecast every minute is worse.

Licence

As a work of the United States federal government, this material is generally in the public domain within the United States and NOAA's policy is one of full and open data sharing without charge. In practice that means you may use, redistribute and build commercial products on it. Two qualifications are worth stating. First, NOAA asks that its products not be presented in a way that implies official endorsement of your derived product, and that you not represent modified data as NOAA's – this is a reputational rather than a licensing constraint, and it is reasonable. Second, some products incorporate inputs from partner agencies and international sources whose own terms may differ, so wholesale republication of a composite product is not automatically as unconstrained as republication of a raw GOES flux file. Attribution costs nothing and resolves nearly all of this. If you are building something commercial on the alerts stream in particular, read the current policy statements rather than relying on the general public domain assumption.

Rate limits and fair use

No published quota, which is not the same as no limits. This is a static file service supporting operational users including national authorities and infrastructure operators, and the etiquette is correspondingly serious. Poll each product at its actual update cadence and not faster: one-minute products once a minute at most, three-hour indices a few times an hour, daily text products a few times a day. Use conditional requests and honour caching headers so unchanged files are not re-transferred. Identify yourself with a descriptive user agent that names your organisation and gives a contact address, so that if your client misbehaves someone can tell you rather than simply blocking the address range. If you need long time series, take them from the archival services rather than by scraping the real-time endpoints repeatedly, which is both slower and inconsiderate. A collector that pulls a dozen products on sensible schedules is invisible; one that hammers the same file continuously is a problem for a public service that other people's safety depends on.

Licensing changes, and it changes without warning. A dataset that was free for research this year may not be free for commercial or evidential use next year. Confirm the current terms before you build a dependency on it, and record the terms you relied on alongside the data — the licence in force at the time of collection is part of the provenance.

Collecting it

How NOAA Space Weather Prediction Center (SWPC) is actually pulled, in the order you would set it up. Prefer the bulk or export interface over per-item lookups wherever one exists: it is kinder to the publisher, faster for you, and gives a reproducible snapshot rather than a series of point-in-time answers you cannot reconstruct later.

Method	Format	Cadence	Notes
Real-time solar wind pull	JSON	every one to five minutes	Magnetic field and plasma products from the L1 monitor. The core signal for storm onset. Store the raw time series rather than derived summaries, because the shape of the arrival matters.
Geomagnetic index pull	JSON	every few minutes for the estimate, three-hourly for the index	Planetary K index and its real-time estimate. Record which variant you took, because the two are not interchangeable and the definitive value arrives later still.
Solar X-ray and particle flux	JSON	every one to five minutes	GOES X-ray irradiance for flare classification and proton flux for radiation storms. These are the products that explain HF blackouts and polar route impacts.
Alerts, watches and warnings stream	JSON	poll every few minutes	The operational product. Each message has an issue time and a condition, and the pair is what supports lead time analysis after an event.
Forecast text products	HTML	daily	Human-authored three-day forecast, discussion and longer outlook. The forecaster's reasoning is genuinely informative and is not recoverable from the numeric products.
Archival retrieval	bulk	one-off per study	For any analysis spanning more than a few days of history, take the data from the national geophysical archives rather than from the rolling real-time files.

Ingesting it into the platform

Every step below is idempotent and cursor-based: interrupt one and it resumes from where it stopped rather than duplicating rows or losing progress. Collection is recorded per source, so a feed that quietly stops publishing shows up as a stale timestamp instead of silently thinning your coverage.

Register the product families separately — Record each product family in sources.php as its own feed with its real cadence, because a single SWPC source entry hides the fact that solar wind, indices, flux and alerts have completely different freshness and failure behaviours.
Schedule per-product polling — Configure collect.php with one job per product at its native cadence and let cron.php own the schedule, so that a product that stops updating shows up as a stalled job rather than as a flat line an analyst mistakes for quiet conditions.
Validate ranges on ingest — During ingest.php, apply physical plausibility bounds to every numeric field and quarantine values outside them. Instrument anomalies propagate into these files, and a spurious density spike will drive a false storm detection all the way to an alert.
Preserve the observed, estimated and predicted distinction — Store the qualifier as a first-class column rather than flattening a mixed series into one time line. Almost every serious misreading of space weather data comes from treating a forecast segment as observation.
Derive the scale values rather than trusting a cached one — Compute G, S and R levels from the underlying flux and index values at analysis time, and keep the published scale value alongside as a cross-check. Where they disagree, the disagreement is itself informative about revision.
Correlate against operational events — Use correlate.php to place space weather conditions on the same axis as RF, GNSS, satellite and infrastructure events already in the platform, so that the natural explanation is tested before any interference or attribution hypothesis is written.
Maintain a standing background layer — Push the indices and scale values into timeline.php as a continuous background series available to every case, so an analyst working an interference report sees the space weather context without having to think to look for it.
Alert on thresholds you chose — Configure alert rules in alerts.php on the conditions that actually matter to your mission – a storm level, a flare class, a proton event – rather than relaying every SWPC message, and record the issue time so lead time can be reconstructed later.

Registered sources and their last-collected state are listed in sources.php, and the scheduled chain that keeps them current is in automation.php.

How it is wrong, and how to tell

Every dataset is wrong in characteristic ways. Knowing which ways is the difference between using a source and being used by one, and it is the part of source evaluation most often skipped because it is the part that takes work.

This is among the most trustworthy sources in the catalogue, and the reasons are structural rather than reputational. The observations come from calibrated instruments on government spacecraft and from an international network of ground magnetometers with decades of continuity. The forecasts are produced by staffed forecasters against documented methods, and the centre operates under an operational mandate with real customers who would notice systematic error. What requires care is not accuracy but interpretation. Real-time indices are explicitly provisional and get revised, and an analysis that cites a real-time estimate as if it were the definitive index is overstating its own precision. Forecast products carry genuine and often large uncertainty, particularly for coronal mass ejection arrival timing, and the centre is honest about this in its discussion text even where the numeric products cannot express it. Model outputs such as the aurora forecast are model outputs, driven by upstream measurement, and should never be described as observations. The right posture is to trust the measurements, respect the forecasts as forecasts, read the human discussion text when the situation is complex, and always state which variant of an index you used and when you retrieved it.

Characteristic false positives

Real-time index estimates are cited as definitive values. The provisional planetary index can differ from the final one published later by the international custodian, so an event classified from real-time data may be reclassified afterwards and your report will not have moved.
Forecast values are read as observations. Several products mix observed, estimated and predicted segments in one series, and an analyst plotting the whole thing without the qualifier produces a chart in which the future is indistinguishable from the past.
Instrument anomalies propagate as physically impossible values – density spikes, field reversals, flux discontinuities – that look exactly like a dramatic event and will drive a false alert unless range validation is applied on ingest.
Upstream measurement times are treated as Earth impact times. Solar wind observed at L1 arrives at Earth roughly thirty to sixty minutes later, and correlating an Earth-side event against the raw upstream timestamp misaligns the whole analysis.
Global indices are applied to local conditions. A planetary index of five is compatible with severe disturbance at one high-latitude site and negligible effect at another, so drawing a local impact conclusion from a planetary number is a category error.
Coronal mass ejection arrival predictions are quoted as times rather than windows, and the uncertainty is frequently many hours, which turns a defensible forecast into an indefensible claim when it is written down without the range.
Aurora and other model products are described as observed conditions. They are physics models driven by upstream input, they are forecasts for the coming half hour or so, and their skill varies with the situation.
Space weather is invoked as a universal explanation. Not every HF failure or GNSS anomaly during a storm is caused by it, and the presence of elevated activity is a hypothesis to test rather than a conclusion, particularly where deliberate interference is the alternative.

None of these make the source unusable. They make it a source that requires corroboration before an assertion built on it goes into a product, which is true of every source and admitted by few.

Ageing

Ageing here is measured in minutes for the real-time products and is the whole point of them: a solar wind value is operationally meaningful for the transit time to Earth and then becomes history. The K index has a three-hour cadence and a real-time estimate that exists precisely because three hours is too long to wait, and both are superseded by a definitive value published later – so the same interval has three different numbers attached to it depending on when you asked. Forecast text ages on its issue cycle and is explicitly superseded by the next issuance; citing a three-day forecast without its issue time is meaningless. Alerts have a defined condition period and expire. The subtler ageing problem is your own derived conclusions: an assessment that a communications failure was solar in origin, written from real-time indices, should be revisited when the definitive index is published, because the revision occasionally changes the story. A stale record here looks like a chart with a flat line at the end, which is almost always a stalled collector rather than quiet space weather, and the only reliable defence is to monitor the age of the newest record in each product rather than the values themselves.

What this source feeds

A source is only worth what it lets you conclude. These are the disciplines that collect through it, the mission domains it serves and the data points it yields — every one is a tag, so you can follow any thread from here into the rest of the library.

Collected by these intelligence disciplines

Meteorological Intelligence

Serves these mission domains

Yields these data points

How each sector uses NOAA Space Weather Prediction Center (SWPC)

The same dataset is worked very differently depending on who you are, what authority you hold, and what you are ultimately producing. A military analyst is supporting a commander’s decision; a journalist is meeting a publication standard; an NGO caseworker is protecting a person. The records are shared — the constraints, thresholds and outputs are not.

🎖 Military and defence

This is baseline environmental intelligence for any force that depends on HF, satellite communications, GNSS or over-the-horizon radar, and it is the first check before any electronic warfare or interference assessment. A radio blackout following an X-class flare degrades HF across the entire sunlit hemisphere simultaneously; a solar radiation storm degrades polar HF and drives route changes for high-latitude flights; a geomagnetic storm produces GNSS scintillation and position errors that look like spoofing to anyone who did not check. Running these products as a standing layer means the natural explanation is excluded before an adversarial one is asserted, which is the difference between an assessment and an embarrassment. It also supports planning: knowing when propagation conditions will favour or degrade a given band, and knowing that satellite drag conditions are elevated and orbital predictions will drift faster than usual.

🕵 National intelligence

For METOCINT and for the RF domain generally, this is the environmental substrate under a large fraction of technical reporting. Its most valuable analytical function is negative: it lets you rule out solar causes for observed degradation, which is what makes an interference or jamming assessment defensible. It also supports counterintelligence-adjacent work, because an adversary conducting interference during a genuine geomagnetic storm gains natural cover and the only way to see through that is to have the environmental baseline. For collection management, elevated activity changes what your own sensors can do, and building the index series into the same timeline as your collection lets you distinguish a gap caused by the environment from a gap caused by the target. Note that only the platform's Summarise function involves a language model, and it writes about records that already exist – every index and flux value here is measured, not generated.

👮 Law enforcement

The law enforcement uses are narrow and specific but they arise more often than expected. Cases involving GNSS-derived evidence – vehicle tracking, vessel position, device location – occasionally turn on whether the position could have been in error, and geomagnetic conditions at the time are a documented, citable environmental factor. Cases involving communications failures, alarm system faults or electrical infrastructure anomalies during a storm period benefit from the same check. In maritime enforcement, HF propagation conditions bear on whether a vessel's claimed inability to communicate is plausible. In all of these the data is corroborative context obtained from an authoritative government source, which makes it unusually easy to introduce and unusually hard to dispute.

🔍 Private investigation and corporate security

Corporate security and private investigators encounter this mostly as an explanation for anomalies that would otherwise look suspicious: a site's satellite link degrading, a fleet's tracking data going erratic, an alarm or telemetry system reporting faults across many locations at once. Checking space weather first is cheap and occasionally saves a client from paying for an intrusion investigation into a solar storm. On the other side, physical security planning for critical infrastructure clients has a genuine space weather component, because geomagnetically induced currents affect grids and pipelines and the severe end of the scale is a documented business continuity risk rather than a theoretical one.

📰 Journalism and OSINT media

For science and infrastructure reporting this is a primary source that can be cited directly, and its scales give readers a framework that flux values do not. The reporting discipline is to use the right vocabulary – flare class, storm level, radiation storm, radio blackout are distinct things with distinct causes and distinct impacts, and they are constantly conflated in coverage. Cite the issue time of any forecast, describe arrival predictions as windows rather than times, and be aware that the real-time index you quoted may be revised. The forecast discussion text is a genuinely good source for the reasoning behind a forecast and is written by working forecasters rather than by a press office, which makes it far more useful than a summary.

🌍 NGO, humanitarian and human rights

Humanitarian and development organisations operating in remote areas depend on HF radio and satellite communications, and space weather is the environmental factor that takes both away simultaneously and without warning. The practical use is operational: monitoring the alerts stream so that a field team's loss of communications is recognised as an expected environmental condition rather than treated as a security incident, and planning around predicted radiation storms for high-latitude flights. For organisations doing early warning work, the severe end of the geomagnetic scale is a genuine infrastructure hazard affecting power and communications in the countries they work in, and it belongs in a risk register alongside more familiar hazards.

🎓 University and research

Space physics, ionospheric research, power systems engineering and satellite operations research all use these products, and the real-time service is only the operational face of a much larger data ecosystem. For research purposes the essential discipline is to use the definitive, archived indices from their custodial institutions rather than the real-time estimates, and to take long time series from the geophysical archives rather than scraping rolling operational files. The service is also a good object of study in its own right for work on warning systems, operational forecasting skill and the sociology of infrastructure risk, because its products, thresholds and customer relationships are unusually well documented for an operational forecasting centre.

Playbook: working NOAA Space Weather Prediction Center (SWPC) end to end

A repeatable sequence from first pull to finished product. Each phase states what you are trying to establish, not merely what to click — the objective is a defensible chain of reasoning, not a completed checklist.

Phase 1 — Establish the baseline before you need it

Stand up continuous collection of the index, flux and alert products before any specific investigation, so that when an anomaly arrives you already have the environmental context rather than reconstructing it under time pressure. This source is nearly worthless as a reactive lookup and extremely valuable as a standing layer.

Phase 2 — Fix the time base explicitly

Everything here is UTC, and the one non-obvious conversion is that solar wind measured at L1 reaches Earth roughly thirty to sixty minutes later depending on speed. Decide whether your timeline is in observation time or impact time, apply the transit offset consistently, and label it. Half the failed correlations in this domain are this mistake.

Phase 3 — Classify the event type before assessing impact

A flare, a solar radiation storm and a geomagnetic storm are different physical events with different onset times, different durations and different affected systems. X-rays arrive in minutes and hit the sunlit side; energetic particles arrive in tens of minutes to hours and hit polar regions; ejecta arrive in a day or more and drive geomagnetic storming. Getting the event class right determines everything downstream.

Phase 4 — Separate observation from forecast in your own store

Split the mixed series on the observed, estimated and predicted qualifier and never plot them as one line without visual distinction. Reviewers will find this if you do not, and the finding will be that your evidence included the future.

Phase 5 — Take the local view seriously

Planetary indices are global averages. If your question concerns a specific location – a receiver, a substation, a ground station – the planetary value is only a first indication, and you need local magnetometer or ionospheric data to say anything about conditions there. State the limitation explicitly rather than letting the global number stand in for a local claim.

Phase 6 — Run the deconfliction test on any interference report

Before writing that a GNSS or HF anomaly was deliberate, check the flare, radiation storm and geomagnetic conditions in the window, applying the correct transit offsets. If elevated conditions coincide, the natural explanation must be positively excluded on grounds beyond timing – geographic pattern, frequency selectivity, affected system mix – and the exclusion reasoning must appear in the report.

Phase 7 — Reconstruct the warning timeline for after-action work

Pull the alerts stream with issue times and condition times and build the sequence of what was known and when. This is the analysis that answers whether an impact was foreseeable and whether the lead time was actionable, and it is far more useful to an operational customer than the physics.

Phase 8 — Read the forecast discussion when the situation is complex

The human-authored text carries reasoning, confidence and alternative scenarios that the numeric products cannot express. When several solar sources are in play or an arrival time is uncertain, the discussion is where the forecaster tells you what they are unsure about, and it is the single most underused product in the set.

Phase 9 — Extend the analysis to satellite and orbital effects

Elevated geomagnetic activity heats and expands the upper atmosphere, increasing drag on low-orbit objects. That accelerates decay, degrades orbital element set accuracy, and explains prediction drift and unexpected re-entry timing. If your work touches the space catalogue, the two sources belong in the same timeline.

Phase 10 — Revisit conclusions when definitive indices publish

Real-time index values are provisional. Once the definitive planetary index and ring current index are published by their custodial institutions, re-check any classification your assessment rested on. Most of the time nothing changes; when it does, you want to be the one who found it.

Phase 11 — Translate into the customer's language

Operational decision makers act on the storm scales and the impact statements, not on flux values in scientific notation. Report at the scale level with a plain statement of the affected systems, keep the numbers in an annex, and be explicit that the impact statements are written for one national infrastructure context and may need adjustment for yours.

Phase 12 — Archive what you used

Record the exact product files, their retrieval times and the index variants your conclusion rested on. The live endpoints roll off within days, the values get revised, and an assessment that cannot be reconstructed six months later cannot be defended when someone challenges it.

The platform ships this as a step-checked workflow in playbooks.php, so progress is recorded against a case rather than held in someone’s head.

What to pair it with

No single source carries a finding. These are the datasets that corroborate, extend or contradict this one — and a source that contradicts is worth more than one that agrees, because it is the only thing that will tell you when you are wrong.

Source	Relationship	What it adds
GFZ geomagnetic index service	supersedes	The custodial producer of the definitive planetary K index. Where a classification matters, the definitive value supersedes the real-time estimate you acted on.
World Data Center for Geomagnetism, Kyoto	extends	Authoritative ring current index used to characterise geomagnetic storm intensity, with provisional and final versions on different schedules.
NASA Community Coordinated Modeling Center	extends	Model runs, arrival time predictions and the integrated space weather analysis system, useful for understanding the modelling behind operational forecasts.
ESA Space Weather Service Network	corroborates	An independent European operational service with its own products and impact framing, valuable precisely because it is not built around United States infrastructure assumptions.
NOAA National Centers for Environmental Information	extends	The archival home for long geophysical time series, which is where multi-year studies should draw data from rather than from rolling operational files.
CelesTrak	corroborates	Orbital element sets whose accuracy degrades measurably during geomagnetic storms through increased atmospheric drag, making the two sources mutually explanatory.
NOAA NESDIS	prerequisite	The satellite operator behind the GOES instruments supplying X-ray and particle measurements, and the reference for instrument status when values look wrong.
International Telecommunication Union	extends	Radio propagation recommendations and the regulatory framework for the HF services most directly affected by ionospheric disturbance.

Legal, ethical and operational constraints

There are essentially no legal restrictions on obtaining or using this data – it is public domain US government output published for exactly this purpose, and no jurisdiction restricts knowing what the sun is doing. The legal questions attach to what you build on it. Where space weather data is used to support a claim about deliberate interference with radio or satellite services, that claim engages telecommunications regulation and potentially international attribution processes, and it needs to be right. Where it is used in a safety-critical operational decision – flight routing, grid operation, offshore work – the constraint is professional and regulatory rather than legal in the abstract: you are supplying environmental data into a decision chain with consequences, and misrepresenting a forecast as an observation, or an estimate as a definitive value, carries real liability. Where you republish or resell derived products, do not present them in a way that implies official status or endorsement. Finally, if your organisation relies on these products for operational decisions, understand that a foreign government service under appropriation-based funding is a dependency that could be interrupted, and that is a continuity question your risk register should carry.

Operational security

Fetching public files from a national weather service is about as unremarkable as network traffic gets, and the traffic itself discloses almost nothing about you. The exposure is in the pattern rather than the content: a client that suddenly begins polling flare and index products at high frequency, or that starts pulling archival data for a specific past window, is a weak but real signal of an operational concern, particularly to anyone monitoring your outbound traffic rather than the service's logs. The mitigation is trivial and should be default – collect continuously on a fixed schedule regardless of whether anything is happening, so there is no cadence change to observe. More significant is the reverse exposure: if you publish assessments that cite specific space weather windows to explain or exclude interference at a named location, you are disclosing that you were monitoring that location's RF environment at that time, which may be the more sensitive fact. Consider the geographic specificity of what you publish separately from the source, which is entirely benign.

Two rules that hold regardless of jurisdiction. Collection that is lawful is not automatically proportionate, and a dataset assembled for one purpose does not carry consent for another. Where the records concern identifiable people, the question is not only whether you may hold the data but whether holding it serves the purpose you are accountable for.

Is it earning its place?

Sources accumulate. Feeds get added during an incident and are never reviewed again, and a decade later the pipeline is carrying dead weight that nobody dares remove. These are the measures that show whether NOAA Space Weather Prediction Center (SWPC) is contributing anything, and they are worth baselining now so the answer is available later.

Freshness of the newest record in each collected product, monitored continuously, because a stalled collector renders as calm conditions and calm conditions are the most dangerous false negative this source can produce.
Number of interference or anomaly investigations in which the space weather check changed the working hypothesis, which is the clearest evidence that the standing layer is earning its place.
Proportion of your alert-triggering events whose classification survived comparison against the definitive indices published later, as a measure of how much your real-time thresholds overcall.
Warning lead time achieved on events that mattered to your operation, computed from alert issue time to observed impact, tracked to see whether the source gives you actionable notice or merely explanation.
Rate of quarantined out-of-range values on ingest, which is a direct read on upstream instrument health and a leading indicator that a product is about to become unreliable.
Share of published assessments touching RF, GNSS or satellite performance that cite the environmental condition explicitly, as a process measure of whether analysts are actually consulting it.
Count of cases where an operational customer acted on a scale-level statement you provided, which is the only measure that tests whether your translation into decision language works.

Beware of volume. Indicator counts rise easily and say almost nothing. Unique contribution — findings this source produced that no other source in your stack would have — is the measure that matters, and it is usually far lower than anyone expects.

Tradecraft notes

The distinctions that separate a competent analyst from a fast one:

Know which of the three storm types you are looking at before you say anything. Radio blackouts, radiation storms and geomagnetic storms have different causes, different onset delays, different geographies and different affected systems, and treating them as one phenomenon called a solar storm produces confident nonsense.
Apply the transit offset. Solar wind measured upstream reaches Earth later, and the offset varies with speed. An analyst who correlates Earth events against raw upstream timestamps will find no correlation and conclude wrongly that space weather was not involved.
Say which index variant you used. The real-time estimate, the operational index and the definitive index are three different numbers for the same interval, and a report that does not distinguish them cannot be checked.
Global indices do not describe local conditions. If the question is about a specific site, the planetary number is a screening value only, and pretending otherwise is the most common overreach in applied space weather analysis.
Read the forecaster's discussion. It contains the uncertainty, the alternative scenarios and the reasoning, none of which survives into the numeric products, and it is written by people whose job is to be right about this.
Range-validate everything on ingest. Instruments fail in ways that produce dramatic, physically impossible values, and an unvalidated pipeline will turn a sensor fault into a storm alert and then into a report.
Exclude, do not assume. Coinciding space weather is a hypothesis for an observed anomaly, not a conclusion; and equally, elevated activity does not preclude deliberate interference. State how you distinguished the two.
Forecast arrival times are windows. Quote the range, and if the product does not give one, say that the uncertainty is typically many hours rather than implying precision the physics does not support.
Carry space weather into orbital work. Elevated geomagnetic activity increases drag, degrades element set accuracy and shifts re-entry predictions, so an unexplained tracking drift often has its explanation in this dataset rather than in the catalogue.

Questions analysts actually ask

How much warning does this give me before a geomagnetic storm?

It depends on the driver. A coronal mass ejection observed leaving the sun gives roughly one to three days of notice with an arrival window uncertain by many hours. Once the disturbance reaches the upstream monitor at L1, you have roughly thirty to sixty minutes of high-confidence warning. A flare's radio blackout gives no warning at all, because the X-rays arrive at the speed of light.

Which index should I use, and why do I see different values?

For real-time operations use the estimated planetary index, understanding it is provisional. For classification and any published analysis use the definitive index from its custodial institution, which is produced later and can differ. Quoting one and calling it the other is the standard error, and it is easy to avoid by always naming the variant and the retrieval time.

Can space weather explain a GNSS outage?

It can explain degradation – increased position error, loss of lock, scintillation at low and high latitudes – particularly during geomagnetic storms and radiation storms. It rarely explains a clean, geographically bounded, single-constellation outage, which looks much more like interference. The value of checking is that it lets you exclude the natural cause on evidence rather than by assertion.

How far back does the data go?

The live product files hold a short rolling window, typically hours to a few days. Long historical series live in the national geophysical archives and in the custodial index institutions, and that is where any study spanning more than a few days should get its data. Building a multi-year dataset by repeatedly scraping the real-time endpoints will produce gaps you will not notice until you analyse them.

Do I need an API key?

No. The service is open files over HTTPS with no registration and no key. That does not mean unlimited polling is acceptable – match your request rate to each product's actual update cadence, use conditional requests, and identify your client with a contactable user agent.

Why did a satellite's orbit prediction go wrong during a storm?

Geomagnetic storms heat and expand the upper atmosphere, which increases drag on low-orbit objects. Orbital element sets fitted before the storm no longer describe the object's motion, decay accelerates, and predicted positions drift. This is a documented operational effect, and it is why space weather and the satellite catalogue belong on the same timeline.

Are the aurora maps observations?

No. They are model output driven by upstream solar wind measurements, forecasting the auroral oval for roughly the next half hour. They are useful and they are not pictures of the aurora, and describing them as observed conditions in a report is a straightforward factual error.

Are the impact statements applicable outside the United States?

The physics is global, the impact language is not. The scale descriptions are written for United States infrastructure at mid-latitudes, and effects at high latitudes are typically more severe at the same index value while equatorial ionospheric effects follow a different pattern altogether. Use the scale as a common vocabulary and adjust the impact assessment for your own geography.

Is the data reliable enough to cite in a report that will be challenged?

Yes, with two conditions. Cite the specific product and retrieval time, and use definitive rather than provisional index values where the classification matters. This is an operational government forecasting centre with calibrated instruments and named methods, which makes it one of the easier sources in this catalogue to defend.

Standards, formats and interoperability

What this source speaks natively, and what it has to be translated into before a partner can consume it. Work that arrives in a recognised format is easier to defend, easier to hand over and easier to automate against:

The NOAA space weather scales – G for geomagnetic storms, S for solar radiation storms, R for radio blackouts, each from one to five – are the common operational vocabulary and map directly to underlying physical thresholds.
Solar flare classification by peak soft X-ray irradiance, in the lettered classes, is the definition from which the radio blackout scale is derived.
The planetary K index is an internationally standardised geomagnetic activity measure derived from a defined network of observatories, with a custodial institution responsible for the definitive values.
The ring current index characterising storm main phase intensity is produced by a separate international custodian on provisional and final schedules.
Geocentric solar magnetospheric and geocentric solar ecliptic coordinate frames define what the magnetic field components mean; the north-south component's sign convention depends on the frame and misreading it inverts the analysis.
ITU radio propagation recommendations describe ionospheric effects on HF and satellite links and are the bridge from these geophysical products to communications engineering.
ICAO space weather advisory arrangements define how this class of information reaches aviation operationally, which is the clearest example of the data being used in a safety-critical decision chain.
The platform exports derived environmental events and correlations in STIX 2.1, MISP, CSV, JSON and JSONL, so a space weather condition can be attached to a case as context alongside technical observables.

References

Primary documentation and authoritative references for this source. Publishers revise and retire material, so treat the retrieval date as part of the citation and re-check before relying on any of it in a formal product.

NOAA Space Weather Prediction Center — NOAA National Weather Service. The centre itself: current conditions, forecast products, alerts and the explanatory material behind every number in the data service.
SWPC data services — NOAA National Weather Service. The machine-readable product host. Browse the directories directly to establish what exists and at what cadence rather than trusting a hardcoded path list.
NOAA space weather scales explanation — NOAA National Weather Service. The definitions and impact descriptions behind the G, S and R scales. Read this before translating any number into an operational statement.
GFZ Helmholtz Centre for Geosciences — GFZ. Custodian of the definitive planetary geomagnetic index. The authority when a real-time estimate and a published classification disagree.
World Data Center for Geomagnetism, Kyoto — Kyoto University. Ring current index production and archive, the standard measure of geomagnetic storm intensity in the research literature.
NASA Community Coordinated Modeling Center — NASA Goddard Space Flight Center. Model runs and integrated analysis tools that show how operational arrival predictions are actually produced and how uncertain they are.
ESA Space Weather Service Network — European Space Agency. An independent operational service with different products and a non-US impact framing, useful as a cross-check and for European infrastructure questions.
NOAA National Centers for Environmental Information — NOAA. Archival geophysical data holdings – the correct source for long time series rather than the rolling real-time files.
NOAA NESDIS — NOAA. The satellite service operating the GOES instruments behind the X-ray and particle products, and the reference for instrument status and transitions.
International Telecommunication Union — ITU. Radio propagation recommendations and the regulatory context for the HF and satellite services these disturbances affect.

Link integrity: every reference above was verified with a live request when this page was generated. Where a publisher had moved or withdrawn a document, the link was repointed at a preserved copy in the Internet Archive and marked as archived. Anything with no reachable copy anywhere had its link removed rather than left to rot — the source is still credited, it simply cannot be linked.

Put it into practice

The Quantus Intel threat intelligence platform operationalises this source: it collects each SWPC product family on its own cadence, range-validates on ingest, keeps observed and forecast values distinct, and carries the geomagnetic and radio-blackout background as a standing layer under every RF, GNSS and satellite case so the natural explanation is tested before an adversarial one is written.. Browse the full source catalogue, or follow any tag above into the rest of the library.