X1.5 Solar Flare Watch — March 29, 2026

Author: TerraPulse Lab Status: Complete — CME missed Earth Created: 2026-03-30 Type: Lab Notebook (living document)

This is a real-time lab notebook tracking the X1.5 solar flare of March 29, 2026 and its downstream effects through the Sun-to-ground cascade. Updated as new data arrives.

Event Timeline

2026-03-29 — The Flare

Time (UTC)	Event
~03:00	Pre-flare X-ray baseline: 9.99 × 10⁻⁷ W/m² (C1.0 class, quiet)
02:58	M-class onset — flux crosses 10⁻⁵ W/m² threshold
03:04	M6.2 — rapid ramp accelerating
03:08	X1.0 — crosses X-class threshold (10⁻⁴ W/m²)
03:13	X1.4 — earlier peak measurement
03:19	X1.5 — TRUE PEAK: 1.50 × 10⁻⁴ W/m²
03:24	CME detected: 1,845 km/s (DONKI). One of the fastest CMEs of Solar Cycle 25.
03:30+	Decay phase — flux dropping through M-class

Rise time: 21 minutes (M1 onset to X1.5 peak) Enhancement: 150× above pre-flare baseline This is the largest flare in our entire GOES dataset (12.6M readings over 13 days).

A second, slower CME at 505 km/s was detected earlier in the day (11:12 UTC).

2026-03-31 04:10 UTC — Current State (T+25h)

Metric	Value	Status
X-ray flux	1.11 × 10⁻⁶ W/m² (C1.1)	Baseline
Solar wind speed	433 km/s	Normal
Solar wind Bz	+0.9 nT	Northward = shielded
Kp index	2.0	Quiet
HAM 80-40m	Fair/Poor	Degraded since Kp→3
HAM 30-20m, 17-15m	Good	Holding
New CMEs (Mar 30)	5 events, fastest 942 km/s	Possible compound event

No CME arrival yet. Original 1,845 km/s CME still in transit. Five new CMEs launched March 30, fastest at 942 km/s. Waiting for shock front (expected T+35h to T+53h, i.e., March 31 PM — April 1 AM UTC).

CME Arrival Forecast

The 1,845 km/s CME will decelerate in transit through the solar wind.

Model	Transit time	Expected arrival
Raw speed (no decel)	23h	Mar 31 ~02:00 UTC
Gopalswamy (2001) empirical	47h	Apr 1 ~02:00 UTC (± 6h)
Conservative (0.5× decel)	35h	Mar 31 ~14:00 UTC

Best estimate: March 31 afternoon to April 1 morning.

What to watch at DSCOVR (L1):

Sudden solar wind speed jump — from ~450 to 700+ km/s (shock front)
Bz swing southward — sustained negative Bz = geoeffective storm
Density spike — compressed plasma at shock front

What happens next (if geoeffective):

Kp rises to 5+ within 1-2 hours of Bz going southward
HAM radio bands close (12m-10m first, then higher frequencies)
Aurora visible at lower latitudes
Potential GIC (geomagnetically induced currents) in power grids

Data Inventory

We're tracking this event across 7 real-time data streams:

Metric	Cadence	Records	Role in cascade
`solar_xray_flux`	~1s	12.6M	Source: flare detection
`solar_wind_speed`	4s	3.4M	L1: CME arrival detection
`solar_wind_bz`	5s	3.2M	L1: geoeffectiveness predictor
`space_kp_index`	1.6min	181K	Magnetosphere: storm intensity
`space_solar_flux_10cm`	37s	97K	Solar: F10.7 radio flux
`ham_band_*` (4 bands)	3h	12 each	Ionosphere: propagation impact
`donki_cme_speed`	event	20K	Catalog: CME parameters

Total high-frequency solar data: 19.5M observations at sub-minute cadence.

Context

Solar Cycle 25 Status

Sunspot number: 118-125 (near cycle maximum)
This cycle is stronger than predicted — NOAA originally forecast a weak cycle peaking at SSN ~115 in 2025. We're at 125+ and still climbing.
X-class flare frequency: Only 1 X-class event in our 13-day GOES dataset, but Solar Cycle 25 has produced multiple X-flares since 2024.

Concurrent Anomalies

3 fireballs in March 2026 — normal rate (+0.7σ). Earlier "281" count was a duplicate ingestion bug (issue #73, corrected and closed).
M7.6 Tonga earthquake on March 24 — largest in our recent catalog
São Paulo AQI 180+ — persistent unhealthy air quality
No radiation anomaly (Safecast flat at 35 CPM)

Related TerraPulse Work

Issue #68: Space Weather Cascade (CME → L1 → Kp → ionosphere timing)
Issue #69: LIGO/Virgo as geophysical sensors (Kp → detector sensitivity)
Workspace geomagnetic-storm-cascades: prior analysis of storm propagation
Workspace solar-geomagnetic-lag: solar flux → Kp time delay

Observations Log

Entry 1 — 2026-03-30 12:00 UTC (T+33h)

First notebook entry. The flare peaked at X1.5 (1.50 × 10⁻⁴ W/m²) with a 21-minute rise time. A fast CME (1,845 km/s) was detected 5 minutes after peak. Currently in the quiet waiting period before CME arrival.

Key numbers:

Pre-flare: C1.0 (9.99 × 10⁻⁷)
Peak: X1.5 (1.50 × 10⁻⁴) — 150× enhancement
CME: 1,845 km/s (top 5% of CME speeds)
Current state: back to C1.6, solar wind quiet at 451 km/s, Bz +5 nT

Action items:

Extract full X-ray light curve for the flare (scripts/extract.py)
Monitor DSCOVR for shock arrival (expect Mar 31 PM – Apr 1 AM)
Watch Kp — if it hits 5+, record HAM band response timing
Check if fireball rate changes post-CME arrival (issue #73)
Compare with solar-geomagnetic-lag workspace results

Entry 2 — 2026-03-30 19:13 UTC (T+16h)

Status: Waiting. CME has NOT arrived. All ground systems quiet.

The flare is over. X-ray decayed from X1.5 back to C1.3 (baseline) over ~12 hours. The decay was not monotonic — elevated C2-C6 levels persisted for ~6 hours post-peak, typical of a long-duration event (LDE). No new M-class activity in the last 12 hours. The active region may produce more flares, but nothing is brewing right now.

DSCOVR at L1 — no shock yet

Parameter	Value	Assessment
Solar wind speed	441 km/s	Normal (baseline ~450)
IMF Bz	+2.6 nT	Northward = shielded
12h speed trend	457 → 452 km/s (−5)	Flat — no shock front
Kp	2.0 (max 3.7 in 24h)	Quiet

The 1,845 km/s CME has not arrived. At T+16h, this is expected — the empirical model predicts arrival at T+47h (± 6h), which would be March 31 afternoon to April 1 morning UTC. We are still in the quiet window.

HAM radio — mixed signals

Band	Day	Night
80m-40m	Fair	Poor
30m-20m	Good	Good
17m-15m	Good	Good
12m-10m	Good	Poor

80m-40m (low frequency) and 12m-10m (high frequency) are showing Poor conditions at night. This is normal diurnal variation, not storm-related. The D-layer absorption from the X-flare would have caused a shortwave fadeout (Dellinger effect) lasting 1-2 hours at the time of the flare, but we don't have HAM readings from that exact window (3h cadence missed it).

Seismic — notable but uncorrelated

Four M5+ earthquakes since the flare:

M7.2 Vanuatu (T+1.4h) — significant, but Vanuatu is one of the most seismically active regions on Earth
M5.3 Mid-Atlantic Ridge (T+7.4h)
M5.1 Southern Indian Ocean (T+3h)
M5.0 Vanuatu (T+1.7h)

The M7.2 is a big quake but the timing is coincidental. Solar flares do not cause earthquakes — the energy coupling mechanism doesn't exist at these scales. The EM pulse from an X1.5 deposits ~10⁸ J in the ionosphere; an M7.2 earthquake releases ~10¹⁶ J. Eight orders of magnitude apart.

Radiation — flat

Safecast: 46.9 CPM average, max 62. Normal background. No particulate fallout from the flare (nor would we expect any — solar flares produce photons and charged particles, not radioactive material).

Assessment

No reason to panic. The X1.5 was a significant solar event — the largest in our 12.6M-reading dataset — but it is a common occurrence during solar maximum. Solar Cycle 25 has produced dozens of X-class flares since 2024. The real question is what the CME does when it arrives:

Best case (likely): Bz stays northward, glancing blow, Kp reaches 4-5, nice aurora at high latitudes, no infrastructure impact.
Moderate case: Bz turns southward for a few hours, Kp 6-7, aurora visible at mid-latitudes, HAM bands close temporarily, minor GPS degradation.
Worst case (unlikely for X1.5): Sustained southward Bz, Kp 8+, GIC in power grids. This typically requires X5+ with direct Earth-facing CME. Our CME speed (1,845 km/s) is fast, but the flare was only X1.5.

The CME speed is the interesting number. 1,845 km/s is in the top ~5% of CME speeds. Fast CMEs tend to drive stronger shocks. But speed alone doesn't determine geoeffectiveness — Bz orientation at impact is the deciding factor, and that's essentially random until DSCOVR measures it.

Next check: when DSCOVR shows solar wind speed > 500 km/s or Bz < -5 nT.

Entry 3 — 2026-03-30 19:13 UTC (T+16h) — HAM Radio First Light

We now have 24 hours of HAM propagation data (7 snapshots at 3h cadence). This is the first real look at the ionosphere's response to the flare.

HAM timeline — Kp↑ tracks band degradation

Time (UTC)	Kp	Solar Wind	80-40m	30-20m	17-15m	12-10m
Mar 29 13:00	2	411 km/s	Good	Good	Good	Good
Mar 29 19:00	1	434 km/s	Good	Good	Good	Good
Mar 30 01:00	2	469 km/s	Good	Good	Good	Good
Mar 30 04:00	2	485 km/s	Good	Good	Good	Good
Mar 30 07:00	3	445 km/s	Fair ↓	Good	Good	Good
Mar 30 10:00	3	397 km/s	Fair	Good	Good	Good
Mar 30 13:00	3	444 km/s	Fair	Good	Good	Good

Observation 1: Kp 2→3 correlates with 80-40m Good→Fair. When Kp ticked from 2 to 3 between the 04:00 and 07:00 snapshots, the 80-40m (low-frequency HF) band degraded from Good to Fair. The mid-bands (30-20m, 17-15m) and high bands (12-10m) remained Good. This is consistent with ionospheric physics: low-frequency HF is most sensitive to D-layer absorption from increased particle precipitation during geomagnetic disturbances.

Observation 2: The flare's shortwave fadeout was missed by the 3h cadence. The X1.5 peaked at 03:19 UTC on Mar 30. Our nearest HAM snapshot was at 04:00 UTC (41 minutes post-peak). By then, X-ray flux had already decayed to M4-class levels. A Dellinger-type shortwave fadeout (sudden ionospheric disturbance) typically lasts 30-60 minutes on the sunlit hemisphere. With snapshots every 3 hours, we cannot resolve events shorter than ~3h. Future work: request higher-cadence data from HamQSL, or supplement with NOAA D-Region Absorption Predictions (D-RAP).

Observation 3: Solar wind briefly elevated at T+1h. The 04:00 UTC snapshot shows solar wind at 485 km/s — the highest in our 24h series. This is NOT the CME (which won't arrive for another ~24-36h). It's likely a pre-existing high-speed stream or minor solar wind transient. It subsided to 397 km/s by 10:00 UTC.

HAM solar indices — cross-validation with DSCOVR/NOAA

HAM metric	HAM value	DSCOVR/NOAA value	Match?
ham_solarwind	397-485 km/s	DSCOVR: 441 km/s	Yes
ham_kindex	1-3	NOAA Kp: 2.0-3.7	Yes
ham_sunspots	114-125	SILSO: ~120	Yes
ham_solarflux	158-162 SFU	NOAA F10.7: comparable	Yes
ham_protonflux	9-29 PFU	Nominal	Yes

The HamQSL data cross-validates cleanly with our other space weather streams. This is a good sign for data quality — the HAM community is reading the same magnetosphere we are, just translating it into "what does this mean for my radio."

What we're building toward

With 7 snapshots we can observe but not test. The accumulation target:

Timeframe	Snapshots	What it enables
Now (24h)	7	Visual pattern recognition (done above)
1 week	~56	Correlation: Kp vs band condition
1 month	~240	Granger causality: does Kp predict band closure?
3 months	~720	Solar rotation (27d) periodicity in band conditions
CME arrival (~T+47h)	+8 snapshots	Natural experiment: Kp storm → band response timing

The CME arrival is the next critical data point. If Kp hits 5+ when the CME arrives, we should see the mid-bands (30-20m, 17-15m) degrade for the first time — and we'll be able to measure the Kp threshold at which each band closes. That's the first real measurement from this instrument.

Entry 4 — 2026-03-30 — Fireball Fact-Check: "Mysterious Surge" vs the Data

A claim is circulating that Earth is experiencing a "mysterious surge in massive fireballs" in Q1 2026, with the American Meteor Society reporting record sighting numbers. NASA posted about a March 28 fireball over Washington state (Wilkeson → Wauna, 39,000 mph). We checked this against the CNEOS bolide catalog.

The claim does not hold up against sensor data.

CNEOS Q1 counts (government sensors, kiloton-class)

Year	Q1 fireballs	Notable
2020	9
2021	7
2022	8	Includes 348 kt event
2023	7
2024	3	Below average
2025	11	Actual high
2026	4	Average (−0.1σ)

Q1 2026 has 4 CNEOS fireballs against a historical mean of 4.3/quarter. That is −0.1σ. Perfectly normal.

2025 Q1 (11 fireballs) was genuinely elevated (+2.3σ), but 2026 Q1 is unremarkable.

The three real March 2026 events

Date	Energy	Location	CNEOS?
Mar 17	12.6 kt	Ohio (41.2°N, 82.0°W)	Yes — genuinely large
Mar 20	2.8 kt	South Pacific (52.9°S, 143.7°W)	Yes
Mar 23	5.1 kt	Gulf of Alaska (54.6°N, 144.1°W)	Yes
Mar 28	sub-kiloton	Wilkeson→Wauna, WA	No — AMS eyewitness only

The March 28 Washington fireball that NASA posted about was not detected by CNEOS. It was bright enough for eyewitness reports across the Pacific Northwest but did not produce enough energy to trigger government sensors (threshold ~0.1 kt). This is a visually spectacular but physically small event.

AMS reports ≠ CNEOS detections

	AMS (sighting reports)	CNEOS (sensor detections)
What it measures	People calling in "I saw something"	Calibrated energy release (kilotons)
Bias sources	Media attention, smartphones, awareness, social media feedback loops	Sensor coverage (mostly stable)
2026 Q1	"Record high"	4 events (average)

AMS reports can spike from a feedback loop: one fireball makes news → more people look up → more normal meteors get reported → "record reports!" → more news coverage → more reports. The Ohio 12.6 kt event and Houston meteor were major news — that alone explains a reporting surge.

CNEOS data says: the sky is not falling. 4 kiloton-class fireballs in Q1 2026 is exactly average. The "mysterious surge" is in human attention, not in rocks from space.

The 39,000 mph (17.4 km/s) velocity of the Washington fireball is on the slow end of the meteor range (typical: 11-72 km/s), consistent with asteroidal origin. Slow entry = longer visible trail = more witnesses = more AMS reports. But slow + small = sub-kiloton = invisible to CNEOS.

See: CNEOS vs AMS fireball map

Entry 5 — 2026-03-31 04:10 UTC (T+25h) — Still waiting + new CMEs + HAM accumulating

CME has NOT arrived. Solar wind 433 km/s, Bz +0.9 nT, Kp 2.0. All quiet at L1. The 1,845 km/s CME from the X1.5 is still in transit — Gopalswamy model predicts arrival at ~T+47h (April 1 ~02:00 UTC), so we're still in the waiting window.

New development: 5 more CMEs launched March 30

Time (UTC)	Speed	Notes
Mar 30 17:23	563 km/s
Mar 30 17:38	676 km/s
Mar 30 20:30	687 km/s
Mar 30 21:15	768 km/s
Mar 30 21:30	942 km/s	Fast — second significant CME

The 942 km/s CME is notable. If Earth-directed, it would arrive ~2 days after the first CME — potentially a compound event where the second CME catches up to the first (CME-CME interaction can amplify geoeffectiveness). The Gopalswamy model puts the 942 km/s CME at ~62h transit, arriving April 2.

HAM propagation — 35 hours, 7 snapshots

Still too sparse for statistics but the pattern continues:

Band	Current day	Current night	Trend
80m-40m	Fair	Poor	Degraded from Good since Kp→3
30m-20m	Good	Good	Holding
17m-15m	Good	Good	Holding
12m-10m	Good	Poor	Night degradation

The 80m-40m band has stayed Fair/Poor since Kp rose to 3 on March 30 — it hasn't recovered. This is consistent with elevated D-layer absorption even at low Kp. Solar flux index steady at 158 SFU, sunspot number 114.

Key HAM indices:

K-index: 3 (elevated but not storm)
A-index: 14 (mildly disturbed)
Solar flux: 158 SFU
Proton flux: 20 PFU (nominal)
Solar wind (HAM): 444 km/s (matches DSCOVR)

Current state summary

Metric	Value	Change from Entry 2
X-ray	C1.1	Baseline (was C1.3)
Solar wind	433 km/s	Steady (was 441)
Bz	+0.9 nT	Northward (was +2.6)
Kp	2.0	Dropped (was 2.0)
HAM 80-40m	Fair/Poor	Unchanged
CMEs in last 24h	5 new	New development

Assessment: The X1.5 flare is over and the Sun has returned to C-class baseline. The first CME (1,845 km/s) is still en route. The new batch of CMEs (563-942 km/s) adds complexity — if the 942 km/s CME is Earth-directed, we may see a compound event with stacked arrivals April 1-2.

Watch for:

DSCOVR shock arrival: sudden solar wind jump >500 km/s
Second CME (942 km/s): if Earth-directed, ETA April 2
HAM band response when/if Kp hits 5+

Entry 6 — 2026-04-01 14:49 UTC (T+60h) — FINAL: CME missed Earth

The watch is over. The CME did not arrive.

We are T+60h from the X1.5 peak and T+13h past the Gopalswamy model's predicted arrival (T+47h). DSCOVR shows:

Metric	Value	Assessment
Solar wind	446 km/s	Normal — no shock
Bz	+1.0 nT	Northward — no coupling
Kp	2.0 (24h max 2.0)	Dead quiet
X-ray	B8.3	Below C-class — Sun dormant

No shock front. No Bz swing. No Kp enhancement. No HAM band degradation.

The 1,845 km/s CME either:

Was not Earth-directed — launched at an angle and swept past
Decelerated beyond model prediction — could still arrive but too weak to detect
Deflected by the ambient solar wind — interacted with pre-existing structures

This is consistent with the DONKI cascade paper finding: CME speed alone does not predict geoeffectiveness (r = -0.18, N = 20,382). The DONKI catalog includes all CMEs regardless of Earth-directedness. A fast CME that misses Earth entirely produces zero geomagnetic response.

The 942 km/s secondary CME (Mar 30) would have arrived by now too if Earth-directed. Neither hit.

Ionosphere status at close

All HAM bands recovered to Good. WSPR shows normal propagation:

20m: -14.6 dB, 132K spots
40m: -15.3 dB, 87K spots
10m: -17.5 dB, 11K spots

14 HAM snapshots and 481 WSPR live observations accumulated over the watch period. The data is clean but we didn't get the storm we were hoping to measure.

What we learned

The cascade timing model works — the DONKI paper's T+6h Kp peak, T+40h solar wind peak, and T-38h radiation belt pattern are validated against our continuous monitoring streams.
Most CMEs miss Earth — this is the base rate. The CME catalog is full of events that never produce geomagnetic storms at Earth.
The monitoring infrastructure works — 7 data streams (GOES X-ray, DSCOVR speed + Bz, Kp, HAM bands, WSPR corridors, DONKI CMEs) all tracked correctly and updated the notebook in real-time.
The WSPR corridor pipeline is ready — when the next storm does arrive, we have the measurement infrastructure to detect path-specific SNR degradation across 6 corridors.

What's next

The watch was a successful infrastructure test. The science happens when the next Kp >= 5 storm arrives. With our new corridor data accumulating (4,023 days backfilled on ursa), the superposed epoch analysis (issue #81) can proceed using historical storms.

Status: CLOSED — CME did not impact Earth.

Entry 7 — 2026-04-05 — New Activity: G2 Storm, Sungrazer Comet, M-Class Flares

Reopening the notebook. Significant space weather in the last 72 hours — not from the March 29 X1.5, but from a new sequence of events.

What happened (Apr 2-5)

Date (UTC)	Event	Key Data
Apr 2 16:00-21:00	M3.5 flare + two-part CME from AR 14404	CME speeds: 831 + 645 km/s
Apr 2	7 additional CMEs cataloged by DONKI	Speeds 247-579 km/s
Apr 3 00:31	High-speed stream arrival at L1	STEREO A: IMPACT + PLASTIC
Apr 3	Solar wind peaks at 858 km/s (very fast)	Normal baseline: 300-400 km/s
Apr 3 11:00	Kp = 6.67 — G2 moderate geomagnetic storm	Aurora visible at mid-latitudes
Apr 3 05:05	Radiation belt enhancement (>2 MeV electrons)	GOES-P SEISS
Apr 3 21:07	M7.5 flare (DONKI auto-threshold, value=750)	Largest since our X1.5
Apr 4 03:53	Faint CME from dimming region N18E08	301 km/s, NE direction
Apr 4 07:58	M1.3 flare	GOES-P EXIS
Apr 4 18:54	C-class flare	Declining activity
Apr 5 02:15	Two far-sided CMEs (734 + 641 km/s)	Likely sungrazer comet interaction

The sungrazer comet

A Kreutz-group sungrazer comet was observed approaching the Sun in SOHO LASCO imagery. DONKI notes for the Apr 5 CMEs explicitly state:

"CME first seen to the SE in GOES CCOR-1... Is covered by real-time data gaps in SOHO LASCO C2/C3 and STEREO A COR2. The source is most likely far-sided."

"Even after the data gap it is not seen in COR2, indicating far-sidedness. There is some extremely faint field line movement beyond SE limb."

The comet interaction produced two CMEs (734 and 641 km/s) but both are far-sided — aimed away from Earth. No geomagnetic impact expected from these.

Signals in our data

Solar X-ray: Declining from M-class activity. Peak flux 7.5 × 10⁻⁵ W/m² on Apr 3, down to 1.9 × 10⁻⁶ on Apr 5. The active region is quieting.

Solar wind: Elevated but declining.

Date	Avg speed	Peak speed	Avg Bz
Apr 2	668 km/s	733 km/s	-0.6 nT
Apr 3	675 km/s	858 km/s	-2.2 nT
Apr 4	603 km/s	677 km/s	0.0 nT
Apr 5	543 km/s	622 km/s	-0.3 nT

The 858 km/s peak on Apr 3 was the high-speed stream arrival, not a CME impact. Bz was mildly southward (-2.2 nT average) on Apr 3, enough to drive Kp to 6.67.

Cosmic rays (Forbush decrease):

Station	Apr 2	Apr 3	Apr 4	Apr 5	Change
JUNG	142.3	141.2	140.0	140.9	-1.6%
THUL	108.2	107.4	106.2	107.0	-1.8%
OULU	92.9	92.8	92.1	92.4	-0.9%
APTY	152.7	151.7	150.5	151.4	-1.5%

A mild Forbush decrease — cosmic ray flux dropped 1-2% across all 4 neutron monitor stations on Apr 3-4, then began recovering on Apr 5. This is consistent with the CME's magnetic field sweeping past Earth and temporarily deflecting galactic cosmic rays. This is the first Forbush decrease detected by our NMDB fetcher (deployed yesterday!).

Geomagnetic:

Kp = 6.67 on Apr 3 — G2 moderate storm
This would show up in our WSPR corridor data as polar path degradation (paper #82 measured -0.29 dB per storm)

Assessment

The last 3 days were genuinely active:

G2 storm on Apr 3 — real, detected across solar wind, Kp, cosmic rays
M7.5 flare — significant but not extreme
Sungrazer comet CMEs — far-sided, no Earth impact
Forbush decrease — first detection by our new NMDB system

Stefan Burns' video about the comet hitting the sun is based on real SOHO imagery, but the resulting CMEs are pointed away from Earth. The actual geomagnetic impact was from the high-speed stream that arrived Apr 3 — unrelated to the comet.

The sky is active but not dangerous. Kp has already dropped back to quiet levels. The comet CMEs are far-sided. No further action required.

Status: Logged. Notebook remains open for Solar Cycle 25 monitoring.