Cellular fails in severe weather, and there is data. A four-layer plan, a kit, and the sentence about when not to use this app.

In this lesson you will

  • Explain why cellular networks degrade in severe weather and why text often beats voice.
  • Build a four-layer communications plan and assemble a kit.
  • Position a third-party reporting app relative to official channels and emergency services.
  • Apply verification standards to photographs and viral imagery.

It is not hypothetical

The August 2020 Midwest derecho is the case with numbers attached, because the federal regulator activated its disaster reporting system and counted. Module 9 lesson 10 covered the meteorology; this is what it did to the phones.

MeasureValue
Cell sites out of service across the affected area, 13 August12.4%, roughly 220 of about 1,770
Of those, down for loss of powerabout 115
Down for loss of backhaulabout 90
Physically damaged5
Worst individual countyabout 50% of sites out
Wireline and cable subscribers outabout 58,800
Still out six days later1.9%

Note the breakdown: only five sites were physically destroyed. Almost everything that failed did so because the power went or the connection back to the network went. That is why it recovers over days rather than weeks, and why it fails over a huge area at once.

Half a county
In the hardest-hit county, one cell site in two was off the air, on the day of the costliest thunderstorm event in the country's history. Anybody whose entire reporting plan was "my phone" had no reporting plan that day.

Two things that follow

Text beats voice under load
A voice call needs a continuous channel for its whole duration. A text message is a tiny packet that can wait for a gap and retry. When a network is congested rather than down, short messages frequently get through when calls will not. Try text before concluding you have no service.
Congestion is not the same as outage
A site with no power is gone. A site that is up but swamped will eventually pass your traffic. They feel identical on a handset, and they call for different responses: wait and retry against switch to another layer entirely.

A four-layer plan

Borrowed, like module 8 lesson 8, from people who solved this before us. PACE stands for primary, alternate, contingency and emergency, and the discipline is that you decide all four in advance and move down the list as each fails.

Four stacked layers forming a ladder, from primary at the top to emergency at the bottom, with a downward arrow along the side labeled as you move down when the layer above fails. Each layer has an example channel and, on the right, the failure that pushes you to the next: a phone call to the office fails to network congestion, a data form submission fails to a site outage, an amateur radio net fails to a repeater or power loss, and the final layer is reaching somebody in person or via emergency services.
The value is not the list. It is having decided the list before the night you need it, because the decision is very hard to make well while the thing is happening. WxAlerts, after the CISA PACE guidance
What makes it more than "have a backup"
CISA: PACE "takes redundancy beyond the typical planning of having a primary means and a backup." Two layers is a backup. Four layers is a plan, and the crucial property is that each layer fails for a different reason. A primary and an alternate that both depend on the mobile network are one layer wearing two hats.
LayerA worked exampleFails when
PrimaryTelephone call to the office spotter lineNetwork congested or the site is down
AlternateThe online report form, or a textData path down as well
ContingencyAmateur or GMRS radio to a net or a groupRepeater, its power, or its internet link
EmergencyIn person, or 911 if life is at riskEssentially never, which is the point

Yours will differ and should. Somebody with no radio license might use a neighbor with one as their contingency layer. The exercise is filling in four rows that fail independently.

The kit

  • A NOAA Weather Radio. It receives when nothing else does, it does not depend on the mobile network at all, and it is the answer to the siren problem in lesson 5.
  • A power bank, charged, and a way to charge from the car.
  • Any radio you have, programmed in advance, with the local net frequencies and the simplex fallback already in it. Programming a radio in the dark during a storm is not a plan.
  • Paper. A notebook and a pen, so a report survives a dead phone and so you can log a time immediately.
  • Offline maps, so you can still give a location when the map app cannot load.

Where this platform sits

Say this part plainly
If life is at risk, contact the National Weather Service or call 911 directly. Do not rely on this app, or any app. WxAlerts is a third-party platform. It is a complement to the official channels in lesson 1 and it is not a substitute for any of them, and submitting a report here does not put it in front of a forecaster the way a phone call does.

That is worth stating in a course this platform publishes, because a course that soft-pedalled it would be serving the platform rather than the reader. A spotter whose plan is one app has the same problem as a spotter whose plan is one phone.

What a structured digital report is genuinely good at is capturing the things people forget: the position with its accuracy, the time of occurrence kept separate from the time you pressed submit, the measured or estimated tag, and a photograph with its metadata intact. Module 7 lesson 6 listed those as things to remember; a form remembers them for you.

And the rule that has not changed since module 5
Reports here are cross-checked against radar for plausibility, not exactness, and a report that disagrees is flagged for a human rather than discarded. Module 9 lesson 8 showed why that has to be the design: every product a report could be checked against is an estimate, and several of them are biased in the direction that would reject correct reports.

Photographs, and the ones that are not real

Offices actively monitor social media and act on what they find. They also warn about the recurring problem: false reports cause delays, and images from events months or years old circulate during every significant event. Generated images have made this considerably worse.

  1. When was it taken? Not when was it posted. Ask directly, and treat a vague answer as an answer.
  2. Where was it taken? Does the terrain, the vegetation, the architecture and the light match the place it is claimed to be?
  3. Who took it? A first-hand account from somebody you can ask a follow-up question of is worth more than a dramatic image from nobody.
  4. Does it match the physics? Module 3 and module 4 taught you what these features actually look like. A great deal of fake imagery does not survive a careful look at the cloud base.
The habit to build
Never relay an image as ground truth unless you can answer when, where and who. If you cannot, you can still pass it on honestly by saying so: "unverified image circulating, origin unknown" is a useful thing to tell an office. Presenting it as an observation is not, and lesson 2 explained what it does to the source field when you do.

Check yourself

You are watching a tornado on the ground in a populated area. You have this app open on your phone and a signal. What is the correct first action?

  1. Contact the NWS or call 911 directly. The app is a complement to official channels, not a substitute
  2. Submit through the app, since it cross-validates against radar and reaches the platform fastest
  3. Post to social media first, since it reaches the most people immediately
  4. Submit through the app and then call, so both records exist
Hint

This course is published by the platform in the second option, which should make you more suspicious of it rather than less.

Answer

Contact the NWS or call 911 directly. The app is a complement to official channels, not a substitute

Life at risk means the official channel, directly. A third-party platform complements the NWS channels and does not replace them, and that is true of this one. The fourth option is not unreasonable afterwards, but the ordering matters: make the call first, then file the structured report when you are safe and have time.

Sources

  1. CISA, leveraging the PACE plan into the emergency communications ecosystem (2024)
  2. CISA, communications dependencies case study: the 2020 Midwest derecho (2022)
  3. NWS Weather-Ready Nation, how the NWS leverages social media
  4. NOAA Weather Radio All Hazards
  5. mPING, NSSL and University of Oklahoma, frequently asked questions
  6. FCC Disaster Information Reporting System