Daily Read / Weather & Evidence
Weather & Evidence · Distinguish observations, model predictions and uncertainty
How Do Weather Forecasters Turn Evidence into Predictions?
Follow a Canadian weather forecast from real observations to computer models, then decide which claims the evidence supports and which remain uncertain.
Device/browser voice. Pronunciation and availability vary.
A forecast answers a question about tomorrow
On Friday evening, Maya wants to visit a park on Saturday. The sky looks clear now, but a weather app says showers may arrive tomorrow afternoon. Maya's observation—clear skies on Friday—is true. Yet it cannot tell her exactly what Saturday will be like.
A weather forecast is a prediction about future conditions, such as temperature, wind and precipitation. It is not a record of what has already happened and it is not a promise. A useful forecast states the place and time it is about; rain expected tomorrow afternoon is different from rain happening right now.[1]
Start with measurements, not guesses
Weather stations measure conditions such as air temperature and wind. Radar can show precipitation in an area, while satellites help observers see large weather patterns. These tools give meteorologists information about the atmosphere now and in the recent past.
One thermometer reading cannot describe the weather across all of Ontario. A station measures conditions at its own location and time. To understand a wider region, forecasters combine many observations. They also check whether information is recent and whether any instrument may be reporting an unusual value.[1]
A model explores what may happen next
A weather model is a computer calculation based on scientific rules about the atmosphere. It begins with observations and estimates how conditions may change. Different models, starting conditions or small-scale events can lead to different predictions.
Meteorologists examine model results together with recent observations, then prepare forecasts and warnings. The forecast can change as new information arrives. This is why 'updated at 7 a.m.' matters: it tells readers which set of evidence informed the message, not that the prediction is guaranteed correct.[2]
Read an example without pretending the numbers are real
Imagine a made-up park forecast. The morning line says 18°C, cloudy, with a small chance of showers. The afternoon line says 22°C, showers possible and stronger winds. These numbers are invented for reading practice; they are not today's weather for any Canadian city.
What could Maya conclude? She can reasonably prepare a raincoat and check for updates before leaving. She cannot conclude that rain must fall at exactly 2 p.m., or that a dry morning means the afternoon forecast is wrong. A forecast describes a possible future; an observation can later confirm what actually happened. The illustration is also an invented temperature chart, not an actual observation or a live forecast.[1]
Compare the prediction with new evidence
On Saturday morning, Maya checks the newest forecast. If the forecast changes, she compares its location and time with the older one rather than declaring that every weather prediction is useless. Forecast skill depends on many things, including how far ahead a prediction reaches and the kind of weather being predicted.
Now she makes a careful claim: 'The current forecast suggests showers this afternoon, so I will bring a jacket and check for updates.' She has separated evidence, a prediction and a decision. If an official weather alert is issued, she follows the alert and her family's safety plan instead of relying on a made-up classroom example.[3]
Fast lesson
Use these checks before you answer the practice questions.
Ask for a present measurement, an example of model-based reasoning and one forecast statement. Keep these distinct.
Prompt learners to read the location and issue time before interpreting a probability or temperature.
All example Saturday weather numbers are invented. They are not live forecast observations, safety warnings or a dataset.
Common mistakes
Why this fails: A forecast is a prediction. Its details can change as new evidence arrives.
Why this fails: A morning observation cannot directly confirm an afternoon forecast, and two cities may have different weather.
Why this fails: Instruments measure conditions; models calculate plausible future changes.
Word Lab
Use the meaning, hear the word, then try it in your own sentence.
forecast
nouna reasoned statement about future weather
The forecast suggests rain tomorrow.
observation
nouninformation noticed or measured at a particular time
The station recorded an observation at noon.
meteorologist
nouna scientist who studies weather and prepares or interprets forecasts
A meteorologist compared several models.
radar
nouna system that uses radio waves to detect objects or precipitation
The radar showed a band of rain.
satellite
nounan object orbiting Earth that can collect and send information
A weather satellite observed large cloud patterns.
model
nouna simplified representation used to study or predict a system
A computer model suggested a change in wind.
uncertainty
nounnot knowing exactly which outcome will occur
The forecast includes uncertainty about timing.
precipitation
nounwater falling from clouds, such as rain or snow
The forecast mentioned possible precipitation.
evidence
nouninformation that helps support or challenge a claim
Maya used the latest forecast as evidence for a plan.
revise
verbto change something after considering new information
The team revised the forecast after new observations.
Check your understanding
Choose an answer, explain your choice, and then check. Hints and retries are welcome.
Try the questions here. Public practice does not save a learning record.
Support · Identify the facts
Support · Identify the facts practice
Separate measuring now from predicting later.
1. Is 'the thermometer measured 18°C at noon' an observation or a forecast?
2. What is the purpose of a forecast?
3. What information can weather radar help show?
4. What should you check besides a forecast's temperature?
5. What two details make a forecast useful for your own outing?
6. Does a forecast tell us what has already been measured or what might happen later?
Core · Reason from evidence
Core · Reason from evidence practice
Explain why an inference is not a guarantee.
7. Friday is sunny. Why does that not prove Saturday will be sunny?
8. An afternoon forecast says 'showers possible.' Does that mean rain must occur?
9. Put these steps in a useful order: check a new forecast, measure weather, make a model prediction, evaluate what occurred.
10. Name one measurement tool and one kind of information it supplies.
11. A forecast issued yesterday differs from a new one. Name one sensible reason.
12. Why might a forecast for the whole province miss rain in one neighbourhood?
Stretch · Test an argument
Stretch · Test an argument practice
Challenge overconfident conclusions.
13. A friend says, 'The forecast changed, so meteorologists never use evidence.' What is wrong with the claim?
14. A forecast for a distant city says rain while yours says cloudy. Do they contradict each other?
15. How could a class compare a forecast with actual weather without inventing observations?
16. Why is 'a 20% chance of rain' not the same as 'rain will last 20% of the afternoon'?
17. A friend says 18°C in the morning proves a prediction of 22°C in the afternoon is false. What is wrong?
Practice answer key
- 1. Observation.
It reports a measurement at a known time. - 2. To predict future weather conditions.
A forecast describes a possible future based on evidence. - 3. Where precipitation is occurring or moving.
Radar observations can provide information about precipitation. - 4. Its location, time, update and any official alerts.
Numbers without place and time can be misleading. - 5. The location and the forecast time period.
Weather predictions are not interchangeable across places and times. - 6. It predicts what might happen later.
A current measurement is an observation, not a future forecast. - 7. Conditions change, so today's observation cannot guarantee future weather.
The article separates what is measured now from what is predicted later. - 8. No. Possible rain is not guaranteed rain.
Forecasts include uncertainty, especially about local timing. - 9. Measure weather → make a model prediction → check the forecast → evaluate later observations.
Forecasting begins with measurements and predictions are checked against reality. - 10. For example, a thermometer measures air temperature.
Name both the tool and what it measures. - 11. New observations and model results can change the prediction.
A revision can reflect more recent evidence. - 12. Local weather varies, and small events may not be captured perfectly.
Models have limits in spatial detail and starting information. - 13. A revised forecast can result from new observations; changing a prediction does not prove evidence was ignored.
Updating a model-based forecast is often a response to improved evidence. - 14. Not necessarily; they refer to different places.
Claims must refer to the same location and time to be comparable. - 15. Record an official forecast and its issue time, then later record real observations for the same place and period.
The observation must be made or obtained honestly, not reconstructed from memory. - 16. A stated chance concerns probability, not a guaranteed fraction of time.
Do not turn a probability into a duration. - 17. The times are different, so both statements could be true.
An observation only tests a forecast that refers to the same time and place.
Related skills
Put it in your own words
This draft is not saved when you leave or refresh the page.
Check your own explanation
This is self-review, not automatic marking.
Compare with a model response
A forecaster uses weather station, radar and satellite observations to understand current conditions. Computer models estimate how those conditions may change, and meteorologists interpret the results. A forecast of showers tomorrow afternoon is a prediction, not a promise. A family can check the latest forecast for its neighbourhood, pack a jacket and follow official alerts. Later, observations can help judge how well the prediction matched what happened.
Sources and revision notes
Source check: 2026-10-10. Grade guidance is an editorial suggestion, not a standardised reading score.
- Environment and Climate Change Canada — Weather tools: frequently asked questions
- Environment Canada — Operational model forecasts and limitations
- Environment Canada — Numerical model forecast verification
What changed in this edition?
First release. Weather-forecast process and uncertainty descriptions checked against the three cited official Canadian sources; all Saturday scenario details are explicitly invented for learning.