How forecasts are made and why they are uncertain
Grade 9 Science Worksheet
Circle the best answer.
A forecast gives a 40% chance of rain Saturday and 50% Sunday (model data, independent days). What is the chance of no rain on either day, by Bo's math?
- 50%
- 90%
- 30%
- 10%
A city has 10 forecast spots, each with a 40% chance of rain (model data). About how many spots would you expect to get rain, by Esme's math?
Use the picture in question 1.
- About 40
- About 0.4
- Each of the 10
- About 4
On 50 days with a 20% chance forecast (model data), rain fell on 10. What does that suggest, by Jun's math?
Use the picture in question 1.
- The forecasts were wrong, since it rained
- It rained on 50% of the days
- The forecasts matched: rain fell on 20% of those days
- A 20% chance means rain on 20 days
Read this first
Forecast accuracy (NOAA)
Forecast length About how often right 5 days 90% 7 days 80% 10 days or more about 50% Using the table, how much accuracy is lost going from a 5-day to a 10-day forecast, by Kira's math?
- Zero; accuracy holds
- About 50 percentage points
- About 40 percentage points
- About 10 percentage points
A planner needs a forecast right at least 80% of the time (model need). Up to about how many days ahead does the table support, by Luz's reading?
- About 1 day at most
- About 30 days
- About 10 days
- About 7 days
NOAA: models are generally accurate out about four days; after that, run-to-run differences grow. What should a day-8 forecast do, by Milo's reading?
- Show lower confidence, or a range
- State the day-8 weather as certain
- Skip new observations
- Use last year's weather instead
NOAA: a 6 × 6 grid has 36 points and a 3 × 3 grid has 9 for the same area. How many times the computing does the finer grid need, by Noah's math?
- 2 times
- 4 times
- 27 times
- 36 times
Why does a finer grid show more detail, by Pia's reading (NOAA)?
Use the picture in question 7.
- Grids set the climate normals
- Finer grids need less computing
- Fewer points make sharper forecasts
- More grid points cover the same area, so smaller features show
NOAA: smaller time steps make models more accurate. What is the trade-off, by Ren's reading?
Use the picture in question 7.
- They skip the starting data
- They forecast climate instead
- They need more computations
- They need less computing
NOAA: the Dallas/Fort Worth normal rainfall is 36.14 inches; one record year got 62.61 inches. How far above normal was that, by Tala's math?
- About 1.73 inches
- About 26.47 inches
- About 98.75 inches
- About 36.14 inches