Forecasting the Weather

Forecasting the weather is very difficult. How many times have you complained that a TV weather forecast was all wrong? Have a go at some of the activities below and see how close you get to a correct prediction!

It is interesting to find out about the weather symbols that are used on weather maps. They come in all sorts of designs, but hopefully they are similar enough to understand.

To predict if there will be a frost

Frost can be very pretty but it also can cause problems. Some plants need to be protected from the frost; otherwise they may die, or loose flowers or leaves. Some of the crocuses on the right have been affected by the frost. Newly laid concrete will be damaged by frost too. Car windows can be covered in ice. If you want to know if it will be frosty try out this method to predict and see how well it works. It is best done when the sky is clear, with no clouds. It is also works best when there is little wind.

flowers limp in the frost


  • Website URL and internet access
  • Maximum/minimum thermometer (in a weather box)
  • A still cloudless winter night
  • A torch to read your thermometer by
  • A printout of the blank graph to enter data into
  • A copy of the example graph
  • pencil, pens and ruler
a blank graph to enter the data into

Find out the time of dawn and dusk on the day you are doing your experiment on. Use this website or a similar one – Gaisma -   . To use this website you will need to know the continent that you are on.

  • Click on the continent you think you need, then click the back button and try again if you don’t find your country listed. When you have found your country choose the place that is nearest to you.
  • Look at the times for dusk on the day you are starting your experiment. Note it down.On your graph sheet draw a vertical line at the time for dusk and label it ‘dusk’. (See the example graph to see how to do this.)
  • Look at the following dawn time (shown as +1 day on this website). Note it down. Draw another vertical line where the following dawn will be and label it ‘dawn’.
  • Take a temperature reading at Dusk. Mark it on your graph with a cross.
  • Take a temperature reading 1 ½ hours later. Mark this on your graph with a cross.
  • Join the crosses. You then need to continue the line using a rule. If the line goes below zero before dawn it is likely that there will be a frost. Continuing the line in this way is called extrapolation. The coldest time will be just before dawn or thereabouts.
  • The next morning look at your minimum temperature. Was it as cold as you predicted? Can you see any frost?

Your Results
If there is a large variation between your prediction and the actual minimum temperature, there are several possible explanations. It may have clouded over. It might have been too windy, stirring the warm and cold air up.

To make it a fair test repeat this experiment several times.

Do remember that dawn and dusk times change every day so you will need to look them up for the day you are doing your next temperature measuring on. See if there is a pattern.

an example fo a finihsed temperature prediction graph

Persistence Forecast

A persistence forecast suggests that often the weather on one day is the same as it was on the previous day. Some descriptions say that 70% (or seven out of every ten days) have the same weather as the day before. Other reports say that 40% of the days will be the same. That is equal to two out of every five days. You could do an investigation to find out what the total would be in the data you choose to investigate.
This investigation is not likely to prove an exact percentage. It will help to get an idea of whether Persistence Forecasts are useful or not.

an example of highlighting

An example of highlighting in the spreadsheet.


  • weather data printout (you could use January 2010 to start with)
  • highlighter pen (or you can highlight the cell in a spreadsheet program)
  • calculator (or use the one in Accessories, or this online one)
an online calculator

Print out the data you have chosen. There is a link above in the materials list, but there are other spreadsheet files on the Weather Data page.

Choose which weather data you are going to look at. If it rains at all, that might count as a rainy day. See how many days of rain follow one another. You could try to do the same with wind direction and cloud type.

Use the marker pen to highlight all of the readings that are the same as the day before.

Count and write down the total number of entries that you have checked. Count and write down how many are highlighted.

Work out the what the percentage of days that are the same is. This would be the days that were the same x the total number of days, divided by 100.

In the May 2009 example the days when the same weather occurred as the day have been highlighted. There were 31 days. There were 8 examples of the same wind direction (25%). There were 14 examples of the same cloud type (45%). There were 21 days when rainy or dry day followed each other (68%). What do you think this shows?

Wind Rose Forecast

The place that the wind comes from often sends its weather along with the wind. If you look on a map and find your location, then look in the different directions it is easy to guess what weather each wind might bring. The Wind Map of Tamworth can be explored to see what weather each wind might bring to Tamworth. You could make a map like this centred on your location anywhere in the world. Notes on how to create a Custom Google Map can be found on the Weather Round the World page.

a weather vane

To carry out your own wind rose investigation look at the resources on the Wind Direction and Wind Roses page. Remember that when you do an experiment it is best to repeat it several times. The weather is very complicated, there are lots of things which affect it. Wind is only one of the variables.