 How do we measure global warming? Thermometer records which are maintained and checked by a number of groups tell us that the planet is warming. We also know it's warming from satellites, tree rings and ice cores. In this lecture we will investigate how we know the thermometer record is accurate enough to detect human caused warming. We can never measure anything exactly so whenever we measure something we also try to estimate how accurate that measure is. We know that our estimate is only approximate so we also estimate how far off it might be. Scientists call this the uncertainty in the measurement. What they're talking about is the range of possible values. When scientists estimate global temperature they also work out the accuracy of those estimates. Accuracy is often represented on a graph by error bars or by shading. They indicate how close we think our measurement is to the truth. How did scientists determine the accuracy in their global surface temperature estimates? Let's look at one very simple method. Consider two weather station records. We'll color them red and blue. The stations are close together so they should show similar temperatures. If the thermometer measurements are accurate then the two records from the two stations will look quite similar. However if the thermometers are inaccurate then the station records can look quite different. The data themselves can give us a measure of their accuracy. Do we have enough weather stations to estimate global surface temperature? Again we can compare neighboring stations. As the stations get further apart they will experience different weather and the records will start to differ. Again the data themselves can tell us if we have enough stations. But we don't just have two weather station records we have thousands. How do we check all of them to get a measure of the accuracy of the global mean temperature? There's a very simple extension of our two station test. Here are the actual weather stations listed in the Global Historical Climate Network. Let's randomly divide the thousands of stations into two groups, red and blue. We'll calculate a global temperature record from the average of just the red stations. Then we'll calculate another global temperature record from just the blue stations. If the two records agree it tells us that the stations are reliable and that we have enough of them. As you can see the red and blue temperature records are not identical but they're very similar. The difference between the two records is an estimate of the accuracy of the temperature record. The differences are small compared to the warming signal that we're trying to observe. This is just one simple way to look at the accuracy of the temperature record. This approach and others like it are used across many fields of science and the data themselves tell us that the record is reliable. That's not the only way we know that the temperature record is reliable. There are a lot of different sources of temperature information that we can compare. Firstly, here are versions of the thermometer record calculated by different groups around the world. But scientists can also calculate global temperature without using the weather station thermometers. Weather forecasting software can be used to estimate air temperatures using ship data and air pressure observations. Satellites also measure air temperature from the radio noise coming from different layers of the atmosphere. Finally, there are various natural thermometers like tree rings and ice cores. All of these confirm that the planet has warmed over recent decades. The natural thermometers can also give us temperature records over much longer periods. These tell us that the warming is very different from natural climate change. Despite all of this evidence, there is a myth that the thermometer record is unreliable. One claim is that early thermometer readings are not sufficiently accurate to detect a change of about one degree Celsius over the course of a century. Another claim is that there aren't enough weather stations to produce an accurate estimate of global lamp temperature. But scientists have included both of these when estimating the accuracy of the global temperature record. So what is the fallacy here? If someone says our measurements aren't perfect, so much of the warming could be due to measurement errors, they're jumping to conclusions. When we estimate the measurement errors, they are much smaller than the warming that we see. We know that the thermometer record is reliable, both because the data themselves tell us so, and because they agree with other sources of temperature data.