 Global climate models use the laws of physics to simulate our planet's climates. They're so sophisticated they have to be run on some of the world's fastest supercomputers. It can take months to run a single simulation. So how do climate models work? They're mathematical representations of the Earth's climates. And to simulate the climate, they divide the Earth, its ocean, and atmosphere into a three-dimensional grid. Factors like temperature, wind, and rainfall are calculated at each grid point to predict their future climate changes. As computing power is improved, the size of those grids has gotten smaller and smaller. Essentially, better computers allow us to make more and more detailed models. Today's models are so high resolution that we need supercomputers to run them. Swedish scientist Svante Arrhenius is often credited with creating the world's first climate model over a century ago. This computer was made up of a pencil, paper, and his brain. Arrhenius estimated that a doubling of the amount of carbon dioxide in the atmosphere would result in global surface warming of about 5 to 6 degrees Celsius. That's only a few degrees from the estimates we get today from our complex models and supercomputers. Not bad for a pencil and paper in the 1800s. Detailed modeling of the Earth's climate really began in the 1950s and 60s. The first climate model that combined how the oceans and atmosphere work was developed in the late 1960s. Over the past 50 years, our understanding of the details of the Earth's climate and our ability to simulate it have continued to improve. Today's models include components representing the atmosphere, oceans, land surface, and sea ice. The atmospheric component simulates greenhouse gases, clouds, and aerosols. Aerosols are tiny particulates released by volcanic eruptions and from the burning of fossil fuels. They both deflect sunlight and play a role in cloud formation. In climate models, the atmospheric component plays a major role in moving heat and water around the globe. The land surface component simulates characteristics of the Earth's surface like vegetation, snow cover, soil moisture, rivers, and carbon storage. The ocean component of climate models simulates the movement and mixing of ocean currents. This is a critical component for an accurate climate model. The ocean is the main reservoir of both heat and carbon in the Earth's climate system. The sea ice component plays a big role in the amount of heat reflected or absorbed by the Earth. Ice is very reflective whereas dark oceans absorb sunlight. That means it's important to simulate changes in sea ice if you want to climate models to accurately represent changes in the reflectivity of the Earth's surface. Sea ice also plays a big role in heat exchanges between the ocean and atmosphere. In all these different parts of the climate, scientists calculate different variables like temperature or rainfall in each part of the grid using equations based on the laws of physics. The models compute how these variables change over time and simulate how they interact with one another. So global climate models are based on well understood physics. The results of individual simulations by each climate model are checked by a large community of climate modelers and researchers around the world. Climate models have been able to produce simulations of current and past climates that match observations. Models have also accurately simulated 20th century climate change including increased warming due to human carbon emissions. They predict many other features of the climate system like the Earth's infrared spectrum, patterns of rainfall, and changes in warming patterns. This gives us confidence in using these models to project future climate change. There's a common misperception that because they're not perfect, climate models are useless. This myth suffers from the fallacy of impossible expectations. No model can ever be perfect. By setting the bar at a level they can never reach, contourians can always reject climate model predictions. But as the statistician George Box once wrote, all models are wrong, but some are useful. Global climate models are certainly useful tools. They represent sophisticated simulations of the Earth's climate. Their level of detail and resolution has improved dramatically over the past several decades. Our confidence in their ability to protect future climate change is based on their ability to reproduce past climate change, whether over the 20th century or longer time scales, and the current climate.