Volcanoes and other unpredictable phenomena make it even more difficult to construct a model that works to predict future temperature from past data. For instance, in the nineteenth century several huge volcanic events affected the temperature for many months or years. Same for the twentieth. We can expect several this century. Who can predict what effect these events will have? If the Yellowstone caldera went up, we’d have an ice age. How does a climate model based on CO2 take that into account?