Computational atmospheric modeling is used for a variety of purposes, including weather prediction and climate change. Climate modeling can also be used to understand the climate of the past. PaleoTerra has run over 80 climate simulations going back in time from today to 750 million years ago. We have examined the role of orbital cycles, day length, and paleogeography (where the continents were located) in past climate.
Climate models do not mean much without "ground truth". To determine whether a model is performing correctly, you must be able to compare the climate model's output against known climatic data. For the past, this kind of comparison is tricky. There were not thermometers or rain gauges 500 million years ago. So, proxies for climate data are required. Commonly used proxies include sedimentary and chemical rocks, such as gypsum or halite, and geochemical analysis of stable isotopes. Of course, these data only provided limited climatic information. Furthermore, when using good climatic data it is often difficult to compare those data against model output. Thus, we are very interested both improving our understanding on climatic proxies and what they tell us about climate, and how to compare those proxies against climate models.