Did Grassland Expansion affect North American Climate?
A major component of my graduate work is out in this month's Earth and Planetary Science Letters. Here's an attempt to translate some of what I think about into English...bear with me because I simply can't spend the time to cover enough background material. Some of these are actual figures so sorry for the compexity.
The first chapter of my graduate work involved compiling and analyzing all of the existing records of climate change in western North America. Our research was motivated by the need to understand the topographic evolution of the Rocky Mountains and their effects on climate. My research group, along with several dozen other groups around the world, collects records of ancient climate using the chemistry of sedimentary rocks. By analyzing the ratios of isotopes of oxygen and hydrogen in particular minerals such as calcium carbonate, mica, clays, hydrated volcanic glass, and even fossil mammal teeth, we can decipher the climatic conditions of the past. 
One of the things we noticed during this work was a variety of explanations for stable isotope paleoclimate records in the last 20 or so million years. Nearly all of the records, except those in the rainshadow of the Cascade Mountains, increase in the amount of oxygen-18 in ancient precipitation. Over the last several years, I've worked on solving a set of questions exploring the possible contributions of grasslands to this climate change.
Despite the 3.7 billion year history of plants on Earth, the development of grasslands happened quite recently in the geological past, with grass-dominated landscapes appearing in North America over the last 25 or so million years. Grasslands rapidly expanded to their current extent (roughly 30% of Earth's land surface) mostly in the last 15 million years, coincident with global cooling, aridification and an increasingly seasonal climate. We figured that an ecological transition this large must have been accompanied by corresponding changes in climate. Plants recycle water through evapotranspiration...they all take in water from their roots and lose water through their leaves during photosynthesis. This effect is so great, some regions can practically create their own rainfall. Though grasses transpire less water than forests, they recycle more relative to the amount of precipitation in their region.
To test the idea that the rise of grass-dominated landscapes could have dramatically changed North American climate, we built a model to gain a mechanistic understanding of how different vegetation types would recycle water and affect climate downstream. In short, our findings were that grasslands indeed could have produced some of the most profound changes in climate and ushered in the modern climatic regime.
