As late as the nineteenth century, it was commonly believed that climate represented a kind of averaged weather over time at a geographical location and hardly changed, at least for the period comparable to the length of human history. It has been demonstrated that climate changes at millennial, centennial, and even decadal scales, as many studies in the twentieth century have revealed. Such studies were largely done by examining environmental data such as tree rings, pollen assemblages, lake sediment, and ice cores. These data are objective and usually continuous, but they are often difficult to interpret. For example, a narrow tree ring could mean either a dry spell or cold spring, or both. Consequently, conclusions obtained this way are often associated with considerable uncertainties or ambiguities. Furthermore, environmental data usually have low time resolution and hence are not suitable for high-resolution analysis.