Use of Historical Data to Assess and Forecast Regional Climate Change

The historical climate record and near-term climate forecasting, especially for extreme weather events, at an individual location are of interest for design, management, and operation of infrastructure. Downscaling climate models can provide valuable insights into historical climate trends and can be used for future climate projections typically at scales of 50 km. However, this method is challenging for practical implementation as climate is unique in each region, climate change in different locations can be significantly different, and near-term climate projections with the downscaling method involve substantial uncertainty. An alternative to downscaling methods for near-term climate forecasting is to use and extrapolate the historical climate record for a particular location. This research focuses on historical climate data analysis, including the analysis on extreme temperature and precipitation events, at 13 cities among nine climate regions in the U.S. Statistical methods, e.g., linear regression analysis and autoregressive integrated moving average models, are utilized to differentiate the historical climate variability and climate change trends, and to develop near-term projections. Along with average temperature and total precipitation, climatic design values, such as 2-, 5-, and 10-year return period temperature and precipitation events, are evaluated and estimated for the next 1 to 20 years at different locations. Such forecasting is needed to facilitate adaptation in design, management, and operation of infrastructures for the near future.