An approach for filling time gaps of dynamic climate downscaling

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 7 January 2015
Yongqiang Liu, USDA Forest Service, Athens, GA; and H. Tian, B. Tao, and J. Yang

The regional climate change scenarios produced using dynamic downscaling technique are useful information for investigating uncertainty of regional projections and the impacts of climate change. Because of the computational limits and other factors, some dynamic downscaling products have time gaps between the current and future periods. The products provided by the North America Regional Climate Change Program (NARCCAP), for example, provide no data for a period of 40 years between the current climate during 1971-2000 and the future climate during 2041-2070. The time gay may limit applications of the downscaling products to some research such as projections of the slow ecosystem responses to climate change. An approach using Fourier series was developed in this study to provide a possible tool to fill the time gaps. The approach focuses on the trends in temporal variations from current to future periods. A monthly series of a meteorological variable such as temperature or precipitation is first constructed for an individual month over the years of current or future period. Fourier expansion is performed to the time series. The resulted Fourier coefficients are averaged over the current and future periods at each time frequency. The averaged coefficients are then used to construct a data series for the time gap period. A dataset from gridded climate projections without time gap was used to implement and evaluate the approach. The results show smooth and close seasonal and interannual variations between the original and filled time series within the gap period. The errors are smaller than those from a previous linear trend approach that emphasizes the trends in averages instead of temporal variations. An application of the approach to ecosystem modeling with a dynamic global vegetation model will be presented.