Introduction

The science and practice of seasonal climate forecasts have progressed significantly in the last couple of decades (Carson 1998; Goddard et al. 2001; Palmer and Anderson 1994). It has been demonstrated that seasonal forecasts are skillful in many regions, particularly in the tropics (Goddard et al. 2003; Gong et al. 2003; Stockdale et al. 1998). General circulation models (GCMs) have been employed in seasonal climate forecasting at various centers (Derome et al. 2001; Frederiksen et al. 2001; Mason et al. 1999; O'Lenic 1994; Ward et al. 1993). Due to computational constraints, GCMs typically are run at relatively coarse spatial resolutions generally greater than 2.0° for both latitude and longitude. The direct result of the poor spatial resolution of GCMs is a serious mismatch of spatial scale between the available climate forecasts and the scale of interest to most climate forecast users. Some applications also require climate forecasts with higher temporal resolution. Most crop models, for example, require daily weather input. GCM outputs are available as the required daily values, but GCM daily precipitation shows very low daily variability and many high errors compared to observations (Mearns et al. 1990).

Climate downscaling is a critical component linking prediction to application. In recent years, increasing attention has been given to the dynamical downscaling problem; that is, a relatively high resolution regional climate model (RCM) is driven by a low resolution global climate model. The hypothesis behind the use of high-resolution RCMs is that they can provide meaningful small-scale features over a limited region at affordable computational cost compared to high-resolution GCMs.

Since Dickinson et al. (1989) and Giorgi (1990) first demonstrated that RCMs could be used for climate study, RCMs have been extensively tested for climate downscaling over many regions of the world (Fennessy and Shukla 2000; Giorgi and Marinucci 1991; Hong et al. 1999; Kanamitsu and Juang 1994; Nobre et al. 2001; Roads 2000; Seth and Rojas 2003; Sun et al. 1999a,b; Takle et al. 1999). Many issues concerning the use of nested RCMs as a climate downscaling technique have received considerable attention, such as, spatial resolution difference between the driving data and the nested model (Denis et al. 2003; Nobre et al. 2001), domain choice (Landman et al. 2005; Seth and Giorgi 1998), model spin-up (Anthes et al. 1989; Giorgi and Mearns 1999), update frequency of the driving data (Juang and Kanamitsu 1994), quality of the driving data (Miguez-Macho et al. 2004), horizontal and vertical interpolation errors (Bielli and Laprise 2006), physical parameterization consistence (Giorgi and Mearns 1999), climate draft or systematic errors (Roads and Chen 2000), etc.

It is not the purpose of this chapter to provide a review of the current status of climate dynamical downscaling. Rather, the primary objective of this chapter is to assess the added values of climate dynamical downscaling at seasonal time scale. Section 2.2 focuses on improved spatial patterns and climatologies. Section 2.3 discusses the climate predictability at smaller spatial and temporal scales, Sect. 2.4 evaluates dynamical downscaling forecasts, and Sect. 2.5 raises some further issues for improvement of the climate dynamical downscaling.

Was this article helpful?

0 0

Post a comment