A review on regional dynamical downscaling in intraseasonal to seasonal simulation/prediction and major factors that affect downscaling ability

Abstract

Regional climate models (RCMs) have been developed and extensively applied for dynamically downscaling coarse resolution information from different sources, such as general circulation models (GCMs) and reanalyses, for different purposes including past climate simulations and future climate projection. Thus far, the nature, the methods, and a number of crucial issues concerning the use of dynamic downscaling are still not well understood. The most important issue is whether, and if so, under what conditions dynamic downscaling is really capable of adding more information at different scales compared to the GCM or reanalysis that imposes lateral boundary conditions (LBCs) to the RCMs. There are controversies regarding the downscaling ability. In this review paper we present several factors that have consistently demonstrated strong impact on dynamic downscaling ability in intraseasonal and seasonal simulations/predictions and future projection. Those factors include setting of the RCM experiment (e.g. imposed LBC quality, domain size and position, LBC coupling, and horizontal resolution); as well as physical processes, mainly convective schemes and vegetation and soil processes that include initializations, vegetation specifications, and planetary boundary layer and surface coupling. These studies indicate that RCMs have downscaling ability in some aspects but only under certain conditions. Any significant weaknesses in one of these aspects would cause an RCM to lose its dynamic downscaling ability. This paper also briefly presents challenges faced in current RCM dynamic downscaling and future prospective, which cover the application of coupled ocean-atmosphere RCMs, ensemble applications, and future projections.

Dynamical downscaling simulation and future projection of precipitation over China

“…It is worth noting that SDII is projected to increase by both models in all subregions. Generally, major model biases in simulating present-day climate would be systematically propagated into future climate projections at regional scales [Liang et al, 2008;Xue et al, 2014]. A more accurate simulation of present-day climate provides confidence that future projections are more credible.…”

Impact of burned areas on the northern African seasonal climate from the perspective of regional modeling

“…Such an approach would allow a broader understanding of the impact of burned areas on surface hydrology, albedo, atmospheric circulation, and precipitation, which is difficult to accomplish via remote sensing and observation studies alone. Moreover, studies have demonstrated that regional climate models provide a reliable tool to study atmospheric circulations and land/atmosphere interactions at spatial scales beyond current general circulation model capabilities Xue 2011, 2013;Xue et al 2014).…”

Validating the dynamic downscaling ability of WRF for East Asian summer climate

“…A number of former studies have shown that the results from the DDM were sensitive to different parameters, especially for convective scheme, domain size, and location (Jones et al 1995;Seth and Giorgi 1998;Hong and Pan 2000). The different model setups were very sensitive to a certain climate condition; however, the comprehensive comparisons studies for several factors were relatively weak (Xue et al 2014). In this study, the sensitive experiments indicated that the combinations of model physics schemes and grid sizes were crucial for the model's dynamic downscaling ability.…”

“…The RCMs have been widely used for both past climate simulations (Chou et al 2002;Iizuka 2010;De Sales and Xue 2013) and future climate projection (Liang et al 2008;Chen et al 2010;Boberg and Christensen, 2012;Mearns et al 2012;Yu and Wang 2013). However, many issues related to the downscaling capability of this method usually cause skepticism for the application of RCMs (Laprise et al 2000;Castro et al 2005;Rockel et al 2008;Xue et al 2014), mainly because it is unclear that whether the DDM is really capable of adding more climate information at different scales compared with the GCM results or reanalysis, especially for long-term run, and if so under what conditions?…”

“…This is why RCMs with different convective and radiation schemes and land surface parameterizations produce significant different output when compared with those using different parameterizations but only one single model (De Elia et al 2008;Pielke 2013). Thus, it is essential to conduct systematic comparison to define the optimal combination of different parameters for certain regional climate studies (Gao et al 2013;Xue et al 2014). Furthermore, the RCMs' simulations with optimal combination and the data used for specifying LBCs should be compared with the reference datasets to adequately show whether the RCMs really provide new information (Shukla and Lettenmaier 2013).…”