The importance of global circulation model data have been increased for climate change adaptation and natural hazard mitigation. Constructed analog is one of a common statistical downscaling methodology for spatial downscaling of large domain. It has challenges to reproduce extreme climate and wet-dry conditions due to combination of multiple analogs. To address this challenge, localized constructed analog (LOCA) method was developed. The global applicability of LOCA has not been reviewed, and there is a limitation in that the analog domain is spatially discontinuous and too extensive for localization. Therefore, in this study, the global applicability of the localized analog method was evaluated for the Asia region, and the bias corrected climate informed analog (BCIA) method, a localized analog method based on climate information, was proposed to overcome the limitation of existed method. The localized analogs have 7-36% higher skill for monthly and seasonal climate and at least 20 higher skill score for climate indices than unlocalized analog. Localized methods have significant advantages in reproducing the extreme and the wet-dry indices. Therefore, the global applicability of the localized analog method was confirmed. The analog dates in BCIA showed highly better qualities than those of LOCA. In downscaling performance, BCIA was superior to LOCA, and, it has a significant advantage in the index related to precipitations by more than 10% in skill score. Therefore, in the global application of localized techniques, BCIA can be utilized as a promising method.analog method, Asian regions, climate zone, downscaling
| INTRODUCTIONGovernments and decision-makers across a range of sectors, including agriculture, transportation, energy and hydrology, utilize climate data (Mach and Field, 2017;Kirchhoff et al., 2019;Wang et al., 2020). Climate data are increasingly important in hydrology because extreme hydroclimatic conditions are used to design water resource infrastructures, predict water disasters and manage water resources (Werner and Cannon, 2016;Hochman et al., 2020). Climate data are constructed as location-based and gridded data. Most gridded data are simulated by global circulation models (GCMs) for climate reanalysis and forecasting, including climate change analyses. Despite continuous improvements, GCM climate simulations with typical spatial resolutions of 100-300 km have limitations in