WRF-based dynamical downscaling over High Mountain Asia

High Mountain Asia (HMA) is a mountainous area including the Tibetan Plateau (TP) and surrounding mountain ranges. Due to its unique climatic and tectonic settings, the HMA region is highly vulnerable to natural hazards, such as landslides and Glacial Lake Outburst Floods (GLOFs). Under climate change, the frequency of natural hazards is expected to increase, posing further threats to society and human lives in HMA. Due to the lack of meteorological data, the triggering mechanisms of atmospherically induced natural hazards, especially landslides, are still not fully understood in HMA, which hinders the development of early-warning systems. To overcome this issue, a new atmospheric data set: the High Asia Refined analysis version 2 (HAR v2), was developed and is presented in this thesis. The HAR v2 was generated by dynamical downscaling of ERA5 reanalysis data using the Weather Research and Forecasting Model (WRF). The HAR v2 provides atmospheric data at 10 km grid spacing and hourly temporal resolution. It is currently available from 2000 to 2020 and will be extended back to 1979. Compared to the old version, the HAR v2 covers a broader area and a longer temporal range. To find the optimal model configuration of the HAR v2, several sensitivity experiments were conducted. Validation of the HAR v2 against in-situ stations from the Global Surface Summary of the Day (GSOD) shows that the HAR v2 fits well with observations and outperforms its forcing data ERA5. In addition, the HAR v2 and a version of the HAR v2 run with 2 km grid spacing (HAR v2 2 km) were compared with other commonly used gridded precipitation data sets (reanalysis data, satellite retrieval, and interpolated in-situ observations), over a sub-region in HMA with rugged terrain. Results indicate the added value of the HAR v2 and HAR v2 2 km since they are the only products that can reproduce orographic precipitation and capture more extreme events. One of the primary goals of the HAR v2 is to provide atmospheric data for landslide researches in Kyrgyzstan and Tajikistan, one of the landslide hot spots in the HMA region. The newly developed HAR v2 was combined with historical landslide inventories to investigate the atmospheric triggering mechanisms of landslides in this region. Results reveal the crucial role of snowmelt in landslide triggering in this region and the added value of climatic disposition derived from atmospheric triggering conditions in landslide susceptibility mapping. Furthermore, the majority of previous studies applied rainfall estimates from in-situ gauges or satellite retrievals. This study also highlights the potential of dynamical downscaling products generated by regional climate models in landslide prediction. Dynamical downscaling has already been extensively applied to understand the present-day climate and also future climate. In the last part of this thesis, the applicability of dynamical downscaling in the context of paleoclimate is demonstrated. Here, two global climate simulations for the present day and the mid-Pliocene ( ∼ 3 Ma) were dynamically downscaled to 30 km grid spacing over HMA, using the same model configuration as the HAR v2. By keeping land surface conditions the same in both downscaling experiments, this study was able to isolate the influence of large-scale climate states and reveal its role in maintaining the Qaidam mega-lake system during the mid-Pliocene. An increase of the water balance (∆S), i.e., the change in terrestrial water storage was found, when the mid-Pliocene climate is imposed on the Qaidam Basin (QB) with its modern land surface settings. This imbalance of ∆S induced solely by the changes in large-scale climate state would lead to an increase of lake extent until a new equilibrium state is reached. The estimated equilibrium lake extent is 12%-21% of the maximum lake extent approximated from proxy data.

Author dc:creator

• Wang, Xun

Advisor dc:contributor.advisor

• Scherer, Dieter

Licence dc:rights.uri

• https://creativecommons.org/licenses/by/4.0/

Language dc:language.iso

en

Identifier URI

http://dx.doi.org/10.14279/depositonce-14933

OAI identifier oai:identifier

oai:depositonce.tu-berlin.de:11303/16159