- Our Research
As a research specialist in the past few years, I considered my role as a linkage for climate model development between RCEC and other research centers, between software engineer and scientist, and between model output and scientific understandings. Through the collaboration with RCEC colleagues, domestic and international researchers and scientists, my research activities focus on three aspects: (1) implementation and development of high resolution GCM for climate research; (2) utilization of high resolution GCM for climate simulations and projections; and (3) implementation and development of FV3-based weather forecast model. These research activities cover climate and weather, model applications and developments. Recently RCEC proposed the development of seamless unified model, a single model family of the atmosphere used across a range of temporal and spatial scales. The seamless unified model is suitable for numerical weather prediction, seasonal forecasting and climate modelling with forecast times ranging from a few days to hundreds of years. Furthermore, seamless unified model can be used for air pollution research with additional consideration of chemical transport and reaction. My recent and future research interests focus on the development of Taiwan’s indigenous seamless unified model.
Arakane, S., H.-H. Hsu, C.‑Y. Tu, H.-C. Liang, Z.-Y. Yan, S.-J. Lin, (2019): Remote Effect of a Tropical Cyclone in the Bay of Bengal on a Heavy Rainfall Event in Subtropical East Asia. Clim Atmos Sci, Vol.2(25), DOI: 10.1038/s41612-019-0082-8
Chen, C., H.-H. Hsu, Hong, C, P.-G. Chiu, C.‑Y. Tu, S.-J. Lin, A. Kitoh, (2019): Seasonal precipitation change in the Western North Pacific and East Asia under global warming in two high‑resolution AGCMs. Clim Dyn, Vol. 53, pp.5583–5605, DOI: 10.1007/s00382-019-04883-1
Tseng, Wan-Ling., Huang-Hsiung Hsu, Noel S. Keenlyside, Chiung-Wen June Chang, Ben-Jei Tsuang, Chia-Ying Tu and Li-Chiang Jiang, (2017): Effects of Surface Orography and Land–Sea Contrast on the Madden–Julian Oscillation in the Maritime Continent: A Numerical Study Using ECHAM5-SIT, J. Climate, Vol.30, pp9725-9741, DOI: 10.1175/JCLI-D-17-0051.1
Freychet, N., H.-H. Hsu, A. Duchez and C.-Y. Tu, (2017): Projection in snowfall characteristics over the European Alps and its sensitivity to the SST changes: results from a 50 km resolution AGCM. Atmos. Sci. Let., Vol.18, pp261-267, DOI: 10.1002/asl.751
Lucas M. Harris, Shian-Jiann Lin and ChiaYing Tu, (2016): High-Resolution Climate Simulations Using GFDL HiRAM with a Stretched Global Grid, J. Climate, Vol.29, pp4293-4314, DOI: 10.1175/JCLI-D-15-0389.1
Tu, C.-Y.; and B.-J. Tsuang, (2005): Cool-skin simulation by a one-column ocean model, Geophys. Res. Lett., 32, L22602, doi:10.1029/2005GL024252..
Tsuang, B.-J. and Tu, C.-Y., (2002): Model structure and land parameter identification: an inverse approach. J. Geophys. Res. 107(D10), 10.1029/2001JD000711, ACL 15
Techniques & Development
Implementation and development of high resolution GCM for climate research HiRAM, the High-Resolution Atmospheric Model developed by GFDL, utilizes the finite-volume dynamical core using a cubed-sphere grid topology (FV3). Through cooperation between RCEC and GFDL in the past few years, I implemented HiRAM on different computing platforms in Taiwan. Other than the model implementations I also worked on the development of the variable-horizontal-resolution GCM with stretched-grid method. The great advantage of applying stretched-grid HiRAM is the enhancement of horizontal resolution without adding additional grids and extra computing costs. This makes HiRAM a powerful GCM for investigating synoptic and mesoscale phenomena. Another model development is the coupling of HiRAM with the one-dimensional ocean model SIT (Snow/Ice/Thermocline). The initiative motivation of developing HiRAM-SIT is to utilize the advantage of high resolution in both horizontal and vertical dimensions. This unique model structure of HiRAM-SIT can be applied to toggle the climatic characters in the vicinity of Taiwan.
Reference: Harris et al., 2016; Tu and Tsuang, 2005
Utilization of high resolution GCM for climate simulations and projections Utilizing GCM for long-term high-resolution climate simulations and projections for climate change study is one of the major activities for RCEC climate research. My colleagues and I designed a series of time-sliced climate experiments that HiRAM is forced using three sets of prescribed SSTs (sea surface temperature) and SICs (sea ice concentration) for three periods of present time (1980~2015), near future (2040~2065), and end of century (2075~2100) as the lower boundary conditions. In the past climatologists in Taiwan acquiring data of climate experiments from abroad was frustrated by the limitation of large amount data transfer. HiRAM time-slice multi-initial-condition ensembles are the first high resolution climate simulations and projections produced on local computing platforms. It provides opportunity for domestic researchers to explore regional climate in global climate system. Domestic scientists either directly utilize HiRAM data or further downscale with regional climate model driven by HiRAM output.
Reference: Chen et al., 2019; Freychet et al., 2017
Implementation and development of FV3-based weather forecast model Since 2016 I started to work on the implementation and development of FV3-based weather forecast system in Taiwan. Later from 2017 I started to cooperate with Central Weather Bureau (CWB) on implementing fv3GFS, the global weather forecast system with cubed-sphere dynamical core and NCEP GFS physics, on CWB’s Fujitsu supercomputer. Cooperation between RCEC and CWB on the implementation and development of fv3GFS can be distinguished according to respective expertise. While CWB focuses on the operational maintenance and data assimilation (DA) development, RCEC focuses on model performance enhancement and horizontal resolution refinement for the need of Taiwan vicinity. For the application of fv3GFS in Taiwan, I further introduced the two-way nested domain in Taiwan area. The model setting in the nested domain is individual from its setting in the global domain. This can reduce the disadvantage of resolution-dependent physics in the stretched-grid coordinate.
Reference: Arakane et al., 2019.