APEC 기후센터(APCC) 기후변화팀의 박경원 박사, 전종안 박사, 강광민 박사, 기후예측팀의 손수진 박사, 기후분석팀의 Hongwei Yang 박사가 오는 12월 3일에서 7일, 미국 샌프란시스코에서 개최되는 미국지구물리학회 가을 회의(2012 American Geophysical Union Fall Meeting)에 참석해 연구 분야를 발표한다.

이번 회의에 참석하는 APCC 연구원은 아래의 제목과 내용으로 연구를 제시한다.

► Dr. Kyungwon Park
Development of a Rainfall Retrieval Algorithm using COMS data

This study presents the description, results, validation, and progress of rainfall rate for Korea's first geostationary multi-purpose satellite COMS (Communication, Ocean and Meteorological Satellite), which is joint project by the Korea Meteorological Administration (KMA), the Ministry of Education, Science & Technology (MEST), the Ministry of Land, Transport and Maritime Affairs (MLTM), and the Korea Communications Commission (KCC), using rainfall retrieval estimation. The described technique uses the infrared (IR) band to compute real-time precipitation amounts based on a combined rainfall retrieval algorithm. This method is derived from statistical analysis of TRMM/PR-derived rainfall estimates and TRMM/VIRS-derived black body temperature, collocated in time and space. We examined the accuracy of the rainfall rate estimates for storm systems and applied the algorithm to a case study of heavy rainfall over the Korean peninsula.

► Dr. Jong Ahn Chun
Modeling the basin-scale water balance impact of different irrigation systems with the Land Information System

Food security can be improved by increasing the extent of agricultural land or by increasing agricultural productivity, including through intensive management practices, such as irrigation. The objectives of this study were to incorporate in-practice irrigation schemes into the land surface models of the NASA Land Information System (LIS) and to apply the tool to estimate the impact of irrigation on land surface states and fluxes—including evapotranspiration, soil moisture, and runoff—in the Murray-Darling basin in Australia. Here we present results obtained using the Noah Land Surface Model v3.2, housed within LIS, without simulated irrigation (IR0) and with three irrigation simulation routines: flood irrigation (IR1), drip irrigation (IR2), and sprinkler irrigation (IR3). The Moderate Resolution Imaging Spectrometer (MODIS) vegetation index was used to define crop growing seasons. Simulations were performed for a full year (July 2002 to June 2003) and evaluated against hydrologic flux estimates obtained in previous studies. Irrigation amounts during the growing season (August 2002 to March 2003) were simulated as 104.6, 24.6, and 188.1 GL for IR1, IR2, and IR3, respectively. These preliminary results showed water use efficiency would be highest from a drip irrigation scheme and lowest from a sprinkler irrigation scheme, with a highly optimized version of flood irrigation falling in between. Irrigation water contributed to a combination of increased evapotranspiration, runoff, and soil moisture storage in the irrigation simulations, relative to IR0. Implications for water management applications and for further model development will also be discussed.

► Dr. Kwangmin Kang
Analysis of Climate Change on Hydrologic Components by using Bayesian Neural Networks

Representation of hydrologic components under climate change is a challenging task. Regional climate models (RCMs) from general circulation models (GCMs) have difficulty representing hydrologic outputs due to several uncertainties in hydrologic impacts of climate change. To overcome this problem, this study presents practical options for hydrological climate change using Bayesian and Neural networks to approach regional adaption to climate change. Bayesian and Neural networks analysis to project climate hydrologic components is one of the new frontiers in considering the expected impacts of climate change. A strong advantage of Bayesian Neural networks is the ability to detect time series in hydrologic components, which is complicated due to data, parameter, and model hypothesis under climate change scenarios, through changing steps by removing and adding connections in the Neural network process that combine Bayesian concepts from the parameter, predict and update processes. As a case study, the Mekong River Watershed, which is surrounded by four countries (Myanmar, Laos, Thailand and Cambodia), was selected. Results will show an assessment of hydrologic component trends in climate model simulations through Bayesian Neural networks.

► Dr. Soo-Jin Sohn
Long-lead multi-model ensemble prediction for a drought index sensitive to global warming

Given the changing climate, advance information on hydrological extremes, such as droughts, will prove helpful in planning for disaster mitigation and facilitating better decision-making for water availability management. A precipitation deficit for long-term time scales beyond 6 months impacts hydrological variables, such as ground water, streamflow, and reservoir storage. However, there are difficulties in predicting such long-lead precipitation anomalies, especially over in-land extratropical areas, even using state-of-the-art multiple coupled model ensembles. The potential of prediction of long-lead hydrological variations based on climatic water balance with multi-coupled model statistics has been investigated. A multi-scalar hydrological index, based on both precipitation and temperature, i.e. the newly proposed standardized precipitation evapotranspiration index (SPEI), is used not only to appropriately define the hydrological extremes but also to consider the hydrological balance between precipitation and evapotranspiration. Furthermore, since it includes the role of temperature, it is sensitive to linear trends, such as global warming, and can properly respond its consequent extremes, unlike the standardized precipitation index (SPI). To predict long-lead, district-level, multi-model ensemble (MME)-based hydrological extremes, a six-month downscaled MME (DMME) prediction system was developed for 60 stations in South Korea. DMME, in conjunction with variance inflation, can generate predictions of hydrological extremes with reasonable skill, in terms of SPI and SPEI. The results could potentially improve hydrological extreme predictions using meteorological forecasts for policymakers and stakeholders in the water management sector for better climate adaption.

► Dr. Hongwei Yang
Reduction of Biases in Regional Climate Downscaling: Application of Bayesian Model Averaging on Large-scale Forcing

Simulations of the 1998 East Asian summer monsoon were carried out using the Weather Research and Forecast (WRF) model forced by four reanalysis datasets (NCEP-R2, ERA-40, ERA-IN, and JRA-25), their equally weighted ensemble mean, and their ensemble mean, based on Bayesian model averaging (BMA). Large discrepancies were found among experiments forced by individual reanalysis datasets. The uncertainties in the moisture field of large-scale forcing over oceans were responsible for the discrepancies. The control experiment, forced by the equally weighted ensemble forcing, reduced the biases in the simulated circulation to a large extent, but only reduced the biases in the simulated precipitation in some cases. The experiment forced by the BMA ensemble forcing outperformed not only the experiments forced by individual reanalysis datasets, but also the control experiment. The results suggest that the uncertainties in lateral boundary forcing can be reduced through the BMA ensemble method based on satellite data and more three-dimensional satellite data is urgently demanded.
 

* 미국지구물리학회 가을 회의(2012 AGU Fall Meeting)란?
미국지구물리학회 가을 회의는 20,000 명 이상의 지구와 우주 과학자, 교육자, 학생 및 기타 리더급 인사를 유치하는 세계적 컨퍼런스로 매년 미국 샌프란시스코에서 개최된다.