동아시아 지역의 위성 토양수분 데이터 활용을 위해 Soil Moisture Ocean Salinity (SMOS) 위성에 탑재된 Microwave Imaging Radiometer using Aperture Synthesis (MIRAS) 센서와 Global Change Observation Mission-Water (GCOM-W1) 위성에 탑재된 Advanced Microwave Scanning Radiometer 2 (AMSR2) 센서 기반 토양수분 데이터를 자료동화 데이터인 Global Land Data Assimilation System (GLDAS)를 기준 값으로 Cumulative Distribution Function (CDF) 기법과 회귀식을 활용하여 보정하는 연구를 수행하였다. 동아시아 지역에서 발생하는 전파간섭의 영향을 고려하여 토양수분 산출에 적합하다고 판단되는 Radio Frequency Interference (RFI), Data Quality indeX (DQX) 한계값과, 합성일수를 제시하였다. 보완된 위성 토양수분 데이터를 지점 토양수분 데이터와 비교한 결과 상관계수가 평균 27%, 11% 증가하였고, Root Mean Square Deviation (RMSD, 평균제곱근 편차)는 평균 61%, 57% 감소하였다. 추가적으로, 보정된 위성데이터를 GLDAS 토양수분 데이터와 비교했을 때, 보정된 MIRAS 및 AMSR2 데이터는 한반도의 80% 및 90%의 지역에서 상관계수가 증가하였으며, 한반도 전역에서 RMSD가 감소하였다. 본 연구를 통해 향후 MIRAS 및 AMSR2 위성 데이터를 융합하여 각 위성의 토양수분 데이터를 보완 할 수 있는 가능성을 제시하였다.
In this study the Microwave Imaging Radiometer using Aperture Synthesis (MIRAS) sensor onboard the Soil Moisture Ocean Salinity (SMOS) and Advanced Microwave Scanning Radiometer 2 (AMSR2) sensor onboard the Global Change Observation Mission-Water (GCOM-W1) based soil moisture retrievals were revised to obtain better accuracy of soil moisture and higher data acquisition rate over East Asia. These satellite-based soil moisture products are revised against a reference land model data set, called Global Land Data Assimilation System (GLDAS), using Cumulative Distribution Function (CDF) matching and regression approach. Since MIRAS sensor is perturbed by radio frequency interferences (RFI), the worst part of soil moisture retrieval, East Asia, constantly have been undergoing loss of data acquisition rate. To overcome this limitation, the threshold of RFI, DQX, and composite days were suggested to increase data acquisition rate while maintaining appropriate data quality through comparison of land surface model data set. The revised MIRAS and AMSR2 products were compared with in-situ soil moisture and land model data set. The results showed that the revising process increased correlation coefficient values of SMOS and AMSR2 averagely 27% 11% and decreased the root mean square deviation (RMSD) decreased 61% and 57% as compared to in-situ data set. In addition, when the revised products’ correlation coefficient values are calculated with model data set, about 80% and 90% of pixels’ correlation coefficients of SMOS and AMSR2 increased and all pixels’ RMSD decreased. Through our CDF-based revising processes, we propose the way of mutual supplementation of MIRAS and AMSR2 soil moisture retrievals.
1. 서 론 2. 연구지역 3. 자료 및 분석방법 4. 결과정리 5. 결 론 References
