표제지
연구보고서
목차
Contents 5
요약문 21
Abstract 26
제1장 서론 31
제1절 연구의 배경 31
제2절 연구의 필요성 31
제3절 국외 기술개발 현황 40
제2장 2006년 ARGO 플로트 현장관측 44
제1절 동해 현장 관측 44
제2절 북서태평양 현장 관측 47
제3장 ARGO 자료 및 기후평균 해양 자료 등 통합 시스템 구축 51
제1절 개요 51
제2절 표출시스템 개선 53
제3절 RTQC 시스템 개선 56
제4장 해양 순환 모델 연구 58
제1절 전구 해양 순환 모델 연구 58
제2절 지역 해양 순환 모델 연구 68
제5장 ARGO 자료 처리 기법 개발 연구 75
제1절 ARGO 자료처리 방법 자동화 75
제2절 기상연구소 ARGO 자료의 Quality Control 86
제3절 해양의 열용랑 변동 분석 88
제4절 인공위성 SST와 ARGO float로부터 얻은 혼합층 수온 비교연구 94
제5절 결론 및 요약 101
제6장 장기 기상예측에 대한 해양 자료동화 영향 분석 103
제1절 서언 103
제2절 접합대순환 모형에 의한 자료동화 반응 실험 105
제3절 자료동화에 따른 접합대순환 모형의 예측성 향상 실험 122
제4절 해양 자료동화 방안에 관한 비교 188
제5절 자료동화된 해양대순환 모형의 운동장 평가 204
제6절 결언 209
제7장 한반도 주변 지역 해양 모사 개선 연구 211
제1절 서언 211
제2절 3차원 지역 해양 순환 모델 실험 212
제3절 한반도 주변 수온의 연 변화 227
제4절 Ensemble Kalman Filter를 이용한 자료동화 231
제8장 요약 및 향후 계획 244
제1절 요약 244
제2절 향후 계획 247
제9장 참고문헌 249
Table 2.1. Field map of 2006-ARGO floats in East Sea. 46
Table 2.2. Field map of 2006-ARGO floats in Northwestern Pacific. 48
Table 3.1. Structure of the system directory. 52
Table 3.2. Web menu of the ARGO system. 53
Table 5.1. Configuration Parameter for Wong's DMQC. 78
Table 5.2. Coefficient for Automatic Quality Control Test in the East/Japan Sea. 84
Table 5.3. Coefficient for Automatic Quality Control Test in the North Pacific. 84
Table 5.4. Statistics of ARGO float data before and after quality control. 86
Table 6.1. Hit rate and False alarm rate of sea temperature for warm events (+), near normal condition (0) and cold events (-) for July start Run. 153
Table 6.2. Hit rate and False alarm rate of sea temperature for warm events (+), near normal condition (0) and cold events (-) for October start Run. 154
Table 6.3. Hit rate and False alarm rate of air temperature for warm events (+), near normal condition (0) and cold events (-) for July start Run. 161
Table 6.4. Hit rate and False alarm rate of air temperature for warm events (+), near normal condition (0) and cold events (-) for October start Run. 162
Table 6.5. Comparison of time needed to 100 times iteration and matrix space for each method. 194
Table 7.1. Experiment cases. 236
Figure 1.1. A plot of Optimal Mapping for salinity data from KODC. Before salinty QC (left), After salinity QC (right). 32
Figure 1.2. Abnormal temperature and salinity data shown in the ARGO data of the North Pacific. 33
Figure 1.3. An example of salinity drift in the ARGO data of the North Pacific. 34
Figure 1.4. Results of quality control in the East Sea. Before QC (left). after QC (right). 35
Figure 2.1. Deployment position of five East Sea floats. 45
Figure 2.2. Capture screen of uplink test for 25822 float. 46
Figure 2.3. Deployment position of 10 Northwestern Pacific floats. 49
Figure 2.4. Deployment ARGO floats on VOS in 2006 year. 50
Figure 3.1. Web page of the model introduction. 54
Figure 3.2. Display of the model results to control the parameters (temperature, salinity and sea level hight), top is global and bottom is regional. 55
Figure 3.3. DB of the Climate data (Levitus, WOA, HR wind, COADS). 56
Figure 4.1. Map of bottom topography. 59
Figure 4.2. Annual mean (a) river discharge (kg/m2*s), (b) sea ice thickness (m), (c) maximum speed of the M2 tidal component (cm/s), and chlorophyll concentration (mg/m3), which are used in MOM4 as surface boundary conditions. 60
Figure 4.3. Global sea surface temperature (left panel) and salinity (right panel) simulated by MOM4. 61
Figure 4.4. Difference of sea surface temperature (left panel) and salinity (right panel) with WOA01 climatology. Units are in ℃ and psu, respectively. 61
Figure 4.5. Left panels show mean distributions of (a) zonal geostrophic flow(cm/s), temperature (℃), and salinity (psu) between Hawaii and Tahiti (about 150˚W and 160˚W), frow April 1979 to March 1980 (After Wyrtki and Kilonsky (1984)). Rights are simulated results from MOM4. 62
Figure 4.6. Left panels show vertical sections of observed temperature and salinity along 137˚E in May 2002 (After Hosoda and Minato (2003)). Rights are simulated results from MOM4. 63
Figure 4.7. Meridional WOCE section (A16) of potential temperature and salinity through the Atlantic Ocean (After Siedler et al. (2001)). 64
Figure 4.8. The same as Figure 4.7. except for P15 section. 64
Figure 4.9. Temperature-Salinity diagram in 10˚ × 10˚ grid of North Pacific Ocean in winter and summer season, respectively. Blue lines are optimal estimates based on observed ARGO data, green lines are climatological Levitus data, and red lines are simulated results from MOM4. 66
Figure 4.10. ROMS model description and daily mean RDAPS wind field in August. 69
Figure 4.11. Surface currents, temperature, zeta and salinity in 2005. 71
Figure 4.12. Surface currents and temperature with NCEP monthly mean wind. 73
Figure 4.13. Surface currents and temperature with RDAPS daily mean wind. 73
Figure 4.14. Time series of measured SST (KMA's buoy) and model results. 74
Figure 5.1. High resolution CTD stations from CREAMS experiment. (a) Data used for salinity calibration from July 1993 to March 1999. (b) Data used for validation from July 1999 to April 2004. 77
Figure 5.2. Flow chart of delayed mode quality control for ARGO float data. 78
Figure 5.3. Temperature (a) and salinity profiles (b) averaged within 145~160˚E and 35~45 ˚N in the North Pacific (red) and 127~142 ˚E 35~45 ˚E in the East/Japan Sea (blue). The broken lines present the thestandard deviations of the temperature and the salinity profiles. 81
Figure 5.4. Flow chart for automatic quality control. 82
Figure 5.5. Effects of spike data on salinity calibration by Wong's method. (a) Salinity offset calibration before removing spike. Blue line:un-calibrated salinity of float ID 229 on 0.4 ℃ isothermal surface. Red line:salinity mapped onto the float track from high resolution CTD data. Green line:calibrated salinity. (b) Salinity offset calibration after removing spike. (c) T-S diagram from 239 float before removing spike. (d) T-S diagram after removing spike. 85
Figure 5.6. Delayed Mode Quality Control for the floats in the Ease Sea. (a) station map, (b) raw temperature profiles, (c) raw salinity profiles, (d) DMQC temperature profiles, and (e) DMQC raw salinity profiles. 87
Figure 5.7. Delayed Mode Quality Control for the floats in the North Pacific. (a) station map, (b) raw temperature profiles, (c) raw salinity profiles, (d) DMQC temperature profiles, and (e) DMQC raw salinity profiles. 87
Figure 5.8. Annual means (solid line) and seasonal means (dots) of the number of ARGO profiles over tile North Pacific. 89
Figure 5.9. Annual means (solid line) and seasonal means (dots) of heat contents over the North Pacific. 90
Figure 5.10. Annual means of heat contents (*2.5e+10 J/(m^2)) of every 10 degree longitude and 5 degree latitude over the North Pacific in 2002 (upper left), 2003 (upper right), 2004 (lower left), and 2005 (lower right). 91
Figure 5.11. Seasonal means of heat contents (*2.5e+10 J/(m^2)) of every 10 degree longitude and 5 degree latitude over the North Pacific in 2005 (winter (JAN.~MAR.):upper left, spring (APR.~JUN.):upper right, summer (JUL.~SEP.):lower left, autumn (OCT.~DEC.):lower right). 91
Figure 5.12. Annual means (solid line) and seasonal means (dots) of heat contents over the East Sea. 92
Figure 5.13. Annual means of heat contents (*1.5e+10 J/(m^2)) over the southwestern part of East Sea during 2000 to 2005. 93
Figure 5.14. The vertical profile of SST according to day and night. 95
Figure 5.15. Wind-induced skin-bulk temperature difference during daytime (upper panel) and nighttime (lower panel). 95
Figure 5.16. Hourly variation of skin-bulk SST difference as a function of wind speed. 96
Figure 5.17. Distribution of the SST errors (AMSR_E/SST-ARGO/MLT) as a function of wind speed (m/s) for (a) ascending pass and (b) descending pass. 98
Figure 5.18. Predicted response of SST difference to changes in wind speed (m/s) estimated from the model of Saunder (1967). 99
Figure 5.19. Wind speed averaged over the period 1999-2006 from QuikSCAT. 100
Figure 5.20. Frequency probability in % of weak wind speed of less than 6m/s during the period 1999-2006 from QuikSCAT. 100
Figure 6.1. A schematic diagram of data assimilation procedure. 106
Figure 6.2. Global Sea Temperature climatology fields of the observation from OISST (OBS), the analysis field (DA), the background field (noDA), and its differences between observation (DA-OBS, noDA-OBS) for July 1st. 110
Figure 6.3. Global Sea Temperature climatology fields of the observation from OISST (OBS), the analysis field (DA), the background field (noDA), and its differences between observation (DA-OBS, noDA-OBS) for October 1st. 111
Figure 6.4. Equatorial Pacific sea temperature climatology fields of the observation from OISST (OBS), the analysis field (DA), the background field (noDA), and its differences between observation (DA-OBS, noDA-OBS) for July 1st. 113
Figure 6.5. Equatorial Pacific sea temperature climatology fields of the observation from OISST (OBS), the analysis field (DA), the background field (noDA), and its differences between observation (DA-OBS, noDA-OBS) for October 1st. 114
Figure 6.6. Cross section of East-west sea temperature climatology along the equator for observation (OBS), analysis field (DA), background field (noDA) and its differences between observation (DA-OBS, noDA-OBS) for July 1st. 115
Figure 6.7. Cross section of East-west sea temperature climatology along the equator for observation (OBS), analysis field (DA), background field (noDA) and difference between observation (DA-OBS, noDA-OBS) for October 1st. 116
Figure 6.8. Cross section of Worth-south sea temperature climatology along 140W for observation (OBS), analysis field (DA), background field (noDA) and difference between observation (DA-OBS, noDA-OBS) for July 1st. 117
Figure 6.9. Cross section of North-south sea temperature climatology along 140W for observation (OBS), analysis field (DA), background field (noDA) and difference between observation (DA-OBS, noDA-OBS) for October 1st. 118
Figure 6.10. Northeast Asia sea temperature climatology fields of the observation from OISST (ObS), the analysis field (DA), the background field (noDA), and its differences between observation (DA-OBS, noDA-OBS) for July 1st. 120
Figure 6.11. Northeast Asia sea temperature climatology fields of the observation from OISST (OBS), the analysis field (Da), the background field (noDA), and its differences between observation (DA-OBS, noDA-OBS) for October 1st. 121
Figure 6.12. Monthly sea temperature climatology fields of observations (OBS), analysis fields (DA), and background fields (noDA) after 1~6 month of Run (July start). 123
Figure 6.13. Monthly sea temperature climatology fields of observations (OBS), analysis fields (DA), and background fields (noDA) after 1~6 month of Run (October start). 125
Figure 6.14. Monthly sea temperature climatology difference fields of analysis fields (DA-OBS), and background fields (noDA-OBS) after 1~6 month of Run (July start). 127
Figure 6.15. Monthly sea temperature climatology difference fields of analysis fields (DA-OBS), and background fields (noDA-OBS) after 1~6 month of Run (October start). 128
Figure 6.16. Monthly air temperature climatology fields of observations (OBS), analysis fields (DA), and background fields (noDA) after 1~6 month of Run (July start). 130
Figure 6.17. Monthly air temperature climatology fields of observations (OBS), analysis fields (DA), and background fields (noDA) after 1~6 month of Run (October start). 132
Figure 6.18. Monthly air temperature climatology difference fields of analysis fields (DA-OBS), and background fields (noDA-OBS) after 1~6 month of Run (July start). 134
Figure 6.19. Monthly air temperature climatology difference fields of analysis fields (DA-OBS), and background fields (noDA-OBS) after 1~6 month of Run (October start). 135
Figure 6.20. Monthly sea temperature correlations between analysis fields and observations (DA&OBS), and background fields and observations (noDA&OBS) after 1~6 month of Run (July start). 137
Figure 6.21. Monthly sea temperature correlations between analysis fields and observations (DA&OBS), and background fields and observations (noDA&OBS) after 1~6 month of Run (October start). 139
Figure 6.22. Monthly sea temperature correlations between assimilated results and observation (DA&OBS), and background and observation (noDA&OBS) during 6 month of Run (July start) over equatorial Pacific region (Latitude 0). 141
Figure 6.23. Monthly sea temperature correlations between assimilated results and observation (DA&OBS), and background and observation (noDA&OBS) during 6 month of Run (October start) over equatorial Pacific region (Latitude 0). 142
Figure 6.24. Monthly air temperature correlations between analysis field and observations (DA&OBS), and background field and observations (noDA&OBS) after 1~6 month of Run (July start). 144
Figure 6.25. Monthly air temperature correlations between analysis field and observations (DA&OBS), and background field and observations (noDA&OBS) after 1~6 month of Run (October start) over equatorial Pacific region. 146
Figure 6.26. Monthly air temperature correlations between assimilated results and observation (DA&OBS), and background and observation (noDA&OBS) during 6 month of Run (July start) over equatorial Pacific region (Latitude 0). 148
Figure 6.27. Monthly air temperature correlations between assimilated results and observation (DA&OBS), and background and observation (noDA&OBS) during 6 month of Run (October start) over equatorial Pacific region (Latitude 0). 149
Figure 6.28. Time series of air temperature correlation coefficient (a) and sea temperature correlation coefficient (b) over equatorial tropics (Lon=180~150W, Lat=5S~5N) for July start Run. 150
Figure 6.29. Time series of air temperature correlation coefficients (a) and sea temperature correlation coefficients (b) over equatorial tropics (Lon=180~150W, Lat=5S~5N) for October start Run. 151
Figure 6.30. Hit rates (a) and False Alarm rates (b) of sea temperature for July start Run. 156
Figure 6.31. Hit rates (a) and False Alarm rates (b) of sea temperature for October start Run. 157
Figure 6.32. Heidke scores of sea temperature for July start Run. 158
Figure 6.33. Heidke scores of sea temperature for October start Run. 159
Figure 6.34. Hit rates (a) and False Alarm rates (b) of air temperature over equatorial Pacific (Lon:180~150W, Lat:5S~5N) for July start Run. 163
Figure 6.35. Hit rates (a) and False Alarm rates (b) of air temperature over equatorial Pacific (Lon:180~150W, Lat:5S~5N) for October start Run. 164
Figure 6.36. Heidke scores of air temperature for July start Run. 165
Figure 6.37. Heidke scores of air temperature for October start Run. 166
Figure 6.38. Sea temperature climatology field over the northeast Asia of observation (a), analysis field (b), background field (c) for SON (July start Run). 168
Figure 6.39. Sea temperature climatology field over the northeast Asia of observation (a), analysis field (b), background field (c) for DJF (October start Run). 169
Figure 6.40. Standard deviation distributions of sea temperature over northeast Asia for SON (July start Run). 170
Figure 6.41. Standard deviation distributions of sea temperature over northeast Asia for DJF (October start Run). 171
Figure 6.42. Air temperature climatology field over the northeast Asia of observation (a), analysis field (b), background field (c) for SON (July start Run). 173
Figure 6.43. Air temperature climatology field over the northeast Asia of observation (a), analysis field (b), background field (c) for DJF (October start Run). 174
Figure 6.44. Standard deviation distributions of air temperature over northeast Asia for SON (July start Run). 175
Figure 6.45. Standard deviation distributions of air temperature over northeast Asia for DJF (October start Run). 176
Figure 6.46. Monthly sea temperature correlations between analysis field and observations (DA&OBS), and background field and observations (noDA&OBS) after 1~6 month of Run (July start) over the northeast Asia. 178
Figure 6.47. Monthly sea temperature correlations between analysis field and observations (DA&OBS), and background field and observations (noDA&OBS) after 1~6 month of Run (October start) over the northeast Asia. 180
Figure 6.48. Monthly air temperature correlations between analysis field and observations (DA&OBS), and background field and observations (noDA&OBS) after 1~6 month of Run (July start) over the northeast Asia. 182
Figure 6.49. Monthly air temperature correlations between analysis field and observations (DA&OBS), and background field and observations (noDA&OBS) after 1~6 month of Run (October start) over the northeast Asia. 184
Figure 6.50. Taylor diagram of sea temperature for July start (a) and October start (b) Run over the northeast Asia. 187
Figure 6.51. Time series of background error cost function (Jb), observation error cost function (Jo) and analysis error cost function (J) of VAN and VAF for each iteration processes. 193
Figure 6.52. Temperature difference between observation and analysis field (VAN), background field, and different VAFs (VAF-filter:3, VAF-filter:10, VAF-filter:15) for January 2004. 195
Figure 6.53. Salinity difference between observation (Levitus) and analysis field (VAN), background field, and different VAFs (VAF-filter:3, VAF-filter:10, VAF-filter:15) for 2004 January. 197
Figure 6.54. East-west sea temperature cross sections along equator of observation from analysis field (VAN), background field, and different VAFs (VAF-filter:3, VAF-filter:10, VAF-filter:15) for 2004 January. 198
Figure 6.55. North-south sea temperature cross sections along 160E from observation, analysis field (VAN), background field, and different VAFs (VAF-filter:3, VAF-filter:10, VAF-filter:15) for 2004 January. 200
Figure 6.56. East-west salinity cross sections along equator from observation (Levitus), analysis field (VAN), background field, and different VAFs (VAF-filter:3, VAF-filter:10, VAF-filter:15) for 2004 January. 201
Figure 6.57. North-south salinity cross sections along 160E from observation, analysis field (VAN), background field, and different VAFs (VAF-filter:3, VAF-filter:10, VAF-filter:15) for 2004 January. 202
Figure 6.58. 3-month mean meridional streamline for Jan~Mar (JFM), Apr~Jun (AMJ), Jul~Sep (JAS), Oct~Dec (OND) of 2004 along 140W. Upper is for background field, and lower analysis field. 206
Figure 6.59. Vertical latitudinal cross section of 3~month averaged sea temperature along 140W for Jan~Mar of 2004. 207
Figure 6.60. Vertical longitudinal cross section of 3-month averaged sea temperature along equator for Jan~Mar of 2004. 208
Figure 7.1. Bottom topography of model area. Numbers indicate water depth. 214
Figure 7.2. Discharge of Changjiang and Huang He river. 216
Figure 7.3. Discharge of Yangze river from 1993 to 1995 (Yang, 2006). 216
Figure 7.4. Tidal ellipse and Amplitude of major tides (M2, S2, O1, K1) in northwestern Pacific. 217
Figure 7.5. Difference of wind speed between monthly mean wind and climate wind in February, May, August and November 1995. (unit:m/s) 219
Figure 7.6. Difference of wind speed between monthly mean of daily wind and climate wind in February, May, August and November 1995. (unit:m/s). 220
Figure 7.7. Comparison of daily mean wind speed (solid line), monthly mean wind speed (dashed line) in 1995 and climatology wind speed (solid-dot line). 221
Figure 7.8. Monthly mean temperature section along the 311 line of KODC (top panel) for monthly mean wind (upper panels) and daily mean wind (lower panels) in February, May, August and November 1995. (units:℃). 222
Figure 7.9. Comparison of horizontally mean sea surface temperature among satellite surface temperature and model results (red line:monthly wind, blue line:daily wind, filled circle:satellite data) in the Yellow Sea from January 1993 to December 2002. 224
Figure 7.10. Correlation of sea surface temperature taken by satellite and model result using daily mean wind in the Yellow Sea from January 1993 to December 2002. 225
Figure 7.11. Sea surface temperature difference between model results using daily mean wind with tide and without tide in February, May, August and November in 1995. Negative means surface temperature decrease by tide. 226
Figure 7.12. Time series of horizontally mean sea surface temperature (blue), solar radiation (cyan) and air temperature (red) in the Yellow Sea from January 1993 to December 2002. 227
Figure 7.13. Time series of sea surface temperature anomaly (blue), solar radiation anomaly (cyan) and air temperature anomaly (red) in the Yellow Sea from January 1993 to December 2002. 228
Figure 7.14. Correlation of sea surface temperature anomaly and air temperature anomaly in the Yellow Sea from January 1993 to December 2002. 229
Figure 7.15. Correlation of sea surface temperature anomaly and solar radiation anomaly in the Yeller Sea from January 1993 to December 2002. 230
Figure 7.16. Example of Schur Product for localization of background error covariance. 233
Figure 7.17. Schematic diagram for linearity test of model. 235
Figure 7.18. Result of linearity test with time. 235
Figure 7.19. Schematic diagram of the assimilation experiments. 236
Figure 7.20. Horizontal distributions of error covariance, (a) auto covariance respect to station M (blue cross) (b) auto covariance after localization (c) multivariate covariance of SST to SSH (d) multivariate covariance of SST to SSH after localization. 238
Figure 7.21. Sections of error covariance. (a) auto covariance respect to station M (blue cross) (b) auto covariance after localization (c) multivariate covariance of SST to SSH (d) multivariate covariance of SST to SSH after localization. 239
Figure 7.22. Comparisons of model results:(a) true, (b) RFE, (c) E32, (d) ASSH. 240
Figure 7.23. RMS error (thin line) and EnSP (dotted line) of (a) SSH and (b) SST for experiments REF, E32, VNLC, CNINF and ASSH. 243
Figure 8.1. Objective and schematic plan of ARGO project carried by METRl/KMA. 248