Assessing the Quality of Regional Ocean Reanalysis Data from ENSO Signals

ATMOSPHERIC AND OCEANIC SCIENCE LETTERS, 2012, VOL. 5, NO. 1, 55 61 Assessing the Quality of Regional Ocean Reanalysis Data from ENSO Signals WANG Lu 1, 2 and ZHOU Tian-Jun 1 1 State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China 2 Graduate University of the Chinese Academy of Sciences, Beijing 100029, China Received 22 June 2011; revised 12 September 2011; accepted 12 September 2011; published 16 January 2012 Abstract The quality of regional ocean reanalysis data for the joining area of Asia and the Indian-Pacific Ocean (AIPO) has been assessed from the perspective of ENSO-related ocean signals. The results derived from the AIPO reanalysis, including SST, sea surface height (SSH), and subsurface ocean temperature and currents, are compared with those of Hadley Center Sea Ice and Sea Surface Temperature (HadISST) data set and Simple Ocean Data Assimilation (SODA) reanalysis data. Both the spatial pattern and the characteristics of evolution of the ENSO-related ocean temperature anomalies are well reproduced by the AIPO reanalysis data. The physical processes proposed to explain the life cycle of ENSO, including the delayed oscillator mechanism, rechargedischarge mechanism, and the zonal advection feedback, are reasonably represented in this dataset. However, the westward Rossby wave signal in 1992 is not obvious in the AIPO data, and the magnitude of the heat content anomalies is different from that of the SODA data. The reason for the discrepancies may lie in the different models and methods for data assimilation and differences in wind stress forcing. The results demonstrate the high reliability of the AIPO reanalysis data in describing ENSO signals, implying its potential application value in ENSOrelated studies. Keywords: reanalysis data, ENSO, ocean assimilation, AIPO data Citation: Wang, L., and T.-J. Zhou, 2012: Assessing the quality of regional ocean reanalysis data from ENSO signals, Atmos. Oceanic Sci. Lett., 5, 55 61. 1 Introduction The joining area of Asia and Indian-Pacific Ocean (AIPO) refers to the ocean area that surrounds both the East Asian and South Asian continental area and that connects to both the Western Pacific and the Indian Ocean. AIPO is one of the key areas affecting the short-term climate variation over China. The air-sea interaction in the AIPO area is one of the key factors leading to the largescale, persistent or explosive meteorological disasters in China. Therefore, it is urgent that the characteristics of the air-sea interactions be explored and that the mechanisms in the AIPO area be explored (Wu et al., 2006). To accomplish this goal, a large amount of reliable re- Corresponding author: WANG Lu, wanglu2007@mail.iap.ac.cn analysis data covering the whole AIPO region is required. However, there are no specific synthesis data for this region. For example, regional ocean-assimilated data such as the Multivariate Ocean Variational Estimation (MOVE) (Fujii and Kamachi, 2003) and the China Oceanic Re- Analysis (CORA) (Han et al., 2011) only focus on small areas of the northwestern Pacific and the coastal region of China, which are not sufficient for the stated purpose. The only data that can be used are several sets of global ocean reanalysis data, such as the Simple Ocean Data Assimilation (SODA) (Carton and Giese, 2008), the Global Ocean Data Assimilation System (GODAS) (Behringer and Xue, 2004), and the products from the Geophysical Fluid Dynamics Laboratory (GFDL) (Zhang et al., 2007). The Institute of Atmospheric Physics (IAP) has recently developed an ocean reanalysis system and delivered a set of high-resolution data for the region of AIPO region with the use of the Hybrid Coordinate Ocean Model (HYCOM) (Bleck and Smith, 1990) and the Ensemble Optimal Interpolation (ENOI) (Evensen, 2003) scheme (Yan et al., 2010). This combination is unique. The system uses new methods in the assimilation process. The assimilated data include the combined sea level anomaly (SLA) data from Collecte Localisation Satellites (CLS) with a resolution of 1/3 degree 1/3 degree Mercator grid. Also included are satellite-based SST data and the in-situ temperature and salinity profiles from the comprehensive data set ENSEMBLES (EN3 version 1d) (Ingleby and Huddlestion, 2007), which includes the World Ocean Database 2005 (WOD05), the Global Temperature-Salinity Profile Program (GTSPP), the Array for Real-time Geostrophic Oceanography (ARGO), and the Arctic Synoptic Basin-wide Oceanography (ASBO) project. The preliminary examination of this data demonstrated its good performance in describing the temperature variability and heat content warming trends (Yan et al., 2010), which implies that the data may be reliable and valuable for understanding the ocean s role in climate variability in the AIPO area. However, the ENSO variability has not yet been examined in this dataset, which is important and may be a good metric for evaluation. The objective of the present study is to make a detailed evaluation of the new ocean reanalysis data in depicting ENSO-related signals. This study includes an analysis of the spatial pattern and evolution fields of ocean temperatures and the primary dynamic processes of ENSO. The

56 ATMOSPHERIC AND OCEANIC SCIENCE LETTERS VOL. 5 organization of this paper is as follows. The datasets employed in this study are described in section 2. The performance of the AIPO data is described in section 3. Finally, concluding remarks are given in the last section. 2 Data description The data evaluated in this study (referred to as AIPO data) have a horizontal resolution of 3/4 degrees by 3/4 degrees, ranging from 51 S to 62 N and 30 E to 69 W. The period studied is from 1992 to 2006. The variables include SST, sea surface height (SSH), and subsurface ocean temperature and velocities. The observed SST data used for comparison are monthly SSTs from Hadley Center Sea Ice and Sea Surface Temperature data set (HadISST) (Rayner et al., 2003), which are independent of the Reynolds SSTs assimilated into the AIPO data. The subsurface and SSH data used for comparison are monthly data from SODA 2.0.2 (Carton and Giese, 2008), in which the altimetry data used for assimilation are the same as those of the AIPO data. The analysis periods are confined to the same period as the AIPO data. 3 Evaluations First, we compared the Niño indexes between the HadISST data and the AIPO data. These indexes show rather high correlations, with the Niño 3 index being 0.99, the Niño 4 index being 0.97, and the Niño 1+2 index being 0.97. Then, we regressed the SST anomalies (SSTAs) from the two datasets onto each standardized Niño 3 index, and we compared the spatial patterns of ENSO (Fig. 1). The SSTAs during the mature phase of El Niño exhibit a striking dipole with a cold polarity located at the western Pacific and a warm polarity in the equatorial eastern-central Pacific. From these polarity centers, anomalous cold and warm bands emanate poleward and eastward toward the extratropics. The anomalous cold zone in the western Pacific exhibits a horseshoe pattern. Note that the SST anomalies in the Indian Ocean basin are warmer Figure 1 Geographical distribution of the regression coefficients of the SST anomalies (units: K) onto the standardized Niño 3 index based on (a) HadISST SST and (b) AIPO SST.

NO. 1 WANG AND ZHOU: ENSO SIGNALS IN AIPO 57 than normal at the same time. In general, the AIPO SSTAs show a nearly identical pattern to the observation. However, the SST warming in the central equatorial Indian Ocean, the South China Sea, and off the western coast of North America is weaker than that of HadISST, and the SST cooling centered at (35 N, 170 W) is also weaker. The average evolution of the SST anomalies and the heat content anomalies in the upper ocean associated with ENSO between the reanalysis data from SODA and AIPO are compared in Fig. 2. In the SODA data (Fig. 2a), the equatorial SSTAs show a nearly stationary growth from the coast of South America to 160 E. The SSTAs develop slightly earlier in the west than in the east (both approximately 9 12 months prior to the event peak) and then decay eastward approximately 9 12 months after the peak. At the Niño 3 SSTA peak, there are cold anomalies in the Figure 2 Lead-lag regressions onto the standardized Niño 3 index of (a) SSTAs (units: K) averaged over 2 S 2 N based on SODA data. Solid (dashed) contours represent positive (negative) values. (b) Temperature anomalies (shaded, units: K) averaged over 2 S 2 N over the top 300 m of the ocean based on SODA data. The contours represent the zonal gradient of the temperature anomalies ( T / x, K m 1 ), and the max value of T / x is marked near its center. (c) Same as (a), but based on AIPO data. (d) Same as (b), but based on AIPO data.

58 ATMOSPHERIC AND OCEANIC SCIENCE LETTERS VOL. 5 far western Pacific. There are also cold SSTAs in the central Pacific, both before and after the warm event. In the subsurface (Fig. 2b), there is a slow eastward propagation of positive heat content anomalies from the equatorial central-western Pacific (135 E 135 W) starting approximately 18 months prior to the event peak. This process results in a reduction of the zonal slope of the thermocline at the peak of El Niño, and the maximum zonal gradient of the temperature anomalies T / x is near 180, with a value of 3.7 10 7 K m 1. The AIPO data (Figs. 2c and 2d) show nearly identical characteristics of the evolution of SSTAs and the heat content anomalies in the upper ocean. However, the source of the eastward positive heat content anomalies is only confined within the western Pacific (135 E 180 ), and the resultant ( T / x) is weaker than that in Fig. 2b. The discrepancies of the temperature evolution are related to those of the zonal current anomalies (u ), which will be shown in Figs. 4b and 4d. As proposed by Guilyardi et al. (2009), the metrics to assess the ENSO phenomenon involve not only the statistical characteristics but also a variety of physical processes responsible for ENSO s evolution. In the following, we made an assessment of the performance of the AIPO data in describing several key physical mechanisms of ENSO through a comparison with the reanalysis data derived from SODA. One of the primary mechanisms explaining the transition of ENSO is called the delayed oscillator, which stresses the importance of oceanic waves (Battisti, 1988; Schopf and Suarez, 1988; Suarez and Schopf, 1988; Battisti and Hirst, 1989). It is argued that the wind anomalies off the equator induced by SST perturbations in the central equatorial Pacific drive Rossby waves on the ocean thermocline, which propagate westward, and, on reaching the western boundary reflect into equatorial Kelvin waves propagating eastward. The Rossby wave can be identified by a Hovmoller diagram of SSH anomalies averaged between 5 N to 8 N (Figs. 3a and 3c), and the Kelvin wave can be identified between 2 S to 2 N (Figs. 3b and 3d). A prominent wave propagation case was found during the 1997/98 El Niño event. First, the upwelling Rossby wave (represented by the negative sea surface height anomalies) was induced in the central off-equatorial Pacific region and began to propagate westward in the summer of 1997 (Fig. 3a). Then at the end of 1997, the Rossby wave reached the western boundary, which in turn induced the equatorial upwelling Kelvin wave propagating eastward (Fig. 3b). The SSH anomalies derived from the AIPO dataset can capture the 1997/98 and 2002/03 cases, but with a weaker amplitude (Figs. 3c and 3d). However, the Rossby wave propagation in 1992 cannot be well represented. As the observed SLA data assimilated in the two products are the same, the reason for the discrepancies may lie in the difference of the models and methods for assimilation. In addition to the delayed oscillator mechanism, the recharge-discharge theory proposed by Jin (1997a, b) is another important mechanism explaining the life cycle of ENSO, and it can be verified through the SODA dataset (Fig. 4a). According to the recharge oscillator theory, during the developing phase, the strong positive heat content prevails in the equatorial region, and the strong negative heat content exists north of the equator. The magnitudes of both heat content anomalies are almost the same, which implies that the equatorial heat content comes from the off-equatorial region. After the peak phase of the El Niño event, the equatorial heat content is rapidly discharged so that a strong negative anomaly appears. At the same time, the zonal-mean heat content anomalies north of the equator develop with a similar magnitude as the equatorial one, but with the opposite sign, which implies that the equatorial heat content is transported to the off-equatorial region. Note that the heat content anomaly south of the equator has the same sign as the equatorial one, which means that the recharge-discharge mechanism only applies to the region between the equator and the region north of the equator, as mentioned by Kug et al. (2003). The AIPO dataset can represent the general characteristic of the heat content anomalies change both in the equatorial and off-equatorial area. However, the magnitude of the negative temperature anomalies at lag 3 around 10 N (Fig. 4d) is only half of that in Fig. 4a. This result is due to the discrepancies of the meridional current anomalies (v, see contour in Figs. 4a and 4d) between the two products. In the SODA data (Fig. 4a), a negative v exits within 0 and 10 N from lag 18 to lag 3, which transports the heat content southward continuously. In the AIPO data (Fig. 4d), however, the v is positive within 0 and 10 N from lag 18 to lag 12. The opposite sign of v would reduce the southward heat content transport from the off-equatorial region, resulting in the weaker negative heat content anomalies at lag 3 near 10 N. Apart from the above two mechanisms responsible for the ENSO evolution, the anomalous zonal advection (i.e., u T / x ) can also make a great contribution to the ENSO life cycle (Picaut et al., 1997; Jin and An, 1999). Figures 4b and 4d show the evolution of the equatorial u at the upper ocean associated with ENSO based on the SODA and AIPO datasets, respectively. It can be seen in Figs. 4b and 4e that there are eastward current anomalies prior to the warmest SST anomalies in the equatorial central eastern Pacific, and there are westward current anomalies after the event. Because the background zonal SST gradient is negative (i.e., T / x 0) in the equatorial Pacific, the eastward advection helps to warm the east, and the westward advection helps to restore the Pacific to equilibrium. Both datasets could represent the mechanism, but discrepancies in both sign and magnitude can be found between Figs. 4b and 4e. A positive u exits within 170 E and 135 W from lag 18 to 12 in Fig. 4b, whereas a negative u is shown in Fig. 4e. The opposite sign of u would result in the opposite sign of heat content anomalies within 180 and 135 W, as seen in Figs. 2b and 2d. In addition, the positive u at lag 3 near 180 in Fig. 4e is weaker than that in Fig. 4b, leading to the weaker zonal contrast of temperature anomalies at lag 0 (Fig. 2d).

NO. 1 WANG AND ZHOU: ENSO SIGNALS IN AIPO 59 Figure 3 Hovmoller diagram of SSH anomalies (units: m) across the Pacific (a) averaged in (5 8 N) based on SODA data; (b) same as (a), but for (2 S 2 N); (c) same as (a), but based on AIPO data; (d) same as (b), but based on AIPO data. Solid (dashed) contours represent positive (negative) values. To further quantify the difference of u between the two products, the time series averaged in three parts of the Equatorial Ocean were compared. The correlations of u between the two products at the western (2 S 2 N, 135 E 180 ) and central (2 S 2 N, 180 135 W) boxes are similar (0.70 vs. 0.72), while that at the eastern box (2 S 2 N, 90 135 W) is lower (0.63). The discrepancies of u can be partly attributed to the different wind stress forcing: the European Center for Medium-Range Weather Forecasts (ECMWF) six-hourly data used in AIPO, and the 40-yr ECMWF atmospheric reanalysis (ERA-40) data (Uppala et al., 2005) used in SODA. As seen in Figs. 4c and 4f, the westward wind stress anomalies within 170 E and 135 W from lag 18 to 12 used in AIPO are stronger than that used in SODA. Another reason may lie in the different ocean models and methods used for assimilation. 4 Conclusions This study evaluates the quality of AIPO regional ocean reanalysis data in terms of ENSO-related signals through comparisons with HadISST data and SODA reanalysis data. The variables assessed include SST, SSH, and subsurface ocean temperature and currents. The results show that the AIPO data captures most of the realistic features, such as the evolution of Niño indexes, the spatial pattern of SSTA and the primary physical processes responsible for the life cycle of ENSO (i.e., the zonal propagation of oceanic waves, the variation of heat content anomalies, and the evolution of the zonal current anomalies). However, discrepancies between the different reanalysis products can also be found. The zonal contrast of the heat content anomalies at the El Niño peak in

60 ATMOSPHERIC AND OCEANIC SCIENCE LETTERS VOL. 5 Figure 4 Lead-lag regressions onto standardized Niño 3 index of (a) temperature anomalies (shaded, units: K) and meridional current anomalies (contour, units: m s 1, interval is 0.0002 m s 1 ) averaged zonally across the Pacific over the top 300 m of the ocean based on SODA data; (b) zonal currents anomalies (units: m s 1 ) averaged over 2 S 2 N over the top 50 m of the ocean based on SODA data; (c) zonal wind stress anomalies (units: 10 3 N m 2 ) from ERA-40 data; (d) same as (a), but for AIPO data; (e) same as (b), but for AIPO data; (f) same as (c), but from ECMWF six-hourly forecast data. Solid (dashed) contours represent positive (negative) values. AIPO is weaker than that in SODA, and the magnitude of the negative heat content anomalies near 10 N prior to the El Niño peak in AIPO, are only half that in SODA. The reason lies in the differences in the ocean current anomalies, which are further due to the different wind stress forcing and different ocean models used for the assimilation. The westward Rossby wave propagation in 1992 can be captured by SSH in SODA, but cannot be well represented in AIPO. As the observed SLA data assimilated in the two data are the same, the reason for the discrepancies may lie in the difference of the models and methods for assimilation. In summary, the assessment shows that the AIPO reanalysis may be highly useful for ENSO-related studies, but the discrepancies between the different reanalysis data should also be noted. Acknowledgements. This study was jointly supported by the Knowledge Innovation Program of the Chinese Academy of Sciences (Grant No. KZCX2-YW-Q11-04), the Public Science and Technology Research Funds Projects of Ocean (Grant No. 201105019-3), and the National Basic Research Program of China (Grant No. 2010CB951904).

NO. 1 WANG AND ZHOU: ENSO SIGNALS IN AIPO 61 References Battisti, D., 1988: Dynamics and thermodynamics of a warming event in a coupled tropical atmosphere-ocean model, J. Atmos. Sci., 45, 2889 2919. Battisti, D. S., and A. C. Hirst, 1989: Interannual variability in a tropical atmosphere-ocean model: Influence of the basic state, ocean geometry and nonlinearity, J. Atmos. Sci., 46, 1687 1712. Behringer, D., and Y. Xue, 2004: Evaluation of the global ocean data assimilation system at NCEP: The Pacific Ocean, in: Eighth Symp. on Integrated Observing and Addimilation Systems for Atmosphere, Oceans, and Land Surface, 84th Annual Meeting, Amer. Meteor. Soc., Seattle, 11 15. Bleck, R., and L. Smith, 1990: A wind-driven isopycnic coordinate model of the north and equatorial Atlantic Ocean 1. Model development and supporting experiments, J. Geophys. Res., 95, 3273 3285. Carton, J. A., and B. S. Giese, 2008: A reanalysis of ocean climate using Simple Ocean Data Assimilation (SODA), Mon. Wea. Rev., 136, 2999 3017. Evensen, G., 2003: The ensemble Kalman filter: Theoretical formulation and practical implementation, Ocean Dyn., 53, 343 367. Fujii, Y., and M. Kamachi, 2003: Three-dimensional analysis of temperature and salinity in the equatorial Pacific using a variational method with vertical coupled temperature-salinity empirical orthogonal function modes, J. Geophys. Res., 108, 3297, doi:10.1029/2002jc001745. Guilyardi, E., A. Wittenberg, A. Fedorov, et al., 2009: Understanding El Niño in ocean-atmosphere general circulation models progress and challenges, Bull. Amer. Meteor. Soc., 90, 325 340. Han, G. J., W. Li, X. F. Zhang, et al., 2011: A regional ocean reanalysis system for coastal waters of China and adjacent seas, Adv. Atmos. Sci., 28, 682 690. Ingleby, B., and M. Huddleston, 2007: Quality control of ocean temperature and salinity profiles Historical and real-time data, J. Mar. Syst., 65, 158 175. Jin, F., and S. An, 1999: Thermocline and zonal advective feedbacks within the equatorial ocean recharge oscillator model for ENSO, Geophys. Res. Lett., 26, 2989 2992. Jin, F.-F., 1997a: An equatorial ocean recharge paradigm for ENSO. Part I: Conceptual model, J. Atmos. Sci., 54, 811 829. Jin, F.-F., 1997b: An equatorial ocean recharge paradigm for ENSO. Part II: A stripped-down coupled model, J. Atmos. Sci., 54, 830 847. Kug, J., I. Kang, and S. An, 2003: Symmetric and antisymmetric mass exchanges between the equatorial and off-equatorial Pacific associated with ENSO, J. Geophys. Res., 108, 3284, doi:10. 1029/2002JC001671. Picaut, J., F. Masia, and Y. Penhoat, 1997: An advective-reflective conceptual model for the oscillatory nature of the ENSO, Science, 277, 663 666. Rayner, N., D. Parker, E. Horton, et al., 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century, J. Geophys. Res., 108, 4407, doi:10.1029/2002jd002670. Schopf, P. S., and M. J. Suarez, 1988: Vacillations in a coupled ocean-atmosphere model, J. Atmos. Sci., 45, 549 566. Suarez, M., and P. Schopf, 1988: A delayed action oscillator for ENSO, J. Atmos. Sci., 45, 3283 3287. Uppala, S., P. Kållberg, A. Simmons, et al., 2005: The ERA-40 re-analysis, Quart. J. Roy. Meteor. Soc., 131, 2961 3012. Wu, G., J. Li, T. Zhou, et al., 2006: The key region affecting the short-term climate variations in China: The joining area of Asia and Indian-Pacific Ocean, Adv. Earth Sci. (in Chinese), 21, 1109 1118. Yan, C.-X., J. Zhu, and J.-P. Xie, 2010: An ocean reanalysis system for the joining area of Asia and Indian-Pacific Ocean, Atmos. Oceanic Sci. Lett., 3, 81 86. Zhang, S., M. J. Harrison, A. Rosati, et al., 2007: System design and evaluation of coupled ensemble data assimilation for global oceanic climate studies, Mon. Wea. Rev., 135, 3541 3564.