Global Upper Ocean Heat Content Estimation: Recent Progress and the Remaining Challenges

CHENG Li-Jing, ZHU Jiang, and John ABRAHAM

1International Center for Climate and Environment Sciences (ICCES), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China

2University of St. Thomas, St. Paul, MN, USA

Global Upper Ocean Heat Content Estimation: Recent Progress and the Remaining Challenges

CHENG Li-Jing1, ZHU Jiang1, and John ABRAHAM2

1International Center for Climate and Environment Sciences (ICCES), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China

2University of St. Thomas, St. Paul, MN, USA

Ocean heat content (OHC) change contributes substantially to global sea level rise, so it is a vital task for the climate research community to estimate historical OHC. While there are large uncertainties regarding its value, in this study, the authors discuss recent progress to reduce the errors in OHC estimates, including corrections to the systematic biases in expendable bathythermograph (XBT) data, filling gaps in the data, and choosing a proper climatology. These improvements lead to a better reconstruction of historical upper (0-700 m) OHC change, which is presented in this study as the Institute of Atmospheric Physics (IAP) version of historical upper OHC assessment. Challenges still remain; for example, there is still no general consensus on mapping methods. Furthermore, we show that Coupled Model Intercomparison Project, Phase 5 (CMIP5) simulations have limited ability in capturing the interannual and decadal variability of historical upper OHC changes during the past 45 years.

ocean heat content, temperature observation, XBT bias, mapping, climatology, climate change, global warming

1　Introduction

Ocean heat content (OHC), as a key indicator of the Earth's energy budget, provides a metric for the current ongoing global warming (Rhein et al., 2013). Since more than 90% of the energy of global warming is buried in th e ocean due to its large heat capacity, it is a vital task for the climate community to estimate the rate of historical OHC change (Abraham et al., 2013). However, due to the insufficient coverage of observations historically, most existing OHC estimates can only go back as far as the 1960s, and extend from the sea surface down to ～700 m only. Existing estimates of historical OHC change show substantial divergence, revealing uncertainties with regard to its value. For instance, IPCC AR5 (the Intergovernmental Panel on Climate Change Fifth Assessment Report; Rhein et al., 2013) provided five independent estimates of historical OHC change from 1970 to 2010 by five different international groups: 74TW (1 TW = 1012Watts) (Smith and Murphy, 2007); 98TW (Ishii and Kimoto, 2009); 108TW (Palmer et al., 2007); 118TW (Levitus et al., 2012); and 137TW (Domingues et al., 2008). Among these values, the minimum is as much as a half of the maximum, implying large divergence in the assessment of the ocean warming rate.

A large number of studies suggest that uncertainties during OHC calculations stem from the following issues: (1) Systematic bias in temperature measurements, such as expendable bathythermographs (XBTs) (Abraham et al., 2012a, b; Cheng et al., 2014; Gouretski and Koltermann, 2007; Gouretski and Reseghetti, 2010; Levitus et al., 2009) and mechanical bathythermographs (MBTs) (Ishii and Kimoto, 2009). (2) Insufficient coverage of in-situ ocean temperature observations, in both horizontal and vertical dimensions (Cheng and Zhu, 2014a, b; Lyman and Johnson, 2008). (3) Choice of key methodologies, such as the climatology (Cheng and Zhu, 2015; Lyman and Johnson, 2013). (4) Quality control of the in-situ data. The accuracy of OHC estimation relies on the proper treatment of these four issues.

In this paper, we briefly discuss the above error sources and review recent progress in solving them. Based on proper solutions to such errors, a new assessment of historical upper 0-700 m OHC change is reported. However, we note that this estimate is not necessarily the best compared with previous estimates, because challenges still remain during OHC calculation. Addressing the challenges requires more detailed studies in the future.

2　Data

Assessment of OHC change relies on in-situ temperature observations. In this study, ocean subsurface temperature profiles for 1970-2014 are from the Institute of Atmospheric Physics (IAP) and the Global Ocean Temperature (IGOT) dataset (Cheng and Zhu, 2014b), which is a quality-controlled and bias-corrected dataset (available at: http://159.226.119.60/cheng/). The in-situ temperature profiles of the IGOT dataset are sourced from the World Ocean Database 2013 (WOD13) (Boyer et al., 2013). The calculation of OHC based on in-situ observations follows the description shown in Fig. 1. Starting from the observed temperature profiles, the first essential step is to quality-control (QC) the data in order to remove erroneous measurements. Different data centers, such as the National Oceanographic Data Center (NODC) (Boyer and Levitus, 1994) and the UK Met Office (Good et al., 2013) have proposed different QC processes; however, there is no consensus on which QC process is the best. Recently, an international project named the “International Quality-controlled Ocean Dataset” (http:// www.iquod.org/index.php/home) proposed communityaccepted QC processes and a QCed dataset. In this study, by generally following the WOD13 QC flags and several additional processes, as described in Cheng and Zhu (2014b), we do not consider the impact of QC processes. After QCing the meta observations, the obtained dataset could be used to calculate historical OHC, as discussed in the following section.

3　Assessment of OHC change

The assessment of historical OHC (0-700 m) change is made according to the methods discussed in section 3.1, and the results are presented in section 3.2.

3.1 Existing problems with OHC estimation and their solutions

3.1.1 XBT bias

It has long been known that there are systematic biases in XBT temperature profiles, including both pure temperature bias in its temperature measurements and depth bias in depth values (Flierl and Robinson, 1977; Hallock and Teague, 1992; Reseghetti et al., 2007). The depth bias is due to the inaccuracy of the fall rate equation used to calculate the depth value from the elapsed time of the free-falling probe in the sea. The pure temperature bias is due to the inaccuracy of the XBT thermistor and the data recording system. Both of the biases have been confirmed to be variable with time (i.e., calendar year), probe manufacturer, probe type, ocean temperature, amongst other factors. The influencing factors of XBT biases are not fully understood, so an optimal correction of XBT biases has still not been achieved. To reduce this XBT bias, many empirical correction schemes have been proposed to correct historical XBT data (e.g., Hanawa et al., 1995; Gouretski and Reseghetti, 2010; Cowley et al., 2013; Cheng et al., 2014). By applying different XBT correction schemes to the same datasets and examining the difference of the results, Lyman et al. (2010) concluded that XBT bias contributes the major error source in OHC calculations in values of ～0.5-2.5 × 1022J (1993-2008). A recent study by Boyer et al. (2014)①Boyer, T., C. Domingues, S. Good, et al., 2014: Sensitivity of global ocean heat content estimates to mapping methods, XBT bias corrections, and baseline climatology, in: 4th XBT Workshop: XBT Science and the Way Forward, oral presentation, 11-13 November 2014, Beijing.updated the Lyman et al. (2010) study and concluded that errors in OHC calculation due to XBT bias are ～0.84-1.79 × 1022J (during 1970-2008). Among all of those schemes, the Cheng et al. (2014) method explicitly takes into account the influence of water temperature on pure temperature bias and depth bias, temporal variations of biases, and probe-type/ manufacturer dependency. In the Fourth XBT Science Workshop, the XBT community suggested that a correction scheme should take account of those influencing factors (http://2014xbtworkshop.csp.escience.cn/dct/ page/ 70005). Therefore, Cheng et al. (2014) is a reasonable choice to correct XBT bias. In this study, we apply this scheme to correct historical XBT data.

3.1.2 Choice of climatology

Once a QCed and bias-corrected temperature profile dataset is obtained, the next step is to remove the monthly climatology from each profile, which results in a temperature anomaly profile dataset. The choice of climatology is an essential step during OHC calculation. There are several different temperature climatologies constructed by different data centers, such as the World Ocean Atlas (WOA) (Locarnini et al., 2010) and Argo climatology (Roemmich and Gilson, 2009). The impact of climatology on OHC calculation has been quantified in Lyman and Johnson (2013), and then further investigated by Cheng and Zhu (2015). Lyman and Johnson (2013) constructed two different climatologies to calculate upper OHC, and indicated that the choice of climatology from which anomalies are estimated can strongly influence the global integral values and their trend. Boyer et al. (2014)①examined the differences in OHC estimates when different climatologies are used, and concluded that error due to the choice of climatology ranges from 0.27 × 1022to 1.52 × 1022J (1970-2008). However, the error due to climatology calculated in Boyer et al. (2014)①might be overestimated because some of the climatologies might be not suitable for OHC calculation. Cheng and Zhu (2015) clarified that a climatology must be constructed using data within a short period and with sufficient global coverage, in order to reduce the error (bias) due to the choice ofclimatology. In our analysis, 12 monthly climatologies are constructed by using all of the observations from 2008 to 2012. All of the temperature profiles in each month within 2008-2012 are interpolated to standard vertical levels and then averaged to 1° × 1° grids. The 5° × 5° median filter is then applied to the obtained field to smooth the climatologies. We do not follow the detailed processes described in WOA2013 because they may lead to a much smoother field, which might induce biases in regions with high spatial variability, such as western boundary current regions (Smith and Murphy, 2007) and coastal areas.

3.1.3 Choice of mapping method

A grid-averaged temperature profile dataset is obtained by interpolating all of the temperature profiles to standard depths and then averaging all of the obtained profiles at each year/depth into a 1° × 1° grid. A complicating factor is that there are regions of ocean grids without any data. Therefore, it is impossible to calculate the global integration, which requires full data coverage. To overcome this data paucity, “mapping” strategies have been proposed to fill those data gaps (i.e., Smith and Murphy, 2007; Domingues et al., 2008; Levitus et al., 2012). Boyer et al. (2014)①calculated OHC using the same methods but eight different mapping strategies, and found that the resultant OHC estimates (i.e., long-term trends) were very sensitive to the choice of mapping method. Therefore, they concluded that the choice of mapping method is another major source of error during OHC calculation. The error due to mapping methods is ～1.65×1022J (1970-2008). However, there is no indication of which is the best method and no clear indication of which method better reconstructs the OHC variability in data gaps.

Alternatively, Cheng and Zhu (2014b) proposed a simple method of dividing the global ocean into a Ship Area (with sufficient data coverage in the past 45 years) and an Argo-Ship Area (with sufficient data coverage only since the Argo Era). The yearly mean OHC in the Ship Area can be directly calculated based on the available observations. However, in the Argo-Ship Area, because of the insufficient data, interannual ocean variability cannot be represented well; instead, only a linear long-term OHC trend is calculated. Therefore, the global-mean OHC is calculated by combining the OHC estimates in the Ship Area and the obtained linear trend in the Argo-Ship Area. By using this strategy, the long-term trend of global OHC can be estimated.

3.1.4 Vertical resolution of historical profile observations

It has been shown that the typical vertical resolution of historical temperature profiles is insufficient (Cheng and Zhu, 2014a), with a global mean of 10-20 m for the upper 100 m and 50-100 m for 300-700 m prior to 2000. The insufficiency of the vertical resolution of temperature data leads to a systematic bias in OHC calculation. This error is calculated by sampling a high-vertical-resolution climatological ocean according to the depth intervals of in-situ subsurface temperature observations, and then the difference between the OHC calculated by subsampled profiles and the OHC of the climatological ocean is defined as the resolution-induced error. On average, in global terms, this bias ranges from 1 × 1022to 2.5 × 1022J, which could offset the long-term (i.e., 1966-2013) OHC trend by ～5%-10%. This bias could be approximately corrected by removing a global mean resolution-induced bias from annual mean OHC, which has been applied in this study.

3.2 A new assessment of historical OHC changes

Based on the processes discussed above, an updated calculation of annual OHC from 1970 to 2014 is obtained, as illustrated in Fig. 2a. We assess the uncertainties using a bootstrap method: we randomly select 60% of the grids to calculate the global mean OHC for each year according to previous methods. This process is repeated 50 times, and the spread of the obtained OHC approximately represents the uncertainties. In this study, the upper 0-700 m OHC change is represented by 0-700 m averaged temperature change for simplicity. The obtained upper 0-700 m OHC trend is 0.0061 ± 0.0018°C yr-1, equal to 0.56 ± 0.15 × 1022J yr-1from 1970 to 2005 on average globally, and 0.0060 ± 0.0018°C yr-1(0.55 ± 0.14 × 1022J yr-1) from 1970 to 2014, where the 95% confidence level of the trend is presented. A total increase of OHC by 0.55 ± 0.14 × 1022J yr-1is equal to 0.78 ± 0.20 mm yr-1of sea level rise. Therefore, the thermosteric sea level rise from 1970 to 2014 for the upper 700 m is about 35 mm, which amounts to 32% of the total sea level rise during the same period compared with the updated global mean sea level assessment of Church et al. (2011).

Besides the long-term trend of OHC change, it is also interesting to examine the interannual and decadal variation of the ocean thermal changes. In Fig. 2b, the annual global-averaged ocean heating rate (0-700 m) is presented, which is represented by the first differences of the global OHC time series (shown in Fig. 2b). The results show a significant year-by-year variation of the upper OHC changes, which is consistent with ENSO variability, as compared with the Niño3.4 index indicated by blue coloring. This observational result confirms that ENSO dominates the ocean energy budget on the interannual scale (Trenberth and Fasullo, 2012). Another remarkable signal is the volcano footprints of two major volcano eruptions: El Chichón in March–April 1982 and Pinatubo in June 1991—as highlighted in Fig. 2b. The OHC decreases due to the volcanic eruptions are likely superimposed on the ENSO signals.

Also, it is also apparent that there is significant decadal variation of the ocean warming, as exhibited by the nine-year trends of OHC at 700 m. Ocean warming decreased during the late 1970s, early 1990s, and early 2000s; however, the reasons are generally unknown. During the most recent 10 years, the upper ocean experienced a significant decrease, indicating a slowdown of upper ocean warming. This slowdown is also revealed by sea surface temperature changes (Trenberth and Fasullo, 2013; England et al., 2014).

The OHC changes in the three major ocean basins are presented in Fig. 3. All of the three basins show significant upper ocean warming. The Atlantic Ocean experiences the fastest and most monotonous warming in the past 45 years, with a long-term trend of 0.0084 ± 0.0020°C yr-1, equal to 0.78 ± 0.18 × 1022J yr-1from 1970 to 2014. Meanwhile, the Pacific Ocean warming is a little slower, with a trend of 0.0054 ± 0.0017°C yr-1(0.50 ± 0.15 × 1022J yr-1), than the Atlantic Ocean. The rate of Indian Ocean warming is similar to the Pacific Ocean, with a trend of 0.0052 ± 0.0016°C yr-1(0.49 ± 0.13 × 1022J yr-1). Although the interannual variability in the Pacific Ocean is dominated by ENSO, a significant decadal variability is also apparent, which is likely to be the signal of the Pacific Decadal Oscillation (PDO). In Fig. 3b, the detrended OHC change in the Pacific Ocean is compared with the PDO index (Zhang et al., 1997), showing a significant negative correlation (r = -0.81). PDO changed from a positive phase to a negative phase around 2000, which is thought to be the cause of the recent climate slowdown (Meehl et al., 2013; Trenberth and Fasullo, 2013; England et al., 2014), which is defined as the slowdown of sea surface temperature change. Such a PDO phase transfer corresponds to the rapid increase of OHC during 1990-2000 in the Pacific Ocean and a slow OHC increase since 2000. In the Atlantic Ocean, the detrended OHC at 700 m sometimes resembles the ENSO variability, as shown in Fig. 3c (i.e., 1985-2005).

4　Discussion on the remaining challenges

A new assessment of historical OHC has been obtained in this study using a new methodology. However, it is not clear that this is the best estimate of OHC because there are still some major challenges in OHC calculation. One of the major remaining challenges is how to infill the OHC data gaps, which requires an appropriate choice of mapping method. Figure 4 provides the OHC time series calculated by applying NODC-mapping, where an objective interpolation method is used to fill the data gaps. NODC-mapping results in a 0.42±0.10×1022J yr-1OHCtrend, ～20% smaller than our estimate. The discrepancy suggests that a comprehensive examination of existing mapping methods is required to understand their performances, and then identify an optimal method.

But is it possible to reconcile the upper OHC change in models and observational estimates? Here, we present the 0-700 m OHC time series calculated by 40 Coupled Model Intercomparison Project, Phase 5 (CMIP5) historical simulations (Taylor et al., 2012) in Fig. 4, compared with our new estimate. The results show that the models are capable of simulating the long-term trend of OHC change: the 36-year OHC trend from 1970 to 2005 of the CMIP5 models is 0.49 ± 0.12 × 1022J yr-1, similar to the observational value of 0.56 ± 0.14 × 1022J yr-1. However, the models show very weak interannual variability, as shown in Fig. 4b. The annual global-averaged upper ocean warming rates are consistent over time, with the exception of those years after the aforementioned major volcanic eruptions of 1982 and 1991. And for decadal variability, there is no apparent decadal variability besides the volcanic signals in the CMIP5 simulations. Therefore, the models might have limited capability in representing the observed interannual and decadal variation. Continuous efforts with regard to model simulations of OHC changes are required in the future.

Acknowledgements. The authors acknowledge the National Oceanic and Atmospheric Administration/National Oceanographic Data Center (NOAA/NODC) for decades of effort in collecting and QCing the historical subsurface data, which was the data source in this study. This work was supported by the Chinese Academy of Sciences project entitled “Western Pacific Ocean System: Structure, Dynamics and Consequences” (Grant No. XDA11010405), and the National Natural Science Foundation of China (Grant No. 41476016).

References

Abraham, J. P., J. M. Gorman, F. Reseghetti, et al., 2012a: Turbulent and transitional modeling of drag on oceanographic measurement devices, Model. Simul. Eng., 8, doi:10.1155/2012/567864. Abraham, J. P., J. M. Gorman, F. Reseghetti, et al., 2012b: Drag coefficients for rotating expendable bathythermographs and the impact of launch parameters on depth predictions, Numer. Heat Transf. A-Appl., 62, 25-43.

Abraham, J. P., M. Baringer, N. L. Bindoff, et al., 2013: A review of global ocean temperature observations: Implications for ocean heat content estimates and climate change, Rev. Geophys., 51, 450-483.

Boyer, T., and S. Levitus, 1994: Quality Control and Processing of Historical Oceanographic Temperature, Salinity and Oxygen Data, National Oceanic and Atmospheric Administration Tech. Rep. NESDIS 81, U.S. Department of Commerce, 64pp.

Boyer, T. P., J. I. Antonov, O. K. Baranova, et al., 2013: World Ocean Database 2013, Sydney Levitus et al. (Ed.), NOAA Atlas NESDIS 72, 209pp.

Cheng, L., and J. Zhu, 2014a: Uncertainties of the ocean heat content estimation induced by insufficient vertical resolution of historical ocean subsurface observations, J. Atmos. Ocean Technol., 31, 1383-1396.

Cheng, L., and J. Zhu, 2014b: Artifacts in variations of ocean heat content induced by the observation system changes, Geophys. Res. Lett., 20, 7276-7283.

Cheng, L., and J. Zhu, 2015: Influences of the choice of climatology on ocean heat content estimation, J. Atmos. Ocean Technol., 32, 388-394.

Cheng, L., J. Zhu, R. Cowley, et al., 2014: Time, probe type and temperature variable bias corrections to historical expendable bathythermograph observations, J. Atmos. Ocean Technol., 31, 1793-1825.

Church, J. A., N. J. White, L. F. Konikow, et al., 2011: Revisiting the Earth's sea-level and energy budgets from 1961 to 2008, Geophys. Res. Lett., 38, 18601-18608 .

Cowley, R., S. Wijffels, L. Cheng, et al., 2013: Biases in expendable bathythermograph data: A new view based on historical side-by-side comparisons, J. Atmos. Ocean Technol., 30, 1195-1225.

Domingues, C. M., J. A. Church, N. J. White, et al., 2008: Improved estimates of upper-ocean warming and multi-decadal sea-level rise, Nature, 453, doi:10.1038/nature07080.

England, H. M., S. McGregor, P. Spence, et al., 2014: Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus, Nature Climate Change, 4, 222-227.

Flierl, G., and A. Robinson, 1977: XBT measurements of thermal gradients in the MODE eddy, J. Phys. Oceanogr., 7, 300-302.

Good, S. A., M. J. Martin, N. A. Rayner, 2013: EN4: Quality controlled ocean temperature and salinity profiles and monthly objective analyses with uncertainty estimates, J. Geophys. Res., 118, 6704-6716.

Gouretski, V., and K. P. Koltermann, 2007: How much is the ocean really warming? Geophys. Res. Lett., 34, 155-170.

Gouretski, V., and F. Reseghetti, 2010: On depth and temperature biases in bathythermograph data: Development of a new correction scheme based on analysis of a global ocean database, Deep-Sea Res., 57, 812-833.

Hallock, Z. R., and W. J. Teague, 1992: The fall-rate of the T7 XBT, J. Atmos. Oceanic Technol., 9, 470-483.

Hanawa, K., P. Rual, R. Bailey, et al., 1995: A new depth time equation for sippican or tsk T-7, T-6 and T-4 expendable bathythermographs (Xbt), Deep-Sea Res., 42, 1423-1451.

Ishii, M., and M. Kimoto, 2009: Reevaluation of historical ocean heat content variations with time-varying XBT and MBT depth bias corrections, J. Oceanogr., 65, 287-299.

Levitus, S., J. I. Antonov, T. P. Boyer, et al., 2009: Global ocean heat content 1955-2008 in light of recently revealed instrumentation problems, Geophys. Res. Lett., 36, 157-163.

Levitus, S., J. I. Antonov, T. P. Boyer, et al., 2012: World ocean heat content and thermosteric sea level change (0-2000 m), 1955-2010, Geophys. Res. Lett., 39, L10603, doi:10.1029/2012GL051106.

Locarnini, R. A., A. V. Mishonov, J. I. Antonov, et al., 2010: World Ocean Atlas 2009, Volume 1: Temperature, S. Levitus (Ed.), NOAA Atlas NESDIS 68, U.S. Government Printing Office, Washington D.C., 184pp.

Lyman, J. M., S. A. Good, V. V. Gouretski, et al., 2010: Robust warming of the global upper ocean, Nature, 465, 334-337.

Lyman, J. M., and G. C. Johnson, 2008: Estimating annual global upper-ocean heat content anomalies despite irregular in situ ocean sampling, J. Climate, 21, 5629-5641.

Lyman, J., and G. Johnson, 2013: Estimating global ocean heat content changes in the upper 1800 m since 1950 and the influence of climatology choice, J. Climate, 27(5), 1945-1957.

Meehl, G. A., A. Hu, J. M. Arblaster, et al., 2013: Externally forced and internally generated decadal climate variability associated with the Interdecadal Pacific Oscillation, J. Climate, 26, 7298-7310.

Palmer, M. D., K. Haines, S. F. B. Tett, et al., 2007: Isolating the signal of ocean global warming, Geophys. Res. Lett., 34, 231-247.

Reseghetti, F., M. Borghini, and G. M. R. Manzella, 2007: Factors affecting the quality of XBT data—Results of analyses on profiles from the Western Mediterranean Sea, Ocean Sci., 3, 59-75.

Rhein, M., S. R. Rintoul, and S. Aoki, et al., 2013: Observations: Ocean, in: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, T. F. Stocker et al. (Eds.), Cambridge University Press, Cambridge and New York, 255-315.

Roemmich, D., and J. Gilson, 2009: The 2004-2008 mean and annual cycle of temperature, salinity, and steric height in the global ocean from the Argo Program, Prog. Oceanogr., 82, 81-100.

Smith, D. M., and J. M. Murphy, 2007: An objective ocean temperature and salinity analysis using covariances from a global climate model, J. Geophys. Res. Oceans, 11, 117-128.

Taylor, K. E., R. J. Stouffer, and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design, Bull. Amer. Meteor. Soc., 93, 485-498.

Trenberth, K. E., and J. T. Fasullo, 2012: Tracking earth's energy: From El Niño to global warming, Surv. Geophys., 33, 413-426. Trenberth, K. E., and J. T. Fasullo, 2013: An apparent hiatus in global warming? Earth's Future, 1, 19-32.

Zhang, Y., J. M. Wallace, and D. S., Battisti, 1997: ENSO-like interdecadal variability: 1990-93, J. Climate, 10, 1004-1020.

Cheng, L.-J., J. Zhu, and J. Abraham, 2015: Global upper ocean heat content estimation: Recent progress and the remaining challenges, Atmos. Oceanic Sci. Lett., 8, 333-338,

10.3878/AOSL20150031.

8 April 2015; revised 26 April 2015; accepted 28 April 2015; published 16 November 2015

CHENG Li-Jing, chenglij@mail.iap.ac.cn