Ocean State Estimation for Climate Research
Signal and uncertainty in ocean reanalyzes
Download 0.6 Mb.
|
- Navigate this page:
- 3. Summary and Challenges ahead
2.3 Signal and uncertainty in ocean reanalyzes As described earlier, one purpose of the intercomparison is to estimate the uncertainty of an estimated quantity from different reanalysis products. Estimating uncertainty is a major challenge. In addition to the three dimensional estimation of the ocean state at a given time (analysis problem), the estimation of the time evolution is also required in a reanalysis. The time evolution represented by an ocean reanalysis will be sensitive to the time variations of the observing system, to the errors of the ocean model, atmospheric fluxes and assimilation system, which are often flow dependent, and not easy to estimate. In cases where the complexity is too large to be tackled analytically, the ensemble methods can offer a helping hand. So, in the same way as the ensemble of multi-model forecasts is used to get an estimation of the uncertainty of future climate projections, and ensemble of different reanalysis products can be used to get a first glance of the time evolution and uncertainty of the climate reconstructions, although the caveats of the ensemble methodology should be taken into account when interpreting the results. This approach was followed by Balmaseda and Weaver 2006 in the comparison of ocean reanalysis organized by the CLIVAR GSOP (http://www.clivar.org/organization/gsop/synthesis/synthesis.php). The results from 20 global ocean reanalysis products, with different models, assimilation methods and forcing fluxes, were gathered together to examine the time evolution of the upper ocean temperature and salinity. One of the reanalysis did not include any ocean model. Some of them did not include any ocean observations, being the result of ocean models forced by atmospheric forcing fluxes. For the early years (1960-1980) there were only seven available products. For the period 1993-2002 there were 20 products, and very few were available for post-2002. The scarcity of reanalyses products for the last few years hampers the assessment of uncertainty on recent climate signals. Figure 5 shows the signal-to-noise ratio of the interannual variability of temperature (left) and salinity (right) in the upper 300m (T300 and S300 in what follows) for different regions. The signal is estimated as the temporal standard deviation of the ensemble mean, and the noise as the ensemble spread. For the period 1960-2002, the North Atlantic (NATL, 30N-60N) stands out as the region where the temperature signal dominates the noise. The signal is also quite clear in Equatorial Pacific (EQPAC, 5S-5N), the Tropical Atlantic (TRATL ,20S-20N) and the GLOBAL ocean. In the most recent period (1993-2002), the signal to noise ratio in temperature is larger than 1 in most areas, and the Equatorial Pacific clearly stands out. In contrast, the upper ocean salinity signal to noise ratio is less than one almost everywhere, except for the Equatorial Pacific, where it is close to 1. The climate signals and source of uncertainty can be investigated further. Figure 6 shows the time evolution of T300 in four regions: GLOBAL ocean, North Atlantic, North Pacific (30-60N) and Equatorial Pacific. The grey shade shows the spread among the 20 different ocean reanalyses. They have been grouped according to their forcing fluxes and use of data. The red shade shows the results for only those reanalysis that use ER40 forcing fluxes (Uppala et al 2005). These have been further subdivided into reanalysis without data assimilation (blue) and with data assimilation (green). The inset in each figure shows the size of the average spread for each of these groups. The difference between grey and red can be taken as a measure of uncertainty due to forcing fluxes. The size of the blue spread is representative of uncertainty due to using different ocean models. The size of the green spread is due to the differences in model an assimilation method. The difference between green and blue is indicative of the ability for the data assimilation to reduce (or on the contrary to increase) the uncertainty of the climate signals. Results indicate that the uncertainty induced by the forcing fluxes is comparable to the uncertainty due to ocean model formulation. In the regions shown here, the assimilation of ocean data reduces the spread of the model-only runs, especially noticeable in the Equatorial Pacific. However, the assimilation can increase the uncertainty in the temperature field in the tropical Atlantic and Indian Ocean and in the salinity field almost everywhere (not shown). The fact that the assimilation can increase the uncertainty points towards the need for more robust assimilation methods that better handle model and data errors and the need for additional observations. In spite of the large uncertainty, results shown in Figure 6 indicate that there are clear climate signals. For instance, the increasing trend in the global upper ocean heat content is visible, as well as decadal variability: lower values in the 60’s, a maxima during at the beginning of the 80’s, another drop in the mid 90’s and an increasing trend ever since. The magnitude of the decadal signals is stronger with assimilation than with the free ocean model. Whether the intensification of the decadal signal is due to the XBT’s bias (e.g., Gouretski and Koltermann 2007) or to other reasons (for instance, to the lack of information in the forcing fluxes) need to be investigated farther. The decadal signals and increasing trend is also visible in the North Atlantic, which shows the largest acceleration in the warming in the late 90’s. After 2000 however warming seems to stabilize, or even decrease. Unfortunately the uncertainty increases tremendously in the last few years. The North Pacific variability is more dominated by decadal variability rather than by secular trends. There is quite a pronounced rapid warming in the late 80’s, which may be related to the changes in the gyre circulation reported by Suga et al 2002, Taneda et al 2000. The origins of this shift and its role in the global climate need further investigation. The Equatorial Pacific is the region where the interannual variability is quite clear, and where the uncertainty has been decreasing steadily with time, probably due to the enhancement of the TOGA-TAO system. The interannual variability in the Equatorial Pacific upper ocean heat content shows a clear downward trend, as has been described by Balmaseda et al (2008). The relation of this trend in heat content with the the weakening of the Walker Circulation (Vecchi et al 2006, Scott and Smith 2007) needs further investigation. 
Figure 5. Signal to noise ratio for the time evolution of temperature and salinity in the upper 300m (T300 and S300) in different regions. The estimates have been done separately for the period 1960-2002 and for the most recent period 1993-2002. 
Figure 6. Time evolution and spread of the upper 300m average temperature in different regions. The spread resulting from using 20 different ocean reanalysis is shown in grey. Those reanalysis that use the same forcing fluxes (ERA40) are shown in red. The difference between grey and red is indicative of uncertainty arising from the forcing fluxes. The blue shading is indicative of the uncertainty arising from the ocean model formulation, since it groups ocean simulations which do not assimilate data and differ only on the ocean model. The same models, but this time using data assimilation, are shown in green. The difference between blue and green is representative of the effect of the assimilation on the signal and on the spread. 3. Summary and Challenges ahead As part of CLIVAR’s and GODAE’s effort in the past decade, significant advancement has been achieved in ocean state estimation efforts that are geared towards climate research. A suite of global ocean reanalysis products have been produced and updated on a routine basis. There have been an ever increasing number of applications of these products for oceanographic and climate-related studies over a wide range of topics. The existence of these products also allows a comprehensive intercomparison to evaluate the consistency and fidelity of the various climate-related diagnostic quantities estimated from different products. The intercomparison help identify areas that need improvement in ocean reanalysis and observing systems. Despite the significant advancement in ocean reanalyses, many challenges lay ahead. The ocean data assimilation groups need to work closely with the modeling community to improve the model physics, especially those associated with the bias of the mean state. The estimates of data and model errors dictate the outcome of the state estimation for a given model. Therefore, the ocean reanalysis community needs to work closely with the observational community in obtaining robust estimates of data errors, an important issue that is often dealt with only by the assimilation groups. The uncertainty in the observations associated with such factors as changing fall rate of XBTs over the past few decades and some biased ARGO data in the past few years are among examples that more effort is needed to estimate data error. Moreover, much research is still required to understand model errors. In particular, we need to identify the sources of model errors and determine if appropriate new control variables can be adopted to account for such error sources (e.g., internal model errors associated with mixing parameterizations). Computational resources remain to be a critical issue for estimation effort based on the ensemble and adjoint methods because it limits the ensemble size and model resolution that one can afford. Finally, the coupled nature of the climate system prompts for a coupled approach for state estimation that includes components of the climate system (such as the atmosphere, cryosphere, hydrosphere, and biogeochemistry) in order to properly account for the potential feedback among different elements of the climate systems. Currently, coupled ocean-atmosphere, ocean-ice, and ocean physics-biochemistry data assimilations are still in their infancy and are expected to pick up momentum in the coming decade. References Balmaseda, M.A. , D. Anderson and F. Molteni (2008). Climate variability from the new ECMWF Ocean Reanalysis System ORA-S3. Third WCRP Conference on Reanalysis. http://wcrp.ipsl.jussieu.fr/Workshops/Reanalysis2008/Documents/G2-264_ea.pdf Balmaseda, M.A., G.C. Smith, K. Haines, et al. 2007: Historical reconstruction of the Atlantic Meridional Overturning Circulation from the ECMWF operational ocean reanalysis. Geophys. Res. Lett., 34, L23615, doi:10.1029/2007GL031645. Cabanes, C., T. Lee, and L.-L. Fu, 2008: Mechanisms of Interannual Variations of the Meridional Overturning Circulation of the North Atlantic Ocean. J. Phys. Oceanogr., 38, 467-480. Carton, J.A., B.S. Giese, 2008: A reanalysis of ocean climate using Simple Ocean Data Assimilation (SODA). Mon. Wea. Rev., 136, 2999-3017. Domingues, C.M., J.A. Church, N.J. White, et al. Improved estimates of upper-ocean warming and multi-decadal sea-level rise. Nature. 453, 1090-U6. Fukumori, I., T. Lee, B. Cheng, and D. Menemenlis, 2004: The origin, pathway, and destination of NINO3 water estimated by a simulated passive tracer and its adjoint. J. Phys. Oceanogr., 34, 582-604. Gouretski V., K. P. Koltermann (2007), How much is the ocean really warming?,Geophys. Res. Lett., 34, L01610, doi:10.1029/2006GL027834. Halkides, D., and T. Lee, 2008: Mechanisms controlling seasonal-to-interannual mixed-layer temperature variability in the southeastern tropical Indian Ocean. J. Geophys. Res., submitted. Kim, S.-B., T. Lee, I. Fukumori, 2007: Mechanisms controlling the interannual variation of mixed layer temperature averaged over the NINO3 region. J. Clim., 20, 3822-3843. Kim, S.-B., T. Lee, and I. Fukumori, 2004: The 1997-99 abrupt change of the upper ocean temperature in the northcentral Pacific. Geophys. Res. Lett., 31, L22304, doi:10.1029/2004GL021142. Köhl, A., and D. Stammer, 2008a: Decadal sea level changes in the 50-year GECCO ocean synthesis. J. Clim., 21, 1866-1890. Köhl, A., D. Stammer, and B. Cornulle, 2007: Interannual to decadal changes in the ECCO global synthesis. J. Phys. Oceanogr., 37, 313-337. Köhl, A. and D. Stammer, 2008b: Variability of the meridional overturning in the North Atlantic from the 50-year GECCO state estimation. J. Phys. Oceanogr., 38, 1913 -1930. Lee, T., and I. Fukumori, 2003: Interannual to decadal variation of tropical-subtropical exchange in the Pacific Ocean: boundary versus interior pycnocline transports. J. Climate. 16, 4022-4042. Masuda, S., T. Awaji, N. Sugiura, T. Toyoda, Y. Ishikawa, K. Horiuchi (2006), Interannual variability of temperature inversions in the subarctic North Pacific, Geophys. Res. Lett., 33, doi:10.1029/2006GL027865. Masuda, S., T. Awaji, T. Toyoda, Y. Shikama, Y. Ishikawa (2008), Temporal evolution of the equatorial thermocline associated with the 1991-2006ENSO: A possible cause of ENSO irregularity, J. Geophys. Res. (submitted). Power, S. B., and I. N. Smith (2007), Weakening of the Walker Circulation and apparent dominance of El Niño both reach record levels, but has ENSO really changed?, Geophys. Res. Lett., 34, L18702, doi:10.1029/2007GL030854. Rabe, B., F.A. Schott, and A. Köhl, 2008: Mean circulation and variability of the tropical Atlantic during 1952-2001 in the GECCO assimilation fields. J. Phy. Oceanpgr., 38, 177-192. Schoenefeldt, R., and F. A. Schott, 2006: Decadal variability of the Indian Ocean cross-equatorial exchange in SODA. Geophys. Res. Lett., 33, L08602, doi:10.1029/2006GL025891. Schott, F.A., L. Stramma, W. Wang , et al., 2008: Pacific Subtropical Cell variability in the SODA 2.0.2/3 assimilation. Geophys. Res. Lett., 35, L10607, doi:10.1029/2008GL033757. Schott, F.A., W.-Q. Wang, and D. Stammer, 2007: Variability of Pacific subtropical cells in the 50-year ECCO assimilation. Geophys. Res. Lett., 34, L05604, doi:10.1029/2006GL028478. Stammer, D., C. Wunsch, and R. Giering, et al. 2003: Volume, heat, and freshwater transports of the global ocean circulation 1993-2000, estimated from a general circulation model constrained by World Ocean Circulation Experiment (WOCE) data/ J. Geophys. Res.-Oceans, 108, Suga, T; Kato, A; Hanawa, K. 2000. North Pacific Tropical Water: its climatology and temporal changes associated with the climate regime shift in the 1970s. Progress in Oceanograph, 47 (2-4): 223-256. Toyoda, T., T. Awaji, S. Masuda, N. Sugiura, H. Igarashi, T. Mochizuki, and Y. Ishikawa: Interannual variability of North Pacific eastern subtropical mode water formation in the 1990s derived from a 4-dimensional variational ocean data assimilation experiment. (submitted to J. Geophys. Res.) Wang, O., I. Fukumori, T. Lee, and B. Cheng, 2004: On the cause of eastern equatorial Pacific Ocean T-S variations associated with El Nino. Geophys. Res. Lett., 31, L15310, doi:10.1029/2004GL02472. Wunsch, C., R.M. Ponte, P. Heimbach, 2007: Decadal trends in sea level patterns: 1993-2004. J. Clim., 20, 5889-5911. Wunsch, C. and P. Heimbach, 2007: Estimated decadal changes in the North Atlantic meridional overturning circulation and heat flux 1993-2004. J. Phys. Oceanogr., 36, 11, 2012-2024. Download 0.6 Mb. Share with your friends:
|