It is valuable to note that the reduction in the uncertainty of estimation of the long-term ice sheet imbalance reported in Sections 11.3.1 and 11.4 came from indirect constraints and the synthesis of information of different types. Such syntheses offer promise for further progress.

Sections 11.3.2.1 and 11.4 reveal significant uncertainty in the analysis of 20th century sea level change. Also, we have little knowledge of the regional pattern of sea level change. Observational determination of such a pattern would be a powerful test of the coupled models required for projections of globally averaged and regional sea level rise. Requirements for reducing uncertainties include:

• A global tide gauge network (the ‘GLOSS Core Network�) (IOC, 1997) for measuring relative change.
• A programme of measurements of vertical land movements at gauge sites by means of GPS (Global Positioning System), DORIS Beacons and/or absolute gravity meters (Neilan et al., 1998).
• Improved models of postglacial rebound.
• A reanalysis of the historical record, including allowing for the impact of variable atmospheric forcing.
• A subset of mostly island tide gauge stations devoted to ongoing calibration of altimetric sea level measurements (Mitchum, 1998).
• An ongoing high-quality (TOPEX/POSEIDON class) satellite radar altimeter mission (Koblinsky et al., 1992) and careful control of biases within a mission and between missions.
• Space gravity missions to estimate the absolute sea surface topography (Balmino et al., 1996, 1999) and its temporal changes, to separate thermal expansion from an increase in ocean mass from melting of glaciers and ice sheets (NASA, 1996, NRC, 1997) and changes in terrestrial storage.

For assessment of possible changes to the severity of storm surges, analyses of historical storm surge data in conjunction with meteorological analyses are needed for the world�s coastlines, including especially vulnerable regions.

Requirements for improved projections of ocean thermal expansion include:

• Global estimates of ocean thermal expansion through analysis of the historical data archive of ocean observations and a programme of new observations, including profiling floats measuring temperature and salinity and limited sets of full-depth repeat oceanographic sections and time-series stations.
• Testing of the ability of AOGCMs to reproduce the observed three-dimensional and time-varying patterns of ocean thermal expansion.
• An active programme of ocean and atmosphere model improvement, with a particular focus on the representation of processes which transport heat into and within the interior of the ocean.

Requirements for improved projections of glacier contributions include (see also Haeberli et al., 1998):

• A strategy of worldwide glacier monitoring, including the application of remote sensing techniques (laser altimetry, aerial photography, high-resolution satellite visible and infrared imagery e.g. from ASTER and Landsat).
• A limited number of detailed and long-term mass balance measurements on glaciers in different climatic regions of the world, with an emphasis on winter and summer balances in order to provide a more direct link with meteorological observations.
• Development of energy balance and dynamical models for more detailed quantitative analysis of glacier geometry changes with respect to mass balance and climate change.
• Glacier inventory data to determine the distribution of glacier parameters such as area and area-altitude relations, so that mass balance, glacier dynamics and runoff/sea level rise models can be more realistically framed.

• Continued observations with satellite altimeters, including the upcoming satellite laser altimeter on ICESat and the radar interferometer on CRYOSAT. Measurements should be continued for at least 15 years (with intercalibration between missions) to establish the climate sensitivities of the mass balance and decadal-scale trends.
• Satellite radar altimetry and synthetic aperture radar interferometry (ERS-1, ERS-2 and Radarsat) for detailed topography, changes in ice sheet volume and surface velocity of the ice sheets (Mohr et al., 1998; Joughin et al., 1999), as well as short-term variability in their flow (Joughin et al., 1996a,b) and grounding line position (Rignot, 1998a,b,c).
• Determination of the Earth�s time-variant gravity field by the Gravity Recovery and Climate Experiment (GRACE) satellite flown concurrently with ICESat to provide an additional constraint on the contemporary mass imbalances (Bentley and Wahr, 1998). This could provide estimates of sea level change to an accuracy of ±0.35 mm/yr.
• Geological observations of sea level change during recent millennia combined with improved postglacial rebound models and palaeoclimatic and palaeoglaciological studies to learn what changes have occurred in the past.
• Further analysis of Earth rotational parameters in combination with sea level measurements.
• Improved estimates of surface mass balance (including its spatial and temporal variability) from in situ observations, accumulation rates inferred from atmospheric moisture budgets and improved estimates of the rate of iceberg calving and the melt-water flux.
• Improved calculation of the surface mass balance within ice sheet models or by atmospheric models, with attention to modelling of changes in sea-ice concentration because of the consequent effect on moisture transports and accumulation.
• Improved understanding and modelling of the dynamics of ice sheets, ice streams and ice shelves (requiring combined studies using glaciological, oceanographic and satellite observations), including the physics of iceberg calving.

Surface and ground water storage changes are thought to be having a significant impact on sea level, but their contribution is very uncertain (Table 11.10, Figure 11.9), and could be either positive or negative. They may become more important in the future, as a result of changes related not only to climate, but also to societal decisions that are beyond the scope of this scientific assessment. There are several general issues in climate-related aspects:

• A more thorough investigative search of historical records could provide addition information on ground water mining, and storage in reservoirs.
• Accurate satellite measurements of variations in the Earth�s gravity (Herring, 1998) to detect changes in land water storage due to water-table variations and impoundments.
• A better understanding of seepage losses beneath reservoirs and in irrigation is required.
• A unified systems approach is needed to trace the path of water more accurately through the atmosphere, hydrologic, and biosphere sub-systems, and to account for various feedbacks (including the use of GCMs and improved hydrologic models).
• Satellite remote sensing offers useful technology for monitoring the global hydrologic budget. A cumulative volume estimate for the many small reservoirs might be possible using high-resolution radar data, targeted ground studies and a classification of land use classes from satellite data and also of changes in deforestation and other land-use transformations (Koster et al., 1999).

Sea level change involves many components of the climate system and thus requires a broad range of research activities. A more detailed discussion of the requirements is given in the report of the recent IGBP/GAIM Workshop on sea level changes (Sahagian and Zerbini, 1999). We recognise that it is important to assign probabilities to projections, but this requires a more critical and quantitative assessment of model uncertainties than is possible at present.

IPCC Homepage