Characteristics Of Sea Level Records

Climate-related variations of sea level, although small compared to some other geophysical phenomena, share a common property in their spectrum, that is, in the way their power is correlated with frequency. In many geophysical data series large changes are associated with long periods. A few common examples are that small earthquakes are more frequent than large ones, great storms occur less often than small ones, and large meteorites hit the earth less frequently than small ones. Temporally, global or regional sea level varies on a scale of about 100 m over the glaciation-déglaciation cycle, but only a few tens of centimeters due to the seasonal heating cycle of the upper ocean, or a seasonal-interannual El Nino event.

Another characteristic of climate-related sea level variations is that they can be spatially correlated at low frequencies, such as seasonal and longer periods. Figure 1.3 displays smoothed and normalized (i.e., divided by the

1905 6

1915 1925

1935

1945

1955 1965 1975 1985

San Diego San Francisco

Figure 1.3 San Diego and San Francisco normalized and detrended relative sea levels (mm/yr).

root-mean-square values of the series) monthly mean sea level records for San Francisco and San Diego, California, during their common observation period that began in 1906. (Their trends of about 2 mm/yr have also been removed for the sake of clarity.) The agreement at interannual and longer periods is very striking, even though these two sites are about 800 km apart. Such high spatial correlation at low frequencies is exhibited nearly everywhere and plays a critical role in attempts to determine a global rate of sea level rise from tide gauge data. Merely adding more tide gauge records to a solution for global sea level does not necessarily make any improvement if the additional data records are highly correlated with existing ones and sample only a water mass already sampled. On the west coast of North and South America, the large interannual-interdecadal variations of sea level are caused by El Niño/La Niña events (note in Fig. 1.3 the especially large excursions during the 1982-83 El Niño period), as pointed out in the classic paper by Chelton and Davis (1982). Sturges and Hong in Chapter 7 of this book show that the low-frequency variations of sea level that occur on the east coast of the U.S. have a origin different from those on the west coast, resulting from wind stress on the Atlantic Ocean. It is important to keep in mind that sea level variations are driven phenomena, not drivers, meaning that they are a result of forcings by other geophysical events. The tides are a perfect example of this, since they are an outcome of the varying gravitational forces of the sun and moon on the earth. Going beyond the tides, a time history of sea level, either global or local, is the sum of many contributors, including thermal expansion, melting of glaciers, wind forcing, and currents.

Sea level variations that occurred relatively soon after the peak of the last glacial maximum about 21,000 years ago are very different from those seen today. Sea level rose about 125 m globally by the conclusion of the melting, but with significant regional variations due to viscoelastic readjustment of the earth and modifications of the geoid from redistribution of mass. The concept of eustasy, that is, a uniform change of sea level occurring everywhere from addition or thermal expansion of water, is an inadequate description for the déglaciation event. This is so because the land rebounded by hundreds of meters in the deeply ice-covered regions, land peripheral to the ice sheets subsided, and the seafloor of necessity sank under the weight of the added meltwater. The redistribution of mass also caused a change of the gravitational potential that altered the geoid and hence sea level. As mentioned earlier, even after the melting was complete, the earth continued to adjust from the removal of the load. The complexities of all of these interactions are very great and are the subject of Chapter 4. We can get at least a feel here for the magnitude of the changes of water level due to Holocene isostatic readjustment from the sea level data shown in Fig. 1.4. Displayed are the elevations relative to modern sea level of the Bristol Channel, United Kingdom, site for the last 8 millennia (data adapted from Roberts, 1994, p. 186). In this figure the origin at 0 BP is the present, and sea level

01 2345678

Millenia BP

Figure 1.4 Bristol Channel mid and late Holocene relative sea levels (data from Roberts, 1994).

01 2345678

Millenia BP

Figure 1.4 Bristol Channel mid and late Holocene relative sea levels (data from Roberts, 1994).

was lower in the past than now. During the time shown in Fig. 1.4, sea level rose in the Bristol Channel area by about 16 m from isostatic readjustment, that is, subsidence of the land, and from changes in water level and the geoid. Note in Fig. 1.4 the nearly exponential nature of the change of sea level. The characteristic decay time (i.e., the time for the level to change by a factor of He) for this curve is about 1/.41 ~ 2.5 millennia. The rate of relative sea level rise at 0 BP due to GIA is still about 0.3 m per millennium, or 0.3 mm/yr.

Although eustasy is a poor approximation to sea level change during and immediately after the deglaciation process, it is not far from an accurate description (within perhaps 0.1 mm/yr) for analyses of contemporary sea level rise. It is observed (see Chapter 3) that properly selected long tide gauge records tend toward a narrow range of values for the trend of sea level regardless of location. This merely reflects that the earth does not adjust viscoelastically instantly to the addition of 1 mm or so of water per year to the oceans, and that the purely elastic response is a small fraction of the ultimate long-term viscoelastic adjustment.

The foregoing has indicated that much analysis is required to determine a rate of global sea level rise that can be used scientifically with confidence. A good place to start is with a blunt instrument, in this case, a histogram of sea level trends derived from tide gauge data from around the globe. Trends of relative sea level from records at least 20 years in length were derived from the sea level data base at the Permanent Service for Mean Sea Level in the United Kingdom and grouped (binned) according to value. Figure 1.5 presents the resulting histogram. As might be expected, the scatter of the trends is

-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 Trend Bins, mm per year

Figure 1.5 Histogram of relative sea level trends derived from individual tide gauge records.

-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 Trend Bins, mm per year

Figure 1.5 Histogram of relative sea level trends derived from individual tide gauge records.

very large and is in addition skewed somewhat in the direction of negative (i.e., falling) values of sea level change. But there is a clear preference for a most probable trend value that is greater than zero. Much of this book will deal with exploring why the range of sea level trend estimates is so broad and which sea level records are appropriate and useful for determining a global trend of sea level.

Was this article helpful?

0 0

Post a comment