15 



simply plotted against latitude as presented in Fig. 2.5. A problem is 

 that the data only encompass latitudes from approximately 25° to 58° and 

 thus it is necessary to extrapolate liberally. At the lower latitudes, the 

 data were extrapolated uniformly at approximately 3 . 2 mm/yr and at the 

 higher latitudes , due to the uncertainties , two extrapolations were adopted 

 to determine sensitivity as presented in Fig. 2.5. Based on the 

 latitudinal variation, T](<f)) , estimates of the ESLR, r^g, were based on the 

 following 



r/ ~ J rjA i4>) cos(^ d(f) (2.1) 



J 



where j = I, II represents the different high latitude extrapolations. The 



resulting values were 



ri =0.32 mm/yr, Extrapolation I 



ri =0.67 mm/yr, Extrapolation II 



^11 



These results are qualitatively in agreement with those of Lambeck and 

 Nakiboglu. 



2.3 THE NATURE AND ANALYSIS OF SEA LEVEL DATA 



From the standpoint of extracting eustatic sea level change, it is 

 useful to represent the total RSL, rjj_(t) , as measured by the i^"^ tide gage 

 as 



r?i(t) = r/E(t) + r7N.(t) (2.2) 



in which rj^(t) is the eustatic sea level at time t and rjj^. (t) is the total 

 "noise" at the i'-'^ tide gage. The noise can contain many components 

 including vertical ground motion, effects of freshwater in the vicinity of 

 the gage, coastal currents, long waves, barometric pressure anomalies, wave 

 effects, etc. Several obvious results follow from Eq . 2.2. First, if 

 there were a uniform coverage of tide gages on the oceans, an average of 

 the elevations from all such tide gages would yield the eustatic sea level. 

 Additionally, the eustatic sea level change rate need not be constant, but 



