| Tide gauges have been used over the last century for measuring sea level in certain locations. The inherent problem with the tide gauges is that they only provide measurements at a single location.; In this work, the errors in different sea level reconstructions are quantified using several methods. A sensitivity analysis is performed to determine appropriate parameters to be used in the reconstruction process. These parameters include an appropriate number of Empirical Orthogonal Functions (EOFs) to use, different weighting schemes, the use of EOF 0, a globally constant in height spatial field, and which tide gauge records to use. The Parallel Climate Model provides a dataset to allow quantification of undescribed variability to be made. Errors due to tide gauge distribution are quantified by comparing subset reconstructions to the satellite altimetry. In addition, correlated errors are quantified as the error estimates are not independent in time. All three error sources are combined and used to determine the formal error of the best reconstruction. The resulting sea level reconstruction has a trend of 2.31 +/- 0.05 mm/yr from 1950--2000. (Abstract shortened by UMI.)... |
