Modeling regional sea level rise using local tide gauge data

H Bâki Iz,L Berry,M Koch

doi:10.2478/v10156-011-0039-2

H Bâki Iz, L Berry + Show 1 more

Open Access

https://doi.org/10.2478/v10156-011-0039-2

Copy DOI

Abstract

AbstractCurrently regional mean sea level trends and variations are inferred from the analysis of several individual local tide gauge data that spanonly a long period of time at a given region. In this study, we propose using a model to merge various tide gauge data, regardless of theirtime span, in a single solution, to estimate parameters representative of regional mean sea level trends. The proposed model can accountfor the geographical correlations among the local tide gauge stations as well as serial correlations, if needed, for individual stations’ data.Such a vigorous regional solution enables statistically optimal uncertainties for estimated and projected trends. The proposed formulationalso unifies all the local reference levels by modeling their offsets from a predefined station’s reference level. To test its effectiveness, theproposed model was used to investigate the regional mean sea level variations for the coastal areas of the Florida Panhandle using 26 localtide gauge stations that span approximately 830 years of monthly averages from the Permanent Service for Mean Sea Level repository. Thenew estimate for the regional trend is 2.14 mm/yr with a ±0.03 mm/yr standard error, which is an order of magnitude improvement overthe most recent mean sea level trend estimates and projections for the Florida region obtained from simple averages of local solutions.

Highlights

Long term changes in the mean sea level (MSL) impacts shoreline and beach erosion, coastal and wetlands inundation, storm surge ooding and coastal development
Currently regional mean sea level trends and variations are inferred from the analysis of several individual local tide gauge data that span only a long period of time at a given region
We propose using a model to merge various tide gauge data, regardless of their time span, in a single solution, to estimate parameters representative of regional mean sea level trends

Summary

Introduction

Long term changes in the mean sea level (MSL) impacts shoreline and beach erosion, coastal and wetlands inundation, storm surge ooding and coastal development. With mixed record lengths are rarely averaged to estimate regional trend in the MSLs because the local mean sea level estimates from shorter records are corrupted by interdecadal uctuations (Gornitz et al 1982 and Barnett 1984) and by unmodeled local effects (Iz and Ng 2005). The disturbances of the tide gauge data may exhibit temporal correlations (autocorrelations, the red noise for the rst order autoregressive process) that need to be accounted for in local as well as in potential regional solutions Their omission may cause underestimation of the error estimates of the solution parameters, thereby leading potential Type I errors in null-hypothesis testing for the signi cance of the model parameters. Geographical and serial correlation information will be used to establish a full variance-covariance matrix of the errors, which will be deployed in a generalized least squares solution for estimating the regional means sea level trend parameters. The new model will be tested using 26 local tide gauge stations that span over 830 years (monthly averages) around the Florida Panhandle for the local and regional solutions

Tide Gauge Stations and Data around Florida Panhandle

Local Tide Gauge Model and Local Solutions

Final solution for regional trend using ordinary least squares

Conclusion