Logo Universitat Bremen
Page path:

WP 3.3: Projection analysis of regional sea level changes in high-resolution models in the context of predictable patterns derived from linear inverse models

PIs:
Detlef Stammer, IfM, CEN, Universität Hamburg, detlef.stammer(at)uni-hamburg.de
Armin Köhl, IfM, CEN, Universität Hamburg, armin.koehl(at)uni-hamburg.de

Team:
Frank Siegesmund
Nuno Serra
Correlation of decadal predictions of the sea level (average of 1-10 years after initialization) between the basic LIMs and the 'truth' of the earth system model. The correlation is between 0.0 and 0.8.

Correlation of decadal predictions of the sea level (average of 1-10 years after initialization) between the basic LIMs and the 'truth' of the earth system model.

The project deals with the characterization of regional sea level changes and their predictability. During the first RACE phase, predictability of long-term regional sea level changes were investigated by means of linear inverse models (LIM). Regional changes in sea level were also examined in time slices with a high-resolution model that was forced by CMIP-5 surface fluxes under global warming scenarios.
In the next phase of the RACE project, the two strands of work are now merged and analyzed together. Since only an analysis of the temporal changes of the decadal means was possible due to strong artificial trends in the time slices of the high resolution runs, we will first replace the time slices by a continuous CMIP-5 forced run. On time scales of one year or more, the amplitude and patterns of variability of the sea level from the high-resolution model should be compared with the CMIP-5 runs, regionally and in coastal areas. The interactions between shelf and the open ocean are one particular focus. As an additional aspect, the eddy resolving model allows the study of changes in the eddy statistics and their relation to the changes in the mean state.
In the next phase, the spatial patterns of the LIM are also to be analyzed. It is intended to examine to what extent the predictable modes of LIM are relevant for the variability in the high-resolution model, the LIM will also be applied to real data. For this purpose global sea level anomalies from altimeter and tide level data will be investigated.