Internal Variability Role on Estimating Sea Level Acceleration in Fremantle Tide Gauge Station

Internal Variability Role on Estimating Sea Level Acceleration in Fremantle Tide Gauge Station

Low frequency internal signals bring challenges to signify the role of anthropogenic factors in sea level rise and to attain a certain accuracy in trend and acceleration estimations. Due to both spatially and temporally poor coverage of the relevant data sets, identification of internal variability patterns is not straightforward. In this study, the identification and the role of low frequency internal variability (decadal and multidecadal) in sea level change of Fremantle tide gauge station is analyzed using two climate indices, Pacific Decadal Oscillation (PDO) and Tripole Interdecadal Pacific Oscillation (TPI). It is shown that the multidecadal sea level variability is anticorrelated with corresponding components of climate indices in the Pacific Ocean, with correlation coefficients of −0.9 and −0.76 for TPI and PDO, respectively. The correlations are comparatively low on decadal time scale, −0.5 for both indices. This shows that internal variability on decadal and multidecadal scales affects the sea level variation in Fremantle unequally and thus, separate terms are required in trajectory models. To estimate trend and acceleration in Fremantle, three trajectory models are tested. The first model is a simple second-degree polynomial comprising trend and acceleration terms. Low passed PDO, representing decadal and interdecadal variabilities in Pacific Ocean, added to the first model to form the second model. For the third model, decomposed signals of decadal and multidecadal variability of TPI are added to the first model. In overall, TPI represents the low frequency internal variability slightly better than PDO for sea level variation in Fremantle. Although the estimated trends do not change significantly, the estimated accelerations varies for the three models. The accelerations estimated from the first and second models are statistically insignificant, 0.006 ± 0.012 mm yr−2 and 0.01 ± 0.01 mm yr−2, respectively, while this figure for the third model is 0.018 ± 0.011 mm yr−2. The outcome exemplifies the importance of modelling low frequency internal variability in acceleration estimations for sea level rise in regional scale.