Abstract's details
Rethinking the Modeling of the Mean Sea Surface in the Era of Climate Change
CoAuthors
Event: 2022 Ocean Surface Topography Science Team Meeting
Session: The Geoid, Mean Sea Surfaces and Mean Dynamic Topography
Presentation type: Type Oral
Contribution: PDF file
Abstract:
Since the beginning of the precision satellite altimeter era in the early 1990s, efforts have been focused on computing the mean height of the ocean surface for use in various geodetic and oceanographic studies. With 30 years of satellite measurements now available, it is time to rethink the modeling of the mean sea surface (MSS) in the era of climate change.
There are linear changes in the height of the ocean surface due to melting ice and increasing ocean heat content that will not average to zero when computing the mean height [Fasullo and Nerem, 2018; Hamlington et al., 2020]. Today, there are places in the ocean that are 15 cm higher than they were 30 years ago, and today the average error is around 5 cm for conventional MSS models.
We propose that linear sea level changes be used in a new definition of the MSS correction which is tied to a particular date in time. For example, one definition could be:
MSS(ϕ,λ,t)= MSS(ϕ,λ,2003.0)+(MSS) ̇(ϕ,λ)(t-2003.0)
Where MSS(ϕ,λ,2003.0) is the height of the sea surface in 2003.0 and the epoch used for the most common MSS models, e.g., the DTU and CLS MSS models. (MSS) ̇(ϕ,λ) is the linear rate of change of the sea surface height averaged over 30 years.
Linear changes will not average to zero over time and thus can build up to cause large errors in the conventional definition of the MSS. If one is unconcerned with the trends, then the MSS(ϕ,λ,2003.0) can be used in the same manner as for conventional MSS models. It’s also possible to consider including the acceleration of sea surface height, but these are small [Nerem et al., 2018] and a longer time series is needed before a stable map of the accelerations can be computed [Hamlington et al., 2020].
The linear rate of change (MSS) ̇(ϕ,λ) will of course change in the future as more data are added, but testing shows that it remain reasonably stable over a decade or so, and can be updated when new MSS models are produced. There is considerable evidence that the 30-year trend patterns in sea surface height are driven by the “forced response” of Greenhouse gases and aerosols, and these patterns will be reasonably persistent as we move forward in time [Fasullo and Nerem, 2018].
We will present the testing of a new mean sea surface correction and find the performance to be significant for recent satellites like Jason-3 and Sentinel 3 and 6. It reduces the mean offset for these recent satellites wrt with respect to conventional satellites, but it also performs significantly better in reducing the spatial variability. With stability of the trend. We believe, that the MSS correction will be significant for the upcoming SWOT mission, as the MSS correction can be considered into the near future as sea level has changed considerably over the last 30 years.
There are linear changes in the height of the ocean surface due to melting ice and increasing ocean heat content that will not average to zero when computing the mean height [Fasullo and Nerem, 2018; Hamlington et al., 2020]. Today, there are places in the ocean that are 15 cm higher than they were 30 years ago, and today the average error is around 5 cm for conventional MSS models.
We propose that linear sea level changes be used in a new definition of the MSS correction which is tied to a particular date in time. For example, one definition could be:
MSS(ϕ,λ,t)= MSS(ϕ,λ,2003.0)+(MSS) ̇(ϕ,λ)(t-2003.0)
Where MSS(ϕ,λ,2003.0) is the height of the sea surface in 2003.0 and the epoch used for the most common MSS models, e.g., the DTU and CLS MSS models. (MSS) ̇(ϕ,λ) is the linear rate of change of the sea surface height averaged over 30 years.
Linear changes will not average to zero over time and thus can build up to cause large errors in the conventional definition of the MSS. If one is unconcerned with the trends, then the MSS(ϕ,λ,2003.0) can be used in the same manner as for conventional MSS models. It’s also possible to consider including the acceleration of sea surface height, but these are small [Nerem et al., 2018] and a longer time series is needed before a stable map of the accelerations can be computed [Hamlington et al., 2020].
The linear rate of change (MSS) ̇(ϕ,λ) will of course change in the future as more data are added, but testing shows that it remain reasonably stable over a decade or so, and can be updated when new MSS models are produced. There is considerable evidence that the 30-year trend patterns in sea surface height are driven by the “forced response” of Greenhouse gases and aerosols, and these patterns will be reasonably persistent as we move forward in time [Fasullo and Nerem, 2018].
We will present the testing of a new mean sea surface correction and find the performance to be significant for recent satellites like Jason-3 and Sentinel 3 and 6. It reduces the mean offset for these recent satellites wrt with respect to conventional satellites, but it also performs significantly better in reducing the spatial variability. With stability of the trend. We believe, that the MSS correction will be significant for the upcoming SWOT mission, as the MSS correction can be considered into the near future as sea level has changed considerably over the last 30 years.