Abstract's details
Calibrating Global Mean Sea Level over the Satellite-era
Event: 2014 Ocean Surface Topography Science Team Meeting
Session: Regional and Global CAL/VAL for Assembling a Climate Data Record
Presentation type: Oral
The accuracy of Global Mean Sea Level (GMSL) derived from the three successive satellite altimeter missions, TOPEX/Poseidon, Jason-1 and OSTM/Jason-2, is dependent on the determination of any inter- and intra-mission biases, as well as the stability of these biases over time. The global tide gauge (TG) network offers the only independent source of in-situ data with sufficient degrees of freedom to satisfactorily quantify any systematic error (so called 'bias drift') that may exist in altimeter derived trends in GMSL. Central to the technique is forming the difference in sea surface height between the altimeter and tide gauge at offshore comparison points (CPs). Key components of the technique include the mitigation of energy at tidal frequencies between the TG and CP, as well as accounting for vertical land motion (VLM) at the tide gauge.
Here we present results from a refined strategy to estimate bias drift across a range of altimeter processing treatments for the TOPEX-A, TOPEX-B, Jason-1 and OSTM/Jason-2 missions. We provide an overview of the revised methodology that avoids sensitivity to outlier CPs or any specific small subset of TGs in the network. Of greatest significance to the estimation of the trend in GMSL is that the largest bias drift is found within the ~6 year TOPEX-A record (+0.9 ± 0.4 mm/yr to +1.5 ± 0.4 mm/yr, depending on the VLM correction used, with the solution including VLM derived from Global Positioning System (GPS) data yielding the largest bias drift estimate). The difference between our bias drift estimates for any pair of satellite missions is robust to the range of available VLM corrections applied. Application of our bias drift estimates to derive a 'calibrated' GMSL series changes the overall linear rate as well as the character of the GMSL curve. We present 'calibrated' and 'uncalibrated' GMSL curves showing the transition from a curve with a small deceleration (uncalibrated data) to a curve with a small (non-significant) acceleration regardless of the VLM applied. The linear trend in GMSL over the three missions is reduced from +3.2 ± 0.4 mm/yr (uncalibrated) to +2.5 ± 0.4 mm/yr (calibrated, using a GPS based VLM dataset). We conclude with a discussion of several sensitivity analyses undertaken to assess the robustness of the technique as well as variability in results as a function of different data treatments (altimeter and VLM for example) and processing strategies (tides and across track slopes for example).
Here we present results from a refined strategy to estimate bias drift across a range of altimeter processing treatments for the TOPEX-A, TOPEX-B, Jason-1 and OSTM/Jason-2 missions. We provide an overview of the revised methodology that avoids sensitivity to outlier CPs or any specific small subset of TGs in the network. Of greatest significance to the estimation of the trend in GMSL is that the largest bias drift is found within the ~6 year TOPEX-A record (+0.9 ± 0.4 mm/yr to +1.5 ± 0.4 mm/yr, depending on the VLM correction used, with the solution including VLM derived from Global Positioning System (GPS) data yielding the largest bias drift estimate). The difference between our bias drift estimates for any pair of satellite missions is robust to the range of available VLM corrections applied. Application of our bias drift estimates to derive a 'calibrated' GMSL series changes the overall linear rate as well as the character of the GMSL curve. We present 'calibrated' and 'uncalibrated' GMSL curves showing the transition from a curve with a small deceleration (uncalibrated data) to a curve with a small (non-significant) acceleration regardless of the VLM applied. The linear trend in GMSL over the three missions is reduced from +3.2 ± 0.4 mm/yr (uncalibrated) to +2.5 ± 0.4 mm/yr (calibrated, using a GPS based VLM dataset). We conclude with a discussion of several sensitivity analyses undertaken to assess the robustness of the technique as well as variability in results as a function of different data treatments (altimeter and VLM for example) and processing strategies (tides and across track slopes for example).
Contribution: 29Ball1615-1-OSTST2014_Watson_etal_Bias_Drift.pdf (pdf, 2756 ko)
Back to the list of abstract