Integration Tide Gauge and Satellite Altimetry for Storm Surge and Sea Level change prediction. Ole B. Andersen and Y. Cheng (DTU, Denmark) Xiaoli Deng, M. Steward, N Idris, and Zahra Gharineiat (Uni NewCastle, Australia)
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Overview Why use satellite altimetry Crital issues: Spatio-temporal sampling vs surge/cyclone Availability and accuracy Accuracy degradation (Coastal and rain). Reliability of surge capturing Importance of residual range corr errors: Ocean tide correction Merging with tide gauges (spatial temporal correlation) Hindcast / forecast modelling Test region: North European Shelf Great Barrier Reef (Cyclone Helen and Larry) Conclusions. Queenland surge
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Satellite altimetry is well Esablished for linear sea level change
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Sea Level and Storm Surges Critial issues: Spatio-temporal sampling vs surge/cyclone Availability and accuracy Accuracy degradation (Coastal and rain). Reliability of surge capturing Importance of residual range corr errors: Ocean tide correction Merging with tide gauges (spatial temporal correlation) Hindcast / forecast modelling
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Typical Surges / Cyclones North sea has mainly external surges generated by wind forcing. GBR has numerous cyclones (near seasonal) = Water level recorders
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Correlations, Water level recorders Timescales, ~10-60 h Counterclockwise propagation of sea level anomalies AutocorrelationCrosscorrelation Curtesy of J. Hoyer
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Observations Sampling: Satellite Repeat Period Track spacing Inclination Coverage CryoSat km92°(+/-88°) AltiKa35 days80 km 98°(+/-82°) TOPEX/JASON10 days315 km66.5° 3-4 ongoing missions JASON-2 Cryosat-2 (SAR) AltiKa (French-India)** HY-2 (China) Sentinel-3 (SAR) Near Real time data J2+C2: 4-6 hours Accurcy: 4-6 cm
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Sea Level and Storm Surges Critial issues: √ Spatio-temporal sampling vs surge/cyclone √ Availability and accuracy Accuracy degradation (Coastal and rain). Reliability of surge capturing Importance of residual range corr errors: Ocean tide correction Merging with tide gauges (spatial temporal correlation) Hindcast / forecast modelling
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Heavy Rain can be problematic
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Sea Level and Storm Surges Critial issues: √ Spatio-temporal sampling vs surge/cyclone √ Availability and accuracy √ Accuracy degradation (Coastal and rain). Reliability of surge capturing Importance of residual range corr errors: Ocean tide correction Merging with tide gauges (spatial temporal correlation) Hindcast / forecast modelling
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Last devastating storm surge in Britain – 1953 ”Data for validation?”
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark High Water in Hvide Sande Simple 2 and 3 std. deviation test on ”high water”
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Satellite Obs One versus two satellite (TOPEX/ENVISAT) Degradation not an issue
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Great Barrier Reef – Summer Cyclones Helen (January, 04, 2008) Larry (March 19th, 2006)
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Sea Level and Storm Surges Critial issues: √ Spatio-temporal sampling vs surge/cyclone √ Availability and accuracy Accuracy degradation (Coastal and rain). Reliability of surge capturing Importance of residual range corr errors: Ocean tide correction Merging with tide gauges (spatial temporal correlation) Hindcast / forecast modelling
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Tide Correction Models Remove residual tidal signal using the ”pointwise tide model. This enhances The spatial correlation between tide gauge and
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Spatio-temporal Correlation TOPEX / JASON (17 years)ERS / ENVISAT (12 years)
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Temporal Correlation at Newlyn After Detiding – temporal correlation = 0.93 (0.89 FES 2004 / GOT4.7)
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Sea Level prediction CAN WE PREDICT/WARN ABOUT INCREASED SEA LEVEL ”STORM SURGES” FROM SATELLITE ALTIMETRY????? Main issues: Sampling (fixed tracks, sampling 9 / 17 or 35 days Availability of ACCURATE real time data (1-6 hours) Capture the Surge by the altimeter Merging Tide gauge and Altimetry (observe the same signal) Establish warning – ”hindcast” high water.
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Combing Tide gauges and Altimetry Small time scale, Large spatial scales, high correlation T/P + Gauges Least squares fit in every T/P observation Regression Model
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Integrating using Neural Network Test on using 4 gauges along coast of Norway. Training using one T/P time series Predicting other satellite points. Higher temporal correlation using Neural Network than regression method.
IAHS-IAPSO-IASPEI Assembly, Gothenburg, Sweden, July, DTU Space, Technical University of Denmark Satellite altimetry delivers: –Sea surface height –Wave height –Wind Speed predefined tracks+sampling Satellite altimetry has high temporal/spatial correlation with tide gauges Huge potential for ”high water forecasting” with tide gauges offers huge potential for improved warning in the future Comming satellites (Cryosat + Sentinel-3) might offer even better prediction in the future (higher resolution + across track coverage) Having vacation in Venice this summer……. Conclusions