Metocean Comparisons of Jason-2 and AltiKa—A Method to Develop a New Wind Speed Algorithm

Abstract

As well as range, the AltiKa altimeter provides estimates of wave height, H s and normalized backscatter, σ 0 , that need to be assessed prior to statistics based on them being included in climate databases. An analysis of crossovers with the Jason-2 altimeter shows AltiKa H s values to be biased high by only ~0.05m, with a standard deviation (s.d.) of ~0.1m for 7-point averages. AltiKa's σ 0 values are 2.5-3 dB less than those from Jason-2, with a s.d. of ~0.3 dB, with these relatively large mismatches to be expected as AltiKa measures a different part of the spectrum of sea surface roughness. A new wind speed algorithm is developed through matching a histogram of σ 0 values to that for Jason-2 wind speeds. The algorithm is robust to the use of short durations of data, with a consistency at roughly the 0.1 m/s level. Incorporation of H s as a secondary input reduces the assessed error at crossovers from 0.82 m/s to 0.71 m/s. A comparison across all altimeter frequencies used to date demonstrates that the lowest wind speeds preferentially develop the shortest scales of roughness.

Tropical Cyclone Wind Speed Estimation From Satellite Altimeter‐Derived Ocean Parameters

“…Physically, the higher magnitude of U 10 causes rougher perturbations on the sea surface, reducing the backscatter. Although the backscatter is the best parameter to infer U 10 of which several studies have also included the H S in their estimation (Gourrion et al, 2002;G. D. Quartly, 2015), the descriptive impact of the T b and the derived W L and W V to U 10 can be further explored with the anticipation that these parameters could provide the propagates altimetric signal with real tropical cyclone environment.…”

“…The Ku-band backscatter shows greater sensitivity to higher U 10 , however, other lower frequency signal have proved to be less affected by rain (G. D. Quartly, 2015). Studies indicated that the raindrops attenuate the Ku-band backscatter, at least, about 2 dB in which significant broadening leading edge waveforms was induced in the presence of high liquid water content (Lillibridge et al, 2014;G.…”

“…An early attempt by Lefevre et al (1994) successfully produced accurate U 10 from TOPEX-Poseidon by incorporating the significant wave height. Such idea has been practiced in other studies with great U 10 accuracy in Ku-band (Gourrion et al, 2002) and in SARAL Ka-band signal (Lillibridge et al, 2014;G. D. Quartly, 2015).…”

Observing Sea States

“…In particular, there is not even an agreed universal scale for σ 0 data, with the values recorded by two Ku-band altimeters differing by an offset of more than 2 dB. Furthermore, as mentioned above, there is no simple adjustment between normalized backscatter observations at one radar frequency with those at another (Figures 9E-F), because each is responding to a different scale length of sea surface roughness as illustrated by Figures 9E,F. Since there have been more than 30 years of Ku-band and C-band altimetry, the launch of SARAL operating only at Ka-band, necessitated a new specific effort to produce relevant wind speed algorithms (Lillibridge et al, 2014;Quartly, 2015).…”

“…Data are averaged for a full cycle of Jason-3, SARAL, Sentinel-3B LRM Mode (S-3B LRM, used in 13 days of cycle 10) and Sentinel-3B SAR Mode (S-3B SARM). (E) Backscatter strength (σ 0 ) as a function of wind speed, and (F) differences for different radar frequencies, adapted fromQuartly (2015), as authorized by Taylor & Francis.…”