Evaluation of altimeter undersampling in estimating global wind and wave climate using virtual observation

Abstract

<p>Altimeters can provide global long-duration observations of oceanic wind speed and wave height. However, altimeters face the undersampling problem in estimating wind and wave climate because of their sparse sampling pattern and the changing number of in-orbit satellites. In this study, the undersampling error of altimeters was studied by sampling the ERA5 oceanic wind speed and wave height data using the track information of multiplatform altimeters. Comparisons were made between the statistics (mean, 90<sup>th</sup> and 99<sup>th</sup> percentiles, and long-term trends of them) of the original ERA5 data and the gridded along-track sampling of the ERA5 data. The results show a large discrepancy with respect to the extreme values (90<sup>th</sup> and 99<sup>th</sup> percentiles). The undersampling of altimeters can lead to significant underestimations of monthly extreme values of oceanic wind speed and wave height. Meanwhile, this underestimation is alleviated with the increase of the number of in-orbit altimeters, leading to very large overestimations of long-term trends of these extreme values over the period 1985-2018. In contrast, the annual extreme values of oceanic wind speed and wave height and their long-term trends are more reliable, although slight aforementioned biases of extreme values still exist and the data from GEOSAT are not suitable for computing annual statistics. For altimeter data, the annual values are a better option to compute long-term trends than the monthly data. This study also presents a correction scheme of using model data to compensate for the wind and wave events missed by altimeter tracks. After the correction, the global trends in oceanic wind speed and wave height over 1992-2017 are recomputed using annual statistics. The results show a clear discrepancy between the trends of wind speed and wave height during this period: the wind speed increased, while the wave height decreased. However, the reason for this discrepancy is unknown at this stage.</p>