Exploitation of Electromagnetic Models for Sea Wind Speed Estimation from C-Band Sentinel-1 Images

Among the different available wind sources, i.e. in situ measurements, numeric weather models, the retrieval of wind speed from Synthetic Aperture Radar (SAR) data is one of the most widely used methods, since it can give high wind resolution cells. For this purpose, one can find two principal approaches: via electromagnetic (EM) models and empirical (EP) models. In both approaches, the Geophysical Model Functions (GMFs) are used to describe the relation of radar scattering, wind speed, and the geometry of observations. By knowing radar scattering and geometric parameters, it is possible to invert the GMFs to retrieve wind speed. It is very interesting to compare wind speed estimated by the EM models, general descriptions of radar scattering from sea surface, to the one estimated by the EP models, specific descriptions for the inverse problem. Based on the comparisons, some ideas are proposed to improve the performance of the EM models for wind speed retrieval.

First assessment of Noise-Equivalent Sigma-Zero in GF3-02 TOPSAR mode with sea surface wind speed retrieval

Gaofen-3-02(GF3-02)is the first C-band synthetic aperture radar(SAR)satellite with terrain observation with progressive scans of SAR(TOPSAR)imaging mode in China,which plays an essential role in marine environment monitoring.Given the weak scattering characteristics of the ocean,the system thermal noise superimposed on SAR images has significant interference,especially in cross-polarization channels.Noise-Equivalent Sigma-Zero(NESZ)is a measure of the sensitivity of the radar to areas of low backscatter.The NESZ is defined to be the scattering cross-section coefficient of an area which contributes a mean level in the image equal to the signal-independent additive noise level.For TOPSAR,NESZ exhibits the shape of the SAR scanning gain curve in the azimuth and the shape of the antenna pattern in the range.Therefore,the accurate measurement of NESZ plays a vital role in the application of spaceborne SAR sea surface cross-polarization data.This paper proposes a theoretical calculation method for the NESZ curve in GF3-02 TOPSAR mode based on SAR noise inner calibration data and the imaging algorithm.A method for correcting the error existing in the theoretical curve of NESZ is also proposed according to the relationship between sea surface backscattering and wind speed and the same characteristics of target scattering in the overlapping area of adjacent sub-swaths.According to assessment with wide-swath TOPSAR cross-polarization data,the GF3-02 TOPSAR mode has a very low thermal noise level,which is better than−33 dB at the edge of each beam,and controlled below−38 dB at the center of the beam.The two-dimensional reference curves of the NESZ of each beam are provided to the GF3-02 TOPSAR users.After discussing the relationship between normalized radar cross section(NRCS)and wind speed,we provide a formula for NRCS related to wind speed and radar incidence angle.Compared with the NRCS derived from this formula and the NESZ-subtracted NRCS of SAR images,the bias is−0.0048 dB,the Root Mean Square Error is 1.671 dB and the correlation coefficient is 0.939.

洋面风场的QuikSCAT/SeaWinds遥感探测

QuikSCAT卫星是监测全球气候和海洋环境的重要卫星之一,它已经被世界气象组织列为业务卫星,所得数据对天气预报、气候研究等方面有重要作用。介绍了QuikSCAT卫星的有关背景情况,星载SeaWinds散射测风计观测洋面风的原理和基本算法,说明了利用该资料在热带气旋中心定位及移向预报以及2003年淮河流域暴雨成因分析方面做了尝试性工作。最后,对该数据资料应用中存在的一些问题进行了讨论。

Evaluation of reanalysis and satellite-based sea surface winds using in situ measurements from Chinese Antarctic Expeditions

Sea surface winds from reanalysis （NCEP-2 and ERA-40 datasets） and satellite-based products （QuikSCAT and NCDC blended sea winds） are evaluated using in situ ship measurements from the Chinese National Antarctic Research Expeditions （CH1NAREs） from 1989 through 2006, with emphasis on the Southern Ocean （south of 45°S）. Compared with ship observations, the reanalysis winds have a positive mean bias （0.32 m·s-1 for NCEP-2 and 0.13 m·s-1 for ERA-40）, and this bias is more pronounced in the Southern Ocean （0.57 m·s-1 and 0.45 m·s-1, respectively）. However, mean biases are negative in the tropics and subtropics. The satellite-based winds also show positive mean biases, larger than those of the reanalysis data. All four wind products overestimate ship wind speed for weak winds （〈4 m·s-1） but underestimate for strong winds （〉10 m·s-1）. Differences between the reanalysis and satellite winds are examined to identify regions with large discrepancies.

Effect of sea surface wind on the seasonal variation of sea level in the east of China seas

Sea level seasonal variations in the east of China seas from 2004 to 2006 are simulated by the advanced ROMS model. The results show similar sea level spatial features with TOPEX/Poseidon observations, with annual ranges decreasing gradually from the sea coast to the Kuroshio region. By getting rid of wind stress in ROMS model, the simulated sea level results still show obvious seasonal variations. However, the phenomenon of sea level anomaly disappears in Min Zhe Current Coastwise （MZCF） and Su Bei current coastwise （SBCF）, and the change of it from coastal area to ocean recedes. The seal level difference between Bohai, Yellow Sea （BYS） and East China Sea （ECS） becomes weaker in spring and autumn. The annual differences decrease obviously, and the gradual change of annual ranges from seacoast to the Kuroshio almost disappears. The annual ranges in BYS are nearly identical. The annual range ratio without the wind stress to with the wind stress increases gradually from the sea coast to Kuroshio region.

Sea Surface Roughness Derivation from Wind Speed Estimated by Satellite Altimeter

For open sea conditions the sea surface roughness is described as a function of surface stress and wind speed over sea surface by Charnock relation. The sea surface roughnessn in the North-west Pacific Ocean is derived successfully using wind speed data estimated by the TOPEX satellite altimeter. From the results we find that： （1） the mean sea surface roughness in winter is greater than in summer; （2） compared with other sea areas, the sea surface roughness in the sea area east of Japan （ N30°- 40°, E135°- 150°） is larger than in other sea areas; （3） sea surface roughness in the South China Sea changes more greatly than that in the Bohai Sea, Yellow Sea and East China Sea.

Evaluation of ENVISAT ASAR data for sea surface wind retrieval in Hong Kong coastal waters of China

The C-band wind speed retrieval models, CMOD4, CMOD - IFR2, and CMOD5 were applied to retrieval of sea surface wind speeds from ENVISAT （European environmental satellite） ASAR （advanced synthetic aperture radar） data in the coastal waters near Hong Kong during a period from October 2005 to July 2007. The retrieved wind speeds are evaluated by comparing with buoy measurements and the QuikSCAT （quick scatterometer） wind products. The results show that the CMOD4 model gives the best performance at wind speeds lower than 15 m/s. The correlation coefficients with buoy and QuikSCAT winds are 0.781 and 0.896, respectively. The root mean square errors are the same 1.74 m/s. Namely, the CMOD4 model is the best one for sea surface wind speed retrieval from ASAR data in the coastal waters near Hong Kong.

The seasonal variations in the significant wave height and sea surface wind speed of the China's seas

Long-term variations in a sea surface wind speed （WS） and a significant wave height （SWH） are associated with the global climate change, the prevention and mitigation of natural disasters, and an ocean resource exploitation, and other activities. The seasonal characteristics of the long-term trends in China＇s seas WS and SWH are determined based on 24 a （1988-2011） cross-calibrated, multi-platform （CCMP） wind data and 24 a hindcast wave data obtained with the WAVEWATCH-III （WW3） wave model forced by CCMP wind data. The results show the following. （1） For the past 24 a, the China＇s WS and SWH exhibit a significant increasing trend as a whole, of 3.38 cm/（s.a） in the WS, 1.3 cm/a in the SWH. （2） As a whole, the increasing trend of the China＇s seas WS and SWH is strongest in March-April-May （MAM） and December-January-February （DJF）, followed by June-July-August （JJA）, and smallest in September-October-November （SON）. （3） The areal extent of significant increases in the WS was largest in MAM, while the area decreased in JJA and DJF; the smallest area was apparent in SON. In contrast to the WS, almost all of China＇s seas exhibited a significant increase in SWH in MAM and DJF; the range was slightly smaller in JJA and SON. The WS and SWH in the Bohai Sea, the Yellow Sea, East China Sea, the Tsushima Strait, the Taiwan Strait, the northern South China Sea, the Beibu Gull and the Gulf of Thailand exhibited a significant increase in all seasons. （4） The variations in China＇s seas SWH and WS depended on the season. The areas with a strong increase usually appeared in DJF.

Evaluation on monthly sea surface wind speed of four reanalysis data sets over the China seas after 1988

This study investigates the long-term changes of monthly sea surface wind speeds over the China seas from 1988 to 2015. The 10-meter wind speeds products from four major global reanalysis datasets with high resolution are used: Cross-Calibrated Multi-Platform data set(CCMP), NCEP climate forecast system reanalysis data set(CFSR),ERA-interim reanalysis data set(ERA-int) and Japanese 55-year reanalysis data set(JRA55). The monthly sea surface wind speeds of four major reanalysis data sets have been investigated through comparisons with the longterm and homogeneous observation wind speeds data recorded at ten stations. The results reveal that(1) the wind speeds bias of CCMP, CFSR, ERA-int and JRA55 are 0.91 m/s, 1.22 m/s, 0.62 m/s and 0.22 m/s, respectively.The wind speeds RMSE of CCMP, CFSR, ERA-int and JRA55 are 1.38 m/s, 1.59 m/s, 1.01 m/s and 0.96 m/s,respectively;(2) JRA55 and ERA-int provides a realistic representation of monthly wind speeds, while CCMP and CFSR tend to overestimate observed wind speeds. And all the four data sets tend to underestimate observed wind speeds in Bohai Sea and Yellow Sea;(3) Comparing the annual wind speeds trends between observation and the four data sets at ten stations for 1988-1997, 1988–2007 and 1988–2015, the result show that ERA-int is superior to represent homogeneity monthly wind speeds over the China seaes.

A preliminary assessment of the sea surface wind speed production of HY-2 scanning microwave radiometer

A scanning microwave radiometer（RM） was launched on August 16,2011,on board HY-2 satellite.The six-month long global sea surface wind speeds observed by the HY-2 scanning microwave radiometer are preliminarily validated using in-situ measurements and WindSat observations,respectively,from January to June 2012.The wind speed root-mean-square（RMS） difference of the comparisons with in-situ data is 1.89 m/s for the measurements of NDBC and 1.72 m/s for the recent four-month data measured by PY30-1 oil platform,respectively.On a global scale,the wind speeds of HY-2 RM are compared with the sea surface wind speeds derived from WindSat,the RMS difference of 1.85 m/s for HY-2 RM collocated observations data set is calculated in the same period as above.With analyzing the global map of a mean difference between HY-2 RM and WindSat,it appears that the bias of the sea surface wind speed is obviously higher in the inshore regions.In the open sea,there is a relatively higher positive bias in the mid-latitude regions due to the overestimation of wind speed observations,while the wind speeds are underestimated in the Southern Ocean by HY-2 RM relative to WindSat observations.

Retrieval of sea surface winds under hurricane conditions from GNSS-R observations

Reflected signals from global navigation satellite systems（GNSSs） have been widely acknowledged as an important remote sensing tool for retrieving sea surface wind speeds.The power of GNSS reflectometry（GNSS-R）signals can be mapped in delay chips and Doppler frequency space to generate delay Doppler power maps（DDMs）,whose characteristics are related to sea surface roughness and can be used to retrieve wind speeds.However,the bistatic radar cross section（BRCS）,which is strongly related to the sea surface roughness,is extensively used in radar.Therefore,a bistatic radar cross section（BRCS） map with a modified BRCS equation in a GNSS-R application is introduced.On the BRCS map,three observables are proposed to represent the sea surface roughness to establish a relationship with the sea surface wind speed.Airborne Hurricane Dennis（2005） GNSS-R data are then used.More than 16 000 BRCS maps are generated to establish GMFs of the three observables.Finally,the proposed model and classic one-dimensional delay waveform（DW） matching methods are compared,and the proposed model demonstrates a better performance for the high wind speed retrievals.

Satellite SAR observation of the sea surface wind field caused by rain cells

Rain cells or convective rain,the dominant form of rain in the tropics and subtropics,can be easy detected by satellite Synthetic Aperture Radar（SAR） images with high horizontal resolution.The footprints of rain cells on SAR images are caused by the scattering and attenuation of the rain drops,as well as the downward airflow.In this study,we extract sea surface wind field and its structure caused by rain cells by using a RADARSAT-2 SAR image with a spatial resolution of 100 m for case study.We extract the sea surface wind speeds from SAR image by using CMOD4 geophysical model function with outside wind directions of NCEP final operational global analysis data,Advance Scatterometer（ASCAT） onboard European Met Op-A satellite and microwave scatterometer onboard Chinese HY-2 satellite,respectively.The root-mean-square errors（RMSE） of these SAR wind speeds,validated against NCEP,ASCAT and HY-2,are 1.48 m/s,1.64 m/s and 2.14 m/s,respectively.Circular signature patterns with brighter on one side and darker on the opposite side on SAR image are interpreted as the sea surface wind speed（or sea surface roughness） variety caused by downdraft associated with rain cells.The wind speeds taken from the transect profile which superposes to the wind ambient vectors and goes through the center of the circular footprint of rain cell can be fitted as a cosine or sine curve in high linear correlation with the values of no less than 0.80.The background wind speed,the wind speed caused by rain cell and the diameter of footprint of the rain cell with kilometers or tens of kilometers can be acquired by fitting curve.Eight cases interpreted and analyzed in this study all show the same conclusion.

Retrieval algorithm of sea surface wind vectors for WindSat based on a simple forward model

WindSat/Coriolis is the first satellite-borne polarimetric microwave radiometer, which aims to improve the potential of polarimetric microwave radiometry for measuring sea surface wind vectors from space. In this paper, a wind vector retrieval algorithm based on a novel and simple forward model was developed for WindSat. The retrieval algorithm of sea surface wind speed was developed using multiple linear regression based on the simulation dataset of the novel forward model. Sea surface wind directions that minimize the difference between simulated and measured values of the third and fourth Stokes parameters were found using maximum likelihood estimation, by which a group of ambiguous wind directions was obtained. A median filter was then used to remove ambiguity of wind direction. Evaluated with sea surface wind speed and direction data from the U.S. National Data Buoy Center （NDBC）, root mean square errors are 1.2 rn/s and 30~ for retrieved wind speed and wind direction, respectively. The evaluation results suggest that the simple forward model and the retrieval algorithm are practicable for near-real time applications, without reducing accuracy.

MODELING AND QUANTITATIVE RETRIEVAL OF FINITE FIELD FOR THE TROPICAL SEA SURFACE WIND SPEED OF THE FY-3B MICROWAVE IMAGER

The purpose of this study is to select a suitable sea wind retrieval method for FY-3B(MWRI). Based on the traditional empirical model of retrieving sea surface wind speed, and in the case of small sample size of FY-3B satellite load regression analysis, this paper analyzes the channel differences between the FY-3B satellite microwave radiation imager(MWRI) and TMI onboard the TRMM. The paper also analyzes the influence of these differences on the channel in terms of receiving temperature, including channel frequency, sensitivity and scaling precision. Then, the limited range of new model coefficient regression analysis is determined(in which the channel range settings include the information and features of channel differences), the regression methods of the finite field are proposed, and the empirical model of wind speed retrieval applicable to MWRI is obtained, which achieves robust results. Compared to the TAO buoy data, the mean deviation of the new model is 0.4 m/s, and the standard deviation is 1.2 m/s. In addition,the schematic diagram of the tropical sea surface wind speed retrieval is provided.

Sea surface wind speed retrieval from Sentinel-1 HH polarization data using conventional and neural network methods

Conventional retrieval and neural network methods are used simultaneously to retrieve sea surface wind speed(SSWS)from HH-polarized Sentinel-1(S1)SAR images.The Polarization Ratio(PR)models combined with the CMOD5.N Geophysical Model Function(GMF)is used for SSWS retrieval from the HH-polarized SAR data.We compared different PR models developed based on previous C-band SAR data in HH-polarization for their applications to the S1 SAR data.The recently proposed CMODH,i.e.,retrieving SSWS directly from the HHpolarized S1 data is also validated.The results indicate that the CMODH model performs better than results achieved using the PR models.We proposed a neural network method based on the backward propagation(BP)neural network to retrieve SSWS from the S1 HH-polarized data.The SSWS retrieved using the BP neural network model agrees better with the buoy measurements and ASCAT dataset than the results achieved using the conventional methods.Compared to the buoy measurements,the bias,root mean square error(RMSE)and scatter index(SI)of wind speed retrieved by the BP neural network model are 0.10 m/s,1.38 m/s and 19.85%,respectively,while compared to the ASCAT dataset the three parameters of training set are–0.01 m/s,1.33 m/s and 15.10%,respectively.It is suggested that the BP neural network model has a potential application in retrieving SSWS from Sentinel-1 images acquired at HH-polarization.

Spatio-temporal variability in sea surface wind stress near and off the east coast of Korea

Sea surface wind stress variabilities near and off the east coast of Korea, are examined using 7 kinds of wind datasets from measurements at 2 coastal (land) stations and 2 ocean buoys,satellite scatterometer (QuikSCAT), and global reanalyzed products (ECMWF,NOGAPS,and NCEP/NCAR). Temporal variabilities are analyzed at 3 frequency bands; synoptic (2-20 d), intra-seasonal (20-90 d),and seasonal (>90 d).Synoptic and intra-seasonal variations are predominant near and off the Donghae City due to the passage of the mesoscale weather system. Seasonal variation is caused by southeastward wind stress during Asian winter monsoon. The sea surface wind stress from reanalyzed datasets.QuikSCAT and KMA-B measurements off the coast show good agreement in the magnitude and direction,which are strongly aligned with the alongshore direction.At the land-based sites,wind stresses are much weaker by factors of 3-10 due to the mountainous landmass on the east parts of Korea Peninsula.The first EOF modes(67%-70%) of wind stresses from reanalyzed and QuikSCAT data have similar structures of the strong southeastward wind stress in winter along the coast but show different curl structures at scales less than 200 km due to the orographic effects.The second EOF modes (23%-25%) show southwestward wind stress in every September along the east coast of the North Korea

Comparison of sea surface wind field measured by HY-2A scatterometer and WindSat in global oceans

In this study, we present a comprehensive comparison of the sea surface wind ?eld measured by scatterometer(Ku-band scatterometer) aboard the Chinese HY-2 A satellite and the full-polarimetric radiometer WindSat aboard the Coriolis satellite. The two datasets cover a four-year period from October2011 to September 2015 in the global oceans. For the sea surface wind speed, the statistical comparison indicates good agreement between the HY-2 A scatterometer and WindSat with a bias of nearly 0 m/s and a root mean square error(RMSE) of 1.13 m/s. For the sea surface wind direction, a bias of 1.41° and an RMSE of 20.39° were achieved after excluding the data collocated with opposing directions. Furthermore,discrepancies in sea surface wind speed measured by the two sensors in the global oceans were investigated.It is found that the larger dif ferences mainly appear in the westerlies in the both hemispheres. Both the bias and RMSE show latitude dependence, i.e., they have signi?cant latitudinal ?uctuations.

Comparison among four kinds of data of sea surface wind stress in the South China Sea

By using remote sensing (ERS) data, FSU data, COADS data and Hellerman & Rosen-stein objective analysis data to analyze the sea surface wind stress in the South China Sea, it is found that the remote sensing data have higher resolution and more reasonable values. Therefore we suggest that remote sensing data be chosen in the study of climatological features of sea surface wind stress and its seasonal variability in the South China Sea, especially in the study of small and middle scale eddies.

Assessment of wind products obtained from multiple microwave scatterometers over the China Seas

Sea surface winds （SSWs） are vital to many meteorological and oceanographic applications, especially for regional study of short-range forecasting and Numerical Weather Prediction （NWP） assimilation. Spaceborne seatterometers can provide global ocean surface vector wind products at high spatial resolution. However, given the limited spatial coverage and revisit time for an individual sensor, it is valuable to study improvements of multiple microwave scatterometer observations, including the advanced scatterometer onboard parallel satellites MetOp-A （ASCAT-A） and MetOp-B （ASCAT-B） and microwave scatterometers aboard Oceansat-2 （OSCAT） and HY-2A （HY2-SCAT）. These four scatterometer-derived wind products over the China Seas （0°-40°N, 105°-135°E） were evaluated in terms of spatial coverage, revisit time, bias of wind speed and direction, after comparison with ERA-Interim forecast winds from the European Centre for Medium-Range Weather Forecasts （ECMWF） and spectral analysis of wind components along the satellite track. The results show that spatial coverage of wind data observed by combination of the four sensors over the China Seas is about 92.8% for a 12-h interval at 12：00 and 90.7% at 24：00, respectively. The analysis of revisit time shows that two periods, from 5：30-8：30 UTC and 17：00-21：00 UTC each day, had no observations in the study area. Wind data observed by the four sensors along satellite orbits in one month were compared with ERA-Interim data, indicating that bias of both wind speed and direction varies with wind speed, especially for speeds less than 7 m/s. The bias depends on characteristics of each satellite sensor and its retrieval algorithm for wind vector data. All these results will be important as guidance in choosing the most suitable wind product for applications and for constructing blended SSW products.

The Application of Sea-Surface Wind Detection with Doppler Lidar in Olympic Sailing

The mobile incoherent Doppler lidar （MIDL）, which was jointly developed by State Key Laboratory of Severe Weather （LaSW） of the Chinese Academy of Meteorological Sciences （CAMS） and Ocean University of China, provided meteorological services during the Olympic sailing events in Qingdao in 2008. In this study, two experiments were performed based on these measurements. First, the capabilities of MIDL detection of sea-surface winds were investigated by comparing its radial velocities with those from a sea buoy. MIDL radial velocity was almost consistent with sea-buoy data; both reflected the changes in wind with time. However, the MIDL data was 0.5 m s-1 lower on average than the sea-buoy data due to differences in detection principle, sample volume, sample interval, spatial and temporal resolution. Second, the wind fields during the Olympic sailing events were calculated using a four-dimensional variation data assimilation （4DVAR） algorithm and were evaluated by comparing them with data from a sea buoy. The results show that the calculations made with the 4DVAR wind retrieval method are able to simulate the fine retrieval of sea-surface wind data--the retrieved wind fields were consistent with those of sea-buoy data. Overall, the correlation coefficient of wind direction was 0.93, and the correlation coefficient of wind speed was 0.70. The distribution of retrieval wind fields was consistent with that of MIDL radial velocity; the root-mean-square error between them had an average of only 1.52 m s-1^.