2 College of Marine Sciences, Shanghai Ocean University, Shanghai 201306, China;
3 Shanghai Engineering Research Center on Estuarine and Oceanographic Mapping, Shanghai 201306, China
Given the background of global warming, climate change has become an important scientific problem. Previous related studies have tended to place greater emphasis on changes in temperature and precipitation than on wind field variation (Li et al., 2007; Cheng, 2010). To investigate the trends of wind energy, some earlier studies used meteorological station reports (Ren et al., 2005; Li et al., 2011), ship reports (Ward and Hoskins, 1996), reanalysis wind field data (Mei et al., 2010; Liu et al., 2013; Sun et al., 2016), and satellite observations (Sun et al., 2010; Zheng, 2011; Wang and Sui, 2013; Kuang et al., 2015). In the context of global climate change, long-term changes in wind energy resources is related to the expected benefit of wind power development, which would affect the formulation and implementation of strategies for the reduction of carbon dioxide emissions. At present, the projected operating period of wind farms is generally over 20 years, and the common practice considers the average wind conditions over the past 30 years to predict likely wind conditions for the following 20 years. This is obviously not consistent with the reality and therefore it is necessary to study a long-term trend of wind energy resources.
The trend of wind speed shows significant regional differences. Ward and Hoskins (1996) studied the near-surface wind over the global ocean using monthly means of the Comprehensive OceanAtmosphere Data set for 1949–1988. Their results indicate that there was no globally averaged trend in circulation strength in the corrected wind data, although there were regional patterns of an upward trend (notable in the tropical North Atlantic and extratropical North Pacific oceans) and downward trend (notable in the equatorial and tropical South Atlantic and subtropical North Pacific oceans). Young et al. (2011) used satellite altimeter data sets covering the period 1991–2008 to investigate global trends in wind speed. They found a general global trend of increase of wind speed, except for the areas of a weak positive trend noted above, particularly the North Pacific Ocean. Zheng and Pan (2014) used the CrossCalibrated Multi-Platform (CCMP) data set (Atlas et al., 2011) for the period 1988–2011 to analyze global trends in wind power density. In most of the global ocean, wind power has followed a significant trend of increase and the trend on the western coasts was found stronger than on the eastern coasts. However, the causes of such wind speed variability remain unresolved (Azorin-Molina et al., 2018).
Although previous research has reported a range of results, many studies (Liu et al., 2008; Lin et al., 2013; Jiang et al., 2016) have found an increasing trend in wind speed over the South China Sea. However, almost all studies that found such trends used mean values. In this study, the annual and seasonal (winter (DJF), spring (MAM), summer (JJA) and autumn (SON)) trends in wind power density over the South China Sea were investigated. Following the method proposed by Young et al. (2011), the trends were quantified based on the linear increase or decrease in the mean within the time series of the annual and seasonal mean and the 90th and 99th percentiles and presented in 0.25°×0.25° grid.2 STUDY AREA AND DATA SET 2.1 Study area
The South China Sea (Fig. 1) is one of the largest marginal seas of the western Pacific Ocean, located between 2.5°–23.5°N and 99.0°–122.0°E. It covers approximately 3.5 million square kilometers with an average water depth of 1 200 m. It is bordered to the north by Guangdong, Guangxi, Hainan, and Taiwan Provinces of China, to the east by the Philippine islands and the western side of Luzon Strait, to the southwest by Vietnam and the Malay Peninsula, and it connects the Pacific and Indian oceans through the Bashi Channel, Sulu Sea, and Strait of Malacca.
The South China Sea is affected by the northeast monsoon from mid-October to mid-March. Early in this period, there are frequent invasions of cold air when the strong and stable northeast monsoon brings dry and sunny weather to the northern coast. Later in the period, low temperatures and rainy days are common with increasing fog and poor visibility. The period of the southwest monsoon extends from midMay to mid-September. During this time, with more southwesterly winds, the temperatures and humidity are high, thunderstorms and heavy rain occur on the northern coast, and there are frequent typhoons.2.2 CCMP wind field data set
The CCMP data set of the ocean surface wind field (Atlas et al., 2011) is a result of an investigation funded by the "Making Earth Science Data Records for Use in Research Environments" program of the National Aeronautics and Space Administration (International Pacific Research Center, 2008). It combines data derived from the Special Sensor Microwave Imager, Advanced Microwave Scanning Radiometer Earth observing system, Tropical Rainfall Measuring Mission Microwave Imager, QuikScat, and other missions using a variational analysis method to produce a consistent climatological record of ocean surface vector winds. The data set (January 1, 1988, to December 31, 2011) of the wind field 10 m above the sea surface (0.0°–23.5°N, 99.0°–122.0°E) was selected for this study. The spatial and temporal resolutions of this data set are 0.25°×0.25° and 6 h, respectively.2.3 Data validation
For this study, wind data recorded by three buoys in the South China Sea (Fig. 1, Table 1) were selected as accurate representations of the wind field with which to validate the CCMP ocean surface wind speed. Comparisons of the time series of the buoyderived and CCMP wind speeds with 6-h temporal resolution are illustrated in Fig. 2. It can be seen that the agreement between the measured and CCMP wind speeds during the selected periods was reasonable. Therefore, the CCMP wind field data were considered reliable for assessing the trends of wind energy.
In addition, some error metrics were derived for quantitative evaluation of the performance of CCMP wind speed. These metrics included the correlation coefficient (R), bias, and root mean square error (RMSE). Under the conditions of low wind speed (< 5 m/s), medium wind speed (5≤v < 10 m/s), high wind speed (≥10 m/s) and all wind speeds, the three metrics can be calculated as follows:
where xi and yi represent the measured and CCMP data, respectively, x and y are the mean values of the measured and CCMP data, respectively and N is the total number of data. The error indices of wind speed are listed in Tables 2–4.
Tables 2–4 show the correlation coefficients are reasonable because all are statistically significant at the 95% level, except for low wind speed (buoy #1 and buoy '#3 significant at the 90% level, and buoy #2 not significant). In general, the correlation coefficients for low, medium and high wind speeds are smaller than for the total wind. According to the bias, the CCMP wind speed is lower than that of the buoys. Specifically, the CCMP wind speed is overestimated at low wind speeds (< 5 m/s) and underestimated at high wind speeds (≥10 m/s). Both the bias and the RMSE indicate the CCMP data have higher credibility at medium wind speeds (5 m/s≤ v < 10 m/s).2.4 Sea surface temperature
Sea surface temperature (SST) data were acquired from the European Centre for Medium-Range Weather Forecasts ERA-Interim reanalysis data set. The ERA-Interim data set is the latest global atmosphere reanalysis from 1979, which is continuously updated in real time. In the ERA-Interim data set, a 4-dimensional variational analysis with a 12-h analysis window is employed. The ERA-Interim data set can be downloaded from either of the following two websites: http://apps.ecmwf.int/data set and https://www.ecmwf.int/en/faq/what-mars. The SST data set selected for this study covered the period January 1, 1988, to December 31, 2011, and it encompassed the area 0.0°–23.5°N, 99.0°–122.0°E. The spatial and temporal resolutions of this data set are 0.25°×0.25° and 6 h, respectively.3 STATISTICAL METHOD
The objective of this study was to investigate the trends of mean wind power density over the South China Sea using CCMP wind data. First, the wind power density was calculated from the wind speed using Eq.4:
where Pij is the wind power density of a wind state at time i and grid point j, ρ0 is the standard sealevel air density, i.e., 1.225 kg/m3 (Zheng and Pan, 2014; Jiang et al., 2016), and vij is the wind speed of a wind state at time i and grid point j. The mean wind power density was calculated as follows:
where Pj is the mean wind power density at grid point j and N is the number of wind states (in a normal year, N=1 460).
To calculate the long-term trend (i.e., annual and monthly trends) of the mean, 90th and 99th percentiles (following Young et al. (2011)) of wind power density over the South China Sea during 1988–2011, a linear regression method (Eq.6) was applied between the time series (independent variable) and the wind power density anomaly series (dependent variable) to retrieve the sign and magnitude of the wind power density trend. The slope of the linear regression model represented the trend of wind power density (W/(m2∙a)):
where x is the time series, y is the wind power density anomaly series, a is the regression constant and b is the regression coefficient:
where pi is the time series, p is the mean of the time series (here,
Following both Young et al. (2011) and AzorinMolina et al. (2018), the statistical significance of the trends was reported at three P-level thresholds (significant at P < 0.05, significant at P < 0.10 and not significant at P>0.10) to evaluate the uncertainty of the estimated wind power density trends. Field significance of the detected significant trends at the 90% and 95% confidence levels (CLs) were evaluated by applying the t-test method to detect whether the grid cells or the average over the entire South China Sea series showed significant trends. Moreover, the mean and standard deviation of the wind power density was calculated considering the South China Sea in its entirety.
The Pearson's correlation coefficient (R) was employed to measure the relationship between the SST anomalies and the wind speed anomalies.4 RESULT AND DISCUSSION 4.1 Wind power density climatology
Table 5 shows the mean and standard deviation of the annual and seasonal wind power density. Affected by the northeast monsoon in winter and by the southeast monsoon in summer, there is an obvious seasonal change in the mean wind power density. The curve of variation follows a single-peaked model, which is higher in winter (116.5, 222.8 and 350.7 W/m2 for the mean, the 90th percentile and 99th percentile, respectively) and lower in spring (40.6, 87.9 and 191.9 W/m2 for the mean, the 90th percentile and 99th percentile, respectively). Regarding the standard deviation, the annual change is smaller than the seasonal change, the value of which is greatest in winter, intermediate in summer and autumn and smallest in spring.4.2 Trend variability of wind power density
Table 6 shows the trend of annual and seasonal wind power density over the South China Sea, where the calculated trend is statistically significant at the 90% CL (P < 0.10) and 95% CL (P < 0.05). There is a clear increase in wind power density for all three statistics. The magnitude of the trend of increase in the mean wind power density is smaller than the 90th and 99th percentiles. Such a result indicates that the intensity of extreme events is larger than that of the mean conditions, consistent with the findings of Young et al. (2011). The magnitude is greatest in winter, intermediate in spring, and smallest in summer and autumn.
The wind power density significance increases both annually and seasonally, except for autumn (all three statistics are non-significant) and summer (the 90th and 99th percentiles are non-significant) when the slight trend of increase is not significant.4.3 Temporal and spatial variations of wind power density
The annual and seasonal trend values for the mean, the 90th and 99th percentile wind power density over the South China Sea, based on the 24-year CCMP data, are shown in Figs. 3–5, respectively. Table 7 summarizes the relative frequency statistics of Figs. 3–5. In addition, the correlation coefficients of the trends were calculated for the South China Sea and illustrated by shading in Figs. 3–5. In the regions shaded gray and dark gray, the absolute values of the correlation coefficients are larger than 0.34 and 0.40, respectively, which means these regions are statistically significant at the 90% and 95% CLs, respectively.
As shown in Fig. 3 and Table 7, the mean wind power density over most of the South China Sea has increased over the past 24 years. Overall, 86.7%– 96.3% of grid cells show positive trends for both annual and seasonal means. The weak negative trends distributed mainly around Palawan and the adjacent seas of southwest Luzon in all seasons and the annual means, and the adjacent seas of the Zhongsha Islands in summer and autumn (-1–0 W/(m2∙a)). Affected by the northeast monsoon in winter and by tropical cyclones in summer and autumn, the increasing trend is stronger over the northern parts of the South China Sea than the southern areas. The largest values are found over the seas adjacent to southwest Taiwan, China and Luzon Strait (>8 W/(m2∙a)). The mean wind power density trend is considerably larger in winter than in the other seasons and the annual mean.
The calculated mean wind power density trend over most of the South China Sea is statistically significant annually (85.1%), in winter (78.5%), and in spring (89.4%). In summer and autumn, the trend is statistically significant over approximately 50% of the area. The areas of non-significance distributed mainly in a belt from the southeastern Indochina Peninsula to the Zhongsha Islands and as an ellipse northwest of Palawan annually and in winter. The belt is reduced to an ellipse located to the south of the Indochina Peninsula in spring, and confined to central and southern areas of the South China Sea in summer and autumn.
The spatial distributions of the 90th percentile (Fig. 4) and mean wind power density trends are similar, although the magnitude of the former is obviously larger than the latter. Moreover, the range of the positive trends of the 90th percentile is in accord with the annual and seasonal means, except for the autumn (areas reduced by 4.9%). The weak negative trends distributed mainly in the seas adjacent to west Luzon in the winter, spring, summer, and annual means (values of trends below -1 W/(m2∙a)). In autumn, the areas of negative trend distributed mainly in a belt from the southeast Indochina Peninsula to Luzon (-2 W/(m2∙a)). The magnitude of the 90th percentile wind power density trend is the largest in winter, intermediate in spring and summer, and the smallest in autumn. Areas with strong increasing trend are located mainly to the south of Guangdong Province (5–10 W/(m2∙a) in summer and autumn; 10– 15 W/(m2∙a) in winter, spring, and annual means), the Beibu Gulf (10 W/(m2∙a) in spring and summer; 15 W/ (m2∙a) in annual mean; 20 W/(m2∙a) in winter and autumn), Gulf of Thailand (>10 W/(m2∙a)) in all seasons and annual means, except spring: 5 W/(m2∙a)), adjacent seas to the south of the Indochina Peninsula (>15 W/(m2∙a) in winter, spring and annual means; 8–10 W/(m2∙a) in summer and autumn), northwest of Kalimantan (>5 W/(m2∙a) in all seasons and annual means), northwest of Luzon (>10 W/(m2∙a) in summer; 20 W/(m2∙a) in autumn; 30 W/(m2∙a) in spring and annual means; and 50 W/(m2∙a) in winter), and southwest of Taiwan, China (>10 W/(m2∙a) in summer; 30 W/(m2∙a) in spring, autumn and annual means; and 40 W/(m2∙a) in winter).
The statistical significance of the 90th percentile wind power density trend is in accord with the mean in all seasons and the annual means, except for spring (significant areas reduced by 11.7%). The nonsignificant areas of the 90th percentile wind power density trend in spring comprise the two ellipses combined (the non-significant areas of the mean wind power density trends in spring).
For the three P-level thresholds, the relative frequencies were calculated with respect to the total number of grid cells showing positive or negative tendencies; the spatial distributions are shown in Figs. 3–5.
As Fig. 5 and Table 7 show, the 99th percentile wind power density trend becomes increasingly positive compared with the mean and the 90th percentile, indicating that extreme wind power density is increasing over the South China Sea by at least 5 W/ (m2∙a) (the relative frequencies of grid cells with magnitude >5 W/(m2∙a) are 88.7%, 81.3%, 72.8%, 70.3%, and 85.9% in winter, spring, summer, autumn, and annual means, respectively). The areas with the greatest trends of increase are distributed mainly in the Taiwan (China) and Luzon straits (with the magnitude of 50, 60, 100, 70, and 50 W/(m2∙a) in winter, spring, summer, autumn, and annual means, respectively). The central region of the South China Sea (areas around the Zhongsha Islands), showing a weak trend (-1–3 W/(m2∙a)) for the mean and the 90th percentile in summer, now shows a strong negative trend (10–20 W/(m2∙a)). Moreover, in autumn, the belt from southeast of the Indochina Peninsula to Luzon, showing a weak trend (-2–1 W/(m2∙a)) for the mean and the 90th percentile, now shows a strong negative trend (5–10 W/(m2∙a)).
The statistical significance of the 99th percentile wind power density trend is in accord with the 90th percentile in summer and autumn. The non-significant areas of the wind power density trend in winter, spring, and annual means are much larger than the mean and the 90th percentile (relative frequencies of grid cells with non-significant trends are 38.3%, 36.5% and 30.0% in winter, spring and annual means, respectively).4.4 Influence of SST on wind power density variability
The wind power density variability over China's adjacent sea areas has been attributed to various causes: (1) atmospheric circulation and monsoon changes (Wang et al., 2004), (2) El Niño indices (Zheng et al., 2013) and (3) SST (Hong et al., 2014). Here, the Pearson's correlation coefficient (R) was employed to measure the relationship between SST anomalies and wind speed anomalies (Fig. 6, Table 8).
As shown in Fig. 6 and Table 8, there is a negative correlation between SST and wind power density variability over the majority of the South China Sea in all seasons and annual means during 1988–2011; except winter (relative frequency of grid cells is 41.7%). The areas of positive relationship are greatest in winter (58.3%), intermediate in spring (27.5%) and summer (26.1%), and smallest in autumn (17.8%). In winter, the areas of negative correlation are located mainly in the northern Beibu Gulf, east of the Leizhou Peninsula, north of Kalimantan, and southeast of the Indochina Peninsula. In spring, the areas of positive relationship are located mainly in the southern Beibu Gulf, the northeastern South China Sea, areas surrounding Huangyan Island, northwest of Kalimantan and the Gulf of Thailand. In summer, the areas of positive relationship are located mainly in the Beibu Gulf, northern South China Sea, northwest of Kalimantan and northeast of Malaysia. In autumn, the areas of positive relationship are located mainly to the north of 20°N in the South China Sea.
Regarding the annual means, the areas of positive relationship are located mainly in the Gulf of Thailand, seas surrounding Malaysia, to the northwest and northeast of Kalimantan and to the north of 20°N in the South China Sea, except for areas in the northern Beibu Gulf and to the east of the Leizhou Peninsula.
The areas with the positive significant relationship are located mainly from east of Hainan to south of Taiwan, China in winter (13.6%) and annually (10.2%), and in the Beibu Gulf and a belt near 20°N, 113°E in summer (6.5%). The areas with the negative significant relationship are located mainly in the central and southern South China Sea in summer (42.7%) and autumn (33.6%), areas around the Nansha and Dongsha islands in spring (23.2%) and areas around the Nansha Islands annually (16.5%).5 CONCLUSION
This study investigated the trends of wind power density over the South China Sea using the CCMP data set for the period of 1988–2011, based on which the following conclusions were derived.
During the 24-year study period, averaged over the entire South China Sea, there has been a clear trend of increase in wind power density for all three base statistics (i.e., the mean, the 90th percentile and the 99th percentile) in all seasons and annual means. Furthermore, the wind power density has increased significantly (at the 90% CL) both annually and seasonally. However, during autumn (for all three statistics) and summer (the 90th and 99th percentiles), no significant trend was identified.
The seasonal trends of wind power density over the South China Sea show distinct temporal and spatial variations. During the 24-year study period, wind power density has shown a trend of increase in most areas of the South China Sea, which would benefit further development of the offshore wind energy industry (Zheng and Pan, 2014). The magnitude of the trend was greatest in winter, intermediate in spring, and smallest in summer and autumn. This trend was stronger in northern areas of the South China Sea than in southern parts. The largest value of the significant trend of increase was found in seas adjacent to the southwest of Taiwan, China and the Luzon Strait. At extreme conditions (99th percentile), central parts of the South China Sea showed a strong negative trend (10–20 W/(m2∙a)) in summer. Moreover, in autumn, the belt from southeast of the Indochina Peninsula to Luzon showed a strong negative trend (5–10 W/(m2∙a)).However, all strong negative trends were non-significant.
The magnitude of the annual and seasonal trends over the South China Sea was larger for extreme high events (i.e., the 90th and 99thpercentiles) relative to the mean condition.
A negative relationship between SST and wind power density variability was found over the majority of the South China Sea in all seasons and annual means; except winter (41.7%). The areas with positive significant relationship were located mainly from east of Hainan to the south of Taiwan, China in winter (13.6%) and annually (10.2%), and in the Beibu Gulf and a belt near 20°N, 113°E in summer (6.5%). The areas with negative significant relationship were located mainly in central and southern parts of the South China Sea in summer (42.7%) and autumn (33.6%), areas around the Nansha and Dongsha Islands in spring 23.2%) and areas around the Nansha Islands annually (16.5%).6 DATA AVAILABILITY STATEMENT
The Cross-Calibrated Multi-Platform (CCMP) wind field data set was provided by the Making Earth Science data records for Use in Research Environments (MEaSUREs) programof the National Aeronautics and Space Administration (NASA), and the sea surface temperature data are provided by the European Centre for Medium-Range Weather Forecasts (ECMWF).
Atlas R, Hoffman R N, Ardizzone J, Leidner S M, Jusem J C, Smith D K, Gombos D. 2011. A cross-calibrated, multiplatform ocean surface wind velocity product for meteorological and oceanographic applications. Bulletin of the American Meteorological Society, 92(2): 157-174. DOI:10.1175/2010BAMS2946.1
Azorin-Molina C, Menendez M, McVicar T R, Acevedo A, Vicente-Serrano S M, Cuevas E, Minola L, Chen D L. 2018. Wind speed variability over the Canary Islands, 1948-2014:focusing on trend differences at the landocean interface and below-above the trade-wind inversion layer. Climate Dynamics, 50(11-12): 4 061-4 081. DOI:10.1007/s00382-017-3861-0
Cheng S. 2010. Temporal and Spatial Variety Characteristics of Wind Speed and Wind Energy in China Over the Past Half Century. Nanjing University of Information Science and Technology, Nanjing. p.1-65. (in Chinese with English abstract)
Hong X Y, Fu Y F, Xian T, Wang Y, Yang Y J, Heng Z W. 2014. Relationships between sea surface wind speed and the diurnal variation of sea surface temperature over the tropical ocean. Climatic and Environmental Research, 19(4): 437-451. (in Chinese with English abstract)
International Pacific Research Center.January 16, 2018. Crosscalibrated multi-platform (CCMP) ocean surface wind components. http://apdrc.soest.hawaii.edu/datadoc/ccmp_6hourly.php.
Jiang B, Hou P P, Zhang S, Wu H, Ding J. 2016. Uncertainty of the offshore wind energy resource assessment due to longterm variability. In: Proceedings of the 2nd International Conference on Sustainable Development. Atlantis Press, Xi'an. p.298-303.
Kuang F F, Zhang Y Q, Zhang J P, Jia C. 2015. Comparison and evaluation of three sea surface wind products in Taiwan Strait. Acta Oceanologica Sinica, 37(5): 44-53. (in Chinese with English abstract)
Li Y, Tang J P, Wang Y, Chu H Y. 2011. Prediction of climate change of the near-surface wind energy potential over China. Acta Energiae Solaris Sinica, 32(3): 338-345. (in Chinese with English abstract)
Li Y, Wang Y, Tang J P. 2007. Temporal and spatial variety characteristics in Near-surface wind energy in China. Journal of Nanjing University (Natural Sciences), 43(3): 280-291. (in Chinese with English abstract)
Lin G, Zheng C W, Shao L T, Li Q H, Liu T N. 2013. Long term linear trend of sea surface wind speed in the South China Sea. Journal of PLA University of Science and Technology (Natural Science Edition), 14(6): 704-708. (in Chinese with English abstract)
Liu T J, Zheng C W, Li X Q, Zhang W J. 2013. Long term trend of sea surface wind speed in the Northern Indian Ocean from 1958 to 2001. Marine Forecasts, 30(4): 34-39. (in Chinese with English abstract)
Liu Y J, Yan J Y, Song Y L. 2008. Features of climate change over the Xishan Island over the South China Sea in recent 50 years. Scientia Geographica Sinica, 28(6): 804-808. (in Chinese with English abstract)
Mei Y, Song S, Zhou L. 2010. Annual variation characteristics of wave fields and wind fields over the North Indian Ocean and South China Sea. Marine Forecasts, 27(5): 27-33. (in Chinese with English abstract)
Ren G Y, Guo J, Xu M Z, Chu Z Y, Zhang L, Zou X K, Li Q X, Liu X N. 2005. Climate changes of China's mainland over the past half century. Acta Meteorologica Sinica, 63(6): 942-956. (in Chinese with English abstract)
Sun L, Yu H M, Wang P, Chang X J, Wang S L. 2010. Analysis of the seasonal and interannual variability of sea surface wind in the East China Seas and its adjacent ocean. Marine Forecasts, 27(2): 30-37. (in Chinese with English abstract)
Sun Z Q, Xiang J, Guan Y P. 2016. Wind energy in the offshore areas of China based on ERA-interim reanalysis data. Marine Forecasts, 33(3): 50-56. (in Chinese with English abstract)
Wang H, Sui W H. 2013. Seasonal variation analysis of sea surface winds in China Sea areas with CCMP wind field data. Meteorological Science and Technology, 41(4): 720-725. (in Chinese with English abstract)
Wang Z Y, Ding Y H, He J H, Yu J. 2004. An updating analysis of the climate change in China in recent 50 years. Acta Meteorologica Sinica, 62(2): 228-236. (in Chinese with English abstract)
Ward M N, Hoskins B J. 1996. Near-surface wind over the global ocean 1949-1988. Journal of Climate, 9(8): 1 877-1 895. DOI:10.1175/1520-0442(1996)009<1877:NSWOTG>2.0.CO;2
Young I R, Zieger S, Babanin A V. 2011. Global trends in wind speed and wave height. Science, 332(6028): 451-455. DOI:10.1126/science.1197219
Zheng C W, Pan J. 2014. Assessment of the global ocean wind energy resource. Renewable and Sustainable Energy Reviews, 33: 382-391. DOI:10.1016/j.rser.2014.01.065
Zheng C W, Zhou L, Huang C F, Shi Y L, Li J X, Li J. 2013. The long-term trend of the sea surface wind speed and the wave height (wind wave, swell, mixed wave) in global ocean during the last 44a. Acta Oceanologica Sinica, 32(10): 1-4. DOI:10.1007/s13131-013-0358-5
Zheng C W. 2011. Sea surface wind field analysis in the China Sea during the last 22 years with CCMP wind field. Meteorology and Disaster Reduction Research, 34(3): 41-46. (in Chinese with English abstract)