| Based on three-parameters Weibull distribution model, urban-rural wind speed differences and its affected factors are discussed by using wind speed and temperature data of Jiangsu Province automatic meteorological station hourly data, 1985-2014 surface weather timing data,and NCEP/NCAR Reanalysis-1, NCEP/DOE Reanalysis-2, ERA-Interim reanalysis dataset.Urban-rural wind speed and long-term wind speed probability distribution function was fitted to find the differences of parameters and its changing trends. City size indexes are regressed with wind speed probability distribution parameters and its relative variables for a thorough cognition of urbanization on surface wind speed probability distribution. The conclusions are as follows:1. Trends of city stations average yearly wind speed are decreased overall, average change rate is -0.0555·10a-1. Wind speed of city stations decreased faster than county stations by contrast,which means urbanization makes wind speed decreased obviously. Reanalysis integrated average wind speed is too large but changing smoothly in a small variation compared with city observation data; Automatic station hourly wind speed and long-term wind speed data shows thaturban-rural wind speed difference and rural wind speed is in a negative linear correlation;Urbanization impacts on the surface wind speed probability distribution characteristics by potentiating underlying surface roughness and urban heat island effect. There exists a wind speed critical value, friction effect plays a leading role when wind speed is greater than it, which could weaken the wind speed, otherwise heating effect enlarges the wind speed when speed when wind speed is smaller than the critical value. Long-term data also shows that surface friction effect is strengthen overtime, which means the weaken effect of wind speed influenced by urbanization friction is more distinct. Enlarging function on wind speed of urbanization heating effect is slightly decreased but not significant.2. Fitting effect would be better by changing universal three parameters Weibull distribution function definition domain condition v?r into limited three parameters Weibull distribution function definition domain v?0, which would be more applied to the research of wind speed probability distribution change under the effect of urbanization. Statistic analysis indicates that a good result could be drawn in the fitting of wind speed probability Weibull distribution parameters by using numerical iteration, correlation coefficient of measured value and theoretical value reaches 0.9999. Theoretical analysis indicates that urbanization friction effect makes scale parameter c decreased, location parameter r moved to the positive direction of abscissa axis,shape parameter k is little affected by city underlying surface.3. Certain rules can be drawn by the comparison of urban and rural wind speed probability distribution parameters, among which urban wind speed parameter k is slightly smaller than rural, urban parameter c is smaller than rural, and urban parameter r is greater than rural stations;Most obvious difference of urban-rural parameters is parameter c , next comes parameter r , and urban-rural parameter k difference is not remarkable. City wind speed probability density peak value is greater than country, which implies that wind speed probability density curve of city is more concentrated and wind speed is gathered around the mode. Preliminary data analysis suggests that excepts the weaken effect of friction brought by urbanization, urban heat island effect could strengthen the surface layer turbulence, particularly in a stable atmosphere, so that calm wind probability comes out lesser.4. Long-term data shows that parameter k is slightly decreased, parameter c decreases distinctly, and parameter r increases along with time. Changes of wind speed probability distribution parameters of 8 cities could be divided into Type ? and Type ? approximately. Both types are decreased in scale parameter c and increased in probability density peak value with the development of city, which means the curves would become abrupt. Increment of location parameter r of Type ? is greater than of Type ?. Type ? decreases in calm wind probability but increases in mode and the curve moved to the right, while those statistic changes characteristicsof Type ? are diametrically opposite. Using county station and reanalysis wind speed data to fit the comparison parameters c' , r'and get relative variables c/c'?r-r', which mainly reflect the part affected by urbanization. Analysis of the relative variables shows a similar result with the consequences of parameter c and r. The change of c/c' mainly depends on friction function of the underlying surface, influencing by single factor, while the variation of r-r' is impact by complex elements, which includes friction, heat, and background climate turbulence.5. City size indexes correlate well with parameter c and r, while correlation is not significant with parameter k in comparison. Regressing effect of difference city size indexes and parameters exists slight discrepancy, but trends of 3 parameters are the same. Urban-reanalysis parameter relative variables c/c' is linear decreased with the adding of city build-up area, which means the distribution of wind speed probability density curve would narrow with the extent of the city.Trend of r-r' shows a linear rise with the increase of city build-up area and the value of r-r'changes from negative to positive, which signifies that calm and small wind probabilities would reduce along with city expansion. Strong effects are reflected in city surface wind speed in the rapid urbanization process, While which would be weaken when the city reaches a certain size such as Beijing, even stabilized the wind speed parameters in some cases. Analysis also indicates that latitude correlates well with the variation of r-r' regressed with urban population or build-up area, but the correlation is not significant in c/c'. |
PDF Full Text Request