论文部分内容阅读
利用长江航道局布设在长江航道沿线重庆段39个人工雾情信号台近3a的雾情观测资料,详细分析了重庆段航道雾的时空分布特征,并基于以上雾情及整个长江山区航道(宜宾到宜昌)逐里程的经度、纬度、水道宽度、河面弯曲度、河道变化剧烈度、河道支流岔道等6类地理信息因子,采用神经网络方法模拟了长江山区航道雾情综合指数的地域精细化分布。统计结果显示:长江山区航道雾总体上呈现冬多夏少的季节特点,但也有冬少夏多或四季比较平均的情况存在;大多数雾情形成于0~8时,结束于8~12时,其中大雾开始早、结束晚、持续时间长;重庆段航道雾空间分布差异较大,涪陵的蔺市到丰都段、万州的黄花城分布最多,年均大雾30~50次,最少的安坪至夔峡段年均不到5次。模拟结果表明,利用地理信息因子及神经网络法基本可以模拟出重庆段航道雾情分布状况,此方法推广应用,可获取整个长江山区航道雾情的精细化分布,输出结果与航道雾情资料分析及实地考察调研结果在分布趋势上比较接近,但由于试验存在局部误差收敛及因子选择局限性问题,因此模拟结果还不能完全代表航道雾情的实际分布状况,模拟试验还需更多资料及影响因子加入。
Using the fog observation data of 39 artificial fog information stations in the Chongqing section along the Yangtze River waterway, the temporal and spatial distribution characteristics of fog in the waterway of Chongqing section are analyzed in detail. Based on the above fog and the entire Yangtze River waterway (Yibin To Yichang) by geographic information factors such as longitude, latitude, width of the watercourse, curvature of the river, drastic intensity of the river course and branch of the tributary of the river. By using the neural network method, the geographical distribution of the fog index in the Yangtze River is simulated. . The statistical results show that the mist in the Yangtze River is generally less seasonal in winter and more in summer, but more or less in winter and summer or autumn. However, most of the fog occurs at 0 ~ 8 o'clock and ends at 8 ~ 12 o'clock , Of which the fog started early and ended late and lasted for a long time. Fog distribution in Chongqing section of the fairway was quite different. Fulin's Linfeng-Fengdu section was the most distributed in Huanghuacheng of Wanzhou, with an average annual fog of 30-50 times Anping to 夔 Gap an average of less than 5 times. The simulation results show that the use of geographic information factors and neural network method can basically simulate the distribution of the fog passageway in Chongqing section of the fair, this method can be applied to obtain the fine distribution of the entire Yangtze River mist, the output results and the analysis of waterway fog conditions And field investigation and survey results are relatively close in distribution trend. However, due to the limitations of local error convergence and selection of factors, the simulation results can not completely represent the actual distribution of waterway fog, and more data and influence are needed in the simulation test Factor added.