基于卫星观测的临近空间大气变分数据同化研究

发布时间：2018-01-12 07:40

本文关键词：基于卫星观测的临近空间大气变分数据同化研究　出处：《中国科学院国家空间科学中心》2017年博士论文　论文类型：学位论文

【摘要】：20-100km临近空间的战略价值已经逐渐引起各国的关注和重视。对临近空间环境的研究,可为临近空间的开发和利用提供科学依据和保障,成为当前的热点问题。卫星探测数据分析及其在数据同化中的应用,可帮助我们对大气的特征和规律进行深入了解,同时还可提高数值模式预报精度。近年来已有许多覆盖全球的卫星探测资料,但利用这些资料开展临近空间大气数据同化,以进一步了解其运动特征和变化趋势、促进临近空间数值预报技术发展的工作还很不充分。基于此,本文利用卫星遥感探测资料进行临近空间大气环境特征的研究,同时基于卫星探测数据进行临近空间数据同化的研究,为临近空间大气数值模式提供更为精确的初值场。主要研究内容如下:(1)GPS信号的相位延迟中包含了低层大气和临近空间大气信息。利用地基GPS相位延迟数据,提出一种结合经验模式的一维变分同化获取大气折射率的方法,利用GPS相位延迟模拟数据进行了同化实验,讨论了背景误差的设置对同化结果的影响,并用实测个例对该方法进行了验证,获得了高精度的0-60km大气折射率。结果表明,该一维变分同化方法可行。首次将同化获取的大气折射率应用于无线电波折射修正实验,取得了很好的修正效果,修正精度可达1mm量级。(2)以TIMED\SABER(V1.07)的红外温度探测数据为观测值,NSSC临近空间大气数据同化预报实验系统的预报场为温度背景值,采用三维变分同化方法,获取了2013年10月1日00:00的20-100km临近空间全球大气温度场。利用统计学方法对同化结果进行评估,结果显示,三维变分同化后临近空间全球温度场误差整体减小,三维变分同化前的温度背景场误差最大可达17K,三维变分同化后的温度分析场最大误差减小至7K以内,效果明显。此算法可用于为临近空间大气环境预报模式提供更精确的初值场。(3)基于AURA\MLS卫星温度观测数据和NSSC临近空间大气数据同化预报实验系统,开展了三维变分连续同化试验,获取了20-100km临近空间全球大气温度场。利用统计学方法对同化结果进行评估,结果显示,三维变分同化后,20-100km临近空间全球温度场的误差整体减小,80km以下最大误差由同化前的10K减小至同化后的4K以内,80km以上最大误差由同化前的22K减小至同化后的7K以内,同化效果明显。与单次三维变分同化相比,三维变分连续同化在保证了同化效果的同时,可增加观测数据作用范围。(4)利用AURA\MLS数据(V4.2)和TIMED\SABERSABER数据(V2.0)对20-92km的大气温度进行统计比较分析,计算AURA\MLS减去TIMED\SABER的温度绝对偏差,并对平均温度偏差在不同季节中随经度、纬度和高度的变化特征进行讨论,为卫星数据的应用提供参考依据。结果表明:20-80km的平均温度偏差在6K以内,相对偏差在3%以内,80-90km平均温度偏差为-10k,相对偏差在9%以内。中低纬度地区平均温度偏差廓线的变化趋势较为一致,从20km的-3k左右的负偏差逐渐增加,在45-50km的平流层顶处有较为明显的3k左右的正偏差峰值。平均温度偏差随纬度的变化明显,随经度的变化很小。(5)基于AURA\MLS数据(V4.2)和TIMED\SABERSABER数据(V2.0)偏差的统计分析结果,提出了消除AURA\MLS和TIMED\SABER两种卫星数据的系统误差偏差的方法。将消除了系统误差的两种卫星温度探测数据作为观测值,以NSSC临近空间大气数据同化预报实验系统的温度预报场为温度背景场,进行基于AURA\MLS和TIMED\SABER联合温度观测数据的三维变分同化,获取了20-100km临近空间全球大气温度场。对比三维变分同化前后的临近空间全球温度场分布,变化较为明显,经验证算法可行。利用统计学方法进行同化评估,结果表明,三维变分同化后,20-100km临近空间全球温度场的误差整体减小,最大误差由三维变分同化前的10K减小至三维变分同化后的4K以内,同化后的温度分析场相对于温度背景场更接近真值,同化效果明显。该方法弥补了单颗卫星的探测数据难以覆盖全球的不足。(6)基于AURA\MLS(V4.2)从2004年8月到2016年12月共12年(149个月份)的温度、位势高度、压强等数据,计算分析了临近空间大气温度、大气密度及大气温度标准偏差的变化规律,并着重对大气温度标准偏差变化规律的原因进行了研究。结果表明,在30km平流层,冬半球的中高纬度地区大气扰动显著增大,是行星波和重力波作用的结果。在1月份和7月份,大气扰动在低纬度地区大于夏半球的中高纬度地区,是重力波和行星波作用的结果。重力波作用对赤道地区大气扰动的贡献,在4月份和11月份也有所体现。在70km的中间层,冬半球中高纬度地区的大气扰动依然显著,是行星波和重力波的贡献所致。中低纬度地区的大气扰动相对于平流层有所增加,是重力波活动增强以及大气潮汐开始出现的结果。在92km的低热层,赤道低纬度地区的大气扰动偏大,是非迁移性周日潮汐(DE3)贡献的结果。夏半球的中高纬度的地区大气扰动较强是重力波的贡献所致。
[Abstract]:The 20-100km strategic value of near space has gradually aroused the concern and attention. The research of near space environment, for the development of near space and provide scientific basis for the utilization and protection, has become a hot issue at present. The satellite data analysis and its application in data assimilation, can help us on the characteristics and laws of the atmosphere understand, also can improve the prediction accuracy of numerical model. In recent years there have been many global coverage of satellite data, but the use of these information to carry out the near space atmospheric data assimilation, in order to further understand its movement characteristics and trends, and promote the development of near space forecasting technology are still not fully. Based on this, this paper uses satellite study of near atmospheric environment feature space remote sensing data, and based on the satellite data with the data of near space Study for near space atmospheric numerical model provides a more accurate initial field. The main contents are as follows: (1) GPS signal phase delay is included in the lower atmosphere and the near space atmospheric information. Using GPS phase delay data foundation, put forward a kind of variational assimilation method to obtain the atmospheric refractive index with a dimension of experience the model of delay, simulation data assimilation experiments using the GPS phase, the influence of background error setting on the assimilation results are discussed, and by actual case to verify this method, a high precision 0-60km atmospheric refraction rate. The results show that the one-dimensional variational assimilation method is feasible for the first time. To obtain atmospheric assimilation the refractive index is applied to radio wave refraction correction experiment, achieved good correction effect, correction precision can reach the magnitude of 1mm. (2) to TIMEDSABER (V1.07) infrared temperature detection data for measurements, approaching NSSC Space atmosphere data assimilation and forecast system for temperature field prediction experiment background value, assimilation method using three-dimensional variable, obtained by 00:00 in October 1, 2013 20-100km near the air temperature field of global space. To evaluate the assimilation results using statistical methods results show that three-dimensional variational assimilation after global temperature error near the whole space is reduced, the three-dimensional variational the temperature background before the assimilation field the maximum error is 17K, three-dimensional variational assimilation after the temperature field analysis of the maximum error is reduced to less than 7K, the effect is obvious. This algorithm can be used for the near space environment prediction model provides more accurate initial field. (3) AURAMLS satellite temperature data and NSSC near space atmospheric data assimilation and forecast the experimental system based on the three-dimensional variational continual assimilation test, the 20-100km is near the air temperature field of global space. By using statistical method The assimilation results are evaluated, results show that the three-dimensional variational assimilation, the overall error decreases 20-100km near space global temperature below 80km, the maximum error is reduced from 10K to the initial assimilation after less than 4K, 80km more than the maximum error by assimilation of 22K before assimilation after reduced to less than 7K, compared with assimilation effect is obvious. With a single three-dimensional variational assimilation, three-dimensional variational assimilation in the continuous assimilation effect at the same time, can increase the scope of data. (4) using AURAMLS data (V4.2) and TIMEDSABERSABER data (V2.0) of 20-92km high temperature were compared and analyzed, the absolute temperature deviation calculation AURAMLS minus TIMEDSABER, and the average deviation of temperature in different seasons with the longitude, latitude and altitude variation characteristics are discussed, and provide the reference for the application of satellite data. The results show that the 20-80km average temperature deviation in 6K Inside, the relative deviation was less than 3% 80-90km, the average temperature deviation is -10k, the relative deviation is less than 9%. The change trend of low and middle latitude area average temperature deviation profile is more uniform, increase gradually from negative deviation of 20km about -3k, a positive deviation peak obviously about 3K 45-50km in the stratosphere at the top of average. The temperature deviation with latitude changes significantly, with the longitude changes is very small. (5) based on the data of AURAMLS (V4.2) and TIMEDSABERSABER data (V2.0) deviation of the results of statistical analysis, and puts forward the method of eliminating system error deviation of AURAMLS and TIMEDSABER two satellite data. The system will eliminate the error of the two satellite temperature detection the data as the observed value, with NSSC near space atmospheric data assimilation and forecast experiments the temperature of the system temperature forecast field as background field, 3D AURAMLS and TIMEDSABER temperature observation data based on the same 20-100km, get close to the atmospheric temperature. The global spatial contrast three-dimensional variational assimilation of near space before and after the distribution of global temperature, the more obvious changes, and proved the algorithm is feasible. The assimilation evaluation, using statistical methods. The results show that the three-dimensional variational assimilation, the overall error decreases 20-100km in near space, global temperature, maximum error by 3DVAR before 10K is reduced to three dimensional variational assimilation after less than 4K, the temperature field analysis of assimilation temperature relative to the background field closer to the true value, the assimilation effect is obvious. This method overcomes the shortcoming of detection data of single satellite to cover the world. (6) based on AURAMLS (V4.2) from August 2004 to December 2016 a total of 12 years (149 months) of the geopotential height, temperature, pressure and other data, analysis of the near space atmospheric temperature calculation, atmospheric density and atmospheric temperature scale changes and deviations. Focuses on the variation of air temperature standard deviation reasons were studied. The results showed that 30km in the stratosphere, the winter hemisphere high latitudes in atmospheric disturbance is significantly increased, planetary wave and gravity wave effect. In January and July, the atmospheric disturbance at lower latitudes than in the summer hemisphere high latitudes, is gravity wave and planetary wave effect. The effect of gravity waves on the equatorial atmospheric disturbance contribution, also reflected in the April and November. In the middle layer of 70km, high latitude winter hemisphere in atmospheric turbulence is still significant, planetary and gravity waves. The contribution caused by the low latitude regions relative to atmospheric disturbances the increase is the gravity wave activity enhancement and atmospheric tides began to appear in the results. At low 92KM, the low latitude atmospheric disturbance is relatively large, non migrating tides (DE3 Sunday The result of the contribution is that the atmospheric disturbance in the middle and high latitudes of the summer hemisphere is the result of the contribution of the gravity wave.

【学位授予单位】：中国科学院国家空间科学中心
【学位级别】：博士
【学位授予年份】：2017
【分类号】：P412.27

【相似文献】