ARTIFICIAL NEURAL NETWORK CORRECTION OF ROUGH ERRORS OF OBSERVATIONS*

In the context of tower measured radiation datasets.following the correction principle meeting a diagnostic equation in data quality control and in terms of a technique for model construction on data and ANN (artificial neural network) retrieval for BP correction of radiation measurements with rough errors available,a BP model is presented.Evidence suggests that the developed model works well and is superior to a convenient multivariate linear regression model,indicating its wide applications.     

北京地区雷电定位系统场地误差及其结构分析

本文利用1988,1989两年北京地区雷电定位系统(LLS)探测到的雷电资料,采用作者在文献[10]中提出的参数化方法,分别计算了该系统中各定向仪(DF)的场地误差曲线。结果表明,场地误差曲线不仅是测量方位角的偶周期函数,而且包含着奇周期的成份。同时还表明,同一场地具有相对稳定的误差曲线。 文中还讨论了场地误差产生的原因,并首次提出偶极辐射是产生场地误差的主要原因的观点,推导出了它可能产生的场地误差的函数形式。用此理论可很好地解释实际计算出的各站场地误差曲线的结构特征。     

HEIFE辐射仪器的标定及资料精度分析

本文对HEIFE期间中方3站所用的辐射仪器在试验前和试验中所进行的各次标定结果以及观测资料的精度进行了较详细的分析。结果表明在HEIFE中所使用的辐射仪器性能稳定，所取得的观测资料接近或达到了国际上太阳辐射野外观测的精度要求。文中还对所取得的红外辐射资料的订正进行了讨论。     

Effects of soil moisture on the sensitivity of a climate model to earth orbital forcing at 9000 yr BP

The sensitivity of climate to orbitally-related changes in solar radiation at 9000 yr BP (before present) is examined using fixed and interactive soil moisture versions of a low resolution general circulation model. In both versions of the model increased solar radiation for June–August at 9000 yr BP (compared to present) produced enhanced northern monsoons and warmer continental interiors in comparison to present whereas decreased solar radiation at 9000 yr BP in December–February produced weaker southern monsoons. The increased rainfall in the northern tropics in summer increased soil moisture and runoff at 9000 yr BP in the interactive model; in the southern hemisphere decreased rainfall in summer led to decreased soil moisture and runoff. Conditions in summer became drier (decreased soil moisture and runoff) at 9000 yr BP in the northern extratropics.The experiments showed that the magnitude (but not the sign) of model sensitivity to 9000 yr BP radiation is altered by the effects of interactive soil moisture. Decreased soil moisture (about 20%) over northern Eurasia in the interactive model led to smaller evaporative increases, greater temperature increases and greater reduction of precipitation than for the model with fixed soil moisture. Over northern tropical lands, slightly smaller temperature increases and greater evaporation and precipitation increases in the interactive model are linked to the simulation of increased soil moisture at 9000 yr BP. The differences in sensitivity between the two versions of the model over northern Eurasia are statistically significant at the 95% level whereas those for the tropics are not.Overall, the results of the simulations are generally supported by the geologic evidence for 9000 yr BP; however, the evidence lacks sufficient precision and the model resolution is too coarse for detailed model/data comparisons and for assessment of the relative accuracy of the two 9000 yr BP experiments.The computed sensitivities of temperature and soil moisture to 9000 yr BP radiation differ from those simulated under equilibrium conditions in the various general circulation model experiments for increased atmospheric concentration of CO₂. In contrast to the effects of the enhanced seasonal cycle of solar radiation at 9000 yr BP, a CO₂ increase causes a broad warming of both the ocean and land with little modification of land/ocean temperature difference. The experiments for 9000 yr BP indicate a clearer signal for summer drying than is obtained in the experiments for increased CO₂. The results suggest that the 9000 yr BP climate may be of limited utility as an analog to future warm climates.     

太阳辐射预报滚动订正方法研究

利用2013年1月至2014年12月北京南郊观象台逐时观测总辐射以及BJ-RUC模式系统预报的该站未来24h逐时总辐射、云量、水汽混合比、云水、云冰含量等14个气象要素数据,运用多种线性订正方案对总辐射预报值进行订正,重点分析了不同方法、不同季节、不同样本数的订正效果差异。结果表明:1不同季节辐射订正的影响因素不尽相同,需采用不同的组合订正因子,其中总辐射、2m比湿、2m相对湿度、低云量、中云量、高云量、总云量、云水雨冰雪霰混合比、水汽混合比可作为推荐因子;2最优样本数选取时需考虑季节差异;3逐时滚动订正方案的订正效果较好,明显优于非滚动方案。订正后总辐射误差显著减小,而且79%的时刻有改进,明显减小了预报偏大的系统误差;4冬春季订正效果优于夏秋季,这与云的宏观和微观物理量预报效果的季节差异有关。本文研究结果可为太阳能资源评估、总辐射和光伏电站发电量预报提供有效的改进方法。     

黄淮地区辐射模型适用性评价及参数优化

利用1961—2015年黄淮地区8个辐射站太阳辐射和日照时数等常规气象资料,分别评价6种常用的太阳总辐射和有效辐射估算模型在黄淮地区的适用性,同时采用多元回归分析和迭代等方法,对辐射参数进行优化调整,建立了适合本地区的辐射最优化估算模型。结果表明:童宏良公式和邓根云公式分别在估算太阳总辐射和地面有效辐射时的误差最小,相比其余的辐射估算模型,两者在黄淮地区适用性最好。另外太阳总辐射本地化修正模型的相对误差绝对值(value of Absolute Relative Error, ARE)和均方根误差(Root Mean Squared Error, RMSE)分别为16.28%和1.730 MJ·m~(-2)·d~(-1),优于童宏良公式等常用太阳辐射估算模型;有效辐射本地化修正模型的ARE和RMSE分别为23.19%和1.404 MJ·m~(-2)·d~(-1),优于邓根云公式等常用有效辐射估算模型;因此黄淮地区本地化辐射修正模型适用于当地地表净辐射估算,且具有较好的估算精度。     

东亚地区的大气辐射能的收支(一)——地球和大气的太阳辐射能收支

本文是讨论东亚地区大气辐射能收支研究工作的第一部分,讨论了以下三个问题: (1)本文利用文献[1]的水汽各吸收带的吸收光谱实验资料,求得了一个适合于手算的水汽对太阳辐射的总吸收能量公式(公式(6))。并把式(6)与Mugge—Moller公式进行了比较。 (2)利用公式(6),计算了东亚地区39个测站1,7月自地面到100毫巴各气层对太阳辐射的吸收能量,及其对大气的加温率。本文还进一步考虑了云的订正、大气对地面反射辐射的吸收,而求得了东亚地区对流层大气吸收能量的分布。 (3)利用1958—1960年中国地区的一些地面总辐射和反射率观测资料,以及本文计算的大气中各种吸牧能量,讨论了中国地区行星反射率的分布和地球大气系统中各种太阳辐射能的收支。     

AN ERROR ANALYSIS ON THE SIMULATED DATA OF TOVS HIRS BY RTTOV5 MODEL

The sounding data of meteorological satellites provide not only the real time weatherinformation about the distribution of both cloud and rainfall,but also some others about themovement and state of atmosphere.They are important variables and parameters for NWP modelused to simulate and predict atmospheric state.In order to introduce remote sensing informationfrom satellites into NWP model,there is an efficient way of establishing an RT model by use of theatmosphere radiation sounding data of meteorological satellites to get the variables and parametersvaluable to NWP model.In this paper,we set up profiles of air temperature and water vapor fromthe surface to upper (0.1 hPa) using the radiosounding data and the surface data from May toAugust 1998 atmosphere East Asia.A TOVS RT model (RTTOV5) is provided to compute thevalue of radiation value of HIRS channels in NOAA14.Then the radiation values of 19 HIRSchannels are gotten.After matching these data computed by the RT model and the correspondingvalues coming from satellite sounding in time,the statistic distribution of bias between tile modeloutput and the satellite sounding at each sounding channel can be gotten.At the same time.thedistribution of RMS to every TOVS HIRS channel,the standard biases to different scanning angleto each channel are also obtained.     

A macrophysical model of the holocene intertropical convergence and jetstream positions and rainfall for the Saharan region

Summary Using a macrophysical model, it has been found possible to model the Late Pleistocene and Holocene history of glacial volume and mean temperature as a function of solar radiation seasonality and modulation of the solar radiation by volcanic aerosols (Bryson, 1988). From this model it has been possible to model the Indian monsoon with fair agreement with the field data (Bryson, 1989). The present paper represents an extension of this effort to the modeling of the position of the Intertropical Convergence in North Africa and the latitude of the atmospheric polar jetstream in the same longitudes.From these two latitudes it is possible to simulate the seasonal rainfall history of the Saharan region. The preliminary results suggest that maximum intrusion of summer rains into North Africa should have occurred between 11,000 and 5000 BP, with a lesser intrusion between 30,000 and 28,000 BP. Winter rains, though not abundant, should have been present in the central sahara until 12,000 BP then diminishing to negligible by 8000 BP, according to the model.On the Mediterranean coast of North Africa, the winter rains should have been about 100% greater than the present until about 18,000 BP, diminishing to near the present value by about 8000 BP.Radiocarbon dated occupation sites appear to be most abundant at about the indicated times of greatest rainfall in sub-Saharan Africa, as does the frequency of high lake levels.A collateral purpose of this paper is to test the utility of the present inter-monthly changes in the parameterization of inter-century changes.With 14 Figures     

AN ERROR ANALYSIS ON THE SIMULATED DATA OF TOVS HIRS BY RTTOV5 MODEL*

The sounding data of meteorological satellites provide not only the real time weather information about the distribution of both cloud and rainfall,but also some others about the movement and state of atmosphere.They are important variables and parameters for NWP model used to simulate and predict atmospheric state.In order to introduce remote sensing information from satellites into NWP model,there is an efficient way of establishing an RT model by use of the atmosphere radiation sounding data of meteorological satellites to get the variables and parameters valuable to NWP model.In this paper,we set up profiles of air temperature and water vapor from the surface to upper (0.1 hPa) using the radiosounding data and the surface data from May to August 1998 atmosphere East Asia.A TOVS RT model (RTTOV5) is provided to compute the value of radiation value of HIRS channels in NOAA14.Then the radiation values of 19 HIRS channels are gotten.After matching these data computed by the RT model and the corresponding values coming from satellite sounding in time,the statistic distribution of bias between tile model output and the satellite sounding at each sounding channel can be gotten.At the same time.the distribution of RMS to every TOVS HIRS channel,the standard biases to different scanning angle to each channel are also obtained.     

基于ELM的风电场短期风速订正技术研究

风速预测是风电场风功率预测的基础,其准确度严重影响着风电场的运行效率。为了提高短期风速预测的准确性,本研究采用了WRF中尺度数值模式,对我国东部沿海某风电场的风速进行预报。在此基础上,利用极限学习机算法(ELM)对WRF模式预报的风速进一步订正。实验结果表明,WRF模式对风速、风向等气象要素有着较好的回报效果,利用ELM算法对WRF模式预报风速进行订正后,预报风速的误差进一步减小,相对均方根误差和相对平均绝对误差降低了20%～30%。与其他的智能算法(BP神经网络、SVM算法)对比分析后得出,ELM算法对WRF模式预报风速具有较好的订正效果,能够有效提高风速预报准确率。     

FY-3A微波资料偏差订正及台风路径预报应用

我国极轨气象卫星FY-3A大大增强了对地球系统的综合探测能力,而偏差订正对卫星资料的应用非常必要。试验中FY-3A卫星微波资料的偏差订正方案是在Harris等的TOVS辐射资料偏差订正经验方法的基础上结合WRF-3DVAR系统发展的,偏差订正后微波资料各通道拟合结果基本位于主对角线上,大多数卫星观测数据与观测算子利用背景场计算的亮温值分布趋于合理,偏差得到很大程度的降低。偏差订正后,利用数值模式直接同化FY-3A气象卫星微波资料,通过对2008年和2009年的4个台风进行预报评估表明:同化FY-3A气象卫星资料后路径预报能力提高明显,尤其是36 h后路径预报结果;同化FY-3A气象卫星微波资料后台风预报路径误差平均降低20%,而只同化常规资料路径误差仅仅降低了4%。     

基于SVM-BP神经网络的气象能见度数据缺失值预估

自动气象站能见度检测仪多采用光学装置采样,雨雪、粉尘等天气因素会对部分仪器镜头造成污染,导致能见度要素数据缺测.针对能见度数据缺失问题,本文选用安徽部分气象站的历年数据,首先运用灰色关联分析方法筛选出与能见度密切相关的其他气象要素,通过支持向量机和BP神经网络单一预估方法预估不同地形的能见度缺失值,然后采用最优权重组合将两种方法预估的能见度值进行组合,并与单一预估方法进行对比.结果表明组合方法的预估结果误差均值小、整体准确度高,可以保证台站观测资料的完备性,为短时天气预报、实况分析和气象公共服务工作提供有效依据.     

Estimation of hourly values of downward atmospheric radiation under cloudless skies during day- and night-time conditions

Summary Several atmospheric radiation models have been developed which provide methods for evaluating the effective emissivity of the atmosphere. Most of these methods are derived for night-time data utilizing local empirical coefficients. There are several, however, which are thought to be more universal in their predictive ability and some of these were tested on a data set registered in Granada from 1983 to 1985, inclusive. The models evaluated in this study are the Brutsaert model, the Berdahl and Martin model and the method proposed by Idso; in the first two cases the original coefficients were used. All the equations showed a good behaviour for night-time data but consistently overestimated day-time measured radiation. These deviations are explained in terms of day-night differences in the effective emissivity regime due to the differing vertical structure of the atmosphere during day and night. An empirical correction term for these differences improves the estimates of these models. The Idso model with a reduction in its independent term provided the best adjustment. Nevertheless, the Berdahl and Martin model with the original coefficients and the day-night correction term provided good results too, deviations being within the experimental error, indicating the universality of its coefficients.With 2 Figures     

基于数值预报产品的地面气温BP-MOS预报方法

在山东省临沂市气象局开发的"中尺度数值预报业务系统"的基础上,利用高分辨率的数值预报产品和地面气温观测资料,建立了地面气温的BP神经网络方法预报模型。检验结果表明BP神经网络模型的气温预报准确率高于逐步回归模型和MM5模式输出的气温预报准确率,可应用于实际预报业务中来制作气温的精细化预报。     

基于GNSS的MODIS大气可降水量校正模型

MODIS大气可降水量（PWV）空间分辨率高但易受云雨等环境因素影响,精度不高.GNSS PWV 空间分辨率较低但具有全天候、不受天气影响、精度高的优点.研究表明两者存在显著的线性相关性,结合两者的优点,基于GNSS PWV校正MODIS PWV可获取大面积高精度的PWV.针对传统的线性回归校正模型没有考虑云、气溶胶等的影响使两种数据线性相关性变差的问题,本文在传统的线性校正模型上增加了使用年积日的非线性周期项的方法来构建校正模型.利用2017—2019年香港地区GNSS对流层延迟与MODIS近红外数据,使用频谱分析线性残差项,结果表明残差具有显著的年周期.对比传统模型,本文模型的平均绝对误差、平均相对误差、均方根误差和拟合度都有明显的改善,表明本文模型可行有效且精度较高.     

微波辐射计温湿廓线反演方法改进试验

为提升地基微波辐射计在不同天气条件下, 特别是云天条件下温湿廓线的反演精度, 利用2011年1月—2016年12月中国气象局北京国家综合气象观测试验基地探空数据, 在微波辐射计反演温湿度廓线的过程中通过区分晴天和云天条件并引入全固态Ka波段测云仪云高及云厚信息, 对反演输入亮温进行质量控制和偏差订正, 建立BP神经网络模型, 采用2017年1月—2018年3月微波辐射计探测数据评估检验, 结果表明:在亮温订正前提下, 晴天温度模型、云天温度模型、晴天相对湿度模型和云天相对湿度模型反演结果与探空的相关系数分别为0.99, 0.99, 0.80和0.78, 均方根误差为2.3℃, 2.3℃, 9%和16%, 较微波辐射计自带产品（LV2产品）减小约0.4℃, 0.3℃, 11%和9%, 准确性提升约30%, 28%, 64%和45%;温度模型偏差在±2℃以内、湿度模型偏差在±20%以内的占比分别为68%, 70%和95%, 78%, 较LV2产品分别提高了7%, 5%和27%, 23%, 其中相对湿度改善明显。可见亮温订正、区分天气类型训练反演模型有利于改善地基微波辐射温湿廓线反演精度。     

几种水平面太阳总辐射量计算模型的对比分析

利用中国区域1961-1999年39 a间98个常规气象观测数据,建立6个模型分别以天文辐射、干洁大气总辐射和湿洁大气总辐射为起始数据,进行太阳辐射日总量的模拟,对比分析了6个水平面太阳总辐射量计算模型的性能.结果表明:在三种起始数据中,干洁大气总辐射和湿洁大气总辐射均能较好地体现宏观地势对太阳辐射空间分布的影响,以湿洁大气总辐射为起始数据的计算模型拟合精度相对较高.对6个水平面太阳总辐射量计算模型的对比分析发现:2个以日照百分率为主导因子,气温日较差为修正项的综合模型拟合误差最小,精度最高;经典的日照百分率模型次之,但其模型系数最稳定可靠;3个气温日较差模型拟合效果最差.最终选用经验系数稳定、拟合精度较高的日照百分率模型,制作了2001年中国水平面太阳辐射日总量空间分布图.     

CORRECTION METHODS FOR THERMAL OFFSET ERRORS IN TBQ-2-B PYRANOMETERS

Due to the existence of thermal offsets, global solar irradiances measured by pyranometers are smaller than actual values, and errors are larger in the daytime. Until now, there is no universally-recognized correction method for thermal offset errors. Therefore, it is imperative to identify a convenient and effective correction method. Five correction methods were evaluated based on the data measured from a field experiment from 23 January to 15 November, 2011. Results have shown: 1) Temporal variation characteristics of thermal offsets in the four tested pyranometers are consistent. 2) Among the five methods, non-dimensional quantity method is suggested for use to correct thermal offsets, because it is convenient and no modification of instruments is required. If collocated net longwave radiation and wind speed data are available and their uncertainties are small, the historical solar radiation datasets can also be corrected. And correction effects by the method are better.