Pilot study using SPOT satellite imagery over Apalachicola national forest to determine appropriate spatial scale for area-wide aggregation of surface fluxes

Summary High resolution radiances from SPOT satellite imagery converted to Normalized Difference Vegetation Indices (NDVI) over a 16×16 km² mixed ground cover study-area in the Apalachicola National Forest in northwest Florida, along with in situ measurements from a Bowen ratio surface flux monitoring system and physical modeling techniques, are used to determine the length manifold beyond which degraded resolution satellite imagery fails to capture flux variability over the scene. The investigation is relevant to an understanding of how bias error is generated in methods designed to produce scale-invariant surface flux estimates from satellite measurements. Error estimates are based on assigning characteristic NDVI values to the four predominant types of ground cover found within the study-area. An open site near the center of the study-area, which satisfies the conditions for surface flux monitoring, is used for obtaining input data for a biosphere-atmosphere exchange model designed to calculate representative fluxes for the different ground covers. Continuous 6-minute meteorological and surface flux measurements were made at the monitoring site for a period of 22 days. These measurements are used in conjunction with surface layer theory to provide surface layer profile estimates of wind speed, temperature, and relative humidity at the tops of the forested sites. The measured and derived meteorological parameters, together with representative biophysical parameters, are used as input to the biosphere-atmosphere exchange model. By representing sensible and latent heat flux distributions due to the variable ground cover with characteristic NDVI values at 20-m resolution, baseline area-wide sensible and latent heat flux quantities are calculated. Error-growth curves as a function of spatial resolution for the fluxes are found by degrading the resolution of the SPOT radiances used to calculate NDVI, and rationing the associated area-wide fluxes to the baseline values. The point at which an error-growth curve becomes invariant represents the edge of a length manifold beyond which the satellite input no longer contains information on surface flux variability, even though NDVI variability continues at all scales up to that of the complete SPOT scene. The error-growth curves are non-linear, with all the error build-up taking place between 20 m and 1.6 km. Decreasing the spatial resolution of the NDVI information down to or below 1.6 km, introduces bias errors in the area-wide surface flux estimates of 10% for sensible heat and 8% for latent heat. The underlying assumptions and modeling produce uncertainty in estimating the manifold limits, however, the principal objective is to show that in using satellite data for scale-invariant surface flux retrieval, there is an optimal spatial resolution factor that can be objectively quantified.With 10 Figures     

Comparison of NCAR community climate model (CCM) climates

The simulated mean January and July climates of four versions of the National Center for Atmospheric Research (NCAR) Community Climate Model (CCM) are compared. The models include standard configurations of CCM1 and CCM2, as well as two widely-cited research versions, the Global Environmental and Ecological Simulation of Interactive Systems (GENESIS) model and the Climate Sensitivity and Carbon Dioxide (CSC02) model. Each CCM version was integrated for 10 years with a horizontal spectral resolution of rhomboidal 15 (R15). Additionally, the standard T42 version of CCM2 was integrated for 20 years. Monthly mean, annually repeating climatological sea surface temperatures provided a lower boundary condition for each of the model simulations. The CCM troposphere is generally too cold, especially in the polar upper troposphere in the summer hemisphere. This is least severe in CCM2 and most pronounced in CCM1. CSC02 is an exception with a substantial warm bias, especially in the tropical upper troposphere. Corresponding biases are evident in atmospheric moisture. The overall superior CCM2 thermodynamic behavior is principally compromised by a large warm and moist bias over the Northern Hemisphere middle and high latitudes during summer. Differences between the simulated and observed stationary wave patterns reveal sizeable amplitude errors and phase shifts in all CCM versions. A common problem evident in the upper troposphere is an erroneous cyclone pair that straddles the equatorial central Pacific in January. The overall January stationary wave error pattern in CCM2 and CSCO2 is suggestive of a reverse Pacific-North American teleconnection pattern originating from the tropical central Pacific. During July, common regional biases include simulated North Pacific troughs that are stronger and shifted to the west of observations, and each model overestimates the strength of the anticyclone pair associated with the summer monsoon circulation over India. The simulated major convergence and divergence centers tend to be very localized in all CCM versions, with a tendency in each model for the maximum divergent centers to be unrelistically concentrated in monsoon regions and tied to regions of steep orography. Maxima in CCM-simulated precipitation correspond to the simulated outflow maxima and are generally larger than observational estimates, and the associated atmospheric latent heating appears to contribute to the stationary wave errors. Comparisons of simulated radiative quantities to satellite measurements reveal that the overall CCM2 radiative balance is better than in the other CCM versions. An error common to all models is that too much solar radiation is absorbed in the middle latitudes during summer.The National Center for Atmospheric Research is sponsored by the National Science Foundation     

Statistical estimation of high-resolution surface air temperature from MODIS over the Yangtze River Delta,China

High-resolution surface air temperature data are critical to regional climate modeling in terms of energy balance, urban climate change, and so on. This study demonstrates the feasibility of using Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature (LST) to estimate air temperature at a high resolution over the Yangtze River Delta region, China. It is found that daytime LST is highly correlated with maximum air temperature, and the linear regression coefficients vary with the type of land surface. The air temperature at a resolution of 1 km is estimated from the MODIS LST with linear regression models. The estimated air temperature shows a clear spatial structure of urban heat islands. Spatial patterns of LST and air temperature differences are detected, indicating maximum differences over urban and forest regions during summer. Validations are performed with independent data samples, demonstrating that the mean absolute error of the estimated air temperature is approximately 2.5°C, and the uncertainty is about 3.1°C, if using all valid LST data. The error is reduced by 0.4°C (15%) if using best-quality LST with errors of less than 1 K. The estimated high-resolution air temperature data have great potential to be used in validating high-resolution climate models and other regional applications.     

An online model correction method based on an inverse problem: Part I—Model error estimation by iteration

Errors inevitably exist in numerical weather prediction (NWP) due to imperfect numeric and physical parameterizations. To eliminate these errors, by considering NWP as an inverse problem, an unknown term in the prediction equations can be estimated inversely by using the past data, which are presumed to represent the imperfection of the NWP model (model error, denoted as ME). In this first paper of a two-part series, an iteration method for obtaining the MEs in past intervals is presented, and the results from testing its convergence in idealized experiments are reported. Moreover, two batches of iteration tests were applied in the global forecast system of the Global and Regional Assimilation and Prediction System (GRAPES-GFS) for July-August 2009 and January-February 2010. The datasets associated with the initial conditions and sea surface temperature (SST) were both based on NCEP (National Centers for Environmental Prediction) FNL (final) data. The results showed that 6th h forecast errors were reduced to 10% of their original value after a 20-step iteration. Then, off-line forecast error corrections were estimated linearly based on the 2-month mean MEs and compared with forecast errors. The estimated error corrections agreed well with the forecast errors, but the linear growth rate of the estimation was steeper than the forecast error. The advantage of this iteration method is that the MEs can provide the foundation for online correction. A larger proportion of the forecast errors can be expected to be canceled out by properly introducing the model error correction into GRAPES-GFS.     

多普勒雷达观测在北京7.21暴雨数值模拟中的三维变分同化研究

基于背景误差的特征长度理论,研究调整背景误差水平分辨率对多普勒雷达资料三维变分同化的影响。首先利用NMC方法针对暴雨落区统计不同水平分辨率的背景误差协方差,分析两种不同分辨率的背景误差的结构特征,研究水平分辨率对背景误差特征长度的影响。将其应用于雷达资料同化中,研究背景误差水平分辨率变化对雷达资料同化的影响。结果表明:背景误差水平分辨率由27 km提高到3 km时,在大气低层体现出更细致的动力场信息,其动力场水平特征长度按水平分辨率的二次根递减,而温度场与水汽场水平特征长度变化不明显。在将不同分辨率的背景误差用于三维变分同化时,更高分辨率的背景误差可以在分析场增量中体现更细致的中小尺度信息,能够明显改善雷达径向速度资料同化效果,并在随后的暴雨数值模拟中雨量及其分布形态更接近实况。     

Quantifying the uncertainty of the annular mode time scale and the role of the stratosphere

The impact of high resolution modern vegetation cover on the West African climate is examined using the International Centre for Theoretical Physics Regional Climate Model implementing the NCAR Community Land Model. Two high resolution 25 km long-term simulations driven by the output from a coarser 50-km resolution simulation are performed for the period 1998–2010. One high resolution simulation uses an earlier and coarser-resolution version of plant functional type distribution and leaf area index, while the other uses a more recent, higher-quality, and finer-resolution version of the data. The results indicate that the new land cover distribution substantially alters the distribution of temperature with warming in Central Nigeria, northern Gulf of Guinea and part of the Sahel due to the replacement of C4 grass with corn; and cooling along the coastlines of the Gulf of Guinea and in Central Africa due to the replacement of C4 grass with tropical broadleaf evergreen trees. Changes in latent heat flux appear to be largely responsible for these temperature changes with a net decrease (increase) in regions of warming (cooling). The improved land cover distribution also results in a wetter monsoon season. The presence of corn tends to favor larger precipitation amounts via more intense events, while the presence of tropical broadleaf evergreen trees tends to favor the occurrence of both more intense and more frequent events. The wetter conditions appear to be sustained via (1) an enhanced soil moisture feedback; and (2) elevated moisture transport due to increased low-level convergence in regions south of 10N where the most substantial land cover differences are present. Overall the changes induced by the improved vegetation cover improve, to some extent, the performance of the high resolution regional climate model in simulating the main West African summer monsoon features.     

资料同化中二维特征长度随模式分辨率变化的分析研究

特征长度是资料同化中的重要参量,决定了观测信息在空间的传递特征,而特征长度随模式水平分辨率增减而变化的特点与背景误差湍流功率谱分布特点密切相关。通过对不同来源实际资料计算获得的特征长度数据分析和对理想数据数值试验分析,结果表明随着模式分辨率的提高,特征长度会按照二次根的规律递减。特征长度的这种变化由背景误差湍流功率谱,特别是与次天气尺度（20～60波）到中尺度波（大于60波）的湍流功率谱斜率特征决定。当湍流功率谱斜率从-5/3变化到-4时,特征长度随模式分辨率变化的敏感性降低。作者估计出的温度场的实际背景误差湍流功率谱斜率在次天气尺度到中尺度大约在-2.8左右。对特征长度的估计除传统方法外,可以根据背景误差的湍流功率谱斜率特征来更方便地给出,该方法可作为传统方法的补充来匹配应用。     

利用MERIS和AATSR资料估算黄土高原塬区蒸散发量研究

基于陆面能量平衡原理,通过对搭载在欧洲空间局环境卫星(Environmental Satellite,ENVI-SAT)上中分辨率影像光谱仪(Medium Resolution Imaging Spectrometer,MERIS)2005年6月7,11和27日的遥感观测资料进行大气纠正等预处理后,得到估算瞬时蒸散发量所需要的地表反照率和植被覆盖度等值,并利用分裂窗法和ENVISAT上搭载的先进的沿轨迹扫描辐射计(Advanced Along-TrackScanning Radiometer,AATSR)的观测资料进行了地表温度的反演,进一步估算出黄土高原塬区午间瞬时净辐射、感热通量和土壤热通量。结合与卫星遥感观测资料同期研究区域气象站的太阳辐射、气温、日照时数和风速等气象要素资料,充分考虑到植被冠层和陆地表面对蒸散发量的不同影响,发展了一个可以估算陆面潜热的简化模型,并将瞬时蒸散发量转化为日蒸散发量。对卫星遥感估算的潜热通量,利用黄土高原塬区陆面过程野外观测试验(Loess Plateau land surface process field Experiments,LOPEXs)的地面通量观测资料进行验证,结果表明:二者最大相对差异为10.9%,最小相对差异为4.8%,并对差异误差产生的原因进行了分析和探讨。     

卫星遥感确定沙特阿拉伯吉达地区非均匀地表区域地表参数和能量通量

对临海沙漠地区非均匀地表区域地表能量通量和蒸发(蒸散)的研究,是一个十分重要但又是一个难点问题。本文提出了1个基于卫星遥感和地面观测的参数化方案,并把其应用于沙特阿拉伯吉达地区,利用1个景的陆地资源卫星Landsat-7E^TM^ 资料进行了分析研究,得到了一些有关临海沙漠地区非均匀地表区域地表特征参数、植被参数和地表能量通量的新概念。最后讨论了所提出的参数化方案的适用范围和需改进之处。     

塔克拉玛干沙漠腹地1961-1998年逐月平均气温序列的重建

 利用塔里木盆地周边27个气象站1961-2006年逐月平均气温和塔中气象站1999-2006年逐月平均气温资料,同时选取1961-2006年NCEP/NCAR 2.5°×2.5°经纬度距地表2 m的月平均气温再分析格点资料,分别用逐步回归分析、EOF分解和NCEP资料3种方法对塔中气象站1961-1998年历年逐月平均气温序列进行了恢复与重建,分析了误差,并与周边气象站的变化特征进行对比。结果表明,逐步回归和EOF法都能够作为重建塔中逐月平均气温的方法,但相对而言,逐步回归法重建的序列误差更小,平均拟合绝对误差为0.3℃,最大绝对误差为1.9℃。而NCEP/NCAR资料由于冬季存在明显的系统性误差,数值显著偏高,不能用于塔中气温序列的重建。     

长波区间太阳辐射对气候模拟的影响

长波区间的太阳辐射在气候模式中往往被忽略。利用国家气候中心BCC_AGCM2.0.1大气环流模式,采用矩阵算子辐射传输算法,研究了长波区间太阳辐射对气候模式辐射通量和温度模拟结果的影响。结果表明,以ISCCP和CERES辐射资料为标准,考虑长波区间太阳辐射后,长波区间晴空大气地表向下辐射通量平均误差减小2.05 W/m2,均方根误差减少1.29 W/m2;长波区间晴空大气模式顶向上辐射通量平均误差减小0.70 W/m2,均方根误差减小0.21 W/m2;长波区间有云大气地表向下辐射通量平均误差减小1.38 W/m2,均方根误差减小1.03 W/m2;长波区间有云大气模式顶向上辐射通量平均误差减小0.99 W/m2,均方根误差减小0.30 W/m2。以ECMWF再分析资料为标准,考虑长波区间太阳辐射后,赤道地区上对流层—下平流层区域温度的冷偏差得到改善,对流层顶温度平均误差减小0.27 K,均方根误差减小0.25 K。     

The Initial Errors in the Tropical Indian Ocean that Can Induce a Significant “Spring Predictability Barrier” for La Ni?a Events and Their Implication for Targeted Observations

Initial errors in the tropical Indian Ocean (IO-related initial errors) that are most likely to yield the Spring Prediction Barrier (SPB) for La Ni?a forecasts are explored by using the CESM model. These initial errors can be classified into two types. Type-1 initial error consists of positive sea temperature errors in the western Indian Ocean and negative sea temperature errors in the eastern Indian Ocean, while the spatial structure of Type-2 initial error is nearly opposite. Both kinds of IO-related initial errors induce positive prediction errors of sea temperature in the Pacific Ocean, leading to under-prediction of La Ni?a events. Type-1 initial error in the tropical Indian Ocean mainly influences the SSTA in the tropical Pacific Ocean via atmospheric bridge, leading to the development of localized sea temperature errors in the eastern Pacific Ocean. However, for Type-2 initial error, its positive sea temperature errors in the eastern Indian Ocean can induce downwelling error and influence La Ni?a predictions through an oceanic channel called Indonesian Throughflow. Based on the location of largest SPB-related initial errors, the sensitive area in the tropical Indian Ocean for La Ni?a predictions is identified. Furthermore, sensitivity experiments show that applying targeted observations in this sensitive area is very useful in decreasing prediction errors of La Ni?a. Therefore, adopting a targeted observation strategy in the tropical Indian Ocean is a promising approach toward increasing ENSO prediction skill.     

成都地区城市化对气候影响的模拟研究

为了研究成都地区城市化对当地气候的影响,利用不同时期的下垫面土地利用类型数据和耦合单层城市冠层模型(UCM)的WRF（Weather Research and Forecasting）模式对成都夏季和冬季城市化效应进行了模拟研究,得到以下主要结论:1）成都地区城市化使夏季城区上空出现增温区域。城区地表气温升高约2.8°C,边界层高度升高约150 m,冬季地表气温平均升高约0.6°C,边界层高度升高约25 m。夏冬两季气温日较差均减小。2）受城市化影响,成都地区夏季和冬季2 m相对湿度减小,感热通量增加,潜热通量减小,且夏季变化程度强于冬季。3）城市化使地表的粗糙度增加,进而使夏季和冬季风速在城区减小,减小约0.1~0.6 m s?1,但夏季风速减小区域较冬季更大。城市化还使城市上空低层散度减小,辐合作用增强,垂直速度增大,夏季水汽往高层输送明显。4）夏季,城市化作用使日平均和白天时段降水量在城区的迎风区和下风区均增加,夜间降水量在下风区域增加,对迎风区域影响不明显。     

Flow-distortion effects on scalar flux measurements in the surface layer: Implications for sensor design

Scalar fluxes measured through the eddy-correlation technique are prone to two types of errors caused by the sensor-induced flow distortion: those due to crosstalk from the horizontal flux, and those due to amplification or attenuation due to flow blocking. We show that the crosstalk error can be eliminated by designing the sensor array to be vertically symmetric about its horizontal midplane. In such an array, the flow-blocking effect causes the scalar flux to be overestimated, but this error can be made negligible by designing an array with minimal stagnation loss in streamwise speed at the flux-measurement point.The National Center for Atmospheric Research is sponsored by the National Science Foundation.     

Observations Of The Turbulence Structure Within Two Stratocumulus-Topped, Marine Boundary Layers

Two situations observed during the second Aerosol Characterization Experiment (ACE-2) are analysed from aircraft measurements in the broken stratocumulus (Sc)-topped marine boundary layer. The first one (26 June 1997), characterized by a non-polluted, oceanic air mass, presents a decoupling between the Sc layer (1400–1520 m) and the turbulent mixed layer, this latter extending from the surface up to 580 m. In contrast, the second case (9 July 1997), during which continental air had been advected over the experimental area, presents a well-coupled layer extending from the surface up to the top of the Sc layer(910 m). This coupling, uncommon in this area in the middle of the day, isrelated to the relative shallowness of the boundary layer. For both situations,it is shown that the turbulent fluxes can be computed with reasonably goodaccuracy (better than 10 %), taking into account both the random and thesystematic errors involved in the eddy-correlation technique. Estimationof random error is based on the computation of the integral scale of thecovariance, and systematic error is estimated from the parameterizationof Mann and Lenschow. The fluxes show that the buoyancy, as a sourceof turbulence, is due to latent heat flux rather than sensible heat flux,with values comparable to previous experiments in the Azores-Canariesbasin. In addition, we propose a method to analyse, for coupled situations,the relationship between the fractional cloudiness and the organization ofthe turbulent field below the clouds. This method is based on a conditionalsampling technique. It is shown that this organization cannot be deducedfrom the analysis of the velocity signal, which is dominated by turbulence.However, when the signals are conditionally sampled according to thepresence or absence of clouds, a weak cloud-related organization can beshown, and the cloud-related transports quantified; the values found areof the order of 10 % of the total transfers, i.e. the same order of magnitude asthe errors on the total flux computation. The method developed is thereforepromising, provided that the uncertainties can be reduced by analyzing a highamount of data.     

CMA-GFS 4DVar边界层过程线性化的改进北大核心CSCD

持续发展和优化切线性模式的线性化物理过程,保持与非线性模式一致是改善四维变分同化(4DVar)分析和预报效果的有效方法之一。目前业务系统的CMA-GFS模式采用基于Charney-Phillips(C-P)跳点的边界层参数化方案,而CMA-GFS 4DVar系统中采用基于Lorenz跳点的边界层线性化方案。为改善CMA-GFS 4DVar系统的边界层分析和预报效果,基于C-P跳点的边界层参数化方案研发了新边界层线性化方案,并通过对方案中地表热量通量和水汽通量扰动、自由大气的理查逊系数扰动、边界层的热量和动量交换系数扰动等进行更加精细地规约化约束,在确保CMA-GFS切线性和伴随模式稳定运行的情况下,减少线性化过程对切线性模式预报精度的影响。切线性近似试验检验表明:相较于原方案,新边界层线性化方案可以减少边界层位温和比湿的相对误差,最大可减少10%。批量4DVar循环同化试验表明:新边界层线性化方案可以有效改善切线性模式对低层位温、风场和比湿扰动的预报精度,减少4DVar内外循环目标泛函的相对差异,并提高700 hPa位势高度的可预报时效。     

四川盆地一次强降水过程水汽特征

利用WRF模式、地基GPS资料以及常规气象观测资料,结合模式输出资料的高空间分辨率(10km)和GPS大气可降水量(GPS-PWV)资料的高时间分辨率(30min)的优点,对2008年7月20—22日四川盆地一次暴雨过程的水汽变化特征及各物理量与大气可降水量的关系进行综合分析。结果表明:此次降雨过程是由高原涡和西南涡共同作用引起,WRF模式能够较好地模拟出降雨落区和强度。GPS-PWV反映的大气可降水量增减趋势与WRF模拟的较为一致。水汽密度垂直分布反映了大气可降水量分布,水汽密度随高度增加而递减,降雨初期,水汽密度随高度减小迅速,降雨强盛时期,水汽密度随高度减小的速度减慢。水汽辐合使得水汽密度和大气可降水量增大,风的散度项与水汽通量散度的变化一致,而水汽平流项对水汽辐合贡献较小,水汽的辐合主要由风场辐合造成。     

地形非均匀性对网格区地面长波辐射通量计算的影响

从理论和数值试验两个方面证明地形的非均匀性(如海拔高度)对网格区地面长波辐射通量的计算有重要影响,海拔高度场的区域平均值及其变差系数是影响网格区地面长波辐射通量的主要因素,仅仅用地表均匀假定下的区域平均参量(如平均海拔高度和平均温度)所计算的网格区地表有效辐射通量值与其真实值之间存在着一定的误差。由于地表有效辐射通量是海拔高度的非线性函数,在特定情况下,其影响相当大,可产生不容忽略的误差。相对而言,海拔高度自身非均匀性对误差的影响远大于地表温度非均匀性项及其混合扰动项所产生的误差。对于不同的地形平均高度,地形非均匀性影响的程度并不相同。平均高度较小时,非均匀性的影响几乎可以忽略,但随着地形平均高度的增加,地形非均匀性的影响程度呈非线性增长趋势。因而,在复杂地形区域,考虑次网格地形的热力作用非常必要。     

Impact of coherent eddies on airborne measurements of vertical turbulent fluxes

During the Hydrological-Atmospheric Pilot Experiment (HAPEX)-Sahel, which took place in Niger in the transitional period between the wet and dry seasons, two French aircraft probed the Sahelian boundary layer to measure sensible and latent heat fluxes. The measurements over the Niamey area often revealed organised structures of a few km scale that were associated with both thermals and dry intrusions. We study the impact of these coherent structures using a single day’s aircraft-measured fluxes and a numerical simulation of that day with a mesoscale model. The numerical simulation at high horizontal resolution (250 m) contains structures that evolve from streaks in the early morning to cells by noon. This simulation shows distribution, variance and skewness similar to the observations. In particular, the numerical simulation shows dry intrusions that can penetrate deeply into the atmospheric boundary layer (ABL), and even reach the surface in some cases, which is in accordance with the observed highly negatively skewed water vapour fluctuations. Dry intrusions and thermals organised at a few km scale give skewed flux statistics and can introduce large errors in measured fluxes. We use the numerical simulation to: (i) evaluate the contribution of the organised structures to the total flux, and (ii) estimate the impact of the organised structures on the systematic and random errors resulting from the 1D sampling of the aircraft as opposed to the 2D numerical simulation estimate. We find a significant contribution by the organised structures to the total resolved fluxes. When rolls occur, and for a leg length of about 30 times the ABL depth, the 1D sampled flux is shown to be sometimes 20% lower than the corresponding 2D flux when the 1D sampling direction is the same as the main axis of the rolls, whereas the systematic error is much lower when the direction of the leg is transverse to the rolls. In the case of cells, an underestimate of around 10% can still be observed with the 1D approach independent of direction, due to poor sampling of the energy-containing scales.     

基于全连接神经网络方法的日最高气温预报

为了考察辅助变量、时间滞后变量设置的重要性和神经网络中嵌入层对分类变量处理的有效性,利用2015年1月15日—2020年12月31日欧洲中期天气预报中心（European Centre for Medium-Range Weather Forecasts,ECMWF）高分辨率模式（high resolution,HRES）输出产品及中国2238个国家级地面气象站基本气象要素数据集,在全连接神经网络基础上设计4个试验,构建24 h最高气温预报神经网络模型。结果表明:加入辅助变量、时间滞后变量的特征和带有嵌入层的全连接神经网络结构的深度学习神经网络模型对HRES日最高气温预报误差均有订正效果,均方根误差降低29.72%~47.82%,温度预报准确率提高16.67%~38.89%。加入经过嵌入层处理的辅助变量后,可显著提高青藏高原中南部和西南地区东部的平均绝对偏差不超过2℃的正技巧站点比例（比仅用HRES预报因子建模分别提高21.74%和14.17%）,在此基础上加入时间滞后变量显著提高上述两个地区的平均绝对偏差不超过2℃的正技巧站点比例（比仅用HRES预报因子建模分别提高40.98%和20.33%）,且预报性能更加稳定。