ADVANCED SPECTRAL METHODS FOR CLIMATIC

合集下载

基于导数光谱的小麦冠层叶片含水量反演汇编

中国农业科学 2013,46(1):18-29Scientia Agricultura Sinica doi: 10.3864/j.issn.0578-1752.2013.01.003基于导数光谱的小麦冠层叶片含水量反演梁亮1,2，张连蓬1，林卉1，李春梅1，杨敏华3（1江苏师范大学测绘学院，江苏徐州 221116；2南京大学国际地球系统科学研究所，南京 210008；3中南大学地球科学与信息物理学院，长沙 410083）摘要：【目的】以高光谱技术实现小麦含水量信息的快速、无损与准确获取，为小麦灌溉的精确管理提供科学依据。

【方法】利用水氮胁迫试验条件下小麦主要生长期的导数光谱构建了16种新指数，将其与NDII、WBI以及NDWI等常用指数进行比较分析，筛选小麦叶片含水量反演最佳光谱指数，并利用其建立反演模型进行小麦含水量的遥感填图。

【结果】在各指数中，FD730-955对小麦冠层叶片含水量的估测结果最佳，其估测模型（对数形式）校正决定系数（C-R2）与检验决定系数（V-R2）分别达0.749与0.742，优于NDII等常用指数；FD730-955所建模型对32个未知样的预测结果与实测值相似度较高，其回归拟合模型R2达0.763，RMSE仅为0.024，取得了良好预测结果，且对叶片含水量以及LAI值较高与较低的样本均具备良好的预测能力，可有效避免样本取值范围以及冠层郁闭度等因素对含水量估测的影响；反演模型对OMIS影像的填图结果与地面实测值拟合模型R2达0.647，RMSE仅为0.027，具有较高的反演精度。

【结论】导数光谱可实现小麦冠层叶片含水量信息的准确估测，其中FD730-955系反演的优选指数。

关键词：高光谱遥感；导数光谱；小麦（Triticum aestivum L.）；含水量；叶面积指数Estimating Canopy Leaf Water Content in WheatBased on Derivative SpectraLIANG Liang1,2, ZHANG Lian-peng1, LIN Hui1, LI Chun-mei1, YANG Min-hua3(1School of Geodesy and Geomatics, Jiangsu Normal University, Xuzhou 221116, Jiangsu; 2International Institute for EarthSystem Science, Nanjing University, Nanjing 210008; 3School of Geosciences and Info-Physics, Central South University,Changsha 410083)Abstract: 【Objective】A method for fast, non-destructive and accurately monitoring leaf water content (LWC) of wheat(Triticum aestivum L.) was improved with hyperspectra technology in this paper. 【Method】The canopy leaf spectral reflectance inmain growing seasons of wheat was collected under the condition of water stress experiment. Using the frist-order derivative spectra,16 new hyperspectral indices were developed to quantify the wheat’s leaf water content (LWC). These indices were then comparedwith the commonly used hyperspectral indices including NDII, WBI and NDWI to screen out the spectral index which was sensitiveto LWC for modeling. Using the inversion model, the OMIS image was calculated one pixel by one pixel, and the remote sensingmapping for LWC of wheat was accomplished. 【Result】The accuracy of inversion model (logarithmic function) which built byindex FD730-955 was higher than that by the hyperspectral indices commonly used, as indicated by a calibration determinationcoefficient (C-R2) of 0.749 and a validation determination coefficient (V-R2) of 0.742. The eatimation values of 32 samples inprediction set were close to the measured values, and R2 and RMSE of regression fitted model between two dataset were 0.763 and0.024, respectivly. Furthere more, the prediction accuracy of FD730-955 was least sensitive to the change of LWC and LAI among all ofthe hyperspectra indices and therefore least affected by the range of sample values and canopy density when used to estimate the收稿日期：2012-06-26；接受日期：2012-10-26基金项目：江苏省自然科学基金项目（BK2012145）、江苏省高校自然科学研究面上项目（12KJB420001）、国家自然科学基金项目（30570279）、江苏师范大学博士学位教师科研支持项目（11XLR03）联系方式：梁亮，E-mail：liangliang198119@1期梁亮等：基于导数光谱的小麦冠层叶片含水量反演 19LWC of wheat. The R2 and RMSE of the fitting model for the inversion and measured values were 0.635 and 0.027, respectively, andindicated the similarity between the inversion and measured value was high. 【Conclusion】It is possible to eatimate LWC of wheatby derivative hyperspectra with a high accuracy, and FD730-955 is an optimal index for modeling.Key words: hyperspectra remote sensing; derivative spectrum; wheat (Triticum aestivum L.); leaf water content (LWC); leafarea index (LAI)0 引言【研究意义】叶片含水量是植株水分状况良好的度量和诊断指标，实时、快速地获取叶片含水量信息是进行作物长势监测以及灌溉调控的前提[1]。

bed

Chebyshev super spectral viscosity method for a ﬂuidizedbed modelScott A.SarraDepartment of Mathematics,Marshall University,One John Marshall Drive,Huntington,WV 25755-2560,USAReceived 7May 2002;received in revised form 30October 2002;accepted 16December 2002AbstractA Chebyshev super spectral viscosity method and operator splitting are used to solve a hyperbolic system of con-servation laws with a source term modeling a ﬂuidized bed.The ﬂuidized bed displays a slugging behavior which corresponds to shocks in the solution.A modiﬁed Gegenbauer postprocessing procedure is used to obtain a solution which is free of oscillations caused by the Gibbs–Wilbraham phenomenon in the spectral viscosity solution.Conser-vation is maintained by working with unphysical negative particle concentrations.Ó2003Elsevier Science B.V.All rights reserved.Keywords:Chebyshev collocation;Super spectral viscosity;Pseudospectral;Gibbs–Wilbraham phenomenon;Edge detection;Gegenbauer postprocessing1.IntroductionFluidized beds are used in the chemical and fossil fuel processing industries to mix particulate solids and ﬂuids (gases or liquids).A typical ﬂuidized bed consists of a vertically oriented chamber,a bed of par-ticulate solids,and a ﬂuid ﬂow distributor at the bottom the chamber.The ﬂuid ﬂows upward through the particles creating a force that counteracts gravity at which time a state of minimum ﬂuidization is reached.Stronger gas inﬂows (more than is necessary to maintain minimum ﬂuidization)lead to pockets of gas,or equivalently low particle concentrations,resembling bubbles in a liquid traveling upward through the particles.Each rising bubble pushes a large amount of mass in front of it.Particles move downward through and around the rising bubble until it reaches the top of the bed.A settled bed is reestablished and the cycle repeats.Each set of upward moving particles is referred to as a slug.In this paper we consider only one-dimensional ﬂow.Physically,this corresponds to ﬂow in a narrow diameter ﬂuidized bed.The ﬂuidized bed model was originally solved numerically in [6]by ﬁnite diﬀerence methods.An exact solution to the homogeneous system with Riemann initial conditions has beendevel-Journal of Computational Physics 186(2003)630–/locate/jcpE-mail address:scott@.0021-9991/03/$-see front matter Ó2003Elsevier Science B.V.All rights reserved.doi:10.1016/S0021-9991(03)00089-5oped in[7].Theﬂuidized bed model can be put in the form of system of conservation laws with a source term asu tþfðuÞx¼bðuÞ:ð1ÞSpectral viscosity methods have been successfully applied to homogeneous systems of conservation laws. We use operator splitting to extend the methods to nonhomogeneous systems of conservation laws.If discontinuities are present in the solutions,the spectral viscosity approximations will be contaminated by the Gibbs–Wilbraham phenomenon,but the spectral viscosity solution may be postprocessed to obtain a better approximation.While the spectral viscosity is applied to the solution at every time level,postpro-cessing is only done at times for which a‘‘clean’’solution is desired.Several methods exist for postpro-cessing spectral approximations.They include spectral molliﬁcation[12,25,26,36]methods which involve using a two-parameterﬁlter,the Gegenbauer Reconstruction Procedure(GRP),and a recently developed Fourier–Pad e-based algorithm[9].Spectral molliﬁcation is a fairly robust method which may be used with or without the knowledge of edge locations.However,it will only recover spectral accuracy up to within a neighborhood of discontinuity locations.The GRP is capable of recovering spectral accuracy at every point,even at the locations of the discontinuities.In this paper,all numerical examples have been postprocessed using the GRP.One of our goals was to examine if the GRP,which has shown great promise on some simple examples,could be used to successfully postprocess PDE solutions which were either more detailed than piecewise linear or if the method could be used to postprocess solutions containing varying subintervals of detail.The solutions in the previous ap-plications[15,32]consisted of homogeneous features throughout the computational domain which allowed the parameters of the postprocessing method to be chosen globally.Theﬂuidized bed solutions contain features of varying detail throughout the computational domain and a diﬀerent strategy must be used to choose the postprocessing parameters.Additionally,we examine what remains to be done if the GRP is to be used as a‘‘black box’’postprocessing method for spectral approximations.This paper is organized as follows:In Section2,the Chebyshev collocation method and super spectral viscosity methods are reviewed.Section3summarizes a method to locate edges in the spectral viscosity approximations.Edge locations will be necessary to apply the postprocessing procedure.Section4describes the GRP for non-periodic functions.Section5describes theﬂuidized bed model.Numerical results are presented in Section6.2.Chebyshev super spectral viscosity methodThe standard collocation points for a Chebyshev collocation(pseudospectral)method are usually de-ﬁned byx j¼Àcosp jN;j¼0;1;...;N:ð2ÞThese points are extrema of the N th order Chebyshev polynomial,T kðxÞ¼cosðk arccosðxÞÞ:ð3ÞThe points are often labeled the Chebyshev–Gauss–Lobatto(CGL)points,a name which alludes to the points role in certain quadrature formulas.The CGL points cluster quadratically around the endpoints and are less densely distributed in the interior of the domain.The Chebyshev collocation method is based on assuming that an unknown PDE solution,u,can be represented by a global,interpolating,Chebyshev partial sum,S.A.Sarra/Journal of Computational Physics186(2003)630–651631u NðxÞ¼X Nn¼0a n T nðxÞ:ð4ÞThe discrete Chebyshev coeﬃcients,a n,are deﬁned bya n¼2N1c nX Nn¼0uðx jÞT nðx jÞc j;where c j¼2when j¼0;N;1otherwise:ð5ÞDerivatives of u at the collocation points are approximated by the derivative of the interpolating polynomial evaluated at the collocation points.Theﬁrst derivative,for example,is deﬁned by,d u d x ¼X Nn¼0að1ÞnT nðxÞ:ð6ÞSince að1ÞNþ1¼0and að1ÞN¼0,the non-zero derivative coeﬃcients can be computed in decreasing order by therecurrence relation:c n að1Þn ¼að1Þnþ2þ2ðnþ1Þa nþ1;n¼NÀ1;...;1;0:ð7ÞThe transform pair given by Eqs.(4)or(6)and(5)can be eﬃciently computed by a fast cosine transform. Equivalently,the interpolating polynomial and its derivatives can be computed in physical space using matrix multiplication[4].Special properties of the Chebyshev basis allow for diﬀerentiation via parity matrix multiplication[3](even–odd decomposition[33]),which can be performed by using slightly more than half as manyﬂoating point operations as standard matrix multiplication.More detailed information may be found in the standard references[4,10,11,18,19,37].After the spectral evaluation of spatial derivatives,the system of ordinary diﬀerential equationsd ud t¼Fðu;tÞresults,where u is the vector containing the unknown PDE solution at the collocation points.The system is typically integrated by a second,third,or fourth-order explicit Runge–Kutta method to advance the so-lution in time.A coordinate transformation may be necessary either to map a computational interval to½a;b from the interval½À1;1 ,or to redistribute the collocation points within an interval for the purpose of giving high resolution to regions of very rapid change.Popular maps used to redistribute the CGL points(2)are the Kosloﬀ/Tal-Ezer map[27]x¼gðn;cÞ¼arcsinðcnÞarcsinðcÞ;ð8Þthe center map[1]x¼gðn;cÞ¼ð1:0ÀcÞn3þcn;ð9Þand the two parameter tangent map[2]x¼gðn;c;lÞ¼x0þtanðdnþxÞc;ð10Þwhere j¼arctanðcð1ÀlÞÞ,c¼arctanðcð1þlÞÞ,d¼0:5ðjþcÞ,x¼0:5ðjÀcÞ,and x0¼À1þ2ðlÀaÞ=ðbÀaÞ.632S.A.Sarra/Journal of Computational Physics186(2003)630–651If n denotes the original variable and x¼gðnÞthe new variable,then after a change of variable is performed Eq.(1)becomesu tþ1g0ðnÞfðuÞx¼bðuÞ:ð11ÞIf the PDE solution contains shocks,the spectral collocation method will not converge to the correct entropy solution[35].In this case,a spectrally small viscosity term must be added in order to stabilize the approximation and ensure convergence to the entropy solution.This can be done without sacriﬁcing spectral accuracy and can be accomplished in several diﬀerent ways,with each way being labeled a par-ticular type of spectral viscosity method.We have used the super spectral viscosity(SSV)method of[28], which for a conservation law in one space dimension,can be stated aso o t u Nþoo xfðu NÞ¼eðÀ1Þsþ1Q2s u N;ð12Þwhere the viscosity operator is given byQ¼ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ1Àx2p oo x:ð13ÞIt was shown in[28]that if e¼CN1À2s,with the parameter C chosen large enough to ensure stability and such that06C6N1=2,and with the parameter s chosen such that s6lnðNÞand allowing s to grow with N, that bounded solutions of(12)will converge to the correct entropy solution in the case bðuÞ¼0.Except for the ranges mentioned in order to ensure convergence to the entropy solution,the parameters s and C are problem dependent,depending mainly on the strength of the shocks involved.A direct implementation of(12)amounts to adding2s spatial derivatives to the equation.This would introduce additional stiﬀness which would severely limit the stable time step and increase the computational work involved by requiring the computation of higher order derivatives.Hence,the practical implementation of the SSV method is an important issue.The eﬃcient implementation oﬀthe SSV method wasﬁrst addressed in[8],where the authors recognized that the SSV method could be implemented as a spectralﬁlter.This fact is based on the examination of the viscosity operator Q2applied to the Chebyshev polynomial(3),T kðxÞ.Q2T kðxÞ¼ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ1Àx2p oo xﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ1Àx2p oo xT kðxÞ¼Àk2T kðxÞ:ð14ÞAs a result of applying the viscosity operator to the Chebyshev polynomials,it can be noticed that the Chebyshev polynomials are the eigenfunctions of the operator Q2with eigenvalues k2.Expanding the viscosity term,which is the right-hand side of(12),we notice thateðÀ1Þsþ1Q2s u N¼ÀCNX Nk¼0kN2sa kðtÞT kðxÞ:ð15ÞIf we implement the SSV method via time splitting where in theﬁrst step we solveo o t u Nþoo xfðu NÞ¼0ð16Þand in the second step we solveo o t u N¼eðÀ1Þsþ1Q2s u N;ð17ÞS.A.Sarra/Journal of Computational Physics186(2003)630–651633the second equation,(17),in the split step can be written aso o tX Nk¼0a kðtÞT kðxÞ"#¼ÀCNX Nk¼0kN2sa kðtÞT kðxÞ;which can be solved analytically.Over one time step,the analytical solution modiﬁes the Chebyshev co-eﬃcients asa kðtþD tÞ¼a kðtÞexpðÀCN D tðk=NÞ2sÞ:Thus,the exact solution of the SSV split step can be written as theﬁltered partial sumu NðxÞ¼X Nk¼0rkNa kðtÞT kðxÞ;ð18Þwherer k N¼expÀakNis an exponentialﬁlter of strength a and order b as described in[38].The Chebyshev SSV method is seen to be equivalent to applying the exponentialﬁlter with b¼2s and a¼CN D t.The method can be implemented with little additional cost.It should be stressed that while the SSV method is being implemented via the exponentialﬁltering framework,that it is not a b th orderﬁlter as it does not meet the requirements set forth in[38].The amount of damping of the high modes is signiﬁcantly less with the SSV method than with the application of a b th order exponentialﬁlter.An application of a b th order exponentialﬁlter typically takes a¼Àln e where e is machine zero(on a32-bit machine using double precisionﬂoating point operations, e¼2À52and lnðeÞ’À36:0437).Fig.1compares two exponentialﬁlters of diﬀerent orders with an appli-cation of theﬁlter with the parameters set as a¼0:032and b¼4,which are possible settings that may be used if theﬁltering framework is used to implement the SSV method.634S.A.Sarra/Journal of Computational Physics186(2003)630–6513.Edge detectionThe GRP recovers spectral accuracy up to the discontinuity points in each smooth subinterval of a piecewise analytic function.Thus,the GRP needs the exact location of discontinuities,or edges,in the function.If a PDE solution is being postprocessed and the solution contains rarefaction waves,disconti-nuities in the ﬁrst derivative of the function will exist and need to be located as well.The method used to ﬁnd the edges originated in [14]for periodic and non-periodic functions.The method is specialized to approximations of functions by Chebyshev methods and is summarized below.Denote the location of discontinuities as a j .Let½f ðx Þ:¼f ðx þÞÀf ðx ÀÞdenote a local jump in the function and deﬁneue ðx Þ¼p ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ1Àx 2p N X N k ¼0a k d d x T k ðx Þ;ð19Þwhered d x T k ðx Þ¼k sin ðk arccos ðx ÞÞﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ1Àx 2p :Essentially,we are looking at the derivative of the spectral projection of the numerical solution to de-termine the location of the discontinuities.The series ue ðx Þhas the convergence propertiesue ðx Þ!O ð1=N Þwhen x ¼a j ;½f ða j Þwhen x ¼a j :The series converges to both the height and direction of the jump at the location of a discontinuity.However,for the GRP,we only need the locations and magnitudes of the jumps,not the directions.While a graphical examination of the series ue ðx Þveriﬁes that the series does have the desired convergence prop-erties,an additional step is needed to numerically pinpoint the location of the discontinuities.For that purpose,make a non-linear enhancement to the edge series asun ðx Þ¼N Q =2½ue ðx Þ Q :The values,un ðx Þ,will serve to amplify the separation of scales which has taken place in (19).The series has the convergence propertiesun ðx Þ!O ðN ÀQ =2Þwhen x ¼a j ;N Q =2½½f ða j Þ Q when x ¼a j :By choosing Q >1we enhance the separation between the O ð½1=N Q =2Þpoints of smoothness and the O ðN Q =2Þpoints of discontinuity.The parameter J ,whose value will be problem dependent,is a critical threshold value.Finally,redeﬁne ue ðx Þasue ðx Þ¼j ue ðx Þj if un ðx Þ>J ;0otherwise :With Q large enough,one ends up with an edge detector ue ðx Þ¼0at all x except at the discontinuities x ¼a j .Only those edges with amplitude larger than J 1=Q ﬃﬃﬃﬃﬃﬃﬃﬃﬃ1=N p will be detected.S.A.Sarra /Journal of Computational Physics 186(2003)630–651635Often the series ue is slow to converge in the area of a discontinuity and the nonlinear enhancement has diﬃculty pinpointing the exact location of the edge.If an additional parameter,g,is added to the procedure this problem can be overcome in a simple manner.The parameter speciﬁes that only one edge may be located in the intervalðx½iÀg ;x½iþg Þ,i¼0;...;N,with appropriate one sided intervals being considered near boundaries.The correct edge will be the maximum of ue in this subinterval.The value of g is problem-dependent and is best chosen after the edge detection procedure has been applied once.The edge detection parameters J,Q,and g,are all problem-dependent.Various combinations of the parameters may be used to successfully locate edges represented by jumps of a magnitude in a certain range.4.Gegenbauer reconstructionThe truncation error decays exponentially as N increases when spectral methods are used to approximate smooth functions.However,the situation changes when the function is discontinuous as the spectral ap-proximation no longer converges in the maximum norm.This is known as the Gibbs–Wilbraham phe-nomenon.Several methods exist for removing or reducing the eﬀects of the Gibbs–Wilbraham phenomenon from spectral approximations.Most however,such as spectral molliﬁcation[17,36],only recover spectral accuracy up to within a neighborhood of each discontinuity.To date,the most powerful postprocessing method seems to be the GRP which is capable of recovering spectral accuracy up to and including at the location of discontinuities.Although the GRP has been shown to produce remarkable results on some simple problem,the method lacks robustness due to the fact that two parameters,for which an optimal choice for is currently not known,must be speciﬁed.The GRP was developed in[20–24]for the purpose of recovering exponential accuracy at all points, including at the discontinuities themselves,from the knowledge of a spectral partial sum of a discontinuous, but piecewise analytic function.While the SSV solution serves as a highly accurate approximation to the exact spectral partial sum,only partial theoretical justiﬁcation can be found concerning using the GRP as a postprocessing method for the SSV solution.However,numerical results indicate that spectral accuracy can be achieved by applying the GRP to the SSV solution of homogeneous systems of conservation laws[12,15]. The same can be said about the edge detection method,as the theoretical results are limited to locating the jump discontinuities of a piecewise smooth function uðxÞ.However,numerical evidence also advocates applying the edge detection method to the SSV solution.The GRP works by expanding the function in another basis,the Gibbs complementary basis,via knowledge of the known Chebyshev coeﬃcients and the location of discontinuities.The Chebyshev partial sums are projected onto a space spanned by the Gegenbauer polynomials.The associated weight functions increasingly emphasize information away from the discontinuities as the number of included modes grow. The approximation converges exponentially in the new basis even though it only converged very slowly in the original basis due to the Gibbs–Wilbraham phenomenon.The choice of a Gibbs complementary basis isthe Ultraspherical or Gegenbauer polynomials,C kn .The Gegenbauer polynomials are orthogonal polyno-mials of order n which satisfyZ1À1ð1Àx2ÞkÀ1=2C kkðxÞC knðxÞd x¼h kn;k¼n;0;k¼n;where(for k P0)h k n ¼p1=2C knð1ÞCðkþð1=2ÞÞCðkÞðnþkÞ636S.A.Sarra/Journal of Computational Physics186(2003)630–651withC k n ð1Þ¼Cðnþ2kÞn!Cð2kÞ:Whether the Gegenbauer basis is the optimal choice as the Gibbs complementary basis for the Chebyshev basis remains an open question.In other words,it may be possible to construct another basis in which the slowing converging Chebyshev approximation could be expanded in to obtain an approximation with better convergence properties than those of the Gegenbauer approximation.However,it is shown in[24] that the Gegenbauer basis is a Gibbs complementary basis for the Chebyshev basis.The Gegenbauer expansion of a function uðxÞ;x2½À1;1 isuðxÞ¼X1l¼0b f klC klðxÞ;where the continuous Gegenbauer coeﬃcients,b f k l,of uðxÞareb f k l ¼1h klZ1À1ð1Àx2ÞkÀ1=2C klðxÞuðxÞd x:ð20ÞSince we do not know the function uðxÞ,implementing the GRP requires obtaining an exponentiallyaccurate approximation,b g kl ,to theﬁrst m coeﬃcients b f k l in the Gegenbauer expansion from theﬁrst Nþ1Chebyshev coeﬃcients of uðxÞ.The approximate Gegenbauer coeﬃcients are deﬁned as the integralb g k l¼1hl Z1À1ð1Àx2ÞkÀ1=2C klðxÞu NðxÞd x;ð21Þwhere u N is the Chebyshev partial sum(4).The integral should be evaluated by Gauss–Lobatto quadraturein order to insure suﬃcient accuracy.The coeﬃcients b g kl are now used in the partial Gegenbauer sum toapproximate the original function asuðxÞ%u km ðxÞ¼X ml¼0b g k l C k lðxÞ:In practice,there will be discontinuities in the interval½À1;1 and the reconstruction must be done on each subinterval½a;b in which the solution remains smooth.To accomplish the reconstruction on each subinterval,deﬁne a local variable for each subinterval as xðnÞ¼ nþd where ¼ðbÀaÞ=2,d¼ðbþaÞ=2 and n j¼cosðp j=NÞ:The reconstruction in each subinterval is then accomplished byu k; m ð nþdÞ¼X ml¼0b g k ðlÞC k lðnÞ;whereb g k ðlÞ¼1hl ZÀ11ð1Àn2ÞkÀ1=2C klðnÞu Nð nþdÞd n:Notice that we have used collocation points on the entire interval½À1;1 to build the approximation in ½a;b .This is referred to as a global–local approach[22].The global–local approach seems to be best when postprocessing PDE solutions where u N is obtained from the time evolution of the PDE solution.The point values uðx iÞmay not be accurate,but the global interpolating polynomial u NðxÞis accurate.S.A.Sarra/Journal of Computational Physics186(2003)630–651637In order to show that the GRP yields uniform exponential accuracy for the approximation,it is nec-essary to select k and m such that k¼m¼b N,where b<2e=ð27ð1þ1=2pÞÞ,and p is the distance from ½À1;1 to the nearest singularity in the complex plane,in each subinterval where the function being re-constructed is assumed to be analytic[24].It is not necessary,and usually not advisable,to choose k¼m.In practice,we are often more concerned with obtaining results for aﬁxed N,rather than achieving an ex-ponential convergence rate.If the function to be postprocessed consists homogeneous features,the reconstruction parameters can be successfully chosen as k¼k k N and m¼k m N for each subinterval where k k and k m are user chosen, globally applied parameters.We refer to this strategy as the global approach.In all previous applications of the GRP in the literature,the method was applied to such functions and it was possible to chose the parameters in this way[12,13].However,in problems with solutions with varying detail throughout the computational domain,the reconstruction parameters may need to be chosen independently in each subinterval[30].We refer to this strategy as the local approach.To date there is no known method to choose optimal values of the reconstruction parameters m and k.The parameters remain very problem dependent.Work is under way on choosing optimal parameters and results will be reported in a future paper.5.Fluidized bed equationsThe variable x denotes the vertical height in the bed.Let aðx;tÞdenote the concentration of particles by volume,vðx;tÞthe particle velocity,and mðx;tÞ¼aðx;tÞvðx;tÞthe particle momentum.The parameter a0is the concentration of particles at equilibrium(when v¼0)and a p is the packing concentration which sets an upper limit for a where a2ð0;1Þ.The parameter a0u denotes the particle concentration corresponding to the critical state dividing linearly stable and unstable states(the particle concentration at minimumﬂu-idization).The constant s¼3:5ð1Àa0uÞ2:5ða pÀa0uÞis related to the linear stability of the equilibrium solutions which correspond to states of uniformﬂuidization.The model can be put in the form of a system of conservation laws with a source term asa tþm x¼0;ð22Þm tþðm2=aþFðaÞÞx¼bða;mÞ;ð23ÞwhereFðaÞ¼s2aþs2a2paÀa pþ2s2a p lnðj aÀa p jÞ:The function bða;mÞin the source term is given bybða;mÞ¼Àaþa JÀm ð1ÀaÞ;where J¼ð1Àa0Þ3:5represents the total volumetricﬂux through the bed.Increasing J(or decreasing a0) corresponds to turning up the inﬂowing gas.Values a0<a0u correspond to large gasﬂuxes and have been shown to produce unstable states corresponding to slug-like solutions.Values a0>a0u give rise to stable states.From a mathematical point of view,the non-homogeneous system of conservation laws coincides with the Euler equations for an isentropic gasﬂow,subject to volumetric forces.The variables a,v,and FðaÞplay the role of density,velocity,and pressure respectively,in the Euler equations.638S.A.Sarra/Journal of Computational Physics186(2003)630–651S.A.Sarra/Journal of Computational Physics186(2003)630–6516395.1.Vacuums and unphysical particle concentrationsA vacuum is said to exist at a collocation point if the particle concentration is zero.Numerically,we will assume that a vacuum exists at a grid point if the concentration is either zero or it is very small (j a j<thres).The system becomes meaningless at vacuum points as m2=a is either undeﬁnedða¼0Þor produces unrealistic valuesðj a j<thresÞ.At each vacuum point encountered in the numerical method,the corresponding values of v,and therefore m,are set equal to zero at that collocation point rather than using the spurious valueðj a j<thresÞor NaN valueða¼0Þ.Values of a such that j a j<thres are retained and not set to zero.Stable approximations by the spectral method always produced a<a p.In the spectral method,a must be allowed to take negative values even though a negative concentration in not physically meaningful,as this information is used in the GRP to postprocess the result.When it was attempted to artiﬁcially force the spectral method to work only with a>0,the quality of the postpro-cessed solution was adversely aﬀected.More importantly,even though the spectral collocation method is conservative,if for a<0,a was redeﬁned as a¼0,the conservative properties of the method were destroyed and the method started producing mass.If the values of a were allowed to be negative,the method was conservative and mass was preserved to as many as six decimal places.In all reported re-sults,the parameter thres was taken to be thres¼0:001.After postprocessing the solution,all concen-trations are such that a P0.6.Numerical resultsAll examples were postprocessed using the spectral signal processing[31]suite.In the reported results we have used a0u¼0:55and a p¼0:6.6.1.Homogeneous systemTheﬁrst two problems solve the homogeneous system with Riemann initial data so that the Chebyshev SSV method with GRP postprocessing may be validated against an exact solution.Ourﬁrst example consists of a left-moving shock wave and a right moving rarefaction wave.The initial conditions are vðx;0Þ¼0for all x in a domain of½À0:2;0:2 and aðx;0Þ¼0:3if x<0and aðx;0Þ¼0:55if x P0.Fig.2shows the solution advanced to time t¼0:5with a fourth-order Runge–Kutta method.The grid consists of64points distributed by map(9)with c¼0:25.The use of the coordinate map has the eﬀect of placing more points in the center of the domain.The SSV parameters used were C¼1and s¼4which produced a viscosity parameter of e¼CN1À2s¼2:27EÀ13(or a¼CN D t¼0:16and b¼8in the expo-nentialﬁlter).The rarefaction wave is characterized by the solution having a discontinuousﬁrst derivative,thus edge detection must be applied to theﬁrst derivative of the solution in addition to the solution itself.The edge detection procedure with Q¼1and J¼1locates jumps of magnitude greater than0.125.With these settings,the edge detection procedure locates edges in the function and theﬁrst derivative of the function at x¼À0:0331,x¼0:0331,and x¼0:1374.We were unable to get good postprocessed results by specifying the reconstruction parameters globally through the parameters k k and k m.Global parameter speciﬁcation failed due to the solution containing three intervals of piecewise constant values and a fourth intervalð0:033;0:1374Þconsisting of a function requiring diﬀerent reconstruction parameters.Good results were obtained by specifying the GRP parameters locally in each smooth subinterval as listed in Table1.The postprocessed solution in Fig.3.。

EN 61853-1-2011 PV module performance testing and energy rating - Part 1

EUROPEAN STANDARD NORME EUROPÉENNE EUROPÄISCHE NORM
ICS 27.160
EN 61853-1
March 2011
English version
Photovoltaic (PV) module performance testing and energy rating Part 1: Irradiance and temperature performance measurements and power rating (IEC 61853-1:2011)
-2-
Foreword
The text of document 82/613/FDIS, future edition 1 of IEC 61853-1, prepared by IEC TC 82, Solar photovoltaic energy systems, was submitted to the IEC-CENELEC parallel vote and was approved by CENELEC as EN 61853-1 on 2011-03-02. Attention is drawn to the possibility that some of the elements of this document may be the subject of patent rights. CEN and CENELEC shall not be held responsible for identifying any or all such patent rights. The following dates were fixed: – latest date by which the EN has to be implemented at national level by publication of an identical national standard or by endorsement – latest date by which the national standards conflicting with the EN have to be withdrawn Annex ZA has been added by CENELEC. __________

ICESat-2_数据背景光子特性及滤波方法研究

第 31 卷第 5 期2023 年 3 月Vol.31 No.5Mar. 2023光学精密工程Optics and Precision EngineeringICESat-2数据背景光子特性及滤波方法研究谢欢1，2*，黄佩琪1，徐琪1，叶丹1，孙媛1，栾奎峰3，刘世杰1，2，童小华1，2（1.同济大学测绘与地理信息学院，上海 200092；2.上海市航天测绘遥感与空间探测重点实验室，上海 200092；3.上海海洋大学海洋科学学院，上海 201306）摘要：ICESat-2（Ice， Cloud and Land Elevation Satellite-2）是世界首颗采用光子计数模式的激光测高卫星，可快速获得高精度、大尺度地面三维数据。

光子探测机制使得数据中除了地面信号外，还包含大气散射等背景信号，需要通过滤波才能获得地形等信息。

为分析ICESat-2背景和信号光子的分布特点及点云滤波算法的效果和适用性，本文首先选取了六种地表覆盖类型（城市、海冰、沙漠、植被、海洋及冰盖/冰川）及不同观测条件的数据，对其背景光子率进行统计分析。

分析结果表明：白天观测数据的背景光子率平均为106（点/秒）数量级，远高于夜晚观测数据的背景光子率——104（点/秒）数量级，弱波束的背景光子率与强波束背景光子率相当，六种地表覆盖类型中，冰盖/冰川的背景光子率最高。

然后，根据统计结果筛选出21组测高数据，并选取七种具有代表性的点云滤波对其进行去噪实验，分析精度后得出结论：改进局部密度法的去噪效果最佳，算法召回率、精准度和F值均大于0.90，算法较为稳定。

最后，对所选取各滤波算法的精度、特点与适用性等性质进行了总结与分析，可为后续该数据的使用和滤波算法的选择提供参考。

关键词：ICESat-2；激光测高；光子计数激光雷达；点云去噪；背景光子率中图分类号：TP391.4 文献标识码：A doi：10.37188/OPE.20233105.0631Research on background photon characteristics and filteringmethods for ICESat-2 dataXIE Huan1，2*，HUANG Peiqi1，XU Qi1，YE Dan1，SUN Yuan1，LUAN Kuifeng3，LIU Shijie1，2，TONG Xiaohua1，2（1.College of Surveying and Geo-informatics， Tongji University， Shanghai 200092， China；2.Shanghai Key Laboratory for Planetary Mapping and Remote Sensing for Deep Space Exploration，Shanghai 200092， China；3.College of Marine Sciences， Shanghai Ocean University， Shanghai 201306， China）* Corresponding author， E-mail： huanxie@Abstract： ICESat-2 is the world's first laser altimetry satellite to use photon counting technology. It pro⁃vides high-precision，three-dimensional，large-scale ground data that can be obtained rapidly.Altimetry data obtained using photon detection systems generally contain ground signals and background signals，such as atmospheric scattering， which must be filtered to obtain ground information. To analyze the distri⁃文章编号1004-924X（2023）05-0631-13收稿日期：2022-06-15；修订日期：2022-07-05.基金项目：国家自然科学基金项目（No.41822106）；上海市教育发展基金会曙光计划项目（No.18SG22）第 31 卷光学精密工程bution characteristics of ICESat-2 background and signal photons and the performance of the point cloud filtering algorithm，we selected data from six types of land cover （city，sea ice，desert，vegetation，ocean and ice sheet （or glacier））in different observation conditions.A statistical analysis of the back⁃ground rate was then performed. The results show that the average background rate of daytime observa⁃tion data is 106 points per second， which is much higher than that of nighttime observation data （104 points per second）. The background rate of the weak beam is equivalent to that of the strong beam. Among the six types of land cover， the ice sheet （or glacier） has the highest background rate. Based on the statistical results，21 datasets and seven representative point cloud filters were selected for denoising experiments. After analyzing the accuracy， we can conclude that the improved local density method with recall， accura⁃cy，and F-measure values greater than 0.90 has the best denoising performance and is relatively stable. Finally， the performance， characteristics， and applicability of the seven filtering algorithms were summa⁃rized and analyzed. This can serve as a reference for the subsequent data use and selection of filtering algo⁃rithms.Key words： ICESat-2； laser altimetry； photon counting LiDAR； photon denoising； background rate1 引言激光测高卫星由于速度快、精度高、节省人力物力［1］等优势，成为了当今获取大范围地表三维信息的重要工具之一［2］。

redfit

Computers&Geosciences28(2002)421–426REDFIT:estimatingred-noise spectra directly from unevenly spaced paleoclimatic time series$Michael Schulz a,*,Manfred Mudelsee ba Institut f.u r Geowissenschaften,Uni v ersit.a t Kiel,Olshausenstr.40,D-24118Kiel,Germanyb Institut f.u r Meteorolo g ie,Uni v ersit.a t Leipzi g,Stephanstr.3,D-04103Leipzi g,GermanyReceived20May2000;received in revised form27March2001;accepted2April2001AbstractPaleoclimatic time series are often unevenly spaced in time,makingit diﬃcult to obtain an accurate estimate of their red-noise spectrum.A Fortran90program(REDFIT)is presented that overcomes this problem byﬁtting aﬁrst-order autoregressive(AR1)process,being characteristic for many climatic processes,directly to unevenly spaced time series. Hence,interpolation in the time domain and its inevitable bias can be avoided.The program can be used to test if peaks in the spectrum of a time series are signiﬁcant against the red-noise background from an AR1process.Generated and paleoclimatic time series are used to demonstrate the capability of the program.r2002Elsevier Science Ltd.All rights reserved.Keywords:Spectral analysis;Irregular sampling intervals;Lomb–Scargle Fourier transform1.IntroductionSpectral analysis is an important tool in climateresearch because it allows the variance of a time series tobe separated into contributions associated with diﬀerenttime scales.It thus helps to understand better thephysical processes,which generate the variability re-corded in a time series.Spectra of paleoclimatic timeseries frequently show a continuous decrease of spectralamplitude with increasingfrequency(‘‘red-noise’’).Hasselmann(1976)demonstrated that aﬁrst-orderautoregressive(AR1)process is suﬃcient to explain thisclimatic red-noise signature.Accordingly,the AR1model is often used as null hypothesis to assess whetheror not the variability recorded in a time series isconsistent with a stochastic origin of this type(Gilmanet al.,1963).Such a test involves estimation of an AR1parameter from the time series under consideration.Forevenly sampled time series this is a relatively straightfor-ward procedure(e.g.Percival and Walden,1993).However,most paleoclimatic time series are unevenlyspaced(i.e.,intervals between samplingtimes are notconstant),and the application of estimation techniquesfor evenly spaced time series would require some sort ofinterpolation.Unfortunately,this procedure results in asigniﬁcant bias because interpolation in the time domainalters the estimated spectrum of a time series byenhancingthe low-frequency components at the expenseof high-frequency components.That is,the estimatedspectrum of an interpolated time series becomes too‘‘red’’compared to the true spectrum(e.g.Schulz andStattegger,1997).We present a computer program which estimates theAR1parameter directly from unevenly spaced timeseries,that is,without requiringinterpolation.Theestimated AR1model is then transformed from the timedomain into the frequency parison of thespectrum of the time series with that of the AR1modelallows to test the hypothesis that the time seriesoriginates from an AR1process.Following a brief $Code available from server at /CGEditor/index.htm*Correspondingauthor.E-mail addresses:mschulz@email.uni-kiel.de(M.Schulz),mudelsee@rz.uni-leipzig.de(M.Mudelsee).0098-3004/02/$-see front matter r2002Elsevier Science Ltd.All rights reserved.PII:S0098-3004(01)00044-9description of the numerical procedure and its imple-mentation in a computer program,we apply the program to a synthetic time series and a paleoclimatic record.2.MethodA discrete AR1process r for times t i ði ¼1;2;y ;N Þwith arbitrary spacingis g iven by (Robinson,1977)r ðt i Þ¼r i r ðt i @1Þþe ðt i Þ;ð1Þr i ¼exp @ðt i @t i @1Þ=t ÀÁ:The constant t is the characteristic time scale of the AR1process (a measure of its ‘‘memory’’)and e indicates ‘‘white’’Gaussian noise with zero mean and variance s 2e 1@exp @2ðt i @t i @1Þ=t ÀÁ.This value of s 2e en-sures that the AR1process is stationary and has unit variance.The spectrum G rr ðf j Þcorrespondingto the time-domain process of Eq.(1)is (e.g.Percival and Walden,1993)G rr ðf j Þ¼G 01@r 21@2r cos ðp f j =f Nyq Þþr 2ð2Þwhere f j denotes discrete frequency up to the Nyquist frequency f Nyq (cf.Schulz and Stattegger,1997)and G 0is the average spectral amplitude.The ‘‘average autocor-relation coeﬃcient’’r is calculated from the arithmetic mean of the samplingintervals D t ¼ðt N @t 1Þ=ðN @1Þas r exp @D t =t ÀÁ.The unknown value of t is estimated from an unevenly spaced time series usingthe least-squares algorithm devised by Mudelsee (2002).The spectrum of an irregularly spaced time series is determined without the need for interpolation by means of the Lomb–Scargle Fourier transform (Lomb,1976;Scargle,1982,1989).Schulz and Stattegger (1997)presented a computer program for this purpose which makes additional use of the so-called Welch-overlapped-seg-ment-averaging (WOSA)procedure (Welch,1967).This algorithms splits a time series into n 50segments which overlap by 50%,the ﬁnal spectral estimate is derived from averaging the n 50periodograms.With an estimate for t as well as an appropriate value for G 0it should then be possible to overlay the red-noise spectrum after Eq.(2)and the spectrum estimated from the data.Provided that the probability distribution of G rr at each frequency follows a w 2distribution (e.g.Percival and Walden,1993),it is ﬁnally possible to test if the data spectrum is consistent with a red-noise model.Unfortunately,this approach is hampered by an inherent aspect of the Lomb–Scargle Fourier transform:in contrast to the classical Fourier transform,the individual Lomb–Scargle Fourier components are not necessarily independent of each other and,as aconsequence,an estimated spectrum based on the Lomb–Scargle transform may be biased (Lomb,1976;Scargle,1982).In particular,spectral amplitudes at the high-frequency end of a spectrum are often over-estimated.Therefore,a red-noise spectrum Eq.(2)which is based on an unbiased estimate of t for a given time series will not necessarily coincide with the ‘‘Lomb–Scargle spectrum’’of the same time series.We therefore seek for a bias correction for the Lomb–Scargle Fourier transform.3.Numerical procedureThe systematic deviation between a theoretical red-noise spectrum Eq.(2)and one estimated from an unevenly spaced time series by means of the Lomb–Scargle Fourier transform depends on the distribution of the samplingtimes in the interval [t 1,t N ](Lomb,1976;Scargle,1982).For some arbitrary distribution of samplingtimes the lack of an analytical solution for the deviation prevents a direct bias correction of a Lomb–Scargle spectrum.To circumvent this problem,we turn to a Monte–Carlo technique.Based on the actual samplingtimes,an ensemble of N sim AR1time series is generated after Eq.(1)with ﬁxed t .The deviation of the average spectrum of the ensemble from the known theoretical spectrum is then employed for the required bias correction.The computational steps to obtain a red-noise spectrum of an unevenly spaced time series x ðt i Þwhich is consistent with the estimated value of t are as follows:1.Estimate t from x ðt i Þusingthe time-domain algorithm of Mudelsee (2002).If more than one WOSA segment is used for spectral analysis (n 50>1),an average value for t is calculated from t estimates for each individual segment.The individual t estimates (Mudel-see,in press)are bias corrected,based on the number of data points in each WOSA segment.2.Estimate spectrum #Gxx ðf j Þof x ðt i Þin the interval [0,f Nyq ]usingthe Lomb–Scarg le Fourier transform as described in Schulz and Stattegger (1997).Determine thearea under #Gxx ðf j Þwhich is an estimate for the variance of x ðt i Þ.3.Monte Carlo simulation loop.Repeat N sim times*create AR1time series accordingto Eq.(1),using the samplingtimes of the input data ðt i Þ,the estimated t ,and an independent set of e ðt i Þfor each simulation*estimate spectrum of the generated AR1time series,#Grr ðf j ÞM.Schulz,M.Mudelsee /Computers &Geosciences 28(2002)421–426422*scale #Grr ðf j Þsuch that the area under the spectrum is identical to the area under #Gxx ðf j ÞDetermine arithmetic mean of the N sim independent red-noise spectral estimates #Grr ðf j Þ .4.Calculate theoretical AR1spectrum G rr ðf j Þfor theestimated value of t (Eq.(2)).(Note that G rr ðf j Þis not aﬀected by the bias of the Lomb–Scargle Fourier transform,because the critical parameter t is estimated in the time domain.)5.Select G 0(see Eq.(2))such that the area underG rr ðf j Þis identical to the area under #Gxx ðf j Þ.(This step is required since the true noise variance of the process under consideration is unknown.)6.Calculate a correction factor c ðf j Þfor the bias adjustment of the Lomb–Scargle spectrum as c ðf j Þ¼#Grr ðf j Þ=G rr ðf j Þ:ing c ðf j Þ,determine a bias-corrected version ofthe spectrum of the data as #G 0xx ðf j Þ¼#G xx ðf j Þ=c ðf j Þ:8.For assessingthe statistical sig niﬁcance of a spectralpeak,the upper conﬁdence interval of the AR1noise is calculated for various signiﬁcance levels (based on w 2distribution;degrees of freedom depend on the actual spectral analysis setting;cf.Schulz and Stattegger,1997).In addition,signiﬁcance levels are calculated from percentiles of the Monte Carlo ensemble.9.Check appropriateness of the AR1model to describe x ðt i Þby testingthe equality of G rr ðf j Þand #G 0xxðf j Þusinga non-parametric runs test (Bendat and Piersol,1986).The assumptions underlyingthis procedure are:(i)The noise background recorded in a time series can indeed be approximated by an AR1process (tested in step 9),that is,the potential eﬀect of non-AR1signal components (e.g.harmonic signals)can be neglected.Although it would be possible to identify and subtract harmonic signal components prior to estimating t (see Mann and Lees,1996for evenly spaced time series),this approach may fail if there are quasi-periodic signals (e.g.narrow-band noise),which often occur in climatic time series.For most practical problems such reﬁnement is unwarranted because such signals cover only a small portion of the entire frequency range and have only a marginal eﬀect on the estimated value of t (Gilman et al.,1963).Situations in which non-AR1features doaﬀectthe estimation of t can be identiﬁed by visual inspection of the resultingred-noise spectrum and the runs test of step9.(ii)The distribution of data points alongthe time axis is not too clustered(Horne and Baliunas,1986).A computer program(REDFIT)that performs the above steps is freely available via anonymous ftp from ftp.rz.uni-kiel.de(ﬁle:/pub/sfb313/mschulz/redﬁt35.zip) or from the IAMG server.The zip-archive includes Fortran90source code,binaries for Windows95(or above),program documentation and exampleﬁles. The program oﬀers the same functionality for uni-variate spectral analysis as the SPECTRUM program (Schulz and Stattegger,1997)and uses the same format for inputﬁles.To cope better with the computational demand of the Monte–Carlo simulation,the program is command-line driven and can therefore be run in batch mode.4.Example computationsTheﬁrst test signal is a pure AR1process after Eq.(1) with t¼15yr and N¼324data points(Fig.1A).The uneven time axis is generated by treating the time interval between subsequent samplingtimes as a random variable following a gamma distribution with3degreesof freedom(which is a geologically realistic model;Schulz and Stattegger,1997).The estimated value for tis15yr(90%conﬁdence interval:10o t o20yr).Theuncorrected Lomb–Scargle spectrum of the AR1timeseries,#G xxðf jÞ,does not show the characteristic red-noise shape,instead spectral amplitudes increase slightly forf>0:09(1/yr)(Fig.1B).As expected,the same holdstrue for the mean,#G rrðf jÞ,of the N sim¼1000simu-lated red-noise spectra(Fig.1B).Compared to thetheoretical spectrum of the generated AR1process,G rrðf jÞ,(based on estimated value of t)the Lomb–Scargle Fourier transform clearly overestimates thespectral amplitudes for a large part of the spectrum(Fig.1B).Applying the bias correction(steps6and7)results in a spectral estimate#G0xxðf jÞwhich is,of course, consistent with G rrðf jÞ(Fig.1B).At the low-frequency end of the spectrum we observe that#G0xxðf jÞ>#G xxðf jÞ. This eﬀect is caused by theﬁnite length of the time series,which leads to an underestimation of the spectral amplitudes for periods exceedingthe leng th of the time series(independently of the spectral-analysis technique beingused and the spacingof the time axis).Thus,as a side eﬀect,the bias correction accounts also for this problem inherent in all spectral analysistechniques.In the second example,we investigate the glacial part of the oxygen-isotope record from the GISP2ice core from Greenland(Grootes and Stuiver,1997;Fig.2A), which reﬂects,to a large extent,air temperature above Greenland.In the initial step of the analysis we determine whether or not the spectrum of this time series is consistent with a red-noise model.Based on the periodogram of the time series(n50¼1;rectangular window;cf.Schulz and Stattegger,1997)and N sim¼1000Monte–Carlo simulations,the runs test indicates that the AR1model is indeed appropriate to character-ize this record(5%signiﬁcance level).The estimated mean value of t is310yr with90%conﬁdence interval 240o t o380yr.Next we test if any non-AR1compo-nents can be identiﬁed in the time series.For this purpose we repeat the analysis,but increase the number of WOSA segments in the spectral analysis in order to obtain a consistent spectral estimate(we refer the reader to Schulz and Stattegger,1997for details of the spectral-analysis technique).Setting n50¼4and selectinga Welch spectral window to reduce spectral leakage results in the spectrum depicted in Fig.2B.We scale the theoretical red-noise spectrum by an appropriate per-centile of the w2-probability distribution to obtain a false-alarm level,which marks the maximum spectral amplitude expected if the time series would have been generated by an AR1process.Accordingly,spectral peaks exceedingthe false-alarm level indicate non-AR1 components in a time series,and should be considered signiﬁcant.We follow Thomson(1990)and select a false-alarm level ofð121=nÞÂ100%,where n is the number of data points in each WOSA segment.For the example at hand,a false-alarm level of99.6%results.At this level the spectrum indicates the presence of a single peak at f¼1=ð1470yrÞwhich is not consistent with the red-noise model.This spectral peak is associated with the so-called Dansgaard–Oeschger oscillations,the dominant mode of millennial-scale climateﬂuctuations during the last glacial period(e.g.Grootes and Stuiver, 1997).However,care should be taken when interpreting these results because the assumption of weak stationar-ity of the time series may be violated.5.ConclusionsWe present a computer program(REDFIT)for testingwhether or not the red-noise shape,often observed in paleoclimatic time series,is consistent with the generation by aﬁrst-order autoregressive(AR1) process.In contrast to existingapproaches,REDFIT allows direct processingof unevenly spaced time series and,hence,the usual prerequisite of data interpolation is not required.Since interpolation of an unevenly spaced time series is equivalent to low-passﬁltering,reddening of an estimated spectrum will result and consequently a biased test result may be the outcome.As an aside,by correctingfor the eﬀect of correlation between Lomb–Scargle Fourier components,the program removes the bias of this Fourier transform for unevenly spaced data.A real-world example demonstrates the capability of REDFIT to detect spectral feature not consistent with an AR1origin.Although REDFIT indicates whether or not the main assumption(i.e.,adequacy of the AR1 model)is violated,the program should not be used as black-box tool without checkingthe structure of a time series prior to its analysis.AcknowledgementsThis work received support from the BMBF ‘Klimaforschungsprogramm’(MS)and DFG Research Grant MU1595/1-1(MM).ReferencesBendat,J.S.,Piersol, A.G.,1986.Random Data,2nd edn.Wiley,New York,566pp.Gilman,D.L.,Fuglister,F.J.,Mitchel Jr.,J.M.,1963.On the power spectrum of red noise.Journal of the Atmospheric Sciences20(2),182–184.Grootes,P.M.,Stuiver,M.,1997.Oxygen18/16variability in Greenland snow and ice with10@3–105-year time resolution.Journal of Geophysical Research102(C12), 26455–26470.Hasselmann,K.,1976.Stochastic climate models:Part I.Theory.Tellus28(6),473–485.Horne,J.H.,Baliunas,S.L.,1986.A prescription for period analysis of unevenly sampled time series.The Astrophysical Journal302(2),757–763.Lomb,N.R.,1976.Least-squares frequency analysis of unequally spaced data.Astrophysics and Space Science39, 447–462.Mann,M.E.,Lees,J.M.,1996.Robust estimation of back-ground noise and signal detection in climatic time series.Climatic Change33(3),409–445.Mudelsee,M.,2002.TAUEST:a computer program for estimatingpersistence in unevenly spaced weather/climate time puters a Geosciences,28(1)69–72. Percival, D.B.,Walden, A.T.,1993.Spectral Analysis for Physical Applications.Cambridge University Press,Cam-bridge,583pp.Robinson,P.M.,1977.Estimation of a time series model from unequally spaced data.Stochastic Processes and their Applications6,9–24.Scargle,J.D.,1982.Studies in astronomical time series analysis.II.Statistical aspects of spectral analysis of unevenly spaced data.The Astrophysical Journal263(2), 835–853.Scargle,J.D.,1989.Studies in astronomical time series analysis.III.Fourier transforms,autocorrelation functions,and cross-correlation functions of unevenly spaced data.The Astrophysical Journal343(2),874–887.M.Schulz,M.Mudelsee/Computers&Geosciences28(2002)421–426425Schulz,M.,Stattegger,K.,1997.SPECTRUM:Spectral analysis of unevenly spaced paleoclimatic time series.Computers a Geosciences23(9),929–945.Thomson,D.J.,1990.Time series analysis of Holocene climate data.Philosophical Transactions of the Royal Society of London.Series A330,601–616.Welch,P.D.,1967.The use of fast Fourier transform for the estimation of power spectra:a method based on time averaging over short,modiﬁed periodograms.IEEE Transactions on Audio and Electroacoustics 15(2),70–73.M.Schulz,M.Mudelsee/Computers&Geosciences28(2002)421–426 426。

“蒸发悖论”在秦岭南北地区的探讨

“蒸发悖论”在秦岭南北地区的探讨蒋冲;王飞;刘思洁;穆兴民;李锐;刘焱序【摘要】潜在蒸散量(ET0)是大气蒸发的估计值,已经广泛应用于灌溉管理和无实测蒸发资料地区的估算.分析ET0的时空变化是研究水资源对气候变化响应的基础工作,同时对于农业水资源的优化利用也具有重要意义.根据秦岭南北47个气象站1960-2011年逐日数据,利用FAO Penman-Monteith公式计算出各站的潜在蒸散量(ET0),研究了气温、降水与ET0之间的长期变化趋势关系,对导致ET0下降的主要原因进行了讨论,着重对秦岭南北地区是否存在“蒸发悖论”进行验证.结果表明:(1)秦岭南北整体气温经历了先降后升的变化过程,1993年为突变年份,1960-1993年的降温速率和1994-2011年的升温速率均表现出由南向北递减的规律,1960-2011年整体升温速率由北向南递减.(2)1979年和1993年是ET0变化的转折点,以1979和1993为界ET0经历了“升—降—降”的变化阶段.1960-1979年仅汉水流域和巴巫谷地存在“蒸发悖论”现象,1980-1993、1994-2011和1960-2011年3个时段区域整体和各子区均发现了“蒸发悖论”现象.秋季后18a 和52a整体以及冬季前34a和52a整体均存在“蒸发悖论”现象,冬季最为明显.(3)近52年整体降水表现出不显著的下降趋势,相较于年尺度,夏季降水与ET0逆向变化趋势更为明显.(4)年尺度上,太阳辐射(日照时数)下降引起的潜热通量减少是造成ET0下降即“蒸发悖论”现象的主要原因.季节尺度,春季ET0下降的主导因素为风速,其它季节均为太阳辐射(日照时数).【期刊名称】《生态学报》【年(卷),期】2013(033)003【总页数】12页(P844-855)【关键词】秦岭南北;潜在蒸散量;蒸发悖论;气温;降水【作者】蒋冲;王飞;刘思洁;穆兴民;李锐;刘焱序【作者单位】西北农林科技大学资源环境学院,杨凌712100;西北农林科技大学资源环境学院,杨凌712100;中国科学院水利部水土保持研究所,杨凌712100;北京大学遥感与地理信息系统研究所,北京100871;西北农林科技大学资源环境学院,杨凌712100;中国科学院水利部水土保持研究所,杨凌712100;西北农林科技大学资源环境学院,杨凌712100;中国科学院水利部水土保持研究所,杨凌712100;陕西师范大学旅游与环境学院,西安710062【正文语种】中文潜在蒸散量(ET0)是指在一定气象条件下水分供应不受限制时，某一固定下垫面土壤蒸发量和植物蒸腾量的总和，它是实际蒸散量的理论上限，也是计算实际蒸散量的基础，又被称之为参考作物蒸散量［1-4］。

ICESat-2 ATL03数据预处理及校正方法

〇引言
近些年来，星载激光雷达凭借其全天时、高空间分辨率以及覆盖范围广的特点，成为多个领域研究的有效工具。2003年 1 月，美国 N A S A 发射激光测高仪系统 Geoscience Laser Altimeter System(GLAS)，搭载于以冰、云和陆地高程卫星（Ice, Cloud and land Elevation Satellite, 丨CESat) 平台。该系统被用于测量冰被地形，同时也监测云层和大气的特性，提供云层高度和厚度，有助于提高短期天气及气候预报能力。 2003年到 2009年在轨期间，收集的测高数据促使科学家对格陵兰岛和南极冰盖的总体质量变化有了更为精确的估计 [1]。为了延续 ICESat的后续观测任务，N A S A 于 2018年 9 月 1 5 日发射了地形激光测高仪系统（ Advanced Topographic Laser Altimeter System, ATLAS),搭载于以冰、云和陆地高程卫星 2 号（The Ice, Cloud, and Land Elevation Satellite-2, ICESat-2)平台。ATLAS/ICESat-2将帮助科学家研究地球冰冻圈发生变化的原因和幅度；测量地球温带和热带区域的高程信息，并对全球森林植被信息进行评估 [2]。不同于 GLAS/ICESat 全波形记录，ATLAS/ICESat-2 采用更为灵敏的单光子探测器，以 10 kH z的重复频率发射 532 nm激光脉冲，通过获取高密度光子计数回波数据，实现高精度全球观测 [\

城市热岛气候

Cooling the cities –A review of reﬂective and green roof mitigation technologies to ﬁght heat island and improvecomfort in urban environmentsM.SantamourisGroup Building Environmental Research,Physics Department,University of Athens,Athens,GreeceCommunicated by:Associate Editor Yogi GoswamiAbstractThe temperature of cities continues to increase because of the heat island phenomenon and the undeniable climatic change.The observed high ambient temperatures intensify the energy problem of cities,deteriorates comfort conditions,put in danger the vulnerable population and amplify the pollution problems.To counterbalance the phenomenon,important mitigation technologies have been devel-oped and proposed.Among them,technologies aiming to increase the albedo of cities and the use of vegetative –green roofs appear to be very promising,presenting a relatively high heat island mitigation potential.This paper aims to present the state of the art on both the above technologies,when applied in the city scale.Tenths of published studies have been analysed.Most of the available data are based on simulation studies using mesoscale modeling techniques while important data are available from the existing experimental studies.When a global increase of the city’s albedo is considered,the expected mean decrease of the average ambient temperature is close to 0.3K per 0.1rise of the albedo,while the corresponding average decrease of the peak ambient temperature is close to 0.9K.When only cool roofs are considered,the analysis of the existing data shows that the expected depression rate of the average urban ambient tem-perature varies between 0.1and 0.33K per 0.1increase of the roofs albedo with a mean value close to 0.2K.As it concerns green roofs,existing simulation studies show that when applied on a city scale,they may reduce the average ambient temperature between 0.3and 3K.Detailed analysis of many studies reporting a comparison of the mitigation potential of both technologies has permitted the deﬁ-nition of the limits,the boundaries and the conditions under which the considered technologies reach their better performance,in a syn-thetic way.Ó2012Elsevier Ltd.All rights reserved.Keywords:Heat island;Cool roofs;Green roofs;Mitigation potential1.IntroductionHeat island is the most documented phenomenon of cli-mate change.The phenomenon is known for almost a cen-tury and is related to higher urban temperatures compared to the adjacent suburban and rural areas (Santamouris,2001).Higher urban temperatures are due to the positive thermal balance of urban areas caused by the important release of anthropogenic heat,the excess storage of solarradiation by the city structures,the lack of green spaces and cool sinks,the non-circulation of air in urban canyons and the reduced ability of the emitted infrared radiation to escape in the atmosphere (Oke et al.,1991).During the recent period,intensive research has been carried out on the topic,the impact and the signiﬁcance,as well as the qualitative and quantitative characteristics of the phenomenon are much better documented (Santa-mouris,2007;Mirzaei and Haghighat,2010;Founda,2011;Stewart,2011;Mihalakakou et al.,2002;Mihalaka-kou et al.,2004;Livada et al.,2002;Mohsin and Gough,2012;Klok et al.,in press;Papanastasiou and Kittas,0038-092X/$-see front matter Ó2012Elsevier Ltd.All rights reserved./10.1016/j.solener.2012.07.003E-mail address:msantam@phys.uoa.gr/locate/solenerSolar Energy xxx (2012)xxx–xxx2012).Higher urban temperatures increase the energy con-sumption for cooling and raise the peak electricity demand (Hassid et al.,2000;Cartalis et al.,2001;Santamouris et al.,2001;Kolokotroni et al.,2012;Hirano and Fujita,in press;Akbari and Konopacki,2004;Akbari et al.,1992).As men-tioned by Santamouris et al.(2001),heat island in the city of Athens,Greece,doubles the cooling load of buildings and almost triples their peak electricity demand,while decreasing the Coeﬃcient of Performance (COP),of mechanical cooling systems up to 25%.According to Akbari et al.(1992),for US cities with population larger than 100,000the peak electricity load will increase 1.5–2%for every 1°F increase in temperature.The cooling energy increase is accompanied by intensiﬁcation of pollu-tion patterns in cities and increase of ozone concentrations (Stathopoulou et al.,2008;Sarrat et al.,2006;Taha,2008a ),while the ecological footprint of the cities is increased (Santamouris et al.,2007a ),the outdoor thermal comfort conditions deteriorate (Pantavou et al.,2011),the thermal stress in low income dwellings is increased,the indoor thermal comfort levels are seriously decreased and health problems are intensiﬁed (Sakka et al.,2012;Luber and McGeehin,2008).Research carried out recently has permitted the develop-ment of technological measures to counterbalance the impact of heat island (Rosenfeld et al.,1995;Akbari et al.,2001;Adnot et al.,2007;Kuttler,2011;Gaitani et al.,2007).Mitigation techniques aim to balance the ther-mal budget of cities by increasing thermal losses and decreasing the corresponding gains.Among the more important of the proposed techniques are those targeting to increase the albedo of the urban environment,to expand the green spaces in cities and to use the natural heat sinks in order to dissipate the excess heat (Akbari et al.,2005a;Julia et al.,2009;Mihalakakou et al.,1994).Recent real scale applications involving the use of the above mitigation techniques have resulted into very important climatic ben-eﬁts and a serious reduction of the heat island strength (Gaitani et al.,2011;Fintikakis et al.,2011;Santamouris et al.,in press,2012).Roofs present a very high fraction of the exposed urban area.Estimations given in Akbari and Rose (2008),for four American cities,show that the roof area fraction varies from 20%to 25%for less or more dense cities.Based on these ﬁndings and considering that urban areas occupy almost 1%of all land,it is estimated that the total roof area of the urban world is close to 3.8Â1011m 2(Akbari et al.,2009a ).However,according to Jacobson et al.(2007),the above assumption regarding the size of urban areas is about 2.26times higher than it is estimated using an anal-ysis of satellite data.Given that the available free ground area in the urban environment is quite limited and of very high economic value,it is relatively diﬃcult to implement large scale mit-igation technologies on the ground surface of cities.At the same time,urbanization decreases the proportion of spaces dedicated to plants and trees or other mitigation infrastruc-tures because of new building developments (Mathieu et al.,2007;Smith,2010).On the contrary,roofs provide an excellent space to apply mitigation techniques,given that the relevant cost is limited,while the corresponding techniques are associated to important energy savings for the buildings.Two are the more important mitigation technologies associated to roofs:(a)Those aiming to increase the albedo of the roofs,known as cool or reﬂective roofs (Zinzi,2010;Akbari and Levinson,2008;Synnefa and Santamouris,in press )those that propose roofs partially or completed cov-ered with vegetation,known as green roofs or living roofs (Theodosiou,2009;Santamouris et al.,2007b;Sfakianaki et al.,2009).Both technologies can lower the surface tem-peratures of roofs and thus to decrease the corresponding sensible heat ﬂux to the atmosphere.Cool or reﬂective roofs are typically white and present a high albedo.Products used in cool roofs are single ply or liquid applied (Mac Cracken,2009).Typical liquid applied products involve white paints,elastomeric,polyurethane or acrylic coatings.Examples of white single ply products involve EPDM (Ethylene-Propylenediene-Tetrolymer Membrane)PVC (Polyvinyl Chloride),CPE (Chlorinated Polyethylene),CPSE (Chlorosulfonated Polyethylene),and TPO (Thermoplastic Polyoleﬁn)materials (Mac Crac-ken,2009).A review of the recent developments on the ﬁeld of liquid applied materials used in reﬂective roofs is given in Santamouris et al.(2011).The ﬁrst generation of materials used in cool roofs consisted of natural materials quite eas-ily found in the nature characterized by a relatively high albedo,rarely higher than 0.75(Doulos et al.,2004;Bretz et al.,1992;Reagan and Acklam,1979),while the second generation was based on the development of artiﬁcial white materials designed to present very high albedo values close or higher than 0.85(Synnefa et al.,2006;Kolokotsa et al.,2012;Santamouris et al.,2008).In a later,third phase of development,colored high reﬂective materials have been developed.The overall idea was to develop colored materi-als presenting a high reﬂectivity value in the infrared spec-trum (Levinson et al.,2005a,2005b;Synnefa et al.,2007a ).The speciﬁc materials were characterized by a much higher global reﬂectivity than the conventional ones of the same color and were associated to important energy savings when used in building roofs or urban infrastructures (Synnefa et al.,2007b,2011;Santamouris et al.,2007c ).Quite recently,fourth generation reﬂective materials based on nanotechnological additives like thermochromic paints and tiles (Ma et al.,2001,2002;Karlessi et al.,2009),or PCM doped cool materials (Karlessi et al.,2011;Zhang et al.,2007;Pasupathy et al.,2008;Cabeza et al.,2007)have been developed and likely to be used for future cool roof applications.Many studies have been performed in order to identify the cooling potential and the possible improvements of indoor thermal comfort caused by cool roofs (Synnefa et al.,in press;Akbari et al.,2005a,2005b;Levinson2M.Santamouris /Solar Energy xxx (2012)xxx–xxxet al.,2005c;Tengfang et al.,2012;Kolokotsa et al.,in press;Romeo and Zinzi,in press;Kolokotroni et al.,in press;Bozonnet et al.,2011;Boixo et al.,2012;Takebay-ashi et al.,in press ).Energy beneﬁts vary mainly as a func-tion of the climatic conditions and the characteristics of the building.Typically,peak summer indoor temperatures may decrease up to 2°C in moderately insulated buildings while cooling loads reductions may range between 10%and 40%.At the same time,the heating penalty may range between 5%and 10%as a function of the local climate and building characteristics.In parallel,important simulation studies have been car-ried out to identify the heat island mitigation potential of cool roofs (Savio et al.,2006;Synnefa et al.,2008;Menon et al.,2010;Jacobson and Ten Hoeve,2012).Low temper-atures at the roof level,decrease the sensible heat ﬂux to the atmosphere and add to the mitigation of the urban heat island.Most of the studies have been carried out in USA,using mesoscale simulation models.The speciﬁc results of the above studies are discussed in the following chapters.Increase of the green spaces in cities,contribute to decrease the urban surface and ambient temperatures and mitigate heat island eﬀect.Studies reported by Gill et al.(2007),show that an increase by 10%of the urban green in Manchester,UK,could amortize the predicted increase by 4K,of the ambient temperature over the next 80years.Green or living roofs are partially or fully covered by veg-etation and a growing medium over a waterprooﬁng mem-brane.There are two main types of green roofs:Extensive roofs which are light and are covered by a thin layer of veg-etation and intensive roofs which are heavier and can sup-port small trees and shrubs.Green roofs present a variety of advantages like storm water runoﬀmanagement,increased roof materials durability,decreased energy con-sumption,possible better air quality and noise reduction,oﬀer space for urban wildlife and increased mitigation of urban heat island (Parizotto and Lamberts,2011;Mentens et al.,2006;Teemusk and Mander,2009;Rowe,2010;Ren-terghem and Botteldooren,2011;Brenneisen,2006;Pataki et al.,2011).Several experimental and theoretical studies have been performed to identify the energy conservation potential of green roofs (Kumar and Kaushik,2005;Alexandri and Jones,2007;Wong et al.,2003a;Theodosiou,2003;Eumorfopoulou and Aravantinos,1998;Jaﬀal et al.,2012;Spala et al.,2008;Takakura et al.,2000;Castleton et al.,2010).The speciﬁc energy beneﬁts depend on the local climate,the green roof design and more importantly on the speciﬁc building characteristics.Given that in green roofs heat transfer beneﬁts are mainly provided through latent heat processes,the performance of the system is higher in dry climates.In parallel,the thickness and the thermal characteristics of the vegetative roof largely deﬁne its U value and the corresponding transfer of heat to the building,while the type and characteristics of the plants (LAI),deﬁne the shading levels and the transfer of radia-tion through the layers.Finally,watering is important asit determines the latent heat release and regulates the ther-mal balance of the roof.However,the building characteris-tics also deﬁne the possible contribution of green roofs.In non-insulated buildings the impact of green roofs is much higher than in insulated ones.It is evident that the better the insulation of the roof,the lower the contribution of the green roof.In parallel,the characteristics of the energy load of the building deﬁne the speciﬁc contribution of the roof system.In buildings presenting a high part of their energy load because of the ventilation gains or losses,inter-nal or solar gains,green roofs have a limited contribution.On the contrary,in buildings where the energy load is due to the heat transfer though the opaque parts of the enve-lope,vegetative roofs may contribute signiﬁcantly to reduce heating and cooling loads.Existing studies per-formed for various types of buildings,green roof character-istics and climatic zones,show that expected reduction of the annual energy load may vary between 1%and 40%in extreme cases.In reality,in well insulated modern buildings the energy contribution of green roofs is quite modest.Although the possible energy contribution of green and reﬂective roofs is a quite well investigated area,the avail-able information on the possible mitigation potential of both technologies is relatively limited (Cameron et al.,2012).Most of the existing studies are based on mesoscale simulation modeling and the given results depend com-pletely on the speciﬁc regional characteristics and the assumptions made,while very few experimental studies are available.Some of the reported studies attempt to com-pare the climatic potential of the two roof technologies either using simulation or experimental techniques.How-ever,most of the studies are building and climatic speciﬁc and it is quite diﬃcult to extract general conclusions,although the provided results are very useful.The objective of the present paper is to review,in a crit-ical way,the available scientiﬁc information on the mitiga-tion potential of reﬂective and green roofs.Also,to combine and analyze the existing theoretical and experi-mental data,compare and homogenize the results and if possible,to provide general conclusions and suggestions.2.Increasing the albedo in the city:the role of cool roofs It is well known and documented that large scale change of albedo has a serious impact on the local peak ambient temperature.Multiyear observations reported in Campra et al.(2008),show an important temperature reduction (À0.3K/decade),because of the massive construction of high albedo greenhouses through the Almeria area in Spain.Several simulation studies have been carried out to investigate the impact of various albedo related mitigation techniques on the possible reduction of ambient tempera-ture.Most of the works evaluate the impact of a general increase of the local albedo considering in a combined way cool roofs,cool pavements roadways and parking lots,while few studies consider and evaluate just the mitigation impact of reﬂective roofs.In parallel,many of the mitiga-M.Santamouris /Solar Energy xxx (2012)xxx–xxx3tion studies investigate the combined impact of several and diﬀerent mitigation technologies,i.e.increase of the local albedo and vegetative cover without splitting the results about the speciﬁc contribution of each technique (Taha et al.,1999).The present paper considers only studies aim-ing either to calculate the inﬂuence of reﬂective roofs or to evaluate the combined impact of a general albedo change.2.1.Mitigation potential of cool roofsEvaluation of the heat island mitigation potential of cool roofs in a city is a quite new scientiﬁc subject.Four speciﬁc studies are currently available focusing on the potential decrease of the ambient temperature because of the increase of roof albedo (Savio et al.,2006;Synnefa et al.,2008;Menon et al.,2010;Jacobson and Ten Hoeve,2012).Two of the studies (Savio et al.,2006;Synnefa et al.,2008),examine the local impact of cool roofs in New York,US and Athens Greece,while the studies reported by Menon et al.(2010)and Jacobson and Ten Hoeve (2012),investigate the climatic impact of reﬂective roof on a plan-etary scale.In Savio et al.(2006),the impact of cool roofs on the potential ambient temperature decrease at 2m height above ground,has been evaluated for New York city,US,using simulations performed with the Penn State/NCAR MM5regional climate model (Grell et al.,1994).Runs were performed for the period of three heat waves during the summer of 2002.An average solar reﬂectivity equal to 0.5was used.It was calculated that the daily aver-age temperature decrease in the various parts of the city ranges between 0.18K and 0.36K.In parallel,the average 3PM reduction of peak ambient temperature ranged between 0.31K and 0.62K as a function of the character-istics of the considered areas.The climatic impact of cool roofs has been simulated for the city of Athens,Greece (Synnefa et al.,2008).The study has been performed using the MM5climate model (Grell et al.,1994),for the 15th of August 2005.Two modiﬁed albedo scenarios were considered:a moderate one,where the albedo of the roofs was increased from 0.18to 0.63,and an extreme one where the ﬁnal albedo of the building rooftops was considered 0.85.It was calculated that for the moderate increase of the albedo,the ambient temperature depression at 2m height at 12:00LST varied between 0.5and 1.5K.For the extreme case of albedo increase,the ambient temperature reduction varied between 1K and 2.2K.The temperature depression found to start at 9:00LST and stopped at around 20:00LST.Simulations were carried out using the CSCRC model (Yoshida et al.,2000;Hong et al.,2004),to calculate the impact of cool roofs in medium and high rise neighbor-hoods with medium and high rise buildings (average height 28.6and 67.4m respectively).The albedo of the roofs was changed from 0.2to 0.5.It was calculated that the decrease of the ambient temperature at street level was 0.1K and 0.12K for the areas of high and medium height rise build-ings.It is evident that the impact of reﬂective roofs on the ambient temperature at street level is seriously reduced when the height of buildings where cool roofs are applied is great.The characteristics of the NY and Athens studies are summarized in Table 1.The expected rate of temperature depression per 0.1increase of the albedo of roofs ranges between 0.1and 0.19K for New York and 0.11–0.33K for Athens.The average temperature depression for both studies is close to 0.2K per 0.1increase of roof albedo,while the corresponding minimum and maximum tempera-ture depression are 0.02K and 0.41K,respectively.Simulations have been carried out using the GATOR–GCMOM model (Jacobson et al.,2007),to calculate the climatic impact of cool roofs on a planetary scale.The solar albedo of all roofs was increased from 0.12to 0.65.This corresponds to an overall urban albedo increase by 0.147.It is reported that a worldwide conversion to cool roofs will contribute to decrease populated weighted tem-peratures by 0.02K but to increase the overall earth tem-perature by 0.07K.The results of this study suggesting an overall global warming because of the possible conver-sion of roofs to white have been discussed in Oleson et al.(2008a,b),raising several concerns about the assump-tions and the results of this publication.Another similar simulation study aiming to evaluate the worldwide heat island mitigation potential of reﬂective roofs is reported in Menon et al.(2010).The study was car-ried out using the urban canyon model CLMU coupled with other models (Oleson et al.,2008a,b ).It was consid-ered that the albedo of roofs increases up to 0.9.It is calcu-lated that the daily maximum urban temperatures decrease by 0.6K while the daily minimum ambient temperature decreased by 0.3K.This corresponds to a decrease in the urban diurnal temperature range of 0.3K.2.2.Increasing the albedo of cities –mitigation potential One of the ﬁrst evaluations of the climatic impact of reﬂective surfaces was published in Sailor (1995).The authors simulated the impact of albedo modiﬁcation in Los Angeles,USA.They have used the Colorado State University Mesoscale Model (Mahrer and Pielke,1977).It was assumed that the albedo of downtown surfaces was increased by 0.14while the corresponding increase for the entire basin was close to 0.08.It was calculated that the considered albedo change reduced ambient tempera-tures by at least 0.5K over the majority of the area,while the peak urban temperatures reduction was 1.4K near downtown LA.A second study carried out for the same city,Los Ange-les,USA,is described in Rosenfeld et al.(1995).The Col-orado State University Mesoscale Model (Mahrer and Pielke,1977),was also used.The authors assumed an increase of the average albedo by 0.13,and in particular from 0.13to 0.26for an area of 100,000km 2,in which over 20%of the land was covered by artiﬁcial surfaces.Flat and4M.Santamouris /Solar Energy xxx (2012)xxx–xxxsloped roofs accounted for about 30%of the area.For the ﬂat roofs it was considered that the albedo changes from 0.25to 0.75while for the sloped roofs from 0.25to 0.6.It was calculated that the peak impact of the albedo change occurs in the early afternoon and the potential cooling exceeds 3K at 3pm.Simulations carried out under diﬀer-ent boundary conditions indicate that the expected peak summertime temperature reductions were between 2and 4K.In another study of the same group (Rosenfeld et al.,1998),the cooling potential of a possible albedo change in the city of Los Angeles was evaluated.It was considered that in almost 1250Km 2of roof surface the albedo was increased by 0.35and in particular from 0.15to 0.5.It was also considered that in 1250km 2of pavements,the albedo was increased by 0.25,from 0.05to 0.30.It was cal-culated that the achieved temperature reduction during the peak period was close to 1.5K.Important energy and pol-lution beneﬁts were also estimated.Research carried out by Millstein and Menon (2011),has investigated the impact of albedo change on the tem-perature status of various American lstein and Menon used the Weather Research and Forecasting model (Skamarock et al.,2008).Cool roofs and cool pavement technologies were considered.It was assumed that roofs and pavements represent 25%and 35%of the urban area respectively,while the considered albedo increase for roofs and pavements were +0.25and +0.15,respectively.The average albedo increase for the whole considered area ranges between 0.0and 0.115.It was calculated that cool roofs and cool pavements contribute to decrease average afternoon summertime temperatures by 0.11–0.53K.For some of the studied urban locations not statistically signif-icant temperature reduction was found.Simulations are carried out using the mesoscale model MM5(Grell et al.,1994)to estimate the impact of a possi-ble increase of albedo in Philadelphia,US (Sailor et al.,2002).It was reported that an increase of the local albedo by 0.1is responsible for an average daytime ambient tem-perature depression of about 0.3–0.5K.A simulation study aiming to investigate the impact of various mitigation technologies for the city of Atlanta,US is described in Zhou and Shepherd (2010).The Weather Research and Forecasting WRF-NOAH land surface model was used (Skamarock et al.,2008).Considering an increase of the local albedo by 100%(0.15–0.30),it was cal-culated that the ambient temperature depression was almost negligible.On the contrary,when the local albedo rises to 0.45,which corresponds to an increase by 200%,the peak ambient temperature in the city decreases by 2.5K.Simulations have been carried out to investigate the impact of a moderate increase of local albedo in Houston,US,using the mesoscale model MM5(Taha,2008b ).Roof albedo increased from 0.1to 0.3,wall albedo from 0.25to 0.3and paved surface albedo from 0.08to 0.2.It is found that the change of the albedo may decrease peak tempera-T a b l e 1C h a r a c t e r i s t i c s o f t h e N e w Y o r k a n d A t h e n s s t u d i e s o n t h e m i t i g a t i o n p o t e n t i a l o f c o o l r o o f s .R e f e r e n c e w o r kA r e a o f a p p l i c a t i o nS i m u l a t i o n t o o lF i n a l a l b e d oR e s u l t s –m i t i g a t i o n p o t e n t i a lP e a k t e m p e r a t u r e d e p r e s s i o n p e r 0.1a l b e d o i n c r e a s eS a v i o e t a l .(2006)N e w Y o r k C i t y U SP e n n S t a t e /N C A R M M 50.5T h e d a i l y a v e r a g e t e m p e r a t u r e r e d u c t i o n r a n g e s b e t w e e n 0.18K a n d 0.36K .T h e a v e r a g e 3P M r e d u c t i o n o f p e a k a m b i e n t t e m p e r a t u r e r a n g e d b e t w e e n 0.31K a n d 0.62K 0.10–0.19KS y n n e f a e t a l .(2008)A t h e n s ,G r e e c e M M 5S c e n a r i o a :0.63S c e n a r i o a :T h e a m b i e n t t e m p e r a t u r e d e p r e s s i o n a t 2m h e i g h t a t 12:00L S T v a r i e d b e t w e e n 0.5a n d 1.5K S c e n a r i o a :0.11–0.33KS c e n a r i o b :0.85S c e n a r i o b :T h e a m b i e n t t e m p e r a t u r e d e c r e a s e v a r i e d b e t w e e n 1K a n d 2.2KS c e n a r i o b :0.15–0.33KM.Santamouris /Solar Energy xxx (2012)xxx–xxx5tures up to 3.5K,while in one occasion ambient tempera-tures increase by up to 1.5K.The average temperature reduction varies from site to site and ranges between 0.3and 0.4K in residential urban areas.Cooling occurs during daytime and heating during the night.A similar simulation study aiming to evaluate the cli-matic impact of various mitigation techniques is described in Lynn et al.(2009).The study refers to the Metropolitan New York Area,US,and is carried out using mainly the MM5mesoscale simulation model.As it concerns the eval-uation of the scenario related to the increase of the urban albedo,it was considered that the albedo of the impervious surfaces changes from 0.15to 0.5.It is found that peak ambient temperature at 2m height for 12:00LST 14th August 2001,decreases between 0.25and 0.5K,while a daily average temperature decrease is close to 0.2–0.3K.Simulations studies have been performed to investigate the impact of various mitigation techniques for several cit-ies of California (Taha,2008c ).The study has been carried out using the PSU/NCAR MM5simulation tool (Dudhia,1993).Two scenarios related to the albedo change were considered and evaluated.The ﬁrst scenario considered a moderate increase of the urban reﬂectivity while the second one a stronger increase.In particular,for the existing situ-ation,albedos ranged from 0.117to 0.152,for the moder-ate albedo change scenario from 0.18to 0.252,while for the last scenario from 0.199to 0.374.Data were calculated for Los Angeles,Pomona and San Fernando Valley.It was found that a moderate change of the urban albedo may decrease peak ambient temperatures by up to 1K,whileconcerning the possible albedo change and the correspond-ing decrease of the average ambient temperature are plot-ted.All data given in Millstein and Menon (2011)are also included.As shown,data correlate quite well in a lin-ear regression given below indicating that an albedo change of 0.1in urban areas decreases the average ambient temper-ature by 0.3K:ATD ¼a ALBINð1Þwhere a =3.11,ATD is the average temperature decrease and ALBIN the albedo increase (0<ALBIN <1).The R 2of the regression is equal to 0.85.The calculated relation between the possible albedo change and the average ambient temperature decrease is slightly higher to the one given in Synnefa et al.(2008)for Athens,Greece (2.4<a <2.85).This is logical as the Athens study considered only reﬂective roofs and not a general albedo change.In parallel,it is slightly lower than the one calculated in Menon et al.(2010)between the albedo change and the average urban surface temperature.The calculated reduction of the peak ambient tempera-ture ranges between 1and 3.5K.The estimated decrease of the peak ambient temperature per 0.1of albedo change varies between 0.57and 2.3K.Fig.2plots the existing data concerning the possible albedo change and the correspond-ing decrease of the peak ambient temperature.A linear relation between the two parameters has been calculated,although the correlation coeﬃcient is not quite high because of the important scattering of the data.It is found that for an albedo change of 0.1in urban areas the peak between the possible albedo change and the corresponding decrease of the average ambient temperature M.Santamouris /Solar Energy xxx (2012)xxx–xxx 7。

欧洲常绿阔叶植物Bioclimatic limits and range shifts of cold-hardy evergreen broad-leaved

Phytocoenologia37(3Ð4)523Ð539BerlinÐStuttgart,December21,2007 Bioclimatic limits and range shifts of cold-hardy evergreen broad-leaved species at their northern distributional limit in Europeby Silje Berger1,Gunnar Söhlke1,Gian-Reto Walther1,2and Richard Pott1,Hannover1and Bayreuth2with4figures and1tableAbstract.The few native evergreen broad-leaved species occurring in central Europe have attracted the interest of generations of scientists;thus,the factors limiting their northern distribution have been well studied.For investigation of climate change-driven range shifts,these climate-sensitive species are particularly well suited.We here analyse recent range shifts of some of the cold-hardiest evergreen broad-leaved species,including both native and introduced species in Europe.Based on updated field data and outputs from bioclimatic models,we show that the milder winter conditions of the last few decades are consistent with the northward expan-sion of potential ranges and an increase in the number of evergreen broad-leaved species. At the landscape scale,these species indicate a considerable change in the composition and structure of temperate deciduous forests in various parts of Europe.Keywords:range shifts,Tertiary flora,climate change,global warming,deciduous forests, forest structure.IntroductionThe vegetation of central Europe is relatively poor in native evergreen broad-leaved species compared to temperate regions of other continents. However,these few species have attracted the interest of generations of scientists,and thus,the factors limiting their northern distribution have been well studied in the past.Climate,especially temperature and the length of the growing season,has been pointed out as an important factor deter-mining establishment and survival of evergreen broad-leaved species at their northern range margins,in Europe and worldwide(Box1981,Walter& Breckle1999,Woodward et al.2004,Pott2005).The temperature limits of many evergreen broad-leaved species have been investigated in detail with physiological methods(rcher1954,1970,2000,Sakai1982, Sakai&Larcher1987),as well as by biogeographical comparison(e.g. Iversen1944,Jäger1975).The fact that even the most cold-hardy evergreen broad-leaved species are sensitive to low winter temperatures makes them suitable as climateDOI:10.1127/0340-269X/2007/0037-05230340-269X/07/0037-0523$4.25”2007Gebrüder Borntraeger,D-14129Berlin·D-70176Stuttgart524S.Berger et al. indicators.This fact has been applied to reconstruct past climatic fluctua-tions from geological records(Iversen1944).Recent climate change has left visible‘fingerprints’within different ecosystems all over the world (Walther et al.2001,Parmesan&Yohe2003,Root et al.2003),and it is likely that the warming trend will continue or even increase(IPCC2001). Climatic indicators serving the reconstruction of past climatic conditions may also be used as indicators of recent and near-future climate change. Climate induced range shifts of individual evergreen broad-leaved species have been detected in different parts of Europe(Dierschke2005,Dobber-tin et al.2005,Walther et al.2005).At the regional scale,an assemblage of evergreen broad-leaved species favoured by climate change has also been investigated in detail(e.g.Walther2000).Whereas the aforementioned studies focus on single species only or on a particular habitat,we here analyse how the cold-hardiest evergreen broad-leaved species,a group in-cluding both natives and exotics,shift their northern range margins in con-cert with recent global warming.The potential ranges implied by species’specific bioclimatic limits are compared with updated data on the realised distributions of these species across Europe.Because the evergreen broad-leaved plant type has been scarcely repre-sented in central European plant communities,changes in the diversity and distribution of evergreen broad-leaved species not only lead to changes in the physiognomy of existing forest communities,but may in the long run also form new assemblages in different parts of Europe.MethodsRecent range extensions at the northern distribution boundaries of selected cold-hardy evergreen broad-leaved species were detected by comparing his-torical distribution maps and records with present distribution data.His-torical data were compiled from Adamovic´(1909),Holmboe(1913),Loe-sener(1919),Enquist(1924),Schröter(1936),Schmucker(1942), Iversen(1944),Browicz(1960),Fægri(1960),Schmid(1956),Walter& Straka(1970),Horva´t et al.(1974),Jäger(1975),Meusel et al.(1978) and others(for details see Berger(2003),Söhlke(2006)and Walther et al.(2007)).The data on the current distribution were compiled from nu-merous national and regional vegetation databases and surveys,email sur-veys interviewing experts on plant distribution(cf.acknowledgments),re-cent literature records and our own field observations.Many of the new literature records were verified in the field,especially occurrences forming new distribution boundaries,in order to confirm the correct identification of the species and to assure that the species is still present in the given locations.The resulting updated distribution maps were superposed on potential distribution areas of the species based on various bioclimatic parameters. These parameters were derived from(I)biogeographical literature addressing limiting parameters of the re-spective species,525 Bioclimatic limits and range shifts of cold-hardy evergreen broad-leaved species(II)climatic parameters of the native range of exotic species,and(III)results of bioclimatic modelling,emphasising the functional importance of the relevant particular parameters for the species-specific biological traits(see discussion section).Distribution maps were compiled,analysed and designed using ESRI ArcView3.3and ArcPress.Selected native and introduced evergreen broad-leaved speciesIlex aquifolium is the northernmost evergreen broad-leaved small tree or shrub in Europe,with severe winter frosts limiting its northern distribu-tion.This can be expressed through the mean temperature of the coldest month,which must be aboveÐ0.5∞C(Iversen1944).The potential distri-bution area of Ilex aquifolium was modelled with the bioclimatic model STASH(Sykes et al.1996)for the period1981Ð2000(Walther et al.2005).The potential distribution of Laurus nobilis was modelled by Sven-ning&Skov(2004),based on three key bioclimatic variables,namely growing degree days(GDD),mean temperature of the coldest month and water balance,using a‘bioclimatic envelope’approach.Prunus laurocerasus is native to the Balkans and the coast of the Black Sea and Caspian Sea,and it is frequently cultivated in central Europe as an ornamental shrub.Winter temperature and length of the growing season are important climatic parameters limiting the distribution of Prunus lauro-cerasus.The limiting January mean temperature based on the native range was estimated toÐ1.2∞C.The length of the growing season was expressed through number of months per year with mean temperature above5∞. Hence,the potential range of Prunus laurocerasus is additionally limited by a growing season length of at least8months,and by annual precipitation which must exceed550mm/year.Bioclimatic limits of the palm Trachycarpus fortunei were derived from its native range in China(Walther et al.2007),suggesting that mean tem-perature of the coldest month must exceed+2.2∞C,annual GDD must exceed2300and drought index(Sykes et al.1996,see also Prentice et al. 1992)does not exceed0.26.The potential distributions of Prunus laurocerasus and Trachycarpus for-tunei are based on climate data from the ALARM dataset(Mitchell et al. 2004,Reginster et al.2005).The reference period1991Ð2000was regarded as representative of current climatic conditions.The1990s were the war-mest decade since the start of the measurements(IPCC2001)and,hence of special importance in the context of analyses of global warming-related data.In the same period,most of the new occurrences of all the regarded species outside their former ranges were recorded.The potential ranges of all four of these species were superposed,revealing overlapping areas with especially suitable climatic conditions for evergreen broad-leaved species.In order to assess how range shifts of a set of evergreen broad-leaved species affect the composition and structure of European deciduous forest526S.Berger et al. ecosystems we compared the evergreen broad-leaved species occurring in southern Switzerland and in Great Britain,both areas where a pronounced increase of evergreen broad-leaved species has taken place in the last few decades.ResultsThe current and potential distributions of the selected evergreen broad-leaved species are shown in Fig.1.The realised historical range(Fig.1,top left:hatched area)of Ilex aquifolium is shown according to Walter& Straka(1970)and Meusel et al.(1978).The new occurrences outside the historical range indicate a shift in the distribution of Ilex aquifolium to-wards the north in Norway and northeast in Germany,Denmark(cf.also Ban˜uelos et al.2004)and southern Sweden,where Ilex aquifolium ex-panded into new areas along the southern Swedish coast.Most new occur-rences(Fig.1,top left:black dots)were first recorded in the same time span(1981Ð2000)as the model refers to(Walther et al.2005).The new occurrences on the mainland and larger islands are all within the modelled new potential range.The realised distribution of Prunus laurocerasus is based on Jäger(1975) and BSBI(2006)(Fig.1,top right:hatched area);further new occurrences (Fig.1,top right:dots)are compiled from various sources and our own observations(Söhlke2006).These new records are found within large parts of the potential distribution area.The records forming the new north-eastern boundary of the introduced range in northern Germany,with single occurrences as far north as Rügen and as far east as Berlin,all originate from the last two decades(Söhlke2006).The number of records from Great Britain also increased substantially in the last decades(see Perring& Walters1962and following editions,BSBI2006).The realised distribution of Laurus nobilis is based on Meusel et al. (1978),Brullo et al.(2001),BSBI(2006)and Wohlgemuth et al.(2006); it overlaps with the potential distribution(Svenning&Skov2004)in large areas(Fig.1,bottom left).The historical range of Laurus nobilis encom-passed the Mediterranean,whereas at present,the species occurs as far north as in Britain and in southern Switzerland.However,to our knowl-edge there are no confirmed realised occurrences in the northernmost part of the potential distribution area in mainland Europe.The distribution records of Trachycarpus fortunei from Europe are listed in Walther(2003),BSBI(2006)and Walther et al.(2007).Escaped Tra-chycarpus fortunei palms have been recorded in southern Switzerland in large numbers,but also in France,as far north as Brittany,and rejuvenation within gardens has been observed in southern England.Also the northern-most subspontaneous occurrences of Trachycarpus fortunei correspond with the boundary of the potential distribution area.In addition to individual species distributions,Fig.2shows the overlap of the potential distribution areas of all the species shown in Fig.1.The potential species distributions overlap in substantial parts of central Europe,527 Bioclimatic limits and range shifts of cold-hardy evergreen broad-leaved speciesFig.1.Potential and currently realised distribution of selected evergreen broad-leaved species(for details see text).with gradually decreasing number of species’ranges towards the northeast. Whereas updated distribution data from France are scarce,more current information on evergreen broad-leaved species is available from the British Isles and Switzerland.Both the southernmost part of Switzerland and southern Great Britain have especially favourable winter temperatures, compared with other parts of Europe where deciduous forests occur.This is reflected in the coincidence of potential distribution areas of the four species regarded here,and additional evergreen broad-leaved species,which have also been recorded in these areas.A non-exhaustive list of additional evergreen broad-leaved species recorded in deciduous forests of southern Great Britain and southern Switzerland is given in Table1.Many of the same(exotic)species occur in both areas,but some species are restricted to528S.Berger et al.Fig.2.Superposed potential distribution of the four evergreen broad-leaved species shown in Fig.1.Shading represents the number of evergreen broad-leaved species(Over-lap of all four species’potential ranges indicated by the darkest shading,decreasing gradu-ally with decreasing number of species’potential ranges overlapping).Great Britain(e.g.Rhododendron ponticum)and others to southern Swit-zerland(e.g.Cinnamomum glanduliferum,Pittosporum tobira).In both Britain and southern Switzerland,the number,developmental stage and abundance of evergreen broad-leaved species varies with latitude and altitude.Releve´s from Walther(2000)and Berger&Walther (2006)reveal an altitudinal gradient in the number and abundance of ever-green broad-leaved species on south-facing slopes in the Insubrian region (Fig.3).At low altitudes there are dense stands of evergreen broad-leaved species in the shrub layer,and some individuals of urus nobilis and Cinnamomum glanduliferum have grown up to the tree layer.Towards higher altitudes the evergreen broad-leaved species are restricted to the shrub layer and their cover decreases.Finally there are only single small specimens in the herb layer,and at even higher altitudes no evergreen broad-leaved species are growing at all.In parallel,the number of evergreen broad-leaved species present declines with altitude.529 Bioclimatic limits and range shifts of cold-hardy evergreen broad-leaved speciesTable1.Evergreen broad-leaved species recorded in deciduous forests of Europe.Records from Switzerland according to Walther(1999)and from Great Britain according to BSBI(2006).Frost resistance refers to the experimental work of Sakai(1982),Sakai& Larcher(1987)and Dirr&Lindstrom(1990).Fig.3.Altitudinal distribution of exotic evergreen broad-leaved species on south facing slopes along Lago Maggiore,schematically.530S.Berger et al. DiscussionPrevious studies showed that climate change has favoured the spread of evergreen broad-leaved species in different parts of Europe.For instance,in southern Switzerland and northern Italy an increasing number of evergreen broad-leaved species was recorded as a consequence of climate change (Gianoni et al.1988,Walther2000,2001,2002).Also,the northward range shift of Ilex aquifolium in Scandinavia(Walther et al.2005)and the altitudinal upward shift of Viscum album in the European Alps(Dobber-tin et al.2005)are regarded as consequences of warmer winter tempera-tures.In central European deciduous forests,the change of the growth form of Hedera helix from creeping along the ground to climbing up trees has likewise been linked to recent climate change(Dierschke2005).We here regard this phenomenon more broadly,for a group of species, and present a biogeographic analysis that shows an emerging assemblage of advancing evergreen broad-leaved species over much of central Europe. The specific bioclimatic limits of these species and the impacts of this on-going process at landscape scale are discussed.European distribution and bioclimatic limitsWinter temperature plays an important role in limiting the distribution of the evergreen broad-leaved vegetation towards the poles(Rübel1930, Schimper1935,Schmithüsen1976,Box1981,Sakai&Larcher1987, Woodward1987,Walter&Breckle1999).Although minimum temper-ature ofÐ15∞C is regarded to be an important threshold value for cold-hardy evergreen broad-leaved species in general,the specific climatic threshold differs with the particular biological traits of the given species (e.g.Sakai&Larcher1987,Walther1999).In addition to low winter temperature,other climatic parameters may also limit the species’persist-ence at northern range margins,such as the length of the growing season (as required to produce viable seeds)and total summer warmth(indicated by GDD;as needed for a positive annual carbon balance and accumulation of biomass).The most obvious effect of rising winter temperatures is a reduced risk of lethal frost incidents,which allows survival beyond the former range boundary.Additionally,the frequency of sub-lethal frosts,that nevertheless may be severe enough to damage tissue,decreases.Hence,sub-lethal dam-age,which must be compensated by replacement of damaged tissue,and thus cost energy,will decrease.This favours species at their range margin and thus,survival in formerly unsuitable areas may become possible under climatic warming.Evergreen broad-leaved species are particularly favoured by milder winter temperatures,as they are able to profit from positive net photosynthesis in periods with favourable climatic conditions even in win-ter(Zeller1951,Fischer&Feller1994,Oliveira&Pen˜uelas2004), in contrast to deciduous trees which are leafless in the winter season.531 Bioclimatic limits and range shifts of cold-hardy evergreen broad-leaved species Although all the four species presented in Fig.1are restricted by cold winters(with frost resistance varying considerably between species),their temperature limits show that other climatic parameters may also play a role in limiting their distributions,due to specific biological traits.Whereas the survival and regeneration of Ilex aquifolium is mainly restricted by severe winters,Prunus laurocerasus additionally needs a sufficiently longer grow-ing season in order to produce viable seeds(Adamovic´1909,Söhlke2006). Hence,reproduction of Prunus laurocerasus is limited by the length of the growing season,which restricts it from extending as far north as Ilex aqui-folium,in spite of its ability to survive equally low winter temperatures.The northward extension of Laurus nobilis is also restricted by cold winters,though the limiting winter temperature used in the model to project the potential range of Laurus nobilis may be overestimated(Sven-ning,m.).The importance of winter temperatures for the north-ern distributional boundary is supported by the findings of Sakai& Larcher(1987),who regarded Laurus nobilis as a Mediterranean broad-leaved tree with medium frost resistance,especially pointing out the limited frost resistance of overwintering propagative organs.Low winter temperatures limit the distribution of Trachycarpus fortunei as well.Furthermore,its cold hardiness also depends on sufficiently warm summers(here expressed as GDD).Warm summers allow this palm to ac-cumulate enough biomass to regenerate damaged leaves and thereby to compensate for sub-lethal frost damage of the winter season(Walther et al.2007).The northernmost sub-spontaneous occurrences of Trachycar-pus fortunei coincide well with the northern boundary of the predicted potential distribution area.The lack of naturalised occurrences of Trachy-carpus fortunei in southernmost Europe,even though the species is culti-vated in this area,reflects the summer dryness of the Mediterranean climate with drought too severe for growth of this species(c.f.Berger&Walther 2006).At local scale,not only macroclimatic limits but also other ecologically relevant parameters,e.g.edaphic factors,must be taken into account,as they also influence the diversity and structure of communities with an in-creasing share of evergreen broad-leaved species(Berger&Walther 2006).This is of special importance at species range margins,as the habitat requirements of species become more specific towards range boundaries (Thomas et al.2001).Towards the Mediterranean the amount and the tem-poral pattern of precipitation also plays a role limiting the distribution of evergreen broad-leaved species(Berger&Walther2006,Del Rio& Penas2006).Impacts at the landscape levelVegetation modelling has projected range shifts both by single species and by biomes(e.g.Prentice et al.1992,Bakkenes et al.2002,Woodward et al.2004,Thuiller et al.2005,Ohlemüller et al.2006).The fact that responses to climate change are species-specific,as results presented here532S.Berger et al. show,suggests that climate change in the longer term will lead to a reorgan-isation of existing communities rather than to synchronous shifts of whole vegetation units.Furthermore,as there are only few native evergreen broad-leaved species in northern and central-European plant communities, the aforementioned range shifts may lead to distinct changes in the diversity and physiognomy of existing forest communities,or to assemblages of new species combinations(Chapin et al.1993,Scoles&Van Breemen1997). The substantial overlap of the projected ranges of several evergreen broad-leaved species in southern Great Britain and southern Switzerland(shown in Fig.2)suggests a trend towards more evergreen broad-leaved species in deciduous forests,which is further substantiated by the occurrence of additional evergreen broad-leaved species recorded in these areas(see Table 1).A shift from deciduous forest to deciduous forest with more evergreen species may change some forest ecosystem processes,e.g.carbon cycling and water balance.A new type of deciduous forest,with evergreen understorey,has been described by Delarze et al.(1999)for southern Switzerland.A structural scheme for this forest type with an increasing number of evergreen broad-leaved species is shown in Fig.3.As species approach their bioclimatic limits towards higher altitudes the size of the individuals and the number of species decreases.Analogously,the number of evergreen broad-leaved species declines towards higher latitudes.Also the growth form of the most cold-hardy species changes towards their northern range margins.For ex-ample,Ilex aquifolium forms trees in the core area of its distribution but grows more often as a shrub towards the range margin(Berger2003).The change in growth form of Ilex aquifolium near its range boundary shows that the species is able to grow and survive in a transitional area,but does not reach the size of the individuals in the core area.Similar observations have been registered for Hedera helix(see Dierschke2005)as well as for Prunus laurocerasus in its native range,where it occurs only as vegetatively spreading shrubbery at high altitudes(Nakhutsrishvili1999).On the British Isles there is no pronounced altitudinal gradient,but the evergreen broad-leaved species may constitute a considerable proportion of the shrub and lower tree layer,e.g.in both southwestern England and Ireland,as has been observed in southern Switzerland.Small evergreen for-ests and scrub stands(often with Quercus ilex)have been described as re-cently generated vegetation types with no counterpart in the native vegeta-tion,whereas Rhododendron thickets and understoreys are matched in the native vegetation by evergreen thickets of Ilex aquifolium(Peterken2001).The overall assessment of the impacts of spreading evergreen broad-leaved species in deciduous forest communities should consider existing communities formed by the respective species in their native ranges as well as those reported in the palaeoecological literature.In North America and Asia species-rich mixed broad-leaved forests are a general phenomenon in latitudes with analogous conditions(Fujiwara&Box1994).In Europe, however,repeated glaciations have reduced the formerly diverse Tertiary evergreen flora to a few species able to recolonise central Europe duringthe Holocene(Schmid1939,Lang1994).The assemblage of currently spreading evergreen broad-leaved species might be divided into three sub-groups based on their origin:(I)species native to Europe,expanding their previous ranges by themselves but also facilitated by human introductions, e.g.Ilex aquifolium and Laurus nobilis;(II)species widespread throughout the Tertiary not able to recolonise central Europe in the Holocene,but reintroduced as ornamentals from nearby refugial areas,e.g.Prunus lauro-cerasus;and(III)species introduced from other more distant parts of the world,e.g.Trachycarpus fortunei,Cinnamomum glanduliferum.The now expanding evergreen broad-leaved species represent plant types as well as families characteristic of the European Tertiary flora.The laurel type(Daphno-Macrophanerophyta)was the most important tree form in Tertiary Europe(Mai1995).Most species of this type are shade tolerant (Mai1995)and hence may thrive in the understorey of deciduous forests. This still applies to Ilex aquifolium,which grows in the understorey of deciduous forests in large parts of its distribution area.In north-western Europe Ilex aquifolium grows in different communities of the Querco-Fagetea,also in Atlantic beech and mixed forests like Abieti-Fagetum and Galio-Abietum,and in montane beech forests of Italy(Fagion australo-italicum)(Pott1990).In south-western Europe beech forests (Rusco-Fagetum,Ilici-Fagetum)have different laurophyllous and sclerophyllous species in the understorey(e.g.Ruscus aculeatus,Hedera helix,Buxus sempervirens and Ilex aquifolium)(Durin et al.1967,Tombal 1972,Rameau1985,Ge´hu&Julve1989).Finally,Ilex aquifolium also grows in the understorey of deciduous forests of the Balkan Peninsula, together with Prunus laurocerasus,Hedera helix and Rhododendron ponti-cum,where Fagus orientalis and Fagus moesiaca form the tree layer(Rho-dodendro pontici-Fagetum orientalis)(Horva´t et al.1974).Prunus laurocerasus is associated as an evergreen shrub to the understorey of decid-uous beech forests(e.g.Fageta laurocerasosa,Fageta luzulosa, Fageto-Abieteta laurocerasosa)in its native range(Nakhutsrishvili 1999).Numerous fossil leaves,flowers and fruits from several Cinnamomum species have been recorded from the Tertiary,showing that the genus Cin-namomum(Lauraceae)played an important role in the Tertiary flora of Europe(Staub1905,c.f.also Kovar-Eder et al.2006).All Cinnamomum tree species require ample precipitation,in addition to warm temperatures, which the introduced Cinnamomum glanduliferum enjoys in the Insubrian region under present conditions(Berger&Walther2006).Today the genus is restricted to warm-temperate and tropical east and southeast Asia (monsoon area)and to Australia(two species)(Staub1905),though some species have become naturalised elsewhere(e.g.Fleischmann1997, Ward&Labisky2004).Another life form of importance in Europe throughout the Tertiary was the tropical tuft-tree type Macrophanerophyta scaposa(“Schopfbaum”),a single-stemmed tree with a terminal rosette of leaves that die and extend the trunk,as in palms(Box1981,Mai1995).In the recent European floraFig.4.a)Tertiary flora“Lausanne zur miocaenen Zeit”(Heer1946):Large tree on the left:Cinnamomum,small tree on the left:Laurus(darker shaded).Palms,e.g.Sabal(fan palm),Phoenicites(feather palm).b)Deciduous forest stand with evergreen understorey in southern Switzerland(winter), also with Cinnamomum,Laurus and the fan palm Trachycarpus fortunei.there are few examples of this type,e.g.Chamaerops,Phoenix and Dra-caena,but the introduced naturalising species Trachycarpus fortunei also belongs to this type(cf.Fig.4).Fig.4a illustrates some of the main aspects of the Tertiary flora of central Europe(Heer1946).The large tree on the left is Cinnamomum and next to it is a small Laurus(shaded darker).Also palms like Sabal and Phoeni-cites are represented.Fig.4b shows a deciduous forest stand in southern Switzerland in winter,also with a Cinnamomum tree,Laurus and a palm species(Trachycarpus fortunei).Recent climate change has favoured the evergreen broad-leaved plant functional type within deciduous forests.With continued warming this process is likely to proceed and to induce changes in the composition and structure of temperate deciduous forests in various parts of central Europe. Other human activities also cause changes to the native vegetation,for ex-ample the introduction of new species.The combination of these anthropo-。

Chapter 6 Spectral Methods in Derivatives Pricing

t
τh = inf t
0:
0
h(Xu ) du
e
where e ∼ Exp(1) is an exponential random variable with unit mean and independent of X . Thus, the default time is (x ∧ y := min{x y }): ζ = τ h ∧ τD When default occurs, the holder of the asset looses the promised payoff f (XT ) and possibly receives some recovery payment instead (see the recent monographs Bielecki and Rutkowski, 2002; Dufﬁe and Singleton, 2003; Lando, 2004 for surveys of credit risk modeling). If h ≡ 0 and the process never exits D, then the derivative asset is default-free.
J.R. Birge and V. Linetsky (Eds.), Handbooks in OR & MS, Vol. 15 Copyright © 2008 Elsevier B.V. All rights reserved DOI: 10.1016/S0927-0507(07)15006-4
This research was supported by the US National Science Foundation under grants DMI-0200429 and DMI-0422937.

METHOD FOR TESTING THE WATER CONTENT AND WATER DIS

专利名称：METHOD FOR TESTING THE WATERCONTENT AND WATER DISTRIBUTION ATCELLULAR LEVEL IN FRUIT AND VEGETABLEWEBS ON RAMAN SPECTRUM BASIS发明人：SUN, Da-Wen,LI, Dongmei,ZHU, Zhiwei申请号：EP19933578申请日：20191016公开号：EP3859315A4公开日：20220330专利内容由知识产权出版社提供摘要：The present invention discloses a method for testing cellular level water content and distribution in fruit and vegetable tissues based on Raman spectroscopy. This method comprises preprocessing of samples, acquisition and preprocessing of imaging spectra, Gaussian peak-separation fitting of imaging spectra, pseudocolor imaging according to the fitting results, and visualization of distribution of water content and water bonding state at the cell level. The present invention realizes the visualization of the distribution of water content and water binding state at the cellular level in the fruit and vegetable tissues for the first time, and obtains relatively reliable quantitative analysis results of the content of water with different bonding states according to the visualization imaging results. The present invention, able to be used as a new method for testing cellular level water content in fruit and vegetable tissues, solves the current problem of not being able to detect cellular level water changes in fruit and vegetable processing, and has a good prospect for the research on fruit and vegetables processing.申请人：South China University of Technology更多信息请下载全文后查看。

基于GBT 38924—2020 民用轻小型无人机系统环境试验方法浅析

装备环境工程第19卷第2期·78·EQUIPMENT ENVIRONMENTAL ENGINEERING 2022年2月收稿日期：2021-07-13；修订日期：2021-08-13 Received ：2021-07-13；Revised ：2021-08-13作者简介：徐静（1981—），女，硕士，高级工程师，主要研究方向为环境可靠性试验与电磁兼容试验。

Biography ：XU Jing (1981—), Female, Master, Senior engineer, Research focus: environmental reliability test and EMC test.引文格式：徐静, 赵毅, 潘勇, 等. 基于GB/T 38924—2020民用轻小型无人机系统环境试验方法浅析[J]. 装备环境工程, 2022, 19(2): 078-084.XU Jing, ZHAO Yi, PAN Yong, et al. Simple Analysis of Environmental Test Method for Civil Light and Small Unmanned Aerial Vehicle System Based 基于GB/T 38924—2020民用轻小型无人机系统环境试验方法浅析徐静，赵毅，潘勇，胡德隆，尹江梅（山东省产品质量检验研究院，济南 250100）摘要：为了提高民用轻小型无人机企业产品质量，提高相关环境试验人员的综合试验能力，提升行业检测技术水平，以GB/T 38924—2020系列为基础，对民用轻小型无人机环境试验方法进行了浅析。

描述了民用轻小型无人机系统环境试验受到的影响因素，主要有气候环境因素和机械环境因素。

对气候环境试验方法、机械环境试验方法进行了对比分析和技术归纳。

根据GB/T 38924—2020系列对具体试验参数进行分析，温度试验的温度变化速率均不大于3 ℃/min ；低气压和高气压试验的压力变化速率不大于4.8 kPa/min ；盐雾试验分为连续喷雾方式和间歇喷雾方式；防水性试验的主要参数是雨滴直径和喷水量；砂尘试验的主要参数是尘浓度、砂浓度、吹尘风速和吹砂风速；冲击试验的主要参数是峰值加速度和脉冲时间，振动试验的主要参数是功率谱密度、振幅、加速度、扫频速率，冲击试验和振动试验参数应根据无人机类型和起飞质量来选取。

西门子可靠性培训资料

product delivered
in production failure control field return analysis & threshold monitoring conservative data based on predicted values regular component supplier audits
spare parts don’t contribute to FMA non-installed equipment at certain carriers low channel count for many years temperature for FIT estimation probably higher than in reality margin (e.g., due to low channel count, low PMD, BOL) can
Mechanical Components: • Connectors • Boards • etc.
Negotiations with suppliers: • Modeling • Field experience • Supplier comparison
Standards: • SN 29500 • IEC 61709
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Operating time T in years
For the field replaceable HW-units and their components the failure rate l could be only predicted by random failures!

几种图像缩放算法的研究

几种图像缩放算法的研究李秀英;袁红【摘要】There are several kinds of image scaling methods in time domain: nearest neighbor interpolation, linear interpolation, quadratic interpolation, cubic interpolation, Lagrange interpolation, Gaussian interpolation, etc. The performance of these methods are compared according to passband, stopband and cutoff frequency. The result shows that the nearest interpolation and the linear interpolation should be avoided, and Gaussian basis function whose N value is larger has better performance. The 2-D nonseparable interpolation filter and 2-D separable in-terpolation filter are studied. With the requirements of passband and stopband as optimizing target, and the struc-ture of filter as constraint condition, the filter design for them is converted into an optimization problem. The ex-periment result shows that the 2-D nonseparable interpolation filter has better performance.%时域内常用的几种图像缩放算法有:最近邻插值、线性插值、二次插值、三次插值、拉格朗日插值、高斯插值等,对这些算法的性能进行分析比较,综合通带、阻带及截止频率,最近邻插值和线性插值应该避免,高斯基函数(N较大者)具有较好性能；并且在频域内研究了二维可分离插值滤波器和不可分离插值滤波器,这两种方法以对通带和阻带的要求作为优化目标,以滤波器的结构为约束条件,将滤波器的设计转化为一个约束优化问题进行解决；实验结果表明二维不可分离插值滤波器的方法图像缩放后的效果最好.【期刊名称】《现代电子技术》【年(卷),期】2012(035)005【总页数】4页(P48-51)【关键词】插值滤波器;图像尺度变换;变量分离;采样率转换【作者】李秀英;袁红【作者单位】上海电力学院数理系,上海200090;临沂师范学院数学系,山东临沂276005【正文语种】中文【中图分类】TN911.73-34;TP3910 引言在实际应用中，经常需要对数字图像进行尺度转换，在改变图像尺寸方面，文献[1-8]提出了通过三次样条插值、三次卷积插值、二维快速傅立叶变换进行图像尺寸转换的方法；这些算法比较简单，性能受到了明显的限制。

科学文献

Venus Express Mission Definition ReportESA-SCI(2001)6 ESA-SCI(2001)6 October 20011An Orbiter for the study of the atmosphere, the plasma environment, and the surface of VenusMission Definition ReportEuropean Space Agency Agence Spatiale Européenne3 Venus Express Mission Definition Report ESA-SCI(2001)6ForewordVenus Express, an Orbiter for the study of the atmosphere, the plasma environment, and the surface of Venus, is a mission which was proposed to ESA in response to the Call for Ideas to re-use the Mars Express platform issued in March 2001. Venus Express together with two other missions, Cosmic DUNE and SPORT Express, was selected by ESA’s Space Science Advisory Committee for a Mission Definition Study. The industrial study of the three missions was conducted in parallel by Astrium-SAS (Toulouse, France) from mid-July to mid-October 2001. The payload included in the Venus Express Study comprises 5 instruments (ASPERA/MEx, PFS/MEx, SPICAM/MEx, VeRa/Rosetta, VIRTIS/Rosetta) from the Core payload of the original Proposal and the VENSIS/MEx radar in line with the SSWG recommendation. During the Study it was found scientifically reasonable and technically feasible to replace the standard Mars Express engineering Video Monitoring Camera by a scientific instrument, the Venus Monitoring Camera (VMC). The Mission Definition Report describes the scientific objectives of the Venus Express mission, presents selected payload set, and summarizes the results of the Mission Definition Study. This version of the report covers all science aspects of the mission but contains only a brief summary of the industrial study. The combined industrial study report for all the three missions is published in a separate cover. A complete Venus Express Mission Definition Report, including a comprehensive description of scientific goals, payload, and technical aspects of the spacecraft will be prepared by the end of 2001. The Venus Express Study was directly supported by the Science Study Team listed below.Mission science coordinationD.V. Titov, MPAe, Germany E. Lellouch, DESPA, France F.W. Taylor, Oxford University, UK L. Marinangeli, Universita d’Annunzio, Italy H. Opgenoorth, IRF-Uppsala, SwedenPrincipal InvestigatorsS. Barabash, IRF-Kiruna, Sweden /PI ASPERA J.-L. Bertaux, Service de Aeronomie, France /Co-PI SPICAM/ P. Drossart, DESPA, France /Co-PI VIRTIS/ V. Formisano, IFSI, Italy /PI PFS/ B. Haeusler, Universitaet der Bundeswehr, Germany /PI VeRa/ O. Korablev, IKI, Moscow, Russia /Co-PI SPICAM/ W.J. Markiewicz, MPAe, Germany /PI VMC/ M. Paetzold, Universitaet zu Koeln, Germany /Co-PI VeRa/ G. Picardi, Infocom Dpt. Univ. of Rome, Italy /PI VENSIS/ G. Piccioni, IAS, Italy /Co-PI VIRTIS/ J. Plaut, JPL/NASA, Pasadena, California, USA J.-A. Sauvaud, CESR-CNRS, France /Co-PI ASPERA/ P. Simon, BISA, Belgium /CO-PI SPICAM/The ESA members of the Scientific Directorate responsible for the study were: J-P. Lebreton, Study Scientist, Research and Science Support Department (RSSD), ESTEC M. Coradini, Science Planning and Coordination Office, ESA HQ, Paris G. Whitcomb, Future Science Projects and Technology Office, SCI-PF, ESTEC D. McCoy, Mars Express Project Team, SCI-PE, ESTEC. The Industrial study was lead by: Ch. Koeck (Study Manager), Astrium, France with support from: S. Kemble (Mission Analysis), Astrium, UK4 Venus Express Mission Definition Report ESA-SCI(2001)6L. Gautret (Payload Interface Engineering), Astrium, France P. Renard (System Engineering), Astrium, France F. Faye (Mars Express expertise), Astrium, France. Support was provided by the following colleagues within ESA:ESOC: M. Hechler and J. Rodriguez-Canabal, (Mission Analysis); R. Van Holtz, (Ground Segment definition) ESTEC A. Chicarro (Mars Express Project Scientist), RSSD/SCI-SO P. Falkner, (payload support), RSSD/SCI-ST P. Martin (Mars Express Deputy Project Scientist), RSSD/SCI-SO J. Romstedt (radiation environment analysis & payload support), RSSD/SCI-ST R. Schmidt (Mars Express Project Manager), SCI-PE J. Sorensen (radiation environment analysis), TOS-EMA P. Wenzel (Head of Solar System Division), RSSD/SCI-SO O. Witasse, (Science support), RSSD/SCI-SOThis report is available in pdf format at: http://solarsystem.estec.esa.nl/Flexi2005/ Requests for further information and additional hard copies of this report should be addressed to: Jean-Pierre Lebreton: Marcello Coradini: Jean-Pierre.Lebreton@esa.int Marcello.Coradini@esa.int5 Venus Express Mission Definition Report ESA-SCI(2001)6Executive SummaryThe first phase of Venus spacecraft exploration (1962-1985) by the Venera, Pioneer Venus and Vega missions established a basic description of the physical and chemical conditions prevailing in the atmosphere, near-planetary environment, and at the surface of the planet. At the same time, they raised many questions on the physical processes sustaining these conditions, most of which remain as of today unsolved. Extensive radar mapping by Venera-15,-16 and Magellan orbiters, combined with earlier glimpses from landers, have expanded considerably our knowledge of Venus’ geology and geophysics. A similar systematic survey of the atmosphere is now in order. This particularly concerns the atmosphere below the cloud tops, which, with the exception of local measurements from descent probes, has escaped detection from previous Venus orbiters. Many problems of the solar wind interaction, in particularly those related to the impact on the planetary evolution are still not resolved. The present proposal aims at a global investigation of Venus’ atmosphere and plasma environment from orbit, and addresses several important aspects of the geology and surface physics. The fundamental mysteries of Venus are related to the global atmospheric circulation, the atmospheric chemical composition and its variations, the surface-atmosphere physical and chemical interactions including volcanism, the physics and chemistry of the cloud layer, the thermal balance and role of trace gases in the greenhouse effect, the origin and evolution of the atmosphere, and the plasma environment and its interaction with the solar wind. Besides, the key issues of the history of Venusian volcanism, the global tectonic structure of Venus, and important characteristics of the planet’s surface are still unresolved. Beyond the specific case of Venus, resolving these issues is of crucial importance in a comparative planetology context and notably for understanding the long-term climatic evolution processes on Earth. The above problems can be efficiently addressed by an orbiter equipped with a suite of adequate remote sensing and in situ instruments. Compared with earlier spacecraft missions, a breakthrough will be accomplished by fully exploiting the existence of spectral “windows” in the near-infrared spectrum of Venus’ nightside, discovered in the late ‘80’-s, in which radiation from the lower atmosphere and even the surface escapes to space and can be measured. Thus, a combination of spectrometers, spectro-imagers, and imagers covering the UV to thermal IR range, along with other instruments such as a radar and a plasma analyzer, is able to sound the entire Venus atmosphere from the surface to 200 km, and to address specific questions on the surface that would complement the Magellan investigations. This mission will also tackle still open questions of the plasma environment focusing on the studies of nonthermal atmospheric escape. This issue will be addressed via traditional in situ measurements as well as via innovative ENA (Energetic Neutral Atom) imaging techniques. The instruments developed for the Mars Express and Rosetta missions are very well suited for this task. The following available instruments: SPICAM – a versatile UV-IR spectrometer for solar/stellar occultations and nadir observations, PFS – a high-resolution IR Fourier spectrometer, ASPERA – a combined energetic neutral atom imager, electron, and ion spectrometer, VIRTIS – a sensitive visible spectro-imager and mid-IR spectrometer, a radio science experiment VeRa, a wide-angle monitoring camera VMC, and subsurface and ionosphere sounding radar VENSIS will form the payload of the proposed Venus Express mission. Taken together, these experiments can address all the broad scientific problems formulated above. The Mission Definition Study demonstrated the feasibility of the proposed mission to Venus in 2005. The Mars Express spacecraft can accommodate the above mentioned experiments with minor modifications. The launch with Soyuz-Fregat can deliver this payload to a polar orbit around Venus with a pericenter altitude of ~250 km and apocenter of6 Venus Express Mission Definition Report ESA-SCI(2001)6~45,000 km. This orbit will provide complete coverage in latitude and local solar time. It is also well suited for atmospheric and surface sounding, as well as the studies based on solar and radio occultations. In comparison to the Pioneer Venus spinning spacecraft, Mars Express is an advanced 3 axis stabilised platform which provides significantly enhanced spectroscopic and imaging capabilities. The proposed duration of the nominal orbital mission is two Venus days (sidereal rotation periods) equivalent to ~500 Earth days. The Venus Express mission will achieve the following “firsts”: • First global monitoring of the composition of the lower atmosphere in the near IR transparency “windows”; • First coherent study of the atmospheric temperature and dynamics at different levels of the atmosphere from the surface up to ~200 km; • First measurements of global surface temperature distribution from orbit; • First study of the middle and upper atmosphere dynamics from O2, O, and NO emissions; • First measurements of the non-thermal atmospheric escape; • First coherent observations of Venus in the spectral range from UV to thermal infrared; • First application of the solar/stellar occultation technique at Venus; • First use of 3D ion mass analyzer, high energy resolution electron spectrometer, and energetic neutral atom imager; • First sounding of Venusian topside ionospheric structure; • First sounding of the Venus subsurface. Together with the Mars Express mission to Mars and the Bepi Colombo mission to Mercury, the proposed mission to Venus, through the expected quality of its science results, would ensure a coherent program of terrestrial planets exploration and provide Europe with a leading position in this field of planetary research. The international cooperation formed in the framework of the Mars Express and Rosetta missions will be inherited by the Venus Express and will include efforts of the scientists of European countries, USA, Russia, and Japan. The Venus Express orbiter will play the role of pathfinder for future, more complex missions to the planet, and the data obtained will help to plan and optimize future investigations. Venus studies can have significant public outreach given the exotic conditions of the planet and the interest in comparing Venus to Earth, especially in a context of concern with the climatic evolution on Earth.7 Venus Express Mission Definition Report ESA-SCI(2001)6Table of content1. INTRODUCTION................................................................................................................................................ 8 2. MISSION SCIENCE OBJECTIVES................................................................................................................. 8 2.1 LOWER ATMOSPHERE AND CLOUD LAYER (0 – 60 KM) ................................................................................... 8 2.2 MIDDLE ATMOSPHERE (60 – 110 KM) ........................................................................................................... 12 2.3 UPPER ATMOSPHERE (110 – 200 KM) ............................................................................................................ 13 2.4 PLASMA ENVIRONMENT AND ESCAPE PROCESSES ......................................................................................... 14 2.5 SURFACE AND SURFACE-ATMOSPHERE INTERACTION ................................................................................... 15 3. SCIENTIFIC PAYLOAD ................................................................................................................................. 17 3.1 ASPERA (ANALYZER OF SPACE PLASMAS AND ENERGETIC ATOMS) ......................................................... 17 3.2 PFS (HIGH RESOLUTION IR FOURIER SPECTROMETER) ................................................................................ 18 3.3 SPICAM (UV AND IR SPECTROMETER FOR SOLAR/STELLAR OCCULTATIONS AND NADIR OBSERVATIONS)20 3.4 VERA (VENUS RADIO SCIENCE).................................................................................................................... 22 3.5 VIRTIS (UV-VISIBLE-NEAR IR IMAGING SPECTROMETER) .......................................................................... 23 3.6 VENSIS (LOW FREQUENCY RADAR FOR SURFACE AND IONOSPHERIC STUDIES). ......................................... 25 3.7 VMC (VENUS MONITORING CAMERA) ......................................................................................................... 26 3.8 SYNERGY OF THE PAYLOAD. .......................................................................................................................... 27 3.9 PAYLOAD ACCOMMODATION ......................................................................................................................... 28 3.10 MISSION AND PAYLOAD SCHEDULE ............................................................................................................. 29 3.11 PAYLOAD TEAMS ......................................................................................................................................... 29 4 MISSION OVERVIEW...................................................................................................................................... 36 4.1 MISSION SCENARIO ........................................................................................................................................ 36 4.2 LAUNCH, DELTA-V, AND MASS BUDGETS ...................................................................................................... 37 4.3 OPERATIONAL ORBIT ..................................................................................................................................... 37 4.4 ORBITAL SCIENCE OPERATIONS ..................................................................................................................... 38 4.5 TELECOMMUNICATIONS ................................................................................................................................. 38 4.6 THERMAL CONTROL ....................................................................................................................................... 39 4.7 RADIATION REQUIREMENTS ........................................................................................................................... 39 4.8 GROUND SEGMENT IMPLEMENTATION AND OPERATIONS SUPPORT ............................................................... 39 4.9 MISCELLANEOUS ............................................................................................................................................ 40 5. SCIENCE OPERATIONS, DATA ANALYSIS, AND ARCHIVING ......................................................... 40 5.1 SCIENCE OPERATIONS CONCEPT ................................................................................................................... 40 5.2 PRINCIPAL INVESTIGATORS ........................................................................................................................... 40 5.3 INTERDISCIPLINARY SCIENTISTS (IDS) ......................................................................................................... 40 5.4 SCIENCE WORKING TEAM ............................................................................................................................. 40 5.6 SCIENCE OPERATION PLAN ............................................................................................................................ 41 5.7 DATA ANALYSIS ............................................................................................................................................. 41 5.8 SCIENCE MANAGEMENT PLAN ...................................................................................................................... 41 5.9 COMPLEMENTARY VENUS GROUND-BASED OBSERVATIONS ........................................................................ 41 6. PROGRAMMATIC VALIDITY ..................................................................................................................... 41 7. SCIENCE COMMUNICATION AND OUTREACH ................................................................................... 42 7.1 GOALS ............................................................................................................................................................ 42 7.2 SCIENTIFIC THEMES ....................................................................................................................................... 42 7.3 IMPLEMENTATION .......................................................................................................................................... 43 8. INTERNATIONAL COOPERATION............................................................................................................ 43 9. REFERENCES................................................................................................................................................... 45 10 ACKNOWLEDGMENTS ................................................................................................................................ 468 Venus Express Mission Definition Report ESA-SCI(2001)61. IntroductionSince the beginning of the space era, Venus has been an attractive target for planetary science. Our nearest planetary neighbour and, in size, the twin sister of Earth, Venus was expected to be very similar to our planet. However, the first phase of Venus spacecraft exploration (1962-1985) discovered an entirely different, exotic world hidden behind a curtain of dense clouds. The earlier exploration of Venus included a set of Soviet orbiters and descent probes, Veneras 4–16, the US Pioneer Venus mission, the Soviet Vega balloons, the Venera 15, 16 and Magellan radar orbiters, the Galileo and Cassini flybys, and a variety of ground-based observations. Despite all of this exploration by more than 20 spacecraft, the “morning star” remains a mysterious world. All these studies gave us a basic knowledge of the conditions on the planet, but generated many more questions concerning the atmospheric composition, chemistry, structure, dynamics, surface-atmosphere interactions, atmospheric and geological evolution, and the plasma environment. It is high time to proceed from the discovery phase to a thorough investigation and deep understanding of what lies behind Venus’ complex chemical, dynamical, and geological phenomena. The data from ground-based observations and previous space missions is very limited in space and time coverage, and, prior to the discovery of the near infrared spectral windows, lacked the capability to sound the lower atmosphere of Venus remotely and study the phenomena hidden behind the thick cloud deck from orbit. Thus a survey of the Venus atmosphere is long overdue. Pioneer Venus, Venera-15, -16, and Magellan provided global comprehensive radar mapping of the surface and investigated its properties. The use of penetrating radar can add a third dimension to the earlier investigations. While a fully comprehensive exploration of Venus will require, in the long term, in situ measurements from probes, balloons and sample return, so many key questions about Venus remain unanswered that even a relatively simple orbiter mission to the planet can bring a rich harvest of high quality scientific results. The re-use of the Mars Express bus with the payload based on the instruments available from the Mars Express and Rosetta projects is very appropriate in this regard. It offers an excellent opportunity to make major progress in the study of the planet.2. Mission science objectivesThe proposed Venus Express mission covers a broad range of scientific goals including atmospheric physics, subsurface and surface studies, investigation of the plasma environment and interaction of the solar wind with the atmosphere. For clarity we divided the atmosphere into three parts: lower atmosphere (0-60 km), middle atmosphere (60 – 110 km), and upper atmosphere (110 – 200 km). The physics, methods of investigation, and scientific goals are quite different for each atmospheric region. However they all can be studied by a multipurpose remote sensing and in situ payload in the framework of the proposed orbiter mission.2.1 Lower atmosphere and cloud layer (0 – 60 km)Structure. Existing observations of the lower atmosphere hidden below the clouds are limited to in situ measurements, acquired by 16 descent probes mostly in equatorial latitudes, by radiooccultations on previous orbiters (Venera 9, 10, 15, 16, Pioneer Venus, and Magellan), and brief glimpses provided by the Galileo and Cassini fly-bys. The descent probes showed that the temperature structure below 30 km is quite constant all over the planet (Fig. 2.1). However, the temperature structure in the lower scale height is virtually unknown. Mapping the regions of high elevation in sub-micron spectral “windows” at the nightside will determine the surface temperature as a function of altitude (Meadows and Crisp (1996)). Assuming this is equal to the near-surface air temperature, this will allow a determination of the thermal profile and lapse rate in the 0-10 km range and an investigation of its degree of static stability, constraining the dynamics and turbulence in this region. The thermal structure above 35 km altitude will be obtained from radiooccultations with high vertical resolution. Composition. The Venusian atmosphere consists mainly of CO2 and N2 with small amounts of trace gases (Fig. 2.1). Although there is very little observational data, the chemistry of the lower atmosphere is expected to be dominated by the thermal decomposition of sulfuric acid, and cycles that include sulfur and carbon compounds (SO2, CO, COS etc.) and water vapour.9 Venus Express Mission Definition Report ESA-SCI(2001)6The discovery of the near IR spectral “windows” (Allen and Crawford, 1984), through which thermal radiation from the lower atmosphere leaks to space, allows us to study the composition of the atmosphere below the clouds on the nightside of the planet. The windows at 2.3 and 1.74 µm sound the atmosphere in broad altitude regions centered at 30-35 km and 20 km respectively, while the windows shortward of 1.2 µm (0.85, 0.9, 1.01, 1.10,and 1.18 µm) probe the first scale height and the surface. The detailed appearance of the windows results from the combined effect of composition, cloud opacity, and thermal structure, including the surface temperature (Taylor et al., 1997). Highresolution observations covering all windows simultaneously, along with physical cloud models, should allow retrieval of all the variables.Figure 2.1 Structure and main parameters of the lower atmosphere of Venus.Water vapour is important not only for chemistry but also as a greenhouse gas. The few existing measurements of the H2O abundance in the deep atmosphere show no evidence for variability so far. By mapping simultaneously at several wavelengths, corresponding to radiation originating at different altitudes, it will be possible to probe the H2O profile below the clouds and to search for possible spatial variations, including those that might be the signature of volcanic activity. A precise inventory is also needed to better constrain the origin of the present atmospheric water. The H2O abundance at the surface has strong implications for the stability of some hydrated rocks. Carbon monoxide is very abundant in the upper atmosphere due to the dissociation of CO2 by solar ultraviolet radiation. It is much less common in the troposphere, but it does there show a definite trend of increasing from equator to pole. The source near the poles could be the downward branch of a Hadley cell transporting CO-rich air from the upper atmosphere, an important diagnostic of the mean meridional circulation. More detailed observations of CO at all levels, latitudes and times are needed to confirm this hypothesis and reveal details of the global-scale dynamics. CO is also a key player in the equilibrium between surface minerals and the atmosphere. The study of the lower atmosphere composition by means of spectroscopy in the near IR transparency “windows” is one of the main goals of the Venus Express mission. More specific objectives include abundance measurements of H2O, SO2, COS, CO, H2O, HCl, and HF and their horizontal and vertical (especially for H2O) variations, to significantly improve our understanding of the chemistry, dynamics, and radiative balance of the lower atmosphere, and to search for localized volcanic activity. Cloud layer. Venus is shrouded by a 20 km thick cloud layer whose opacity varies between 20 and 40 in the UV, visible and infrared (Fig. 2.1). The clouds are almost featureless in visible light but display prominent markings in the UV-blue spectral region (Fig. 2.2). Earlier observations showed that at least the upper cloud consists of micron size droplets of 75% H2SO4, which is produced by photochemical reactions at the cloud tops. The physical and chemical processes forming the lower clouds are virtually unknown, including major problems like (1) the nature of the UV-blue absorber which produces the features observed from space and absorbs half of the energy received by the10 Venus Express Mission Definition Report ESA-SCI(2001)6planet from the Sun, and (2) the origin of the large solid particles detected by the PioneerVenus probe. The remote sensing instruments on Venus Express will sound the structure, composition, dynamics, and variability of the cloud layer, including: • Cloud and haze structure and opacity variations; Distribution and nature of the UV• blue absorber; • Measurements of atmospheric composition which constrain models of cloud formation and evolution. Greenhouse effect. The high surface temperature of about 735 K results from the powerful greenhouse effect created by the presence of sulphuric acid clouds and certain Figure 2.2 Venus images in the violet filter taken gases (CO2, H2O, SO2) in the atmosphere by the Gallileo spacecraft (see Crisp and Titov, 1997). Less than 10% of the incoming solar radiation penetrates through the atmosphere and heats the surface, but thermal radiation from the surface and lower atmosphere has a lower probability of escape to space due to the strong absorption by gas and clouds. The result is about 500K difference between the surface temperature and that of the cloud tops, an absolute record among the terrestrial planets (Fig. 2.1). The measurements of outgoing fluxes over a broad spectral range, combined with temporarily and latitudinally resolved cloud mapping and high resolution spectroscopy in the near IR windows will give an insight into the roles of radiative and dynamical heat transport, and the various species, in the greenhouse mechanism. Atmospheric dynamics. The dynamics of the lower atmosphere of Venus is mysterious. Tracking of the UV markings, descent probes, and Vega balloons trajectories all showed that the atmosphere is involved in zonal retrograde super-rotation with wind velocities decreasing from ~100 m/s at the cloud tops to almost 0 at the surface (Fig. 2.3). At the same time, there appears to be a slower overturning of the atmosphere from equator to pole, with giant vortices at each pole recycling the air downwards. What is most puzzling about the regime represented by this scenario is how the atmosphere is accelerated to such high speeds on a slowly-rotating planet. Additional questions include (1) whether the meridional circulation is one enormous 'Hadley' cell extending from the upper atmosphere to the surface, or a stack of such cells, or something else altogether; (2) how the polar vortices couple the two main components of the global circulation and why they have such a complex shape and behaviour; and (3) what the observed (and observable) distributions of the minor constituents in Venus' atmosphere, including the clouds, are telling us about the motions (Fig.2.4). All attempts to model the zonal superrotation have been unsuccessful so far, indicating that the basic mechanisms of the phenomenon are unclear. There is an even Figure 2.3 Zonal winds in the Venus atmosphere。