基于生态化学计量学的草地退化研究
|
摘要
大安市姜家店草场位于松嫩平原中部低平原，该区生态环境比较脆弱。近年来，受到全球变化和人为因素的影响，草地面积逐渐减少，植被覆盖度降低，草地沙化、盐碱化、退化三化现象严重，草地生态环境不断恶化，已经开始制约当地社会经济的发展。研究该区退化草地的相关问题对科学保护草地资源、抑制草地退化、修复草地生态环境具有重要的意义。
本文以植被-土壤作为系统研究，基于生态化学计量学及元素限制性理论，运用SPSS、Matlab等软件分析羊草叶片碳氮磷生态化学计量特征与土壤营养元素、化学性质间相互关系，应用RBF人工神经网络建立羊草叶片碳氮磷含量的预测模型，并采用K均值聚类分析方法对草地退化程度进行评价，进而分析不同退化程度土壤化学性质及羊草叶片生态化学计量特征。研究结果表明：
1.研究区土壤有机质和全氮分别低于全国土壤平均水平（31.00mg·g~(-1)、1.60mg·g~(-1)），根据全国第二次土壤普查分级标准，土壤有机质及全氮含量属三级中等水平，土壤全磷含量属六级极贫乏水平（小于0.4mg·g~(-1)）；土壤C/N值、C/P值和N/P值分别为9.81、41.72和4.39均低于我国平均值；土壤速效氮、速效磷较贫乏，速效钾较丰富；土壤既存在盐化又存在碱化，可溶盐总量和碱化度与土壤营养元素含量显著负相关。
2.大安市姜家店草场羊草叶片碳氮磷的含量分别为443.75mg·g~(-1)，18.70mg·g~(-1)，1.30mg·g~(-1)，皆低于全球陆生植物叶片碳氮磷的平均含量（464mg·g~(-1)，20.6mg·g~(-1)，2.0mg·g~(-1)），并且氮、磷低于我国植物叶片氮、磷平均含量；羊草叶片碳的空间差异性明显低于氮和磷；羊草叶片C/N值、C/P值、N/P值皆高于全球陆生植物叶片C/N值、C/P值、N/P值；羊草叶片14＜N/P＜16，研究区羊草的生长同时受N、P的限制，N、P含量显著正相关；叶片C与土壤碳氮磷正相关，叶片N与土壤碳氮磷含量相关性不显著，叶片P与土壤氮相关性显著。
3.基于人工神经网络的工作原理，采用土壤营养元素及相关化学性质作为输入层，羊草叶片碳氮磷含量作为输出层，利用Matlab软件建立RBF人工神经网络模型用于模拟预测羊草叶片碳氮磷含量与土壤化学性质之间的关系是可行的，经检验相对误差均在20%之内，预测精度较高。
4.通过K均值聚类分析法利用土壤和植物指标将研究区85个采样点划分为四种类型，并与野外调查实际情况相结合确定采样点草地退化程度。结果表明研究区轻度退化草地占5.88%；中度退化草地占9.41%；重度退化草地占35.29%；极度退化草地占49.41%，研究区草地退化情况严峻。
5.随着草地退化程度加重研究区土壤全碳、全氮含量逐渐降低，全磷含量变化幅度较小；土壤C/N值、C/P值和N/P值总体呈现出下降的趋势；速效氮含量中度退化＞轻度退化＞重度退化＞极度退化，速效磷含量逐渐下降，速效钾含量从轻度退化阶段（214.98mg·kg~(-1)）急剧下降到中度退化阶段（144.63mg·kg~(-1)），随后保持相对稳定；pH值、可溶盐总量和碱化度随草地退化程度的加剧，逐渐升高，草地退化程度状况与土壤盐碱化程度状况具有正相关关系。
6.羊草叶片C含量表现为中度退化＞轻度退化＞重度退化＞极度退化；叶片N含量表现为中度退化＞极度退化＞轻度退化＞重度退化；羊草叶片P含量表现为先降低后增加；C/N值呈波状下降，C/P值、N/P值均为先上升后下降；轻度退化阶段羊草生长同时受N、P的限制，在中度退化阶段相对于N主要受到P的限制，在重度退化和极度退化阶段相对于P主要受N的限制；除营养元素以外羊草的生长速率还受其它土壤理化性质的影响。
Jiangjiadian meadow of Da’an city locates at the low plain in the middle of SongnenPlain. The entironment here is weaker. In recent years, influenced by global changing andhuman reason, the grassland area is reducing gradually; meanwhile, vegetation coveragelevel is depressing. Desertification, salinization and degeneration of the grassland aremore serious and its entironment is getting worse. All these have restricted thedeveloping of local society and economy. The research of grassland degenerationquestion in this area has important meaning to protecting grassland resource scientifically,to controlling grassland degeneration, and to repairing grassland entironment.
This paper takes Plant-Soil as a system to research. Based on ecologicalstoichiometry and element theory of constraints, using software such as SPSS and Matlab,I analyzed the interaction between the ecological chemical measurement features of C, Nand P in L. chinensis leaves and soil element and its chemical character. Using RBFArtificial Neural Networks (ANNs), I built the forecast model of the content of C,N andP elements in L. chinensis leaves.Using K value clustering analysis method, Ianalyze thedegeneration level of the grassland. After that, I analyze the soil chemistry character ofdifferent degeneration levels, and ecological chemical measurement features of L.chinensis leaves. The result of the research proves:
1. The contents of soil gein and TotalN in research area are all lower than nationalaverage. Compared with national second time soil survey standard, the contents of soilgein and TotalN belong to middle third level, and the content of soil TotalP belongs to extremely poor level (lower than0.4mg·g~(-1)). In the soil, the value of C/N, C/P and N/Pare9.81、41.72、4.39, all lower than national average. The contents of available N andavailable P are poor, available K is richer. There are both a soil salinization and alkalineexistence. Soluble salt and alkaline level show significant negative correlation with thecontent of soil nutrient elements.
2. The contents of C, N and P in L. chinensis leaves at Da’anJiangjiadian meadoware443.75mg·g~(-1)、18.70mg·g~(-1)、1.30mg·g~(-1). They are all lower than global averagecontents in terraneous plant leaves （464mg·g~(-1)、20.6mg·g~(-1)、2.0mg·g~(-1)）. And thecontent of N and P are even lower than national data. The space difference of CinL.chinensis leaves is obviously lower than N and P. The values of C/N, C/P and N/P inL.chinensis leaves are all higher than those in global terrestrial plant leaves. InL.chinensis leaves,14＜N/P＜16. The growth of L.chinensis in research area islimited by both N and P.The content of N and P show significant positive correlation; Thecontent of C inleaves show significant positive correlation with content of C, N, P andavailable P in soil. There is no significant correlation between the content of N in leavesand C, N, P in soil.
3. Based on the principle of ANNs, I take soil nutrient elements and relevantchemistry character as input layer. The contents of C, N and P in L. chinensis leaves areregarded as output layer. It’s feasible to create RBF ANNS model with Matlab and use itfor calculate the interaction between the contents of C, N and P in L. chinensis leaves andsoil chemical nature. After checking, the relative error is within20%. The forecastingaccuracy is higher.
4. Through K value clustering analysis method, using soil and plants as index, Idivide the85sampling points in research area into four types of degeneration level. Andconfirm the grassland degeneration level combined with the physical truth of field survey.The result shows that in research area, grassland of low-grade degeneration takes5.88%,mid-grade takes9.41%, high-grade takes35.29, and extremely high-grade takes49.41%.The regeneration of grassland in research area is severe.
5. With the aggravation of grassland degeneration degree, the contents of TotalC and TotalN get lower gradually, but the changing of TotalP content is smaller. As a whole,the values of C/N, C/P and N/P all show decreasing trends. In different degenerationlevels, the content sequence of available N in L. chinensis leaves is mid-grade＞low-grade＞high-grade＞extremely high-grade. The content of available P declinesgradually with the aggravation of grassland degeneration degree. And at the same time,the content of available K declines sharply from low-grade of degeneration（214.98mg·kg~(-1)）to mid-grade （144.63mg·kg~(-1)）, then remains comparatively steady.PH, total soluble salt and basification degree ratchet up with the aggravation of grasslanddegeneration. Grassland degradation degreeand soil salinity level condition is correlated.
6. In different degeneration levels, the content sequenceof C in L. chinensis leavesis mid-grade＞low-grade＞high-grade＞extremely high-grade.The content sequenceof Nis mid-grade＞extremely high-grade＞low-grade＞high-grade. Except extremelyhigh-grade degeneration stage, the variation trend of N in leaves is same as available N.The content of P in L. chinensis leaves shows decrease then increase. C/N value declineswavily. In low-grade degeneration stage, the growth of L. chinensis is limited both by Nand P. In mid-grade degeneration stage, it is limited mainly by P. In high-grade andextremely high-grade degeneration stages, it is limited mainly by N. Except nutrientelement, the growth speed of L. chinensis is still influenced by other soil physical andchemical properties.
本文以植被-土壤作为系统研究，基于生态化学计量学及元素限制性理论，运用SPSS、Matlab等软件分析羊草叶片碳氮磷生态化学计量特征与土壤营养元素、化学性质间相互关系，应用RBF人工神经网络建立羊草叶片碳氮磷含量的预测模型，并采用K均值聚类分析方法对草地退化程度进行评价，进而分析不同退化程度土壤化学性质及羊草叶片生态化学计量特征。研究结果表明：
1.研究区土壤有机质和全氮分别低于全国土壤平均水平（31.00mg·g~(-1)、1.60mg·g~(-1)），根据全国第二次土壤普查分级标准，土壤有机质及全氮含量属三级中等水平，土壤全磷含量属六级极贫乏水平（小于0.4mg·g~(-1)）；土壤C/N值、C/P值和N/P值分别为9.81、41.72和4.39均低于我国平均值；土壤速效氮、速效磷较贫乏，速效钾较丰富；土壤既存在盐化又存在碱化，可溶盐总量和碱化度与土壤营养元素含量显著负相关。
2.大安市姜家店草场羊草叶片碳氮磷的含量分别为443.75mg·g~(-1)，18.70mg·g~(-1)，1.30mg·g~(-1)，皆低于全球陆生植物叶片碳氮磷的平均含量（464mg·g~(-1)，20.6mg·g~(-1)，2.0mg·g~(-1)），并且氮、磷低于我国植物叶片氮、磷平均含量；羊草叶片碳的空间差异性明显低于氮和磷；羊草叶片C/N值、C/P值、N/P值皆高于全球陆生植物叶片C/N值、C/P值、N/P值；羊草叶片14＜N/P＜16，研究区羊草的生长同时受N、P的限制，N、P含量显著正相关；叶片C与土壤碳氮磷正相关，叶片N与土壤碳氮磷含量相关性不显著，叶片P与土壤氮相关性显著。
3.基于人工神经网络的工作原理，采用土壤营养元素及相关化学性质作为输入层，羊草叶片碳氮磷含量作为输出层，利用Matlab软件建立RBF人工神经网络模型用于模拟预测羊草叶片碳氮磷含量与土壤化学性质之间的关系是可行的，经检验相对误差均在20%之内，预测精度较高。
4.通过K均值聚类分析法利用土壤和植物指标将研究区85个采样点划分为四种类型，并与野外调查实际情况相结合确定采样点草地退化程度。结果表明研究区轻度退化草地占5.88%；中度退化草地占9.41%；重度退化草地占35.29%；极度退化草地占49.41%，研究区草地退化情况严峻。
5.随着草地退化程度加重研究区土壤全碳、全氮含量逐渐降低，全磷含量变化幅度较小；土壤C/N值、C/P值和N/P值总体呈现出下降的趋势；速效氮含量中度退化＞轻度退化＞重度退化＞极度退化，速效磷含量逐渐下降，速效钾含量从轻度退化阶段（214.98mg·kg~(-1)）急剧下降到中度退化阶段（144.63mg·kg~(-1)），随后保持相对稳定；pH值、可溶盐总量和碱化度随草地退化程度的加剧，逐渐升高，草地退化程度状况与土壤盐碱化程度状况具有正相关关系。
6.羊草叶片C含量表现为中度退化＞轻度退化＞重度退化＞极度退化；叶片N含量表现为中度退化＞极度退化＞轻度退化＞重度退化；羊草叶片P含量表现为先降低后增加；C/N值呈波状下降，C/P值、N/P值均为先上升后下降；轻度退化阶段羊草生长同时受N、P的限制，在中度退化阶段相对于N主要受到P的限制，在重度退化和极度退化阶段相对于P主要受N的限制；除营养元素以外羊草的生长速率还受其它土壤理化性质的影响。
Jiangjiadian meadow of Da’an city locates at the low plain in the middle of SongnenPlain. The entironment here is weaker. In recent years, influenced by global changing andhuman reason, the grassland area is reducing gradually; meanwhile, vegetation coveragelevel is depressing. Desertification, salinization and degeneration of the grassland aremore serious and its entironment is getting worse. All these have restricted thedeveloping of local society and economy. The research of grassland degenerationquestion in this area has important meaning to protecting grassland resource scientifically,to controlling grassland degeneration, and to repairing grassland entironment.
This paper takes Plant-Soil as a system to research. Based on ecologicalstoichiometry and element theory of constraints, using software such as SPSS and Matlab,I analyzed the interaction between the ecological chemical measurement features of C, Nand P in L. chinensis leaves and soil element and its chemical character. Using RBFArtificial Neural Networks (ANNs), I built the forecast model of the content of C,N andP elements in L. chinensis leaves.Using K value clustering analysis method, Ianalyze thedegeneration level of the grassland. After that, I analyze the soil chemistry character ofdifferent degeneration levels, and ecological chemical measurement features of L.chinensis leaves. The result of the research proves:
1. The contents of soil gein and TotalN in research area are all lower than nationalaverage. Compared with national second time soil survey standard, the contents of soilgein and TotalN belong to middle third level, and the content of soil TotalP belongs to extremely poor level (lower than0.4mg·g~(-1)). In the soil, the value of C/N, C/P and N/Pare9.81、41.72、4.39, all lower than national average. The contents of available N andavailable P are poor, available K is richer. There are both a soil salinization and alkalineexistence. Soluble salt and alkaline level show significant negative correlation with thecontent of soil nutrient elements.
2. The contents of C, N and P in L. chinensis leaves at Da’anJiangjiadian meadoware443.75mg·g~(-1)、18.70mg·g~(-1)、1.30mg·g~(-1). They are all lower than global averagecontents in terraneous plant leaves （464mg·g~(-1)、20.6mg·g~(-1)、2.0mg·g~(-1)）. And thecontent of N and P are even lower than national data. The space difference of CinL.chinensis leaves is obviously lower than N and P. The values of C/N, C/P and N/P inL.chinensis leaves are all higher than those in global terrestrial plant leaves. InL.chinensis leaves,14＜N/P＜16. The growth of L.chinensis in research area islimited by both N and P.The content of N and P show significant positive correlation; Thecontent of C inleaves show significant positive correlation with content of C, N, P andavailable P in soil. There is no significant correlation between the content of N in leavesand C, N, P in soil.
3. Based on the principle of ANNs, I take soil nutrient elements and relevantchemistry character as input layer. The contents of C, N and P in L. chinensis leaves areregarded as output layer. It’s feasible to create RBF ANNS model with Matlab and use itfor calculate the interaction between the contents of C, N and P in L. chinensis leaves andsoil chemical nature. After checking, the relative error is within20%. The forecastingaccuracy is higher.
4. Through K value clustering analysis method, using soil and plants as index, Idivide the85sampling points in research area into four types of degeneration level. Andconfirm the grassland degeneration level combined with the physical truth of field survey.The result shows that in research area, grassland of low-grade degeneration takes5.88%,mid-grade takes9.41%, high-grade takes35.29, and extremely high-grade takes49.41%.The regeneration of grassland in research area is severe.
5. With the aggravation of grassland degeneration degree, the contents of TotalC and TotalN get lower gradually, but the changing of TotalP content is smaller. As a whole,the values of C/N, C/P and N/P all show decreasing trends. In different degenerationlevels, the content sequence of available N in L. chinensis leaves is mid-grade＞low-grade＞high-grade＞extremely high-grade. The content of available P declinesgradually with the aggravation of grassland degeneration degree. And at the same time,the content of available K declines sharply from low-grade of degeneration（214.98mg·kg~(-1)）to mid-grade （144.63mg·kg~(-1)）, then remains comparatively steady.PH, total soluble salt and basification degree ratchet up with the aggravation of grasslanddegeneration. Grassland degradation degreeand soil salinity level condition is correlated.
6. In different degeneration levels, the content sequenceof C in L. chinensis leavesis mid-grade＞low-grade＞high-grade＞extremely high-grade.The content sequenceof Nis mid-grade＞extremely high-grade＞low-grade＞high-grade. Except extremelyhigh-grade degeneration stage, the variation trend of N in leaves is same as available N.The content of P in L. chinensis leaves shows decrease then increase. C/N value declineswavily. In low-grade degeneration stage, the growth of L. chinensis is limited both by Nand P. In mid-grade degeneration stage, it is limited mainly by P. In high-grade andextremely high-grade degeneration stages, it is limited mainly by N. Except nutrientelement, the growth speed of L. chinensis is still influenced by other soil physical andchemical properties.
引文
[1] Agren G I.The C∶N∶P stoichiometry of autotropHs-theory and observations [J]. EcologyLetters,2004,7(3):185-191.
[2] Ali Ghodsi, Dale Schurmans. Automatic basis selection techniques for RBF Networks[J].NeuralNetworks,2003,(16):809-816.
[3] C1ements F E.Plant Succession: An Analysis of the Development of Vegetation [M].Washington DC: Carnegie Institute of Washington,1916.242:512.
[4] Chaneton E J,Lemcoff J H,Lavado R S. Nitrogen and pHospHorus cycling in grazed and ungrazedplots in a temperate subhumid grassland in Argentina[J].Journal of AppliedEcology,1996,33:291-302.
[5] Cordell S, Goldstein G, Meinzer F C, et al. M orpHological and pHysiological adjustment to N andP fertilization in nutrient-limited M etrosideros polymorpHa canopy trees in Hawaii[J]. TreePHysiology,2001,21:43-50.
[6] Cordell S, Goldstein G, Meinzer F C, et al. Regulation of life-span and nutrient-use efficiency ofMetrosideros polym orpHa trees at two extremes of a long chronosequence[J]. Oecologia,2001,127:198-206.
[7] D.S. Broomhead, D. Lowe. Multivariable function interpolation and adaptive network[J].Complex Systems,1988(2):33-46.
[8] DeAngelis D.J. Energy flow, nutrient cycling and ecosystem resilience[J]. Ecology,1980,61,764-771.
[9] Douglas Turnbull, Charles Elkan. Fast Recognition of Musical Genres Using RBFNetworks[J].IEEE Transactions on Knowledge and Data Engineering,2005,17(4):580-584.
[10] Elser J J,Acharya K,Kyle M, et al..Growth rate-stoichiometry couplings in diverse biota.EcologyLetters,2003,6(10):936-943.
[11] Elser J J,Sterner R W,Gorokhova E et al. Biological stoichiometry from genes to ecosystems [J].Ecology Letters,2000,3,540-550.
[12] Elser J J,Bracken M E S,Cleland E E,et al..Global analysis of nitrogen and pHospHoruslimitation of primary producers in freshwater，marine and terrestrial ecosystems[J].EcologyLetters,2007,10(12):1135-1142.
[13] Elser J J,Fagan W F,Denno R F, et al. Nutritional constraints in terrestrial and freshwater foodwebs[J]. Nature,2000,408(6812):578-580.
[14] Gastal F, Lemaire G. N uptake and distribution in crops: An agronomical and ecopHysiologicalperspective [J].J. Exp. Bot.,2002,53:789-799.
[15] Güsewell S. N:P ratios in terrestrial plants: variation and functional significance [J]. NewpHytologist,2004,164（2）:243-266.
[16] Güsewell S, Koerselman W. Variation in nitrogen andpHospHorus concentrations ofwetlandplants.Perspectives inPlantEcology[J], Evolution and Systematics,2002,5:37-61.
[17] Haykin S. Neural Networks, a comprehensive foundation[M].USA: Prentice Hall,1999.
[18] H. M. Yao, H. B. Vuthaluru, M. O. Tade, D. D jukanovic.Artificial neural network-basedprediction of hydrogen content of coal in power station boilers[J].Fuel,2005,(84):1535-1542.
[19] Hopfield J. Neural Networks and PHysical Systems with Emergent collective ComputingAbilities [J]. Proc Nati. Acad Sci, USA,1982,79(4):2554-2558.
[20] J.A.Ludweig,et al.Principles，problems and priorities for restoring degraded angelands [J]. AustRangel.1990,12（1）,30-31.
[21] Lemaire G, Avice JC, Kim TH, et al. Developmental changes in shoot N dynamics of Lucerne（Medicago satova L.）in relation to leaf growth dynamics as a function of plant density andhierarchical position within the canopy[J].J. Exp.Bot.2005,56:935-943.
[22] Monisha Kaul, Robert L. Hill, Charles Walthall. Artificial neural networks for corn and soybeanyield prediction[J]. Agricultural systems,2005,(85):1-18.
[23] Q. Zhang, S. X. Yang, G. S. Mittal, et al. Prediction of performance indices and optimalparameters of rough rice drying with neural networks[J],Biosyst.Eng,2002,83(3)：281-290.
[24] Reich P B,Tjoelker M G,Machado J L,Oleksyn J.Universal scaling of respiratory metabolism,sizeand nitrogen in plants[J]. Nature,2006,439(7095):457-461.
[25] Reiners W A. Complementary models for ecosystems. American Naturalist,1986,127:59-73.
[26] Tessier J T,Raynal D J. Use of nitrogen to pHospHorus ratios in plant tissue as an indicator ofnutrient limitation and nitrogen saturation[J]. Journal of Applied Ecology,2003,40(3):523-534.
[27] Xiang W H,Huang Z H,Yan W D, et al. Review on coupling of interactive functions betweencarbon and nitrogen cycles in forestecosystems[J]. Acta Ecologica Sinica,2006,26(7):2365-2372.
[28] Zhang LX, Bai YF, Han XG. Differential responses of N: P stoichiometry of Leymus chinensisand Carex korshinskyi to N additions in a steppe ecosystem in Nei Mongol[J]. Acta BotanicaSinica,2004,46:259-270.
[29] Lichtenthaler H K,王可玢.以体内叶绿素荧光建材植物的逆境因素.生物技术通报,1992:1-4.
[30]曹云雷.大安市土地利用分区与安全格局构建[D].东北师范大学，2010.
[31]程滨，赵永军，张文广，等．生态化学计量学研究进展．生态学报，2010，30（6）：1628-1637.
[32]程建忠，李心清，刘钟龄，等．中国北方草地植物群落碳、氮元素组成空间变化及其与土壤地球化学变化的关系[J]．地球化学，2008，37（3）：265-274．
[33]崔大方．中国琵琶柴属分类、分布、生态和形态解剖学的初步研究[J]．干旱区研究，1988，（01）：65-69．
[34]丁凡，廉培勇，曾德慧.松嫩平原草甸三种植物叶片N、P化学计量特征及其与土壤N、P浓度的关系[J]．生态学杂志，2011，30（1）：77-81．
[35]戴雅婷，那日苏，吴洪新，等，赵海霞.我国北方温带草原碳循环研究进展[J]．草业科学，2009，26（9）：43-48．
[36]干文芝．雀儿山东北坡草地退化评价研究[D]．四川农业大学，2007．
[37]高三平．天童常绿阔叶林不同演替阶段N、P化学计量学研究[D].华东师范大学硕士学位论文.上海：华东师范大学，2008.
[38]葛琳.新疆和田绿洲土壤速效养分特征及其变化研究[D].新疆师范大学，2008.
[39]辜智慧，史培军，陈晋，等.基于植被气候最大响应模型的草地退化评价[J]．自然灾害学报，2010,19（1）：13-20．
[40]韩文轩，吴漪，汤璐瑛，等.北京及周边地区植物叶的碳氮磷元素计量特征[J].北京大学学报（自然科学版），2008，（4）：67-72.
[41]黄葆宁，李希来.高寒草甸退化草地黑土滩植物量测定[J].青海畜牧兽医杂志，1993,23（4）：10-13.
[42]黄昌勇，李保国，潘兴根，等.土壤学[M].北京：中国农业出版社，2000.
[43]黄国如，胡和平，田富强.用径向基函数神经网络模型预报感潮河段洪水位[J].水科学进展，2003,14（2）：158-162.
[44]黄建军，王希华.浙江天童32种常绿阔叶树叶片的营养及结构特征[J].华东师范大学学报（自然科学版），2003，（01）：92-97.
[45]黄丽华，梁正伟，马红媛.苏打盐碱胁迫对羊草光合、蒸腾速率及水分利用效率的影响[J].草业学报，2009,18（5）：25-30.
[46]呼天明，王培，姚爱兴，等.多年生黑麦草/白三叶人工操死的放牧演替及群落稳定性的研究[J].草地学报，1995,3（2）：152-157.
[47]姜玲玲，林年丰，唐晓慧等.吉林西部生态环境研究及质量评价[J].干旱区研究，2005,22（2）：246-250.
[48]芦清水，赵志平.应对草地退化的生态移民政策及牧户影响分析基于黄河源区玛多县的牧户调查[J].地理研究，2009，28（1）：144-152.
[49]卢文喜，马洪云，杨忠平.吉林西部退化草地不同修复措施的效果分析[J].云南农业大学学报，2005，20（1）：111-113.
[50]林年丰，汤洁.松嫩平原环境演变与土地盐碱化、荒漠化的成因分析[J].第四纪研究，2005，25（4）：474-483.
[51]林年丰，汤洁，王娟，等.松嫩平原西部生态环境安全研究[J].干旱区研究，2007,26（4）：867-874.
[52]梁娜.基于神经网络与主成分分析的组合预测研究[D].2007.武汉理工大学.
[53]吕贻忠，马兴旺.荒漠化土壤养分变化的影响因素研究进展[J].生态环境，2003，12（4）：473-477.
[54]李博.中国北方草地退化及其防治对策[J].中国农业科学，1997，30（6）：1-9.
[55]李彬，王志春.松嫩平原苏打盐渍土碱化特征与影响因素[J].干旱区资源与环境.2006，(6):183-191.
[56]李建东,郑慧莹.松嫩平原盐碱化草地治理及其生态机理[M].北京:科学出版社,1997,271.
[57]李绍良，陈有君，关世英，等．土壤退化与草地退化关系的研究[J]．干旱区资源与环境，2002,16（1）：92-95．
[58]刘万德，苏建荣，李帅锋，等.云南普洱疾风常绿阔叶林演替系列之五和土壤C、N、P化学计量特征[J].生态学报，2010,30（23）：6581-6590.
[59]楼文高.基于人工神经网络的三江平原土壤质量综合评价与预测模型[J]．中国管理科学，2002,10（1）：79-83．
[60]鲁如坤.土壤农业化学分析方法[M].北京：中国农业科技出版社，2000.
[61]毛子龙.1890~2029年白城市土地利用/覆被变化与土壤碳库研究[D].吉林：吉林大学环境与资源学院，2010.
[62]潘文杰，Sim on X. Yang，唐远驹，等.烤烟锌含量的预测模型研究[A].中国烟草会2006年学术年会论文集[C].2007.
[63]朴河春，刘丛强，朱书法，等.贵州石灰岩和砂岩地区C4和C3植物营养元素的化学计量对N/P比值波动的影响[J].第四纪研究，2005，25（5）：552-560.
[64]苏凤国.经济发展概况[M].中国城市年鉴,2010:629.
[65]王明君，韩国栋，赵萌莉，等.内蒙古呼伦贝尔草甸草原的草地退化等级数量分析[J].西北植物学报，2007，27（4）：0797-0804.
[66]王建安，韩国栋，鲍雅静，等.我国草地生态系统碳氮循环研究概述[J].内蒙古农业大学学报，2007，28（4）：254-258.
[67]王绍强，于贵瑞.生态系统碳氮磷元素的生态化学计量学特征[J].生态学报，2008，28（8）：3937-3947.
[68]王琪，徐程杨.氮磷对植物光合作用及碳分配的影响[J].山东林业科技，2005，5:59-62.
[69]王彩虹.不同退化阶段下伊犁绢蒿荒漠草地土壤养分变化分析[D]新疆农业大学，2007.
[70]王维奇，仝川，贾瑞霞，等.不同淹水频率下湿地土壤碳氮磷生态化学计量学特征[J].水土保持学报,2010,24（3）：238-242.
[71]王维奇，仝川，曾从盛.不同质地实地土壤碳、氮、磷计量学及厌氧碳分解特征[J].中国环境科学,2010,30（10）：1369-1374.
[72]吴统贵，吴明，刘丽，等.杭州湾滨海湿地3种草本植物叶片N、P化学计量学的季节变化[J].植物生态学报，2010，34（1）：23-28.
[73]杨国栋，王肖娟.基于人工神经网络的土壤养分肥力等级评价方法[J]．土壤通报，2005,36（1）：30-33．
[74]袁曾任.人工神经元网络及其应用[M].北京:清华大学出版社，1999.10
[75]宰松梅，郭冬冬，温季.人工神经网络在土壤含盐量预测中的应用[J]．中国农村水利水电，2010,10：33-35．
[76]曾慧德，陈广生.生态化学计量学：复杂生命系统奥秘的探索[J].植物生态学报，2005,29（6）：1007-1019.
[77]郑淑华，蔡瑞磊，赵萌莉，等.贝加尔针茅+线叶菊草地退化系列的定量划分和动态模型的建立[J].干旱区研究，2008,25（6）：766-771.
[78]郑淑华，赵萌莉，韩国栋，等.不同放牧压力下典型草原土壤物理性质与植被关系的研究[J]．干旱区资源与环境，2005,19（7）：199-203．
[79]郑淑华.不同放牧强度下羊草草原生态系统服务功能价值评估[D]．兰州大学，2006．
[80]张丽霞.植物N：P计量化学：中国高等植物的分异规律与野外实验初步验证[D].中国科学院博士研究生学位论文.北京：中国科学院植物研究所，2003.
[81]张立明.人工神经网络模型及其应用[M].上海:复旦大学出版社1992，7.
[82]张久明，宿庆瑞，迟凤琴，等.嫩江县土壤速效养分差异性分析[J].黑龙江农业科学，2010，(3)：38-41.
[83]张文彦，樊江文，钟华平，等.中国典型草原优势植物功能群氮磷化学计量学特征研究[J].草地学报，2010，18（4）：503-509.
[84]钟诚，何晓蓉，李辉霞．遥感技术在西藏那曲地区草地退化评价中的应用[J]．遥感技术与应用，2003,18（2）：99-102．
[85]中国环境监测总站.中国土壤元素背景值[M].北京：中国环境科学出版社，1990.
[86]赵彦锋，史学正，于东升，等.小尺度土壤养分的空间变异及其影响因素探讨以江苏省无锡市典型城乡交错区为例[J].土壤通报，2006,37（2）：214-219.
[87]赵志红.半干旱黄土区集雨措施和养分添加对苜蓿草地和封育植被生产力及土壤生态化学计量特征的影响[D].兰州大学，2010.
[2] Ali Ghodsi, Dale Schurmans. Automatic basis selection techniques for RBF Networks[J].NeuralNetworks,2003,(16):809-816.
[3] C1ements F E.Plant Succession: An Analysis of the Development of Vegetation [M].Washington DC: Carnegie Institute of Washington,1916.242:512.
[4] Chaneton E J,Lemcoff J H,Lavado R S. Nitrogen and pHospHorus cycling in grazed and ungrazedplots in a temperate subhumid grassland in Argentina[J].Journal of AppliedEcology,1996,33:291-302.
[5] Cordell S, Goldstein G, Meinzer F C, et al. M orpHological and pHysiological adjustment to N andP fertilization in nutrient-limited M etrosideros polymorpHa canopy trees in Hawaii[J]. TreePHysiology,2001,21:43-50.
[6] Cordell S, Goldstein G, Meinzer F C, et al. Regulation of life-span and nutrient-use efficiency ofMetrosideros polym orpHa trees at two extremes of a long chronosequence[J]. Oecologia,2001,127:198-206.
[7] D.S. Broomhead, D. Lowe. Multivariable function interpolation and adaptive network[J].Complex Systems,1988(2):33-46.
[8] DeAngelis D.J. Energy flow, nutrient cycling and ecosystem resilience[J]. Ecology,1980,61,764-771.
[9] Douglas Turnbull, Charles Elkan. Fast Recognition of Musical Genres Using RBFNetworks[J].IEEE Transactions on Knowledge and Data Engineering,2005,17(4):580-584.
[10] Elser J J,Acharya K,Kyle M, et al..Growth rate-stoichiometry couplings in diverse biota.EcologyLetters,2003,6(10):936-943.
[11] Elser J J,Sterner R W,Gorokhova E et al. Biological stoichiometry from genes to ecosystems [J].Ecology Letters,2000,3,540-550.
[12] Elser J J,Bracken M E S,Cleland E E,et al..Global analysis of nitrogen and pHospHoruslimitation of primary producers in freshwater，marine and terrestrial ecosystems[J].EcologyLetters,2007,10(12):1135-1142.
[13] Elser J J,Fagan W F,Denno R F, et al. Nutritional constraints in terrestrial and freshwater foodwebs[J]. Nature,2000,408(6812):578-580.
[14] Gastal F, Lemaire G. N uptake and distribution in crops: An agronomical and ecopHysiologicalperspective [J].J. Exp. Bot.,2002,53:789-799.
[15] Güsewell S. N:P ratios in terrestrial plants: variation and functional significance [J]. NewpHytologist,2004,164（2）:243-266.
[16] Güsewell S, Koerselman W. Variation in nitrogen andpHospHorus concentrations ofwetlandplants.Perspectives inPlantEcology[J], Evolution and Systematics,2002,5:37-61.
[17] Haykin S. Neural Networks, a comprehensive foundation[M].USA: Prentice Hall,1999.
[18] H. M. Yao, H. B. Vuthaluru, M. O. Tade, D. D jukanovic.Artificial neural network-basedprediction of hydrogen content of coal in power station boilers[J].Fuel,2005,(84):1535-1542.
[19] Hopfield J. Neural Networks and PHysical Systems with Emergent collective ComputingAbilities [J]. Proc Nati. Acad Sci, USA,1982,79(4):2554-2558.
[20] J.A.Ludweig,et al.Principles，problems and priorities for restoring degraded angelands [J]. AustRangel.1990,12（1）,30-31.
[21] Lemaire G, Avice JC, Kim TH, et al. Developmental changes in shoot N dynamics of Lucerne（Medicago satova L.）in relation to leaf growth dynamics as a function of plant density andhierarchical position within the canopy[J].J. Exp.Bot.2005,56:935-943.
[22] Monisha Kaul, Robert L. Hill, Charles Walthall. Artificial neural networks for corn and soybeanyield prediction[J]. Agricultural systems,2005,(85):1-18.
[23] Q. Zhang, S. X. Yang, G. S. Mittal, et al. Prediction of performance indices and optimalparameters of rough rice drying with neural networks[J],Biosyst.Eng,2002,83(3)：281-290.
[24] Reich P B,Tjoelker M G,Machado J L,Oleksyn J.Universal scaling of respiratory metabolism,sizeand nitrogen in plants[J]. Nature,2006,439(7095):457-461.
[25] Reiners W A. Complementary models for ecosystems. American Naturalist,1986,127:59-73.
[26] Tessier J T,Raynal D J. Use of nitrogen to pHospHorus ratios in plant tissue as an indicator ofnutrient limitation and nitrogen saturation[J]. Journal of Applied Ecology,2003,40(3):523-534.
[27] Xiang W H,Huang Z H,Yan W D, et al. Review on coupling of interactive functions betweencarbon and nitrogen cycles in forestecosystems[J]. Acta Ecologica Sinica,2006,26(7):2365-2372.
[28] Zhang LX, Bai YF, Han XG. Differential responses of N: P stoichiometry of Leymus chinensisand Carex korshinskyi to N additions in a steppe ecosystem in Nei Mongol[J]. Acta BotanicaSinica,2004,46:259-270.
[29] Lichtenthaler H K,王可玢.以体内叶绿素荧光建材植物的逆境因素.生物技术通报,1992:1-4.
[30]曹云雷.大安市土地利用分区与安全格局构建[D].东北师范大学，2010.
[31]程滨，赵永军，张文广，等．生态化学计量学研究进展．生态学报，2010，30（6）：1628-1637.
[32]程建忠，李心清，刘钟龄，等．中国北方草地植物群落碳、氮元素组成空间变化及其与土壤地球化学变化的关系[J]．地球化学，2008，37（3）：265-274．
[33]崔大方．中国琵琶柴属分类、分布、生态和形态解剖学的初步研究[J]．干旱区研究，1988，（01）：65-69．
[34]丁凡，廉培勇，曾德慧.松嫩平原草甸三种植物叶片N、P化学计量特征及其与土壤N、P浓度的关系[J]．生态学杂志，2011，30（1）：77-81．
[35]戴雅婷，那日苏，吴洪新，等，赵海霞.我国北方温带草原碳循环研究进展[J]．草业科学，2009，26（9）：43-48．
[36]干文芝．雀儿山东北坡草地退化评价研究[D]．四川农业大学，2007．
[37]高三平．天童常绿阔叶林不同演替阶段N、P化学计量学研究[D].华东师范大学硕士学位论文.上海：华东师范大学，2008.
[38]葛琳.新疆和田绿洲土壤速效养分特征及其变化研究[D].新疆师范大学，2008.
[39]辜智慧，史培军，陈晋，等.基于植被气候最大响应模型的草地退化评价[J]．自然灾害学报，2010,19（1）：13-20．
[40]韩文轩，吴漪，汤璐瑛，等.北京及周边地区植物叶的碳氮磷元素计量特征[J].北京大学学报（自然科学版），2008，（4）：67-72.
[41]黄葆宁，李希来.高寒草甸退化草地黑土滩植物量测定[J].青海畜牧兽医杂志，1993,23（4）：10-13.
[42]黄昌勇，李保国，潘兴根，等.土壤学[M].北京：中国农业出版社，2000.
[43]黄国如，胡和平，田富强.用径向基函数神经网络模型预报感潮河段洪水位[J].水科学进展，2003,14（2）：158-162.
[44]黄建军，王希华.浙江天童32种常绿阔叶树叶片的营养及结构特征[J].华东师范大学学报（自然科学版），2003，（01）：92-97.
[45]黄丽华，梁正伟，马红媛.苏打盐碱胁迫对羊草光合、蒸腾速率及水分利用效率的影响[J].草业学报，2009,18（5）：25-30.
[46]呼天明，王培，姚爱兴，等.多年生黑麦草/白三叶人工操死的放牧演替及群落稳定性的研究[J].草地学报，1995,3（2）：152-157.
[47]姜玲玲，林年丰，唐晓慧等.吉林西部生态环境研究及质量评价[J].干旱区研究，2005,22（2）：246-250.
[48]芦清水，赵志平.应对草地退化的生态移民政策及牧户影响分析基于黄河源区玛多县的牧户调查[J].地理研究，2009，28（1）：144-152.
[49]卢文喜，马洪云，杨忠平.吉林西部退化草地不同修复措施的效果分析[J].云南农业大学学报，2005，20（1）：111-113.
[50]林年丰，汤洁.松嫩平原环境演变与土地盐碱化、荒漠化的成因分析[J].第四纪研究，2005，25（4）：474-483.
[51]林年丰，汤洁，王娟，等.松嫩平原西部生态环境安全研究[J].干旱区研究，2007,26（4）：867-874.
[52]梁娜.基于神经网络与主成分分析的组合预测研究[D].2007.武汉理工大学.
[53]吕贻忠，马兴旺.荒漠化土壤养分变化的影响因素研究进展[J].生态环境，2003，12（4）：473-477.
[54]李博.中国北方草地退化及其防治对策[J].中国农业科学，1997，30（6）：1-9.
[55]李彬，王志春.松嫩平原苏打盐渍土碱化特征与影响因素[J].干旱区资源与环境.2006，(6):183-191.
[56]李建东,郑慧莹.松嫩平原盐碱化草地治理及其生态机理[M].北京:科学出版社,1997,271.
[57]李绍良，陈有君，关世英，等．土壤退化与草地退化关系的研究[J]．干旱区资源与环境，2002,16（1）：92-95．
[58]刘万德，苏建荣，李帅锋，等.云南普洱疾风常绿阔叶林演替系列之五和土壤C、N、P化学计量特征[J].生态学报，2010,30（23）：6581-6590.
[59]楼文高.基于人工神经网络的三江平原土壤质量综合评价与预测模型[J]．中国管理科学，2002,10（1）：79-83．
[60]鲁如坤.土壤农业化学分析方法[M].北京：中国农业科技出版社，2000.
[61]毛子龙.1890~2029年白城市土地利用/覆被变化与土壤碳库研究[D].吉林：吉林大学环境与资源学院，2010.
[62]潘文杰，Sim on X. Yang，唐远驹，等.烤烟锌含量的预测模型研究[A].中国烟草会2006年学术年会论文集[C].2007.
[63]朴河春，刘丛强，朱书法，等.贵州石灰岩和砂岩地区C4和C3植物营养元素的化学计量对N/P比值波动的影响[J].第四纪研究，2005，25（5）：552-560.
[64]苏凤国.经济发展概况[M].中国城市年鉴,2010:629.
[65]王明君，韩国栋，赵萌莉，等.内蒙古呼伦贝尔草甸草原的草地退化等级数量分析[J].西北植物学报，2007，27（4）：0797-0804.
[66]王建安，韩国栋，鲍雅静，等.我国草地生态系统碳氮循环研究概述[J].内蒙古农业大学学报，2007，28（4）：254-258.
[67]王绍强，于贵瑞.生态系统碳氮磷元素的生态化学计量学特征[J].生态学报，2008，28（8）：3937-3947.
[68]王琪，徐程杨.氮磷对植物光合作用及碳分配的影响[J].山东林业科技，2005，5:59-62.
[69]王彩虹.不同退化阶段下伊犁绢蒿荒漠草地土壤养分变化分析[D]新疆农业大学，2007.
[70]王维奇，仝川，贾瑞霞，等.不同淹水频率下湿地土壤碳氮磷生态化学计量学特征[J].水土保持学报,2010,24（3）：238-242.
[71]王维奇，仝川，曾从盛.不同质地实地土壤碳、氮、磷计量学及厌氧碳分解特征[J].中国环境科学,2010,30（10）：1369-1374.
[72]吴统贵，吴明，刘丽，等.杭州湾滨海湿地3种草本植物叶片N、P化学计量学的季节变化[J].植物生态学报，2010，34（1）：23-28.
[73]杨国栋，王肖娟.基于人工神经网络的土壤养分肥力等级评价方法[J]．土壤通报，2005,36（1）：30-33．
[74]袁曾任.人工神经元网络及其应用[M].北京:清华大学出版社，1999.10
[75]宰松梅，郭冬冬，温季.人工神经网络在土壤含盐量预测中的应用[J]．中国农村水利水电，2010,10：33-35．
[76]曾慧德，陈广生.生态化学计量学：复杂生命系统奥秘的探索[J].植物生态学报，2005,29（6）：1007-1019.
[77]郑淑华，蔡瑞磊，赵萌莉，等.贝加尔针茅+线叶菊草地退化系列的定量划分和动态模型的建立[J].干旱区研究，2008,25（6）：766-771.
[78]郑淑华，赵萌莉，韩国栋，等.不同放牧压力下典型草原土壤物理性质与植被关系的研究[J]．干旱区资源与环境，2005,19（7）：199-203．
[79]郑淑华.不同放牧强度下羊草草原生态系统服务功能价值评估[D]．兰州大学，2006．
[80]张丽霞.植物N：P计量化学：中国高等植物的分异规律与野外实验初步验证[D].中国科学院博士研究生学位论文.北京：中国科学院植物研究所，2003.
[81]张立明.人工神经网络模型及其应用[M].上海:复旦大学出版社1992，7.
[82]张久明，宿庆瑞，迟凤琴，等.嫩江县土壤速效养分差异性分析[J].黑龙江农业科学，2010，(3)：38-41.
[83]张文彦，樊江文，钟华平，等.中国典型草原优势植物功能群氮磷化学计量学特征研究[J].草地学报，2010，18（4）：503-509.
[84]钟诚，何晓蓉，李辉霞．遥感技术在西藏那曲地区草地退化评价中的应用[J]．遥感技术与应用，2003,18（2）：99-102．
[85]中国环境监测总站.中国土壤元素背景值[M].北京：中国环境科学出版社，1990.
[86]赵彦锋，史学正，于东升，等.小尺度土壤养分的空间变异及其影响因素探讨以江苏省无锡市典型城乡交错区为例[J].土壤通报，2006,37（2）：214-219.
[87]赵志红.半干旱黄土区集雨措施和养分添加对苜蓿草地和封育植被生产力及土壤生态化学计量特征的影响[D].兰州大学，2010.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |