大熊猫干扰下巴山木竹无性系种群生态学研究

【作者】 王逸之；

【导师】 董文渊；

【作者基本信息】 南京林业大学 ， 森林培育， 2013， 博士

【摘要】 巴山木竹(Bashania fargesii)为地下茎复轴混生的中小型竹种，分布区以大巴山脉和秦岭为主，布及川、甘、陕、豫、鄂、渝等六省市，跨暖温带、北亚热带、中亚热带三个气候带，是大熊猫的主食竹种之一。本文选取了陕西佛坪国家级自然保护区核心区大熊猫活动密集斑块内巴山木竹无性系种群为研究对象，同时对比了保护状态下的巴山木竹无性系种群，从生物学特性以及近红外光谱技术应用等几方面系统研究了大熊猫种群活动对巴山木竹无性系种群造成的影响，结果如下：1.巴山木竹发笋存在较为明显的大小年现象，线性回归分析显示大熊猫取食强度与巴山木竹发笋量有显著相关性。干扰样地与非干扰样地巴山木竹竹笋的成竹率与退笋率分别为54.09%，45.91%；63.46%，36.54%。非干扰样地与干扰样地的巴山木竹竹笋出土后完成高生长大约需要60d和65d。非干扰样地幼竹高度测量值与总体平均值均高于干扰样地。2.干扰样地与非干扰样地各年龄段的巴山木竹生物量总量大小依次为：2年生>4年生>3年生>1年生>5年生及以上；2年生>1年生>3年生>4年生>5年生及以上。干扰样地与非干扰样地巴山木竹无性系种群生物量总量差别不大，但地上与地下部分所占比例差异明显，分别为60.74%，39.26%；56.19%，43.81%。3.非干扰样地巴山木竹无性系种群各年龄段所占比例逐年递减，分布有序；干扰样地的巴山木竹无性系种群年龄结构紊乱，没有表现出稳定的金字塔结构。干扰样地的巴山木竹无性系种群，87.20%的分株高度小于500cm；70.82%的分株胸径小于1.70cm。与之相比非干扰样地巴山木竹无性系种群，65.77%的分株高于500cm；64.43%的分株直径大于1.70cm。干扰样地与非干扰样地巴山木竹无性系种群静态生命表生存曲线分别接近于B型(直线型B2)与C型(凹型)。4.巴山木竹竹笋随着出土时间的增长，灰分、粗纤维的含量在逐步增加，而粗脂肪、粗蛋白与氨基酸的含量在逐渐减少。巴山木竹叶中Ca、灰分和粗纤维在一年生叶片中含量最低，随龄级增加呈现递增趋势。P、K、Zn和粗蛋白含量随龄级增长呈现依次递减趋势。Fe和Mg在第二年的叶片中含量最低，一年生叶片中含量最高。粗脂肪含量随叶片龄级别变化不明显。巴山木竹竹秆营养成分随着龄级的不同呈现出不同的变化趋势：Ca、Mg与粗纤维含量随着竹秆龄级增加而增大；K、Zn、粗脂肪和粗蛋白含量，随着龄级增加而减少；Fe和Zn含量随着龄级变化不明显。不同大熊猫取食阶段巴山木竹营养成分比较结果表明：在各个取食阶段，大熊猫主要选择可利用营养最为丰富巴山木竹部位作为主要摄食对象；在取食竹秆阶段，虽然竹秆中营养成分含量低于竹叶，但竹秆中粗纤维含量为全年最低，这可能是大熊猫在该阶段偏好取食竹秆的主导因素。5.应用近红外光谱结合PLS-DA方法对校正集样本建立的判别模型，使用竹叶正面光谱建模效果较好，校正和验证结果与实际分变量的相关系数均超过了0.96，交叉验证均方根误差(RMSECV)和预测均方根误差(RMSEP)都低于0.14，利用模型对验证集中4个竹种叶片正面近红外光谱进行判别，识别率均为100%。竹叶样品蛋白质、氨基酸、粗脂肪与粗纤维最优校正模型的均方差分别为0.351、0.279、0.221、1.603，校正决定系数分别为0.935、0.854、0.917、0.972。使用验证集对蛋白质、氨基酸、粗脂肪与粗纤维最优校正模型验证结果：相对分析误差分别为3.562、2.468、2.671、4.247，预测决定系数分别为0.916、0.734、0.816、0.962，预测标准差分别为0.835、0.637、0.228、0.249。验证结果表明，除氨基酸最优校正模型相对分析误差小于2.5，精度较低之外，其余各最优校正模型精度较好，可以用于巴山木竹竹叶营养成分的实际检测。竹秆样品蛋白质、氨基酸、粗脂肪与粗纤维最优校正模型的的均方差分别为0.172、0.153、0.095、2.550，校正决定系数分别为0.862、0.795、0.883、0.965。使用验证集对蛋白质、氨基酸、粗脂肪与粗纤维最优校正模型验证结果：相对分析误差分别为2.840、2.126、2.285、3.864，预测决定系数分别为0.874、0.720、0.735、0.925，预测标准差分别为0.124、0.162、0.433、0.291。验证结果表明，竹秆氨基酸与粗脂肪最优校正模型相对分析误差小于2.5，精度较低，只能用于估测，蛋白质与粗纤维最优校正模型精度较好，可以用于巴山木竹竹秆营养成分的实际检测。更多还原

【Abstract】 Bashania fargesii is a kind of small and medium-sized amphipodial bamboo species. Thedistribution area is main in Daba mountain range and Qinling mountains such as Sichuan,Gansu, Shanxi, Henan, Hubei and Chongqing and cross warm temperate zone, north subtropicalzone and mid-subtropical zone triple climatic zones B. fargesii is one of bamboo species as themain food of giant panda. B. fargesii clone population under giant panda disturbance and undercontrol in Shanxi Foping national nature reserve were both chosen as the research object. Thebiological characteristics and NIRS technology were investigated for the comprehensive studyof the effects of B. fargesii clone population under giant panda disturbance.The main experimental results are as follows:1. Unstable yielding phenomenon in B. fargesii shooting characteristics was existed. Theresult of linear regression analysis showed that there was significant correlation between theintensity of giant panda foraging behavior and the bamboo shoot emerging amount. The maturebamboo rate and degenerated shoot rate of disturbance and control plots were54.09%，45.91%；63.46%，36.54%. The bamboo shoots of control and disturbance plots needed60d and65d tofinish the height growth stage. The data showed that the average height of bamboo in controlplots were higher than those in disturbance plots.2. The total biomass of B. fargesii of different ages in disturbance and control plots were inthe order of2>4>3>1>5,2>1>3>4>5. There was no significant different between the totalbiomass of B. fargesii in disturbance and control plots. But the biomass ratio of over-groundand underground in disturbance and control plots showed great difference, were60.74%，39.26%；56.19%，43.81%respectively.3. The distribution of age structure of B. fargesii age in control plots was ordered. Instead,the distribution of age structure of B. fargesii age in disturbance plots was unordered.87.20%ramet’s height of the clone population of B. fargesii in disturbance plots was lower than500cmand70.82%ramet’s diameter was lower than1.70cm. on the other hand,65.77%ramet’s heightof the clone population of B. fargesii in control plots was higher than500cm and64.43%ramet’s diameter was higher than1.70cm. The survival curve of static life table of B. fargesiiclone population in disturbance and control plots were B type and C type respectively.4. The contents of ash and crude fiber in B. fargesii shoots increased with increased time.In contrast, the contents of crude protein and crude fat in Bashania fargesii shoots decreasedwith increased time.The contents of Ca、ash and crude fiber was lowest in one year old leaves and increasedwith increased age classes. P、K、Zn and crude protein decreased with increased leaf age classes.The contents of Fe and Mg were lowest in two years old leaves and highest in one year oldleaves. There was no significant difference between age classes of crude fat contents in leaves.The nutrient contents of B. fargesii shoots culms showed different trends with different age classes. The contents of Ca、Mg and crude fiber increased with increased culm age classes. Thecontents of K、Zn、crude fat and crude protein decreased with increased culm age classes. Therewas no significant difference between age classes of Fe and Zn contents in culms.The comparison results of B. fargesii nutrient contents among different giant pandaforaging stages showed that: giant panda mostly chose the part of B. fargesii which had highernutrient contents as the food during different foraging stages. During the culm stage, the crudefiber of culms was lowest in the year. This result showed that the content of crude fiber mightbe the dominant factor which influences the foraging choice of giant panda.5. Near-infrared spectroscopy (NIRS) coupled with Partial Least Squares-DiscriminantAnalysis (PLS-DA) pattern recognition methods were used to identify different bamboo species.The results showed that the correlations between the predicted category variables of calibrationor validation and the measured category variables were all remarkable with a correlationcoefficient(R) over0.96and low RMSECV and RMSEP (<0.14) used the top near-infraredspectroscopy of bamboo leaves. The discrimination accuracy for the bamboo varieties was100％by PLS-DA model based on the validation set of samples.SECV and rcof the optimal calibration models for crude protein, amino acid, crude fat andcrude fiber were0.351,0.279,0.221,1.603and0.935,0.854,0.917,0.972respectively. And thevalidation results were that, RPD were3.562,2.468,2.671,4.247; rpwere0.916,0.734,0.816,0.962; SEP was0.835,0.637,0.228, and0.249. The verification results showed that the RPDvalue were all higher than2.5except the value of amino acid. The optimal calibration model ofcrude protein, crude fat and crude fiber had good accuracy could be used in actual test of B.fargesii leaves nutrient contents..SECV and rcof the optimal calibration models for crude protein, amino acid, crude fat andcrude fiber were0.172,0.153,0.095,2.550and0.862,0.795,0.883,0.965respectively. And thevalidation results were that, RPD were2.840,2.126,2.285,3.864; rpwere0.874,0.720,0.735,0.925; SEP was0.124,0.162,0.433, and0.291. The verification results showed that the RPDvalue of crude protein and crude fiber were higher2.5but the RPD value of amino acid andcrude fat were lower than2.5. The optimal calibration model of crude protein and crude fiberhad better accuracy could be used in actual test of B. fargesii culms nutrient contents.更多还原