發(fā)布時(shí)間：2018-03-08 06:38

  本文選題：丹頂鶴　切入點(diǎn)：生境選擇　出處：《南京林業(yè)大學(xué)》2015年碩士論文　論文類型：學(xué)位論文

【摘要】：識別影響瀕危物種生境選擇的關(guān)鍵環(huán)境變量是進(jìn)行物種生境保護(hù)工作的前提。然而環(huán)境變量存在于不同的空間尺度上,只選取單一尺度的環(huán)境變量不能全面反映物種的生境特征;另外,環(huán)境變量存在的空間自相關(guān)現(xiàn)象(空間上越靠近越相似)和多元共線性問題(變量之間線性相關(guān))分別違背了統(tǒng)計(jì)檢驗(yàn)中樣本點(diǎn)之間相互獨(dú)立的基本假設(shè)以及自變量之間相互獨(dú)立的基本假設(shè),這些問題可能會導(dǎo)致推斷出的影響物種生境選擇的環(huán)境變量不夠準(zhǔn)確。為解決尺度、空間自相關(guān)和多元共線性這3個(gè)問題,本文以鹽城濕地珍禽國家級自然保護(hù)區(qū)為研究地,以越冬期的丹頂鶴為研究對象,結(jié)合野外調(diào)查和遙感信息采集微生境和景觀2種尺度上的環(huán)境數(shù)據(jù),同時(shí)在環(huán)境變量中加入空間變量以描述空間自相關(guān)現(xiàn)象,將環(huán)境變量按照2個(gè)等級水平進(jìn)行劃分,在第一等級水平上將所有環(huán)境變量分為微生境、景觀和空間3組變量,在第二等級水平上將微生境變量分為植被因子和遮蔽物因子,將景觀變量分為植被因子和干擾因子,然后利用方差分解法計(jì)算第一等級水平上微生境、景觀和空間的獨(dú)立效應(yīng)和聯(lián)合效應(yīng),以及第二等級水平上各生境因子對丹頂鶴分布變異的獨(dú)立解釋及聯(lián)合解釋,最后利用等級劃分法驗(yàn)證方差分解法的結(jié)果,并且計(jì)算各環(huán)境變量對丹頂鶴分布變異的獨(dú)立貢獻(xiàn)。通過研究,主要得到以下結(jié)果:(1)第一等級水平的方差分解結(jié)果表明:所有環(huán)境變量共解釋了67.4%的丹頂鶴分布變異。景觀變量總共解釋了52.9%的丹頂鶴分布變異,微生境變量總共解釋了32.2%的丹頂鶴分布變異,空間變量總共解釋了33.7%的丹頂鶴分布變異。微生境、景觀和空間3組變量的獨(dú)立解釋都很小,分別獨(dú)立解釋了7.0%、4.5%和5.7%的丹頂鶴分布變異,而它們聯(lián)合解釋了50.2%的丹頂鶴分布變異,其中景觀變量與微生境變量、空間變量分別聯(lián)合解釋了22.2%和25.0%的丹頂鶴分布變異。(2)第二等級水平的方差分解結(jié)果表明:微生境變量對丹頂鶴分布變異總的解釋方差比例(32.2%)中,遮蔽物因子和植被因子分別獨(dú)立解釋了9.8%和8.6%的丹頂鶴分布變異,聯(lián)合解釋了13.8%的丹頂鶴分布變異。景觀變量對丹頂鶴分布變異總的解釋方差比例(52.9%)中,干擾因子和植被因子分別獨(dú)立解釋了17.4%和39.9%的丹頂鶴分布變異,聯(lián)合解釋了-4.4%的丹頂鶴分布變異。(3)等級劃分結(jié)果表明:微生境變量、景觀變量和空間變量分別獨(dú)立解釋了9.1%、14.9%、3.1%的丹頂鶴分布變異。景觀變量和微生境變量中的植被因子分別解釋了7.1%和6.1%的丹頂鶴分布變異,高于其它生境因子的解釋量。500 hm2范圍內(nèi)的堿蓬面積百分比、50 hm2范圍內(nèi)的水源面積百分比、遮蔽物距離對丹頂鶴分布變異的獨(dú)立解釋最高,它們分別解釋了4.12%、3.34%和3.04%的丹頂鶴分布變異。研究結(jié)果表明:景觀尺度是影響丹頂鶴生境選擇的最重要尺度。景觀尺度上的植被因子是影響丹頂鶴生境選擇的最重要生境因子。堿蓬面積是影響丹頂鶴生境選擇的最重要環(huán)境變量�？臻g變量與景觀變量的聯(lián)合解釋很高(25.0%),表明丹頂鶴分布的空間結(jié)構(gòu)與景觀變量的空間結(jié)構(gòu)有很強(qiáng)的相關(guān)性,因此景觀生境因子的空間分布特征能夠準(zhǔn)確反映丹頂鶴喜好或厭惡的生境類型。方差分解法和等級劃分法對丹頂鶴分布變異的分解結(jié)果基本一致,都得出景觀尺度和植被因子是限制丹頂鶴分布的最重要尺度和生境因子。
[Abstract]:Identifying key environmental variables selection of endangered species habitat is a prerequisite for habitat protection. However, environmental variables exist in different spatial scales, environmental variables can not only choose a single scale fully reflect the habitat characteristics of species; in addition, the environmental variables are the spatial autocorrelation phenomenon (space closer and more similar) multicollinearity problem (linear correlation between variables) are contrary to the basic assumption between the sample points in the statistic test independent basic assumptions and variables independent of each other, these problems will lead to environmental variables of species habitat selection effect inferred is not accurate enough. In order to solve the problem of scale, spatial autocorrelation and multicollinearity of the 3, taking the Yancheng wetland national reserve for the study, to the wintering Red Crowned Crane as the research object, combined with the field The environmental survey data and remote sensing information collection and microhabitat 2 landscape scale, and spatial variables added in environment variables to describe the spatial autocorrelation phenomenon, environment variables are divided according to the 2 level in the first grade level, all environment variables are divided into microhabitat, landscape and space in 3 groups of variables. Second grade level, microhabitat variables are divided into vegetation and cover factor, the landscape variables are divided into vegetation factor and disturbance factor, then decomposition method to calculate the first level of microhabitat use variance, independent effect of landscape and space and combined effect, and the second level of the habitat factors of independent interpretation of the Red Crowned Crane Distribution the variation and joint interpretation, finally by using grading method verify the variance decomposition results, and calculate the environment variables on the distribution of Red Crowned Crane variant of the independent contribution. Through research, the main results are as follows: (1) the first level of the variance decomposition results show that: all the environmental variables explained 67.4% of the variation in the distribution of red crowned cranes. The landscape variables altogether explained 52.9% of the variation in the distribution of Red Crowned Crane, microhabitat variables explained 32.2% of the total variation of Dan crane top distribution, spatial variables altogether explained the Red Crowned Crane 33.7%. Spatial variation of microhabitat, independent landscape and 3 sets of variables are very small, respectively explained 7%, 4.5% and 5.7% of variation in the distribution of Red Crowned Crane, and they jointly explained 50.2% of the variation in the distribution of red crowned crane, the landscape variables and microhabitat variables, combined with spatial variables explained the Red Crowned Crane distribution variation of 22.2% and 25%. (2) the second level variance decomposition results show that microhabitat variables on the distribution of Red Crowned Crane explained variance proportion (32.2%) of the total variation, shelter and vegetation factor Independent factors were explained 9.8% of the variation and distribution of Red Crowned Crane 8.6%, jointly explained 13.8% of the variation in the distribution of red crowned cranes. The landscape variables on the distribution of Red Crowned Crane explained variance proportion the total variation (52.9%), interference factor and vegetation factor respectively explained 17.4% and 39.9% of the top Dan crane variation distribution, joint interpretation of red crowned the distribution variation of -4.4%. (3) classification results show that microhabitat variables, landscape and spatial variables respectively explained 9.1%, 14.9%, 3.1% of the variation in distribution of Red Crowned Crane. The landscape variables and microhabitat variables in the vegetation factor explained 7.1% and 6.1% of variation in the distribution of Red Crowned Crane, the area percentage of Suaeda salsa higher than that of other ecological factors explain the amount of.500 in the range of Hm2, the area percentage of water within the range of 50 Hm2, covering the distance independent distribution variation of Red Crowned Crane is the highest, which accounted for 4.12%, 3. The Red Crowned Crane distribution variation of 34% and 3.04%. The results showed that the landscape scale is the most important influence on scale selection of Red Crowned Crane Habitat. Vegetation landscape scale is the most important influencing habitat selection of Red Crowned Crane. Salsa area influence the selection of Red Crowned Crane and the most important environmental variables. Combined with spatial variables and interpretation the landscape variables is very high (25%), showed a strong correlation between the spatial structure of the Red Crowned Crane Distribution spatial structure and landscape variables, habitat types so landscape habitat factors can reflect the spatial distribution of Red Crowned Crane preferences or disgust. Variance decomposition method and classification method of decomposition of the variation of the basic distribution of Red Crowned Crane get consistent, landscape scale and vegetation factor is to limit the distribution of red crowned crane scale and the most important habitat factors.

【學(xué)位授予單位】：南京林業(yè)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：Q958.1

【共引文獻(xiàn)】

相關(guān)期刊論文 前10條

1 金春愛;;混凝土壩變形觀測資料分析理論研究[J];吉林水利;2014年11期

2 宋小會;郭志忠;郭華平;吳雙惠;王兆慶;鄔長安;;一種基于森林模型的光伏發(fā)電功率預(yù)測方法研究[J];電力系統(tǒng)保護(hù)與控制;2015年02期

3 朱衛(wèi)平;唐書恒;呂建偉;韓繼凡;張海霞;徐建紅;;棗園區(qū)塊煤層氣井產(chǎn)出水化學(xué)特征及動態(tài)變化規(guī)律[J];煤田地質(zhì)與勘探;2015年01期

4 陳鵬;赫圣杰;李振中;;兩軸道路汽車質(zhì)心高度數(shù)據(jù)處理方法[J];汽車工程師;2014年05期

5 李望月;;微觀數(shù)據(jù)和宏觀匯總數(shù)據(jù)在統(tǒng)計(jì)分析上的差異——以C-D生產(chǎn)函數(shù)為例[J];數(shù)學(xué)的實(shí)踐與認(rèn)識;2014年09期

6 李曉琳;惠潔;劉倩;;我國省際SO_2排放量的特異性研究[J];四川理工學(xué)院學(xué)報(bào)(自然科學(xué)版);2014年05期

7 武帥;楊光;;課業(yè)負(fù)擔(dān)檢測與監(jiān)控模型[J];沈陽師范大學(xué)學(xué)報(bào)(自然科學(xué)版);2014年04期

8 陳龍;史學(xué)正;徐勝祥;于東升;王軼虹;宋正姍;王美艷;;基于水稻葉面積指數(shù)的根生物量預(yù)測模型研究[J];土壤;2014年05期

9 王振紅;張昆;朱泓;趙蘇文;徐彬;;微型十字板在海洋軟土中的應(yīng)用[J];水利學(xué)報(bào);2015年S1期

10 李丹丹;金森;;用可見光圖像法測定林火蔓延速率的誤差[J];中南林業(yè)科技大學(xué)學(xué)報(bào);2014年03期

相關(guān)碩士學(xué)位論文 前10條

1 李倩;南方單棟塑料大棚微氣候模擬研究[D];南京信息工程大學(xué);2013年

2 趙沙;統(tǒng)計(jì)模型在國際原油價(jià)格預(yù)測及天然氣定價(jià)中的應(yīng)用[D];清華大學(xué);2013年

3 李長占;收割機(jī)械測產(chǎn)系統(tǒng)的研究[D];黑龍江八一農(nóng)墾大學(xué);2014年

4 談茜;中間視覺下LED光度特性及LED光譜模型研究[D];江蘇科技大學(xué);2013年

5 龔瑞琴;文本分類中特征選擇和分類算法的研究[D];寧夏大學(xué);2014年

6 劉雙艷;長沙市建設(shè)用地供需預(yù)測研究[D];湖南師范大學(xué);2014年

7 王怡達(dá);某企業(yè)收益的計(jì)量分析與預(yù)測[D];長春工業(yè)大學(xué);2014年

8 邴建峰;基于內(nèi)部營銷的飯店員工忠誠研究[D];中國海洋大學(xué);2014年

9 吳晶;一類有理函數(shù)逼近的參數(shù)估計(jì)方法[D];東北師范大學(xué);2014年

10 郝琳;我國人口老齡化問題研究及對策[D];曲阜師范大學(xué);2014年

，

本文編號：1582821