發(fā)布時(shí)間：2018-01-26 09:42

  本文關(guān)鍵詞： 土壤呼吸 分布式監(jiān)測(cè) 土壤溫度 監(jiān)測(cè)點(diǎn)布置 貝葉斯最大熵　出處：《北京林業(yè)大學(xué)》2016年博士論文　論文類型：學(xué)位論文

【摘要】：土壤是一個(gè)巨大的碳庫(kù),總儲(chǔ)存量達(dá)到1394PgC,土壤呼吸占整個(gè)陸地生態(tài)系統(tǒng)呼吸的600%-90%,其微小的變化都可能引起大氣CO2濃度較大的改變,是導(dǎo)致全球氣候變化的關(guān)鍵過程。森林是陸地生態(tài)系統(tǒng)的主體,其土壤中的碳貯量占全球土壤碳庫(kù)的73%。故探索森林土壤呼吸的監(jiān)測(cè)手段和評(píng)價(jià)方法,不僅對(duì)于碳匯計(jì)量、碳交易、緩解氣候變化具有重要的戰(zhàn)略意義,而且對(duì)森林碳儲(chǔ)量的經(jīng)營(yíng)管理具有重要的指導(dǎo)價(jià)值。論文首先研究土壤呼吸監(jiān)測(cè)平臺(tái),然后研究土壤溫度空間分布,以土壤溫度指導(dǎo)監(jiān)測(cè)點(diǎn)布置理論,并把土壤溫度作為軟數(shù)據(jù)結(jié)合貝葉斯最大熵方法提高土壤呼吸預(yù)測(cè)的可靠性,最后研發(fā)土壤呼吸分析與評(píng)價(jià)軟件,為研究者分析與評(píng)價(jià)土壤呼吸提供便利的方法。本論文主要形成以下結(jié)論：(1)研究了基于氣體擴(kuò)散定律的土壤呼吸儀,在儀器內(nèi)設(shè)置多層傳感器采集數(shù)據(jù),建立CO2在儀器內(nèi)的擴(kuò)散模型計(jì)算得到土壤碳通量值,此土壤呼吸監(jiān)測(cè)儀平均誤差為10%左右。然后通過AD-HOC與GPRS無(wú)線通信技術(shù),組建土壤呼吸分布式監(jiān)測(cè)網(wǎng)絡(luò),實(shí)現(xiàn)多點(diǎn)、同時(shí)監(jiān)測(cè)。研究表明在實(shí)驗(yàn)區(qū)域內(nèi)僅用一個(gè)采樣點(diǎn)進(jìn)行土壤呼吸監(jiān)測(cè)平均誤差為42%左右,說明單一點(diǎn)監(jiān)測(cè)會(huì)產(chǎn)生較大誤差,用分布式監(jiān)測(cè)網(wǎng)絡(luò)可以減小監(jiān)測(cè)誤差。(2)研究了基于BP神經(jīng)網(wǎng)絡(luò)的土壤溫度空間分布方法。針對(duì)土壤溫度分布具有不確定性和非線性等特征,提出分組訓(xùn)練、組合優(yōu)化的改進(jìn)BP神經(jīng)網(wǎng)絡(luò)算法,此算法隱含層節(jié)點(diǎn)數(shù)量設(shè)為10,學(xué)習(xí)率設(shè)為0.75,算法達(dá)到較好的性能。并以真實(shí)值與預(yù)測(cè)值進(jìn)行對(duì)比,用R2作為評(píng)價(jià)標(biāo)準(zhǔn),樣條插值算法,克里金算法,BP神經(jīng)網(wǎng)絡(luò)算法的R2分別為0.59,0.73和0.92,而改進(jìn)BP神經(jīng)網(wǎng)絡(luò)算法R2為0.97,說明此算法能有效提高土壤溫度空間分布的可靠性。(3)研究了基于土壤溫度梯度與圖像分割算法的土壤呼吸監(jiān)測(cè)點(diǎn)優(yōu)化布置方法,以土壤溫度為依據(jù),確定實(shí)驗(yàn)區(qū)土壤呼吸監(jiān)測(cè)點(diǎn)數(shù)量設(shè)置5個(gè)比較合理,并利用區(qū)域生長(zhǎng)圖像分割算法,確定這5個(gè)監(jiān)測(cè)點(diǎn)位置。然后以平均誤差率為評(píng)價(jià)標(biāo)準(zhǔn),均勻網(wǎng)格法確定位置和隨機(jī)布置方法確定位置的平均誤差分別為24%和29%,而以區(qū)域生長(zhǎng)圖像分割算法平均誤差為15%,說明此方法能提高土壤呼吸監(jiān)測(cè)點(diǎn)布置的合理性。(4)研究了以土壤溫度為輔助信息的貝葉斯最大熵算法,利用土壤溫度和土壤呼吸之間的函數(shù)關(guān)系,把土壤溫度作為軟數(shù)據(jù)融入到土壤呼吸估計(jì)中,可以提高土壤呼吸空間插值算法的準(zhǔn)確性。貝葉斯最大熵算法(BME)的平均CR為0.745,明顯高于協(xié)同克里金算法(Co-OK)方法的0.505以及普通克里金算法(OK)方法的0.267;BME的平均RMSE結(jié)果為0.568,均低于Co-OK方法的0.851以及OK方法的1.511；BME方法的Bias絕對(duì)值平均值為0.192,相比較Co-OK方法的0.508以及OK方法的1.143,實(shí)驗(yàn)表明BME方法的性能比Co-OK方法和OK方法好。同時(shí),利用BME算法還能減少監(jiān)測(cè)點(diǎn)數(shù)量。(5)利用SQL Server數(shù)據(jù)庫(kù)存儲(chǔ)傳感器數(shù)據(jù),利用ArcGIS建立土壤呼吸相關(guān)圖層,利用Visual Studio 2008研發(fā)土壤呼吸分析與評(píng)價(jià)軟件,為用戶進(jìn)行土壤呼吸研究提供便利。研究結(jié)果為監(jiān)測(cè)與評(píng)價(jià)森林土壤呼吸提供一套優(yōu)化的方法,為進(jìn)一步研究土壤碳的動(dòng)態(tài)變化以及指導(dǎo)碳匯林和森林資源的經(jīng)營(yíng)管理作必要的技術(shù)支撐。
[Abstract]:The soil is a huge carbon pool, the total storage capacity reached 1394PgC, soil respiration accounted for the entire land ecosystem respiration 600%-90%, the tiny change can cause the change of atmospheric CO2 concentration is larger, which is a critical process of global climate change. The forest is the main terrestrial ecosystem, monitoring and evaluation methods of carbon the storage of soil accounted for the global soil carbon pool 73%. therefore exploring the forest soil respiration, not only for carbon accounting, carbon trading, has an important strategic significance for the mitigation of climate change, but also has an important guiding value for the management of forest carbon storage. The paper firstly studies the soil respiration monitoring platform, and then study the soil temperature the soil temperature distribution, to guide the layout of monitoring points and the theory of soil temperature as the reliability of soft data based on Bayesian maximum entropy method to improve the prediction of soil respiration, finally. Analysis and evaluation of soil respiration software, provide a convenient method for the research of analysis and evaluation of soil respiration. The main conclusions as following: (1) study of soil respiration instrument gas diffusion law based on multilayer data acquisition sensor is arranged on the instrument, the diffusion model is established in CO2 instrument in the calculation of soil carbon flux the value of soil respiration monitor, the average error is about 10%. And then through the AD-HOC and GPRS wireless communication technology, the formation of soil respiration in distributed monitoring network, realize multi point simultaneous monitoring. The results show that only one sampling point of soil respiration monitoring the average error is about 42% in the experimental area, the single point monitoring will produce large error, distributed monitoring network monitoring can reduce the error. (2) studied the spatial distribution of soil temperature method based on BP neural network. According to the soil temperature distribution is The uncertainty and nonlinear characteristics, put forward the team training, improved BP neural network algorithm in combinatorial optimization, the algorithm of hidden layer node number is set to 10, the learning rate is set to 0.75 algorithms to achieve better performance. And the true value and predictive value compared with R2 as the evaluation standard, spline interpolation. Kriging algorithm, BP neural network algorithm R2 0.59,0.73 and 0.92 respectively, and the improved BP neural network algorithm R2 is 0.97, indicating that this algorithm can effectively improve the reliability of the soil temperature distribution. (3) studied the soil temperature gradient image segmentation algorithm and the optimization of monitoring points of soil respiration to soil temperature based method according to determine the number of experimentation area soil respiration monitoring point setting 5 is reasonable, and the use of regional growth image segmentation algorithm, determine which of the 5 monitoring points. Then as the evaluation standard to the average error rate, uniform grid Method to determine the location and random layout method to determine the average error of location is 24% and 29% respectively, while in the region growing image segmentation algorithm the average error is 15%, indicating that this method can improve the rationality of the arrangement of soil respiration monitoring. (4) studied the Bias maximum entropy algorithm of the auxiliary information to the soil temperature, the relationship between soil the soil temperature and the soil temperature between breath, as soft data into the estimation of soil respiration, can improve the accuracy of the interpolation algorithm of breathing space soil. Bias maximum entropy algorithm (BME) and the average CR was 0.745, significantly higher than the collaborative Krikin algorithm (Co-OK) method and 0.505 ordinary Krikin algorithm (OK) for 0.267; the average RMSE BME was 0.568, 1.511 and 0.851 were lower than that of Co-OK method OK method; BME method Bias the absolute value of the average value is 0.192, compared to 0.508 Co-OK method 1.143 and OK method, experiments show that the performance of BME method than the Co-OK method and OK method. At the same time, using the BME algorithm can reduce the number of monitoring points. (5) the use of sensor data storage SQL Server database, using ArcGIS to establish soil respiration related layer, using Visual Studio 2008 software R & D analysis and evaluation of soil respiration, provide convenience study of soil respiration for users. The results provide a set of optimization methods for forest soil respiration monitoring and evaluation, the management for further research on dynamic changes of soil carbon and carbon sequestration forest and forest resources as necessary technical support.

【學(xué)位授予單位】：北京林業(yè)大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2016
【分類號(hào)】：S714.3

【相似文獻(xiàn)】

相關(guān)期刊論文 前3條

1 H.Onishi;H.Kawakami;K.Marino;T.Komiyama;K.Kuriyama;M.Araya;焦慧;;用于自主屏氣照像的簡(jiǎn)易呼吸指示器:無(wú)電子材料的一鍵式觸摸設(shè)置[J];國(guó)際醫(yī)學(xué)放射學(xué)雜志;2010年04期

2 杜曉蘭,吳寶明,劉彥,何慶華,肖劍;呼吸、體溫兩生理參數(shù)檢測(cè)電路的設(shè)計(jì)[J];生物醫(yī)學(xué)工程學(xué)雜志;2001年04期

3 ;[J];;年期

相關(guān)會(huì)議論文 前1條

1 陳榮昌;;機(jī)械通氣的床旁呼吸監(jiān)測(cè)[A];呼吸與危重癥醫(yī)學(xué)（2010-2011）[C];2011年

相關(guān)博士學(xué)位論文 前1條

1 胡軍國(guó);森林土壤呼吸監(jiān)測(cè)與評(píng)價(jià)方法研究[D];北京林業(yè)大學(xué);2016年

相關(guān)碩士學(xué)位論文 前1條

1 吳丹;可應(yīng)用于家庭的非侵入式呼吸監(jiān)測(cè)方法的研究[D];中南大學(xué);2010年

，

本文編號(hào)：1465293