本文關(guān)鍵詞：基于神經(jīng)網(wǎng)絡(luò)模型和地統(tǒng)計(jì)學(xué)方法的土壤養(yǎng)分空間分布預(yù)測(cè)，，由筆耕文化傳播整理發(fā)布。

基于神經(jīng)網(wǎng)絡(luò)模型和地統(tǒng)計(jì)學(xué)方法的土壤養(yǎng)分空間分布預(yù)測(cè)

Prediction of soil nutrients spatial distribution based on neural network model combined with goestatistics

[1] [2] [3] [4] [5] [6] [7] [8]

LI Qi-quan,WANG Chang-quan,ZHANG Wen-jiang,YU Yong,LI Bing,YANG Juan,BAI Gen-chuan,CAI Yan（1College of Resources and Environment,Sichuan Agricultural University,Chengdu 611130,China;2St

[1]四川農(nóng)業(yè)大學(xué)資源環(huán)境學(xué)院,成都611130; [2]四川大學(xué)水力學(xué)與山區(qū)河流開發(fā)保護(hù)國家重點(diǎn)實(shí)驗(yàn)室,成都610065; [3]四川農(nóng)業(yè)大學(xué)林學(xué)院,四川雅安625014

文章摘要：采用徑向基函數(shù)神經(jīng)網(wǎng)絡(luò)模型與普通克里格法相結(jié)合的方法,預(yù)測(cè)川中丘陵區(qū)縣域尺度土壤養(yǎng)分（有機(jī)質(zhì)和全氮）的空間分布,并與普通克里格法和回歸克里格法進(jìn)行比較.結(jié)果表明：各方法對(duì)研究區(qū)土壤養(yǎng)分的預(yù)測(cè)結(jié)果相似.與多元回歸模型相比,神經(jīng)網(wǎng)絡(luò)模型對(duì)驗(yàn)證樣點(diǎn)土壤有機(jī)質(zhì)和全氮的預(yù)測(cè)值與樣點(diǎn)實(shí)測(cè)值的相關(guān)系數(shù)分別提高了12.3%和16.5%,表明神經(jīng)網(wǎng)絡(luò)模型能更準(zhǔn)確地捕捉土壤養(yǎng)分與定量環(huán)境因子間的復(fù)雜關(guān)系.對(duì)469個(gè)驗(yàn)證樣點(diǎn)預(yù)測(cè)結(jié)果的誤差分析表明,神經(jīng)網(wǎng)絡(luò)模型與普通克里格法相結(jié)合的方法對(duì)土壤有機(jī)質(zhì)和全氮預(yù)測(cè)結(jié)果的平均絕對(duì)誤差、平均相對(duì)誤差、均方根誤差較普通克里格法分別降低了6.9%、7.4%、5.1%和4.9%、6.1%、4.6%,降低幅度達(dá)到極顯著水平（P〈0.01）;與回歸克里格法相比則分別降低了2.4%、2.6%、1.8%和2.1%、2.8%、2.2%,降低幅度達(dá)顯著水平（P〈0.05）.

Abstr：In this study,a radial basis function neural network model combined with ordinary kriging（RBFNN_OK） was adopted to predict the spatial distribution of soil nutrients（organic matter and total N） in a typical hilly region of Sichuan Basin,Southwest China,and the performance of this method was compared with that of ordinary kriging（OK） and regression kriging（RK）.All the three methods produced the similar soil nutrient maps.However,as compared with those obtained by multiple linear regression model,the correlation coefficients between the measured values and the predicted values of soil organic matter and total N obtained by neural network model increased by 12.3% and 16.5%,respectively,suggesting that neural network model could more accurately capture the complicated relationships between soil nutrients and quantitative environmental factors.The error analyses of the prediction values of 469 validation points indicated that the mean absolute error（MAE）,mean relative error（MRE）,and root mean squared error（RMSE） of RBFNN_OK were 6.9%,7.4%,and 5.1%（for soil organic matter）,and 4.9%,6.1%,and 4.6%（for soil total N） smaller than those of OK（P0.01）,and 2.4%,2.6%,and 1.8%（for soil organic matter）,and 2.1%,2.8%,and 2.2%（for soil total N） smaller than those of RK,respectively（P0.05）.

文章關(guān)鍵詞：

Keyword：:radial basis function neural network model ordinary kriging regression kriging soil nutrient.

課題項(xiàng)目：國家杰出青年科學(xué)基金項(xiàng)目（40825003）; 國家自然科學(xué)基金項(xiàng)目（41201214,40801175）資助

  本文關(guān)鍵詞：基于神經(jīng)網(wǎng)絡(luò)模型和地統(tǒng)計(jì)學(xué)方法的土壤養(yǎng)分空間分布預(yù)測(cè)，由筆耕文化傳播整理發(fā)布。