艾比湖流域鹽漬土含水量光譜特征分析與建模

張海威,張飛,3?,李哲,井云清

(1.新疆大學(xué)資源與環(huán)境科學(xué)學(xué)院,830046,烏魯木齊;2.新疆大學(xué)綠洲生態(tài)教育部重點(diǎn)實(shí)驗(yàn)室,830046,烏魯木齊; 3.新疆智慧城市與環(huán)境建模普通高校重點(diǎn)實(shí)驗(yàn)室,830046,烏魯木齊)

艾比湖流域鹽漬土含水量光譜特征分析與建模

張海威1,2,張飛1,2,3?,李哲1,2,井云清1,2

(1.新疆大學(xué)資源與環(huán)境科學(xué)學(xué)院,830046,烏魯木齊;2.新疆大學(xué)綠洲生態(tài)教育部重點(diǎn)實(shí)驗(yàn)室,830046,烏魯木齊; 3.新疆智慧城市與環(huán)境建模普通高校重點(diǎn)實(shí)驗(yàn)室,830046,烏魯木齊)

鹽漬土表層鹽分累積與土壤含水量有著直接的關(guān)系。為了建立干旱區(qū)鹽漬土含水量高光譜遙感監(jiān)測(cè)模型,本文以艾比湖流域不同含水量的鹽漬土為研究對(duì)象,采用光譜反射率變換和多元統(tǒng)計(jì)分析(MSAM)方法,對(duì)土壤含水量的光譜特征進(jìn)行分析和建模。結(jié)果表明:隨著土壤含水量的增加,土壤反射率呈下降趨勢(shì);在一定范圍內(nèi),波長(zhǎng)越長(zhǎng),土壤光譜反射率與含水量的相關(guān)性越高,其中1 937 nm處的土壤光譜反射率與含水量具有最高的相關(guān)性(r=-0.636)。對(duì)土壤光譜反射率進(jìn)行8種光譜反射率變換后,在此基礎(chǔ)上利用多元統(tǒng)計(jì)方法分析鹽漬土的不同含水量與光譜之間的相關(guān)性,篩選敏感波段,建立關(guān)系模型。得出對(duì)數(shù)一階微分(Logarithm First Order Differential)在波長(zhǎng)2 024和2 357 nm建立的模型以及均方根一階微分(Root Mean Square First Order Differential)在波長(zhǎng)1 972和2 357 nm建立的模型最優(yōu),相關(guān)系數(shù)r分別為0.894和0.865。基于上述模型作者構(gòu)建了一種耦合模型,其相關(guān)性r=0.926比對(duì)數(shù)一階微分模型提高了0.032,比均方根一階模型提高了0.061;因此,所構(gòu)建的鹽漬土含水量估算模型是可行的,可以為遙感反演提供理論參考,對(duì)高光譜遙感反演具有一定意義。

光譜特征;含水量;模型;艾比湖流域

干旱區(qū)鹽漬土研究一直是土壤學(xué)科的熱門問題之一。土壤含水量不僅是微觀氣象學(xué)中一個(gè)重要的能量平衡參數(shù),而且還是旱情監(jiān)測(cè)的重要參數(shù)之一[1],因此土壤含水量對(duì)土壤鹽漬化的研究具有特殊意義。土壤含水量不僅與次生鹽漬化有關(guān),還與鹽漬土的鹽分運(yùn)移有著密切的關(guān)系,是影響綠洲生態(tài)環(huán)境穩(wěn)定性的重要因素之一[2]。國(guó)內(nèi)外一些學(xué)者建立光譜與土壤含水量模型,通常利用微波遙感對(duì)地表層濕度直接進(jìn)行數(shù)據(jù)收集,或利用熱量距平植被指數(shù)、作物缺水指數(shù)等方法,獲取地表能量與作物生長(zhǎng)狀況建立模型[35]。當(dāng)前分析土壤理化性質(zhì)與光譜反射率之間關(guān)系,主要使用主成分分析、逐步多元線性回歸分析、偏最小二乘法回歸分析及支持向量機(jī)和人工神經(jīng)網(wǎng)絡(luò)方法[67]。Lu Ning等[8]采用PHI高光譜圖像數(shù)據(jù),利用多元回歸分析與最小二乘法等方法對(duì)土壤鹽漬化信息進(jìn)行了可視化。孫小艷等[9]使用手持便攜式光譜儀對(duì)小麥的光譜反射率進(jìn)行研究。夏學(xué)齊等[10]對(duì)不同理化性質(zhì)的土壤進(jìn)行了光譜分析與建模。吳亞坤等[11]通過變換光譜反射率與土壤EC之間構(gòu)建回歸模型,分析變化光譜反射率與EC關(guān)系。以上研究均是在野外進(jìn)行的實(shí)驗(yàn)研究,另外也有很多學(xué)者開展了室內(nèi)光譜研究,如劉煥軍等[12]對(duì)黑土土壤水分進(jìn)行了光譜響應(yīng)特征分析與建模,何挺等[13]對(duì)土壤水分光譜特征進(jìn)行對(duì)比分析,翁永玲等[14]基于PLSR方法對(duì)青海茶卡-共和盆地的土壤鹽分進(jìn)行高光譜遙感反演。他們都是在室內(nèi)進(jìn)行土壤含水量研究,均取得一定的成果。目前,利用實(shí)測(cè)光譜與土壤理化性質(zhì)建立估算模型,已成為鹽漬土研究的主流方法。

筆者以室內(nèi)實(shí)測(cè)光譜數(shù)據(jù)和鹽漬土含水量為數(shù)據(jù)源,提出基于不同模型下的耦合建模思想,彌補(bǔ)單一性建模方法的不足。由于針對(duì)新疆典型生態(tài)脆弱區(qū)土壤含水量的光譜特性及建模研究還比較缺乏;因此,選擇新疆艾比湖流域?yàn)槭痉秴^(qū),以鹽漬土為研究對(duì)象,選用多元統(tǒng)計(jì)分析等方法,建立鹽漬土不同含水量估算模型,為治理與監(jiān)測(cè)鹽漬土提供技術(shù)支持。

1 研究區(qū)概況

艾比湖流域地處歐亞大陸腹地,位于E 79°53＇～85°02＇,N 43°38＇～45°52＇,三面環(huán)山,遠(yuǎn)離海洋,位于托里縣和烏蘇南部。研究區(qū)屬于典型的中溫帶干旱大陸氣候,以干旱少雨、氣溫變化劇烈為特點(diǎn)。年平均氣溫6.6～7.8℃,年降水量為116.0～169.2 mm。近40年來,由于入湖水量的減少,艾比湖湖面常年萎縮,湖面曾縮至不足500 km2,水位下降近23 m,干涸湖地也早已淪為鹽漠,成為浮塵天氣發(fā)源地。植被也依次形成旱生、超旱生、沙生、鹽生、水生等多種類型,主要植物類型有梭梭(Haloxylon ammodendron)、胡楊(Populus euphratica)、檉柳(Tamar-rix chinensis)和蘆葦(Phragmites australis)等[1516]。

2 材料與方法

2.1 數(shù)據(jù)的采集與處理

2015年10月,對(duì)艾比湖流域進(jìn)行綜合調(diào)查,整個(gè)流域共布設(shè)有53個(gè)樣點(diǎn),用GPS對(duì)其進(jìn)行坐標(biāo)定位同時(shí)采集距地面0～10 cm的土壤樣品,采集土壤時(shí)用鋁盒裝土樣,這樣可以防止水分的散失。對(duì)采集回來的土樣,立即使用便攜式地物光譜(Field Spec Hi res,ASD,美國(guó))的人工光源采集每個(gè)土樣光譜曲線,該儀器光譜范圍為350～2 500 nm,光譜分辨率3 nm。容器半徑為2.5 cm,垂直探頭,探頭距離為5 cm。測(cè)量步驟為:首先開機(jī)預(yù)熱2～3 min、去除暗電流、利用白板定標(biāo)獲取絕對(duì)反射率、容器土樣表面刮平、收集光譜曲線5條;每采集5個(gè)樣點(diǎn)光譜要利用白板定標(biāo);后利用光譜儀的后處理軟件ASD View Spec Pro對(duì)采集的光譜曲線分析,對(duì)每個(gè)點(diǎn)號(hào)的5條光譜數(shù)據(jù)在Process下的Statistics內(nèi)的Mean命令進(jìn)行平均值計(jì)算,最后光譜數(shù)據(jù)的歸一化;最后在Origin 8.0軟件中進(jìn)行微分變換與多元統(tǒng)計(jì)分析。

將土樣放入烘箱,在105℃下烘干12 h,放于干燥器中冷卻,然后進(jìn)行干土樣稱量,根據(jù)土壤含水量測(cè)定方法[17]計(jì)算出土壤含水量

式中:msc為土壤含水量,g/kg,m為鋁盒質(zhì)量,g;m1為土壤烘干前質(zhì)量,g;m2為土壤烘干后質(zhì)量,g。

2.2 光譜數(shù)據(jù)分析

為更好地分析與展示光譜數(shù)據(jù)與土壤含水量之間的相關(guān)性,對(duì)土壤光譜反射率分別進(jìn)行倒數(shù)、對(duì)數(shù)和均方根等數(shù)學(xué)變換,然后再進(jìn)行一階、二階微分變換消除干擾。一階、二階微分方程[18]分別為:

式中:λi+1、λi-1和λi為波長(zhǎng);R＇λi為波長(zhǎng)λi的一階微分光譜;R″λi為波長(zhǎng)λi的二階微分光譜。將變換后的反射率一階微分、反射率二階微分、倒數(shù)一階微分、對(duì)數(shù)一階微分、均方根一階微分與原始反射率作為鹽漬土的高光譜指數(shù)。

2.3 相關(guān)性分析

通過每個(gè)波段計(jì)算鹽漬土的高光譜指數(shù)與土壤含水量的相關(guān)系數(shù),篩選出鹽漬土含水量的敏感波段[19]即

式中:rj為高光譜指數(shù)與土壤含水量的相關(guān)系數(shù);j為波段;Rij為第i個(gè)樣品第j波段的反射率值;個(gè)土樣在第j波_段反射率的平均值;Wi為第i個(gè)土樣的水分含量,g;為土壤樣本含水量的平均值,g;n為鹽漬土樣品個(gè)數(shù)。

3 結(jié)果與分析

3.1 鹽漬土含水量光譜特征分析

通過室內(nèi)分析發(fā)現(xiàn)艾比湖流域鹽漬土含水量的范圍為2.6%～25.6%。把含水量的范圍分為5個(gè)層次:0～5%、5～10%、10～15%、15～20%、20～25%,分別從各個(gè)層次隨機(jī)選取一條來分析不同含水量的光譜特征。不同含水量下鹽漬土光譜特征曲線趨勢(shì)和態(tài)勢(shì)很相似,但反射率高低不同。如圖1所示,總體來看,光譜曲線從400到1350 nm呈平緩上升趨勢(shì),但不同含水量?jī)A斜度不同。其中含水量為10.2%的光譜曲線在380～600 nm范圍內(nèi),土壤光譜反射率最高。含水量為9.7%的光譜曲線在380～1 350 nm范圍內(nèi),不符合隨著含水量的增加而反射率降低的規(guī)律,這可能是由于土壤中的含鹽量過高所導(dǎo)致的此現(xiàn)象,在600～1450 nm附近反射率曲線斜率變緩,在1 470和1 950 nm左右各有一個(gè)水分反射谷,其中在1 950 nm附近的反射谷比較明顯。不同土壤含水量反射谷程度不同,反射率呈先升高后降低的趨勢(shì)。22.1%、10.2%和9.7%3條不同含水量的光譜曲線,在1 950 nm附近出現(xiàn)很明顯的反射谷,且這3條曲線反射率差異很小,而4.6%含水量的曲線差異較高。從整體看來隨著含水量的增加反射率降低,不同含水量增減速率不同,這與顧燕等[20]和劉曾媛等[21]的研究一致。

圖1 不同含水量鹽漬土光譜反射率曲線Fig.1 Spectral reflectance curves of salinized soil with different levels of water content

3.2 敏感波段的選取與分析

在Origin 8.0軟件中進(jìn)行8種數(shù)學(xué)變換:一階微分、二階微分、倒數(shù)、倒數(shù)一階微分、對(duì)數(shù)、對(duì)數(shù)一階微分、均方根、均方根一階微分,之后分別與土壤含水量建立相關(guān)關(guān)系,可以看出在不同波段上變化差異比較大(圖2)。

圖2 土壤含水量和光譜指數(shù)的相關(guān)系數(shù)(R為反射率)Fig.2 Correlation coefficients between soil water content and spectral indices(R represents reflectance)

由圖2可得,對(duì)光譜反射率進(jìn)行微分變換后,其與鹽漬土含水量的相關(guān)性要明顯高于原始反射率,既有明顯正相關(guān),又有負(fù)相關(guān),且均通過顯著性檢驗(yàn)。這由于經(jīng)過數(shù)學(xué)變換后,放大某些原始光譜數(shù)據(jù)中細(xì)小的波段信息,所以微分變換后的波段與鹽漬土含水量相關(guān)性比較高。據(jù)此可以篩選出鹽漬土含水量的敏感波段,確定相關(guān)系數(shù)最高值對(duì)應(yīng)的波段(表1)?？梢?出各種數(shù)學(xué)變換篩選出的敏感波段波長(zhǎng)均＞1 000 nm;一階微分、倒數(shù)一階、對(duì)數(shù)一階和均方根一階都有相同的敏感波段2 357 nm,且此敏感波段的相關(guān)系數(shù)較高。

3.3 建模與分析

根據(jù)相關(guān)分析結(jié)果得到鹽漬土含水量的敏感波段,建立相應(yīng)的土壤含水量的估算模型。由于實(shí)驗(yàn)誤差或數(shù)據(jù)異常,53個(gè)樣本中只有44個(gè)樣點(diǎn)進(jìn)行了建模與驗(yàn)證,然后從44個(gè)樣本中隨機(jī)選取34個(gè)樣品構(gòu)建土壤含水量的估算模型,以敏感波段作為自變量,實(shí)測(cè)的土壤含水量作為因變量,并選擇相關(guān)系數(shù)r較大的敏感波段,進(jìn)行多元線性回歸分析,建立鹽漬土含水量估算模型(表2)。

表1 光譜反射率及其變換形式與土壤含水量敏感波段Tab.1 Spectral reflectance and its transformations,and sensitive bands of soil water content

表2 不同自變量回歸模型Tab.2 Prediction models based on different characteristic bands

可見,土壤含水量與所選敏感波段呈良好的線性關(guān)系,原始反射率和經(jīng)各數(shù)學(xué)變換后構(gòu)建的模型均通過了0.05顯著性檢驗(yàn)。其中原始反射率的回歸模型回歸系數(shù)最低(r=0.618)。一階微分回歸模型回歸系數(shù)最高(r=0.858)。依據(jù)回歸系數(shù)和線性回歸方程顯著性檢驗(yàn)統(tǒng)計(jì)量F值最大、均方根誤差最小的原則,本文選取一階微分,對(duì)數(shù)一階,均方根一階進(jìn)行構(gòu)建模型。

3.3.1 模型驗(yàn)證

模型的驗(yàn)證主要考慮估算模型的穩(wěn)定性和效果。利用剩下的10個(gè)真實(shí)值作為驗(yàn)證數(shù)據(jù),對(duì)上述所選模型對(duì)其進(jìn)行驗(yàn)證,對(duì)反演模型的估算效果進(jìn)行檢驗(yàn)其中一階微分模型精度r=0.554,對(duì)數(shù)一階微分模型精度r=0.894,均方根一階微分精度r= 0.865。根據(jù)以上結(jié)果剔除一階微分模型,選取對(duì)數(shù)一階微分和均方根一階微分模型進(jìn)行鹽漬土含水量的反演。從模型的估算值和真實(shí)值的線性擬合(圖3)可以看出,對(duì)數(shù)一階模型樣點(diǎn)基本上聚集在對(duì)角線附近,樣本相關(guān)系數(shù)r=0.894。相比之下均方根一階樣點(diǎn)沒有對(duì)數(shù)一階樣點(diǎn)那樣緊湊,樣點(diǎn)相關(guān)系數(shù)r=0.865。除個(gè)別異常點(diǎn),估算值和實(shí)測(cè)值的線性擬合程度均較好。本文建立的鹽漬土含水量遙感監(jiān)測(cè)模型估算能力通過了檢驗(yàn),在一定程度上可以用來估算鹽漬土含水量。

3.3.2 多元線性回歸與構(gòu)建的耦合估算模型

與原始反射率相比,不同變換后的光譜波段與含水量相關(guān)性有著不同程度的提高。經(jīng)過以往研究模型的驗(yàn)證可知:構(gòu)建的模型比較單一性,都是一種數(shù)學(xué)變換下的模型構(gòu)建,有一定的不足;因此本文基于以上模型選取不同變換下的敏感波段構(gòu)建模型。為了保證模型數(shù)據(jù)的一致性,模型選取的樣本和上述模型一致。構(gòu)建模型為y=-6.147x1+53.535x2-181.344x3+0.044,式中:y為鹽漬土含水量,x1為倒數(shù)一階變換下2 032 nm處的光譜反射率,x2為對(duì)數(shù)一階變換2 357 nm處的光譜反射率,x3為均方根一階變換2 357 nm處的光譜反射率。通過10個(gè)樣點(diǎn)對(duì)模型進(jìn)行精度驗(yàn)證,可以得到,模型的精度有所提高,相關(guān)系數(shù)r=0.926,比對(duì)數(shù)一階微分模型相關(guān)性r提高了0.032,比均方根一階微分模型相關(guān)系數(shù)如提高了0.061(表3)。

圖3 模型估算值與實(shí)測(cè)值線性擬合圖(N為樣本數(shù),下同)Fig.3 Linearfit of measured values and predicted values with the model(N:number of samples,the same below)

表3 不同回歸模型對(duì)比Tab.3 Comparison of different regression models

相比之前的模型,此模型基于對(duì)數(shù)一階微分模型與比均方根一階微分模型的相關(guān)性最優(yōu)波段,且經(jīng)過驗(yàn)證發(fā)現(xiàn)耦合模型對(duì)鹽漬土含水量的估算是可行的(圖4)。

圖4 模型估算值與實(shí)測(cè)值線性擬合圖Fig.4 Linearfit of measured values and predicted values with the model

4 結(jié)論與討論

1)鹽漬土不同土含水量光譜曲線的形態(tài)與趨勢(shì)基本一致;在可見光波段范圍內(nèi),不同含水量的鹽漬土反射率呈規(guī)律性變化,即隨著土壤含水量的增加,土壤反射率呈下降走勢(shì)。

2)對(duì)鹽漬土含水量與光譜建模與驗(yàn)證得:經(jīng)建模與驗(yàn)證后得出對(duì)數(shù)一階的模型最好(r=0.804)其次是均方根一階(r=0.798)。對(duì)數(shù)一階微分模型的估算值與實(shí)測(cè)值的相關(guān)系數(shù)最高(r=0.894),其次是均方根一階微分模型,相關(guān)系數(shù)為r=0.865,即鹽漬土含水量模型估算良好。

3)基于均方根一階微分模型與對(duì)數(shù)一階微分模型建立耦合模型,其方程:y=-6.147x1+ 53.535x2-181.344x3+0.044,其模型效果良好(r= 0.926)。證明了選取不同變換后的敏感波段進(jìn)行建模,對(duì)鹽漬土含水量估算是可行的。

筆者提出的基于均方根一階微分模型與對(duì)數(shù)一階微分模型下的耦合模型,對(duì)艾比湖流域鹽漬土的治理與監(jiān)測(cè)有著重要意義和作用;但本研究只是在室內(nèi)的一種嘗試,并沒有在室外進(jìn)行驗(yàn)證,因此耦合模型有待于進(jìn)一步檢驗(yàn)與改進(jìn)。

[1] 李美婷,武紅旗,蔣平安,等.利用土壤的近紅外光譜特征測(cè)定土壤含水量[J].光譜學(xué)與光譜分析,2012, 32(8):2117. LI Meiting,WU Hongqi,JIANG Pingan,et al.Measuring soil water content by using near infrared spectral characteristics of soil[J].Spectroscopy and Spectral Analysis, 2012,32(8):2117.

[2] 中國(guó)土壤學(xué)會(huì).中國(guó)土壤學(xué)在前進(jìn)[M].北京:中國(guó)農(nóng)業(yè)科技出版社,1995:9. Soil Science Society of China.Progress of soil science in China[M].Beijing:China's Agricultural Science and Technology Press,1995:9.

[3] ROBINOVE C J,CHAVEZ P S,GEHRING D,et al. Arid land monitoring using Landsat albedo difference images[J].Remote Sensing of Environment,1981,11:133.

[4] HENRICKSEN B L.Reflections on drought:Ethiopia 1983-1984[J].International Journal of Remote Sensing,1986,7(11):1447.

[5] 劉志明,張柏,晏明,等.土壤水分與干旱遙感研究的進(jìn)展與趨勢(shì)[J].地球科學(xué)進(jìn)展,2003,18(4):576. LIU Zhiming,ZHANG Bai,YAN Ming,et al.Some research advances and trends on soil moisture and drought monitoring by remote sensing[J].Advance in Earth Sciences,2003,18(4):576.

[6] LIU Huanjun,ZHANG Yuanzhi,ZHANG Bai.Novel hyper spectral reflectance models for estimating black-soil organic matter in Northeast China[J].Environmental Monitoring and Assessment,2009,154(1/2/3/4): 147.

[7] LI Y,DEMETRIADES-SHAH T H,KANEMASU E T, et al.Use of second derivatives of canopy reflectance for monitoring prairie vegetation over different soil backgrounds[J].Remote Sensing of Environment,1993,44 (1):81.

[8] LU Ning,ZHANG Zhi,GAO Yang.Recognition and mapping of soil salinization in arid environment with hyper spectral data[C].Proceedings.2005 IEEE International Geo science and Remote Sensing Symposium,2005. IGArSS'05.IEEE,2005,6:4520.

[9] 孫小艷,常學(xué)禮,張寧,等.不同取樣單元對(duì)干旱區(qū)綠洲小麥地上生物量光譜估算模型的影響[J].中國(guó)沙漠,2012,32(2):568. SUN Xiaoyan,CHANG Xueli,ZHANG Ning,et al.Impact of different sampling units on ground spectral modelsfor estimating wheat aboveground biomass in oases of arid zone[J].Journal of Desert Research,2012,32(2):568.

[10] 夏學(xué)齊,季峻峰,陳駿,等.土壤理化參數(shù)的反射光譜分析[J].地學(xué)前緣,2009,16(4):354. XIA Xueqi,JI Junfeng,CHEN Jun,et al.Analysis of soil physical and chemical properties by reflectance spectroscopy[J].Earth Science Frontiers,2009,16 (4):354.

[11] 吳亞坤,楊勁松,李曉明.基于光譜指數(shù)與EM38的土壤鹽分空間變異性研究[J].光譜學(xué)與光譜分析, 2009(4):1023. WU Yakun,YANG Jinsong,LI Xiaoming.Study on spatial variability of soil salinity based on spectral indices and EM38 readings[J].Spectroscopy and Spectral Analysis,2009,29(4):1023.

[12] 劉煥軍,張柏,宋開山,等.黑土土壤水分光譜響應(yīng)特征與模型[J].中國(guó)科學(xué)院研究生院學(xué)報(bào),2008,25 (4):503. LIU Huanjun,ZHANG Bai,SONG Kaishan,et al.Soil moisture spectral analysis and its spectral model[J]. Journal of the Graduate School of the Chinese Academy of Sciences,2008,25(4):503.

[13] 何挺,王靜,程燁,等.土壤水分光譜特征研究[J].土壤學(xué)報(bào),2006,43(6):1027. HE Ting,WANG Jing,CHENG Ye,et al.Soil water spectrum characteristic research[J].Acta Pedologica Sinica,2006,43(6):1027.

[14] 翁永玲,戚浩平,方洪賓,等.基于PLSR方法的青海茶卡-共和盆地土壤鹽分高光譜遙感反演[J].土壤學(xué)報(bào),2010,47(6):1255. WENG Yongling,QIE Haoping,FANG Hongbin,et al. PLSR based hyper spectral remote sensing retrieval of soil salinity of Chaka·Gonghe basin in Qinghai province[J].Acta Pedologica Sinica,2010,47(6):1255.

[15] 姜紅濤,特依拜,克里木,等.艾比湖流域NDWI變化及其與降水、溫度的關(guān)系[J].中國(guó)沙漠,2014,34 (6):1678. JIANG Hongtao,TIYIP,KELIMU,et al.Responses of NDVI to the variation of precipitation and temperature the Ebinur Lake Basin[J].Journal of Desert Research, 2014,34(6):1678.

[16] 張小龍.艾比湖流域氣候變化及其徑流響應(yīng)[J].鹽湖研究,2011,19(2):11. ZHANG Xiaolong Climatic variation and runoff response in Ebinur Lake Catchment[J].Journal of Salt Lake ResearcH,2011,19(2):11.

[17] 王昌佐,王紀(jì)華,王錦地,等.裸土表層含水量高光譜遙感的最佳波段選擇[J].遙感信息,2003(4):33. WANG Changzuo,WANG Jihua,WANG Jindi,et al. The choice of best detecting band for hyper spectral remote sensing on surface water content of bare soil[J]. Remote Sensing Information,2003.(4):33.

[18] 關(guān)紅,賈科利,張至椅,等.鹽漬化土壤光譜特征分析與建模[J].國(guó)土資源遙感,2015,27(2)100. GUAN Hong,JIA Keli,ZHANG Zhinan.Research on remote sensing monitoring model of soil salinization based on spectrum characteristic analysis[J].Remote Sensing for Land and resources,2015,27(2):100.

[19] 趙振亮,特依拜,丁建麗,等.新疆典型綠洲土壤電導(dǎo)率和PH值的光譜響應(yīng)特征[J].中國(guó)沙漠,2013,33 (5):1413. ZHAO Zhenliang,TIYIP,DING Jianli,et al.Characteristics of spectral responding to soil electrical conductivity PH in the typical Oasis of Xinjiang[J].Journal of Desert Research,2013,33(5):1413.

[20] 顧燕,張鷹,李歡.基于實(shí)測(cè)光譜的潮灘土壤含水量遙感反演模型研究[J].濕地科學(xué),2013,11(2): 167. GU Yan,ZHANG Ying,LI Huan,et al.Remote sensing retrieval model on soil moisture content of tidal flat based on measured spectra[J].Wetland Science, 2012,11(2):168.

[21] 劉曾媛,貢璐,張雪妮,等.克里雅河流域于田綠洲土壤酶活性c理化因子相關(guān)性分析[J].中國(guó)土壤與肥料,2014,4(7):35. LIU Zeng yuan,GONG Lu,ZHANG Xueni,et al.Soil enzyme activities and their relationship with physicochemical factors in Yutian oasis of the Keriya river watershed[J].Soil and Fertilizer Sciences in China,2014, 4(7):35.

Modeling and analysis of spectral characteristic of soil water content in the salinized soil of Ebinur Lake Watershed

ZHANG Haiwei1,2,ZHANG Fei1,2,3,LI Zhe1,2,JING Yunqing1,2
(1.College of Resources&Environmental Science,Xinjiang University,830046,Urumqi,China; 2.Key Laboratory of Oasis Ecology,830046,Urumqi,China;3.General Institutes of Higher Learning Key Laboratory of Smart City and Environmental Modeling,Xinjiang University,830046,Urumqi,China)

[Background]With the development of spatial information science,the hyperspectral remote sensing becomes more and more important in nowadays.Studying on spectral characteristic of soil water content is an important work,for it is the base of monitoring remote sensing.This study aims to investigate the spectral characteristic of soil water content and the relationship between the hyperspectral data and the soil water content,to explore a rapid and accurate method for estimating soil water contents, and to establish the hyperspectral remote sensing monitoring model for saline soil water in the arid area of the Ebinur Lake Watershed.[Methods]This study took saline soil with different water contents in Ebinur Lake Watershed as the research object,used the spectral reflectance transformation and multivariate statistical analysis methods(MSAM)to analyze the spectral characteristic of soil water content,and built models.[Results]The result showed that with the increase of soil water content,thereflectance of soil declined.In a certain range,the longer the wavelength,the higher the correlation between the soil spectral reflectance and soil water content;the soil spectral reflectance at the wavelength of 1 937 nm(r=-0.636)had the highest correlation with water content.After 8 transformation of soil spectral reflectance,the correlation of logarithmic first order differential sensitive band of 2 357 nm was the best(r=-0.808 6).Subsequently,MSAM were applied to analyze the correlation between spectrum and salinized soil with different water contents,then the spectral sensitive band of the soil was screened,and the correlation model was established.The result indicated that the model set up by Logarithm First Order Differential at the wavelength of 2 024 nm and 2 357 nm and the Root Mean Square First Order Differential at the wavelength of 1 972 nm and 2 357 nm was the best,the correlation coefficientrwas 0.894 and 0.865 respectively.Based on the above established models,the authors constructed a new model,and the correlationrof which was 0.926,increased 0.032 against the Logarithm First Order Differential model,and increased 0.061 compared to the Root Mean Square First Order Differential model.[Conclusions]Therefore,the estimation model of soil water content is feasible.In addition,this study provides a new model for the indoor hyperspectral estimation of soil water content in Ebinur Lake Watershed,and it also provides a theoretical and technical reference for the hyperspectral quantitative estimation of soil water content,which has certain guiding significance for hyperspectral remote sensing.

spectrum characteristic;water content;modeling;Ebinur Lake Watershed

O433.4

:A

:2096-2673(2017)01-0008-07

10.16843/j.sswc.2017.01.002

2016- 07- 03

2016- 11- 27

項(xiàng)目名稱:國(guó)家自然科學(xué)基金“新疆聯(lián)合基金本地優(yōu)秀青年人才培養(yǎng)專項(xiàng)”(U1503302)

張海威(1990—),男,碩士。主要研究方向:干旱區(qū)遙感應(yīng)用。E-mail:1522634426@qq.com

?通信作者簡(jiǎn)介:張飛(1980—),男,博士,副教授。主要研究方向:干旱區(qū)資源環(huán)境遙感。E-mail:zhangfei3s@163.com