Soil salinity inversion in Yutian Oasis based on PALSAR radar data
ZAYTUNGULYakup1,2,, MAMATSawut1,2,, ABDUSALAMAbdujappar1,2, ZHANGDong1,2 1. College of Resources and Environmental Science, Xinjiang University, Urumqi 830046, China2. Ministry of Education Key Laboratories of Oasis Ecology, Xinjiang University, Urumqi 830046, China 通讯作者:通讯作者:买买提·沙吾提,E-mail: korxat@xju.edu.cn 收稿日期:2017-06-12 修回日期:2018-05-28 网络出版日期:2018-10-25 版权声明:2018《资源科学》编辑部《资源科学》编辑部 基金资助:国家自然科学基金项目(41361016;41561089) 作者简介: -->作者简介:再屯古丽·亚库普,女,新疆轮胎县人,硕士生,主要从事干旱区资源与环境遥感应用研究。E-mail: zaytungul1992@163.com
展开
摘要 土壤盐渍化是当今土地退化和荒漠化的主要形式之一,不仅严重制约农业和经济的发展,并且对生态环境和人类生存造成威胁。本研究以新疆于田绿洲为研究区,利用四极化PALSAR (Phased Array type L-band Synthetic Aperture Radar)数据后向散射系数,土壤含水量,土壤pH值以及土壤盐分实测值,采用多元线性回归模型、地理加权回归模型和BP神经网络模型,以土壤含盐量作为因变量建立了定量反演模型。从土壤盐分反演结果图可以看出,反演结果与地面实地考察结果基本一致。经过模型验证得知,3层BPANN模型的均方根误差RMSE=0.99,平均相对误差MRE=0.31,模型性能指数RPD=5.34,其模型预测能力优于前2种传统模型。本文建立的神经网络模型无需考虑复杂的介电常数,在一定程度上能够满足土壤盐渍化监测的需要,促进PALSAR数据在土壤盐渍化监测中的应用。
关键词:土壤盐渍化;PALSAR-2雷达数据;后向散射系数;神经网络;反演;新疆于田绿洲 Abstract Soil salinization and/or desertification is one of the main forms of land degradation and environmental issues. Meanwhile, it causes the destruction of resources, hampers to development of agriculture and threats to the environment and human survival. Yutian Oasis was identified as a study area and soil salinity information was extracted from the PALSAR-2 ALOS-2 data, which exhibited a kind of fine four-polarization SLC (single look complex) format and were bought in 2015 with 5.1 (range resolution)×4.3 (azimuth resolution) ground resolutions. Considering the distribution of saline soil spatial variability, 68 points were designed as sampling points, Hand-held GPS (global position system) receiver was used to record the coordinates of sampling points and 0-10 cm topsoil samples were collected in the field. Soil total soluble salt content was measured in the lab. The four-polarization back-scatter coefficient values corresponding to the sampling points were extracted based on the previous results by the spatial analysis module of ArcGIS. Total salt content was taken as dependent variable and four-polarization PALSAR-2 data back-scatter coefficient values, soil moisture and pH values as independent variables. The multiple linear regression (MLR), geographically weighted regression (GWR) and back propagation artificial neural network (BP ANN) were adopted to establish the quantitative inversion models of soil salt content. Results illustrated that among the ANN (BP), MLR and GWR models employed in this contribution, the ANN (BP) model was identified as the most potential predictive model of soil salinity. Best predictive results were achieved using ANN (BP) with R2=0.84, RMSE=0.99, MRE=0.31 and RPD=5.34. The established ANN (BP) model in this paper can reduce the smoothing effect compared with the two traditional models and improve the accuracy and reliability of model predictions, which meets the needs of soil salinity monitoring to a certain extent. It can promote and develop the application of microwave remote sensing in the soil salinity monitoring.
本研究选用日本宇航局发射的对地观测卫星(Advanced Land Observing Satellite,ALOS-2)获取的2015年4月23 日的PALSAR全极化(包括HH、VV、HV、VH等4种极化方式)数据(具体参数请见表1),其工作波段为L波段(1.2GHz段)[17]。ALOS-2PALSAR-2数据具有多种分辨率成像、更短的重访周期和能在任何大气条件下全天候工作等优点,根据所选的模式发射出水平方向(H)和垂直方向(V)两种极化,每个散射元素(HH、HV、VH、VV)对地表面不同的特征具有不同的敏感性,有助于辨别不同的土地类型[18]。 Table 1 表1 表1全极化ALOS-2 PALSAR-2数据的主要参数[19] Table 1Main parameters of fully polari metric ALOS-2 PALSAR data
参数类型
数据
数据获取日期
2015/4/23
地图投影
UTM
像素间距
12.5m
处理级
L1.1
偏移量
0
天底偏角
30.4
卫星高度
628km
行号
22609
列号
8080
极化方式
HH、HV、VH、VV
获取方式
Fine Quad Polarization
产品类型
HBQ
新窗口打开 本研究利用ENVI®5.3软件进行了预处理。预处理过程包括数据导入(生成单式复图像Single Look Complex,SLC)、多视处理(方位和距离系数分别为8和4,生成强度图像)、图像配准、通过3×3窗口Refined Lee方法进行斑点噪声消除处理、地理编码及辐射定标等。在此基础上得到标准四极化后向散射系数影像图,并用ArcGIS10.2软件提取出采样点后向散射系数值。试验区部分土壤盐分含量、含水量、pH值与四极化后向散射系数如表2所示。为了成像时间与野外采样时间对应一致,本研究选用了2015年4月20日至5月1日的野外采样数据,并根据于田绿洲的地貌特征、气候条件、植被分布情况、土壤盐渍化程度等不同条件在整个研究区内,由绿洲内部到外围,考虑到不同程度盐渍地分布情况,选取具有代表性的68个采样点(采样点分布如图1所示),用手持GPS接收机获取各采样点在WGS84坐标系统下的经纬度,记录采样点周围植被类型和土壤质地情况,并分别拍摄东西南北方向的景观建立野外调查照片库。每个采样点0~10cm土层均匀取3个土壤样品,并编号装入采样袋和铝盒,将土样带回实验室经过自然风干、研磨,并用1mm孔径筛过滤后,配置5:1水土比浸提液,同时参考“土壤农业化学分析方法”一书[20]进行土壤pH值和含盐量的测定。其中土壤含盐量数据用EC200电导仪进行测定,pH值使用pH7310进行测定。为了消除取样代表性误差,每个采样点3个样品的测定值求其均值作为该点的代表值。所取铝盒土壤样本使用烘干称重法测定土壤含水量。 Table 2 表2 表2雷达影像后向散射系数与于田绿洲土壤含盐量数据统计 Table 2Statistics of radar back-scattering coefficients and salt contents of soil in Yutian Oasis
采样点编号
SHH /dB
SHV /dB
SVH /dB
SVV /dB
含盐量/(g/kg)
pH值
含水量/(g/kg)
1
-14.33
-23.71
-23.96
-16.35
0.52
8.23
7.17
2
-14.78
-22.15
-23.25
-17.62
0.28
8.55
5.26
3
-13.48
-22.63
-22.65
-11.01
8.91
8.16
5.56
4
-26.78
-28.86
-27.54
-18.25
1.75
8.18
8.75
5
-23.30
-30.73
-30.64
-23.65
2.58
8.15
4.26
6
-17.59
-27.69
-26.56
-16.52
3.63
8.97
7.15
7
-20.26
-26.72
-28.04
-19.96
6.39
7.95
3.71
8
-15.13
-24.13
-24.79
-12.41
9.91
9.02
4.24
9
-20.21
-28.93
-28.31
-19.05
8.51
8.09
5.36
…
…
…
…
…
…
…
…
68
-21.06
-28.29
-28.12
-19.77
11.40
8.70
3.44
新窗口打开 显示原图|下载原图ZIP|生成PPT 图1于田绿洲位置和采样点分布 -->Figure 1Location of study area and distribution of ground sampling points -->
本研究利用K-S检验方法对研究区土壤含水量和pH值进行了正态检验,其P值均小于0.05,表明该样本不服从正态分布。因此本文利用反距离加权法(Inverse Distance Weighted,IDW)对土壤含水量和pH值进行插值,为了得到模型反演结果,将插值结果的投影、像素大小、图像大小调整到跟PALSAR雷达图像后向散射图一致,并利用上述所提出的MLR模型和GWR模型,通过ENVI软件的 Band Math功能得到土壤含盐量反演(图2)。 显示原图|下载原图ZIP|生成PPT 图2于田绿洲土壤含盐量反演结果 -->Figure 2Results of Soil salinity inversion in Yutian Oasis -->
从反演结果图可以看出两种模型反演后的盐渍化程度变化趋势大致相同,然而MLR模型反演结果出现部分绿洲内部轻度盐渍化区域误分成重度盐渍地的现象。从而可知,GWR模型比MLR模型的预测能力和模型精度都有所提高,反演结果较好。为了验证模型的预测能力,本研究将22个验证样本的土壤盐分模型模拟值和实测值分别为x,y轴,制作出散点分布(图3),并分别算出决定系数(R2)、均方根误差(RMSE)、平均相对误差(MRE)和模型性能指数(RPD)对土壤含盐量实测值和模型预测值进行相关分析来评价3种模型结果(表3)。 显示原图|下载原图ZIP|生成PPT 图3训练样本的实测值与模拟值散点分布 -->Figure 3Scatter diagram of measured values and predicted values of training points -->
Table 3 表3 表3不同模型对于田绿洲土壤含盐量的模拟结果评价 Table 3Assessment results of soil salt content in Yutian Oasis by three different models
本研究利用于田绿洲四极化ALOS-2 PALSAR-2数据,利用多元回归模型、地理加权回归模型、BP神经网络模型建立了土壤含盐量、土壤含水量、pH值、雷达后向散射系数之间的定量反演模型。研究结论如下: (1)从模型反演结果图看出,研究区东北部没有采样点的沙漠地区出现重度盐渍化现象,说明,采样点的分布情况在一定程度上影响模型的反演精度。研究区中部绿洲内以及绿洲和裸露地交错带区土壤含盐量较高,是土壤盐渍化现象比较严重的区域。 (2)通过模型验证,GWR盐渍化监测模型与MLR模型相比,变量之间的相关性有所提高和改进,但GWR模型对变量的依赖性较大,当变量数量较少时,难以建立合理又准确的回归模型,导致因变量的预测精度低,反演结果存在不确定性。 (3)由于BP模型的建立过程中模型匹配了土壤含水量、pH值、PALSAR影像后向散射系数组合值SHH、SHV、SVH、SVV、SHV/SHH、(S2VH+S2HH)/(S2VH-S2HH)作为参数,模型能够映射复杂的非线性关系,反演进度好于线性模型。后向散射系数对盐渍化有一定的影响,如果所选择的参数不同,模型的精度也会有所不同。 (4)本研究建立的三层BP模型的均方根误差为0.99,与MLR模型和GWR模型相比分别降低了0.59和0.69,平均相对误差为0.31,分别减少了25%和21%,模型预测能力RPD为5.43,优于前两种回归模型,说明,模型的可靠性以及预测能力均有提高。 本文所提出的基于PALSAR数据的土壤盐分反演模型在一定程度上能够满足土壤盐渍化监测的需要,为理解盐渍化对该地区生态系统的影响程度和变化格局、防治盐渍化、制定盐渍化土壤治理方案等方面提供了基础资料。影响土壤盐渍化的因素较多,地表粗糙度对后向散射系数的影响也较大,由于知识水平,技术和仪器设备的限制,本研究未能加入到考虑范围内,这也是今后工作和研究中需要解决的问题。 The authors have declared that no competing interests exist.
[DingJ L, ChenW Q, ChenY.Soil salinization disaster warning in arid zones: a case study in the Ugan-Kuqa Oasis [J]. , 2016, 36(4): 1079-1086. ] [本文引用: 1]
[HuangE X.Research progress of remote sensing technique in soil salinization [J]. , 2010, 38(13): 6849-6850. ] [本文引用: 1]
[6]
JimenezL O, Rivera-MedinaJ L, Rodriguez-DiazE, et al. Integration of spatial and spectral information by means of unsupervised extraction and classification for homogenous objects applied to multispectral and hyper spectral data [J]. , 2005, 43(4): 844-851. [本文引用: 1]
[7]
FariftehJ, MeerF V D, AtzbergerC, et al. Quantitative analysis of salt-affected soil reflectance spectra: a comparison of two adaptive methods (PLSR and ANN) [J]. , 2007, 110(1): 59-78. [本文引用: 1]
[LiuQ M, ChengQ M, WangX, et al. Soil salinity inversion in Hetao Irrigation district using microwave radar [J]. , 2016, 32(16): 109-114. ] [本文引用: 1]
[9]
FariftehJ, FarshadA, GeorgerR J.Assessing salt-affected soils using remote sensing, solute modeling, and geophysics [J]. , 2006, 130(3-4): 191-206. [本文引用: 1]
[10]
BrunnerP, LiH T, KinzelbachW, et al. Generating soil electrical conductivity maps at regional level by integrating measurements on the ground and remote sensing data [J]. , 2007, 28(15): 3341-3361. [本文引用: 1]
[GengM.Research on Remote Sensing Inversion of Salt Content in Saline-alkali Soil Based on BP Network in Songliao Plain [D]. , 2008. ] [本文引用: 1]
[15]
HolahN, BaghdadiN, ZribiM, et al. Potential of ASAR/ENVISAT for the characterization of soil surface parameters over bare agricultural fields [J]. , 2005, 96(1): 78-86. [本文引用: 1]
[YuanY Y, HalikW, GuanJ Y, et al. Spatial differentiation and impact factors of Yutian Oasis soil surface salt based on GWR model [J]. , 2016, 27(10): 3273-3282. ] [本文引用: 1]
[GanJ, YangC Z, ZhengW J, et al. Pishan earthquake coseismic deformation field inversion based on ALOS-2Data [J]. , 2017, 40(2): 51-54. ] [本文引用: 1]
[18]
NatsuakiR, NagaiH, MotohkaT, et al. SAR interferometry using ALOS-2 PALSAR-2 data for the MW7. 8 Gorkha, Nepal earthquake [J]. , 2016, 68(1): 1-13. [本文引用: 1]
[19]
RosenqvistA, ShimadaM, SuzukiS, et al. Operational performance of the ALOS global systematic acquisition strategy and observation plans for ALOS-2 PALSAR-2 [J]. , 2014, 155(4): 3-12. [本文引用: 1]
[WuC S, HuangC, LiuG H, et al. Spatial prediction of soil salinity in the Yellow River Delta based on geographically weighted regression [J]. , 2016, 38(4): 704-713. ] [本文引用: 2]
[LiQ Q, WangC Q, YueT X, et al. Method for spatial variety of soil organic matter based on radial basis function neural network [J]. , 2010, 26(1): 87-93. ] [本文引用: 1]
[24]
NakayaT, FotheringhamA S, CharltonM, et al. Semiparametric Geographically Weighted Generalized Linear Modeling in GWR4. 0 [C]. , 2009. [本文引用: 1]
[GuoY L, LiuY Z, BiR T, et al. Study on cultivated land surface soil organic matter spatial characteristics based on GWR model-a case study of Xinfu district [J]. , 2013, 40(13): 187-190. ] [本文引用: 1]