1.Water conservancy and Civil Engineering College, Inner Mongolia Agricultural University, Hohhot, Inner Mongolia 010018, China; 2.Ningxia Institute of Water Resources Research, Yinchuan, Ningxia 750021, China
Mesoscale inversion models of mass percentage of clay, power, sand and organic matter were established by using multiple regression, support vector machine and BP neural network based on the hyper-spectrum of 1 024 samples in 3 scales in Hetao Irrigation District of Inner Mongolia. The mesoscale scale was shifted to large and small scales, and the applicability of scale transformation was assessed for these models. Results showed that the mesoscale inversion models of soil particle composition and organic matter were well applicable at the other two scales with correlation coefficient of 0.33-0.60 in multiple regression, 0.41-0.52 in support vector machine and 0.52-0.72 in BP neural network. Clearly, the models based on BP neural network method showed even better applicability at the other two scales. The correlation coefficients of clay, power, sand and organic matter were 0.44-0.62, 0.37-0.72, 0.42-0.72 and 0.33-0.56, respectively, suggesting the fitting effect of particle compositions was better than that that of organic matter as a whole.
ZHANG Na-,,ZHANG Hong-Ling- et al. Scale transformation of regional soil particle composition and organic matter prediction models based on hyper-spectrum[J]. Journal of Drainage and Irrigation Machinery Engin, 2018, 36(11): 1175-1179.
CHEN Da, LU Fan, LI Qifeng. Development of multi-scale modeling methods for calibration transfer in near infrared spectroscopy[J]. Nanotechnology and precision engineering, 2017,15(2):121-126.(in Chinese)
WU Jianguo, QI Ye. Dealing with scale in landscape analysis: an overview[J]. Geographic information sciences, 2000,6(1):1-5.
WAGENET R J. Scale issues in agroecological research chains[J]. Nutrient cycling in agroecosystems, 1998,50(3):23-34.
ZHANG Na, ZHANG Dongliang, QU Zhongyi, et al. Establishment and evaluation of model for predicting soil texture based on hyperspectal data—case study of Jie-fangzha Irrigation Area in Hetao Irrigation District[J]. Journal of arid land resources and environment, 2014,28(5):67-72.(in Chinese)
ZHANG Na,ZHANG Dongliang,QU Zhongyi, et al. Spatial variability of soil organic matter at different sampling scales in Inner Mongolia Hetao Irrigation Area[J]. Chinese journal of ecology, 2016,35(3):630-640.(in Chinese)
PATEL S S, RAMACHANDRAN P. A comparison of machine learning techniques for modeling river flow time series: the case of upper Cauvery River Basin[J]. Water resources management, 2015,29(2):589-602.