%0 Journal Article
%A Jian Kang
%A Xin Li
%A Rui Jin
%A Yong Ge
%A Jinfeng Wang
%A Jianghao Wang
%+ Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou 730000, China;;Heihe Remote Sensing Experimental Research Station, Chinese Academy of Sciences, Lanzhou 730000, China;;State Key Laboratory of Resources and Environmental Information System, Institute of Geographical Science and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China;;University of Chinese Academy of Sciences, Beijing 100049, China
%T Hybrid Optimal Design of the Eco-Hydrological Wireless Sensor Network in the Middle Reach of the Heihe River Basin, China
%J Sensors
%D 2014
%N 10
%V 14
%K eco-hydrological wireless sensor network;spatial sampling;hybrid optimization criterion;unconditional stochastic simulation
%X The eco-hydrological wireless sensor network (EHWSN) in the middle reaches of the Heihe River Basin in China is designed to capture the spatial and temporal variability and to estimate the ground truth for validating the remote sensing productions. However, there is no available prior information about a target variable. To meet both requirements, a hybrid model-based sampling method without any spatial autocorrelation assumptions is developed to optimize the distribution of EHWSN nodes based on geostatistics. This hybrid model incorporates two sub-criteria: one for the variogram modeling to represent the variability, another for improving the spatial prediction to evaluate remote sensing productions. The reasonability of the optimized EHWSN is validated from representativeness, the variogram modeling and the spatial accuracy through using 15 types of simulation fields generated with the unconditional geostatistical stochastic simulation. The sampling design shows good representativeness; variograms estimated by samples have less than 3% mean error relative to true variograms. Then, fields at multiple scales are predicted. As the scale increases, estimated fields have higher similarities to simulation fields at block sizes exceeding 240 m. The validations prove that this hybrid sampling method is effective for both objectives when we do not know the characteristics of an optimized variables.
%W CNKI