Support vector regression （SVR） method is a novel type of learning machine algorithms, which is seldom applied to the development of urban atmospheric quality models under multiple socio-economic factors. This study...Support vector regression （SVR） method is a novel type of learning machine algorithms, which is seldom applied to the development of urban atmospheric quality models under multiple socio-economic factors. This study presents four SVR models by selecting linear, radial basis, spline, and polynomial functions as kernels, respectively for the prediction of urban dust fall levels. The inputs of the models are identified as industrial coal consumption, population density, traffic flow coefficient, and shopping density coefficient. The training and testing results show that the SVR model with radial basis kernel performs better than the other three both in the training and testing processes. In addition, a number of scenario analyses reveal that the most suitable parameters （insensitive loss function e, the parameter to reduce the influence of error C, and discrete level or average distribution of parameters σ） are 0.001, 0.5, and 2 000, respectively.展开更多
Considering that the noises resulting from low modulation frequency are serious and cannot be totally eliminated by the classic filters,a novel infrared(IR) gas concentration detection system based on the least square...Considering that the noises resulting from low modulation frequency are serious and cannot be totally eliminated by the classic filters,a novel infrared(IR) gas concentration detection system based on the least square fast transverse filtering(LS-FTF) self-adaptive modern filter structure is proposed.The principle,procedure and simulation on the LS-FTF algorithm are described.The system schematic diagram and key techniques are discussed.The procedures for the ARM7 processor,including LS-FTF and main program,are demonstrated.Comparisons between the experimental results of the detection system using the LS-FTF algorithm and those of the system without using this algorithm are performed.By using the LS-FTF algorithm,the maximum detection error is decreased from 14.3 to 5.4,and also the detection stability increases as the variation range of the relative error becomes much smaller.The proposed LS-FTF self-adaptive denoising method can be of practical value for mid-IR gas detection,especially for weak signal detection.展开更多
Satellite observations of atmospheric CO2 are able to truly capture the variation of global and regional CO2 concentration.The model simulations based on atmospheric transport models can also assess variations of atmo...Satellite observations of atmospheric CO2 are able to truly capture the variation of global and regional CO2 concentration.The model simulations based on atmospheric transport models can also assess variations of atmospheric CO2 concentrations in a continuous space and time,which is one of approaches for qualitatively and quantitatively studying the atmospheric transport mechanism and spatio-temporal variation of atmospheric CO2 in a global scale.Satellite observations and model simulations of CO2 offer us two different approaches to understand the atmospheric CO2.However,the difference between them has not been comprehensively compared and assessed for revealing the global and regional features of atmospheric CO2.In this study,we compared and assessed the spatio-temporal variation of atmospheric CO2 using two datasets of the column-averaged dry air mole fractions of atmospheric CO2(XCO2)in a year from June 2009 to May 2010,respectively from GOSAT retrievals(V02.xx)and from Goddard Earth Observing System-Chemistry(GEOS-Chem),which is a global 3-D chemistry transport model.In addition to the global comparison,we further compared and analyzed the difference of CO2 between the China land region and the United States(US)land region from two datasets,and demonstrated the reasonability and uncertainty of satellite observations and model simulations.The results show that the XCO2 retrieved from GOSAT is globally lower than GEOS-Chem model simulation by 2 ppm on average,which is close to the validation conclusion for GOSAT by ground measures.This difference of XCO2 between the two datasets,however,changes with the different regions.In China land region,the difference is large,from 0.6 to 5.6 ppm,whereas it is 1.6 to 3.7 ppm in the global land region and 1.4 to 2.7 ppm in the US land region.The goodness of fit test between the two datasets is 0.81 in the US land region,which is higher than that in the global land region(0.67)and China land region(0.68).The analysis results further indicate that the inconsistency of CO2concentration between satellite observations and model simulations in China is larger than that in the US and the globe.This inconsistency is related to the GOSAT retrieval error of CO2 caused by the interference among input parameters of satellite retrieval algorithm,and the uncertainty of driving parameters in GEOS-Chem model.展开更多
基金The National Key Scientific Instrument and Equipment Development Projects(2012YQ22011902)The 863 National High Technology Research and Development Program of China(2014AA06A503)
基金Supported by the National Key Scientific Instrument and Equipment Development Projects(2012YQ22011902)the Special Foundation for Young Scientists of Hefei Institutes of Physical Science,Chinese Academy of Sciences(YZJJ201502)
基金Projects(2007JT3018, 2008JT1013, 2009FJ4056) supported by the Key Project in Hunan Science and Technology Program, ChinaProject(20090161120014) supported by the New Teachers Sustentation Fund in Doctoral Program, Ministry of Education, China
文摘Support vector regression （SVR） method is a novel type of learning machine algorithms, which is seldom applied to the development of urban atmospheric quality models under multiple socio-economic factors. This study presents four SVR models by selecting linear, radial basis, spline, and polynomial functions as kernels, respectively for the prediction of urban dust fall levels. The inputs of the models are identified as industrial coal consumption, population density, traffic flow coefficient, and shopping density coefficient. The training and testing results show that the SVR model with radial basis kernel performs better than the other three both in the training and testing processes. In addition, a number of scenario analyses reveal that the most suitable parameters （insensitive loss function e, the parameter to reduce the influence of error C, and discrete level or average distribution of parameters σ） are 0.001, 0.5, and 2 000, respectively.
基金supported by the National "863" Project of China (Nos. 2007AA06Z112,2007AA03Z446 and 2009AA03Z442)the National Natural Science Foundation of China (No.61077074)the Science and Technology Department of Jilin Province of China (Nos. 20070709 and 20090422)
文摘Considering that the noises resulting from low modulation frequency are serious and cannot be totally eliminated by the classic filters,a novel infrared(IR) gas concentration detection system based on the least square fast transverse filtering(LS-FTF) self-adaptive modern filter structure is proposed.The principle,procedure and simulation on the LS-FTF algorithm are described.The system schematic diagram and key techniques are discussed.The procedures for the ARM7 processor,including LS-FTF and main program,are demonstrated.Comparisons between the experimental results of the detection system using the LS-FTF algorithm and those of the system without using this algorithm are performed.By using the LS-FTF algorithm,the maximum detection error is decreased from 14.3 to 5.4,and also the detection stability increases as the variation range of the relative error becomes much smaller.The proposed LS-FTF self-adaptive denoising method can be of practical value for mid-IR gas detection,especially for weak signal detection.
基金supported by the National Natural Science Foundation of China(Grant No.41071234)"Strategic Priority Research Program-Climate Change:Carbon Budget and Relevant Issues"of the Chinese Academy of Sciences(Grant No.XDA05040401)the National High Techondogy Research and Development Program of China(Grant No.2012AA12A301)
文摘Satellite observations of atmospheric CO2 are able to truly capture the variation of global and regional CO2 concentration.The model simulations based on atmospheric transport models can also assess variations of atmospheric CO2 concentrations in a continuous space and time,which is one of approaches for qualitatively and quantitatively studying the atmospheric transport mechanism and spatio-temporal variation of atmospheric CO2 in a global scale.Satellite observations and model simulations of CO2 offer us two different approaches to understand the atmospheric CO2.However,the difference between them has not been comprehensively compared and assessed for revealing the global and regional features of atmospheric CO2.In this study,we compared and assessed the spatio-temporal variation of atmospheric CO2 using two datasets of the column-averaged dry air mole fractions of atmospheric CO2(XCO2)in a year from June 2009 to May 2010,respectively from GOSAT retrievals(V02.xx)and from Goddard Earth Observing System-Chemistry(GEOS-Chem),which is a global 3-D chemistry transport model.In addition to the global comparison,we further compared and analyzed the difference of CO2 between the China land region and the United States(US)land region from two datasets,and demonstrated the reasonability and uncertainty of satellite observations and model simulations.The results show that the XCO2 retrieved from GOSAT is globally lower than GEOS-Chem model simulation by 2 ppm on average,which is close to the validation conclusion for GOSAT by ground measures.This difference of XCO2 between the two datasets,however,changes with the different regions.In China land region,the difference is large,from 0.6 to 5.6 ppm,whereas it is 1.6 to 3.7 ppm in the global land region and 1.4 to 2.7 ppm in the US land region.The goodness of fit test between the two datasets is 0.81 in the US land region,which is higher than that in the global land region(0.67)and China land region(0.68).The analysis results further indicate that the inconsistency of CO2concentration between satellite observations and model simulations in China is larger than that in the US and the globe.This inconsistency is related to the GOSAT retrieval error of CO2 caused by the interference among input parameters of satellite retrieval algorithm,and the uncertainty of driving parameters in GEOS-Chem model.
