йѧ
CHINA ENVIRONMENTAL SCIENCE
1999 195ڡVol.19 No.5 1999



ڶȾۺҪԵɷӼϻѡ
Ӿ
ժҪݶԶȾﻷѡ,˵ȾֵҪԺͶȾֵۺҪ¸,ۺҪԵļ㹫ʽ.ݲۺҪ,ɷӼѡѡ.÷ڽԵɷӼѡ,˼,ұ˵ȾѡĲ㲻һʱ,ȷŻ.÷ڳɶдȾ12ѡʵ.
ؼʣȾ⣻Żѡ㣻ɷַ
ͼţX830.1 ױʶ룺A ±ţ1000- 6923(1999)05- 0458- 03
Optimization of environmental monitoring sites by principal component subset selection based on comprehensive relative importance of pollutants.
LI Zuo-yong, DENG Xin-min
(Department of Atmospheric Electrology Engineering, Chengdu Institute of Meteorology, Chengdu 610041, China). China Environmental Science. 1999,19(5)458460
AbstractIn order to select the optimal numbers of environmental monitoring sites, new concepts of the relative importance on monitoring values of single pollutant were advanced; and the formula of comprehensive relative importance was derived. On the basis of comprehensive relative importance of monitoring values of sites, the optimization of environmental monitoring sites was carried out by principal component subset selection. The advantages of this method not only reduce the calculation but also refrain from the difficulty of determining common optimal sites as a result of coincidenceless of optimal sites selected from single pollutant, as compared with the optimization of principal component subset selection based on the cross-correlation. The results of optimization from 12 monitoring sites of atmospheric pollutants in Chengdu Urban coincide with the practical cases.
Key wordsatmospheric pollutionenvironmental monitoringoptimal number of sitesprincipal component analysis
ŻǻһҪ,ⲻѧ߶Դ˽о[1-6].[6]ɷӼѡ(PCSS)ѡ.÷ڵȾѡж֮,ԶȾѡ,㲻֮:ԸȾﶼֱҪӦɷӼŻѡ,ȾĿ϶,ܴ;ԲͬȾֵĽԲ,ɲͬȾѡĲĿв,ѡĲ㲻ͬ,ѴȷȾĹŻ.
ĽȾֵӳϢҪԱʾ,ùʽȾۺϢҪ,ȻۺҪӦɷӼѡ񷨽ѡ.
1 ۺҪԵɷӼ(CRI-PCSS)Ļѡ
1.1 ȾֵϢҪ
ü񻯹ʽȾֵyijΪx'ij,±jʱ,i. 

(i=1, 2  mj=1,2  k)
ĳȾֵ[yij]maxӳϢҪ1,x'ijΪȾֵyijϢϢҪ.Ȼ1x'ij0.
1.2 ȾۺҪԹʽ
ܶȾȾ,ڲi,ʱjĶȾֵӳۺϢҪxij(¼ۺҪ)Ӧ:ڸȾ֮Ĺϵʮָ,ǵҪԼȲ໥ų,Ҳ໥,,xijӦǸȾx'ijļԵ;Ϊͻĳ㴦ҪһȾxijĹ,xijӦxijmax x'ij (l), (1ln).ʽnΪȾ;xijֵ1.ȸJordanʽ,nȾۺҪԹʽӦʽ:

ʽ:l, p, q, sȾ.ԸȾ֮໥Ӱ,ʽ(2)ʽ(3):

Ӧùʽ(1)(2)(3)ۺҪxij,ͿӦ[6]PCSSmۺѡ.
2 CRI-PCSSŻѡʵ
÷ڳɶ(1987~ 1992)12SO2NOxTSP 3ȾͬڼݵļȾֵ.
Ӧʽ(1)ֱ3ȾĲͬݵĸȾֵй񻯴,ʽ(2)12,24ȵ3ȾۺҪxijֵ.
12xijֵط,õ֮ۺҪԵϵR.
Ӧ[6]ɷӼѡ,RֵֵϢۻ.ɷӼѡ޳1;µĲٽɷӼѡ,޳1.˷жɷӼѡ.TӦɷӼ϶µm' ѡ,T޳ǸϢռm'ϢİٷֱΪ:
(4)
ʽ:m'Ϊm'ֵСֵ.µ(m'- 1)Ϣռm'ϢȫmϢİٷֱȦTֱͦΪ:
T=1-RT(5)
(6)
Աʵ̽7ɷӼѡ,޳1#()2#(޻У)3#(ţ)4#()6#()7#(԰)11#(ҽ)7,5#(¥)8#(Ǳ)9#(޳)10#(Ϲ⳧)12#(վ)5.5㱣Ϣռȫ12ϢİٷֱΪ:
(7)
3 ѡȽ
[6]ɷӼϷֱ3ֵȾѡĵԼǵĽѡŻ1ʾ.ӱ1ɼ,ѡCRI-PCSSѡѡĲֻ1в.ǰѡ3#,ѡ8#.3ֵȾɷӼѡĲ,10#3ȾֵԺܺ,3ȾֵԲ,1#2#4#6#㶼ûн.,ɸȾɷӼѡѡ5ĿȻһ,ϢռȫϢİٷֱȲͬ.15ѡѡĵ㱣ϢֱΪ92.58%90.56%93.86%92.33%94.13%.
1 ѡڳɶлŻѡ
Table 1 Results of optimization of environmental monitoring sites in Chengdu Urban
using two optimal methods

ڽԵɷӼѡ
SO2ѡĵ㿼NOxѡĵ㿼TSPѡĵۺϹŻĵCRI-PCSSѡĵ
1#()12#(վ)12#(վ)12#(վ)12#(վ)
3#(ţ)4# ()3#(ţ)3#(ţ)8#(Ǳ)
5#(¥)5#(¥)2#(޻У)5#(¥)5#(¥)
9#(޳)9#(޳)6#()9#(޳)9#(޳)
10#(Ϲ⳧)10#(Ϲ⳧)10#(Ϲ⳧)10#(Ϲ⳧)10#(Ϲ⳧)

4 
Ϊ˼CRI-PCSSѡ5Ƿܴȫ12ӳĻ,Ӧ[7]Ĵ۵ıָʽ:
(8)
(9)
ʽ:cikci0ֱΪiִȾʵŨֵͱŨֵ;aiΪiִȾҪԱ;nΪȾ.ۺϱָKķּ׼Ϊ()(- ,3.5];(Ⱦ)(3.5, 5.5];(Ⱦ)(5.5, ).ֱ12CPI-PCSSѡ5ĸȾֵʽ(8),(9)ıָΪK12=5.90K5=6.25,ͬ.
5 
5.1 жȾ,ñCRI-PCSSѡֻԶȾۺҪԱɷӼϷ,ѡ.[6],ÿȾﶼɷӼŻѡ,Ӷ˼.
5.2 CRI-PCSSѡԱڶȾֵԲõ,ȷۺŻ.
5.3 ԺȾ֮໥,ʽ(3)ʽ(2)ȾۺҪ,.
5.4 ÷Ըֻѡʺ,ҾнϺõĿ͹.
Ŀ塱صƼĿ(96-911-08-03)
߼飺Ӿ(1944-),,Ĵ˱,ѧ˶ʿ,ɶѧԺ,ԪŴ㷨˹缰ͶӰѰټ(PP)ȶеӦûо.ˡ۵·(׽ĴʡίƼȽ)11п.5,180ƪ,9ƪĽ,ıSCI,BA,PA,CA,EI¼û10.
ߵλɶѧԺӹϵ,Ĵ ɶ 610041
οף
[1]Venkatram A. Inherent uncertainty in air quality modeling [J]. Atmos. Environ., 1988, 22(6):1221- 1227.
[2]Grimm J W. Statistical analysis of errors in estimating wet doposition using five surface estimation algorithms [J]. Atmos. Environ., 1991, 25A(2):317- 320.
[3] . Żо [J]. й, 1994, 10(5):1- 3.
[4]̩,̽,žΰ. ѵ㼯ɸŻ [J]. ѧо, 1996, 9(1):12- 15.
[5]Ӿ. Ż˹ģ [J]. йѧ, 1997, 17(1):27- 29.
[6]ׯ,߳,ŷ. ɷӼѡѡ [J]. йѧ,1990, 10(2):148- 151.
[7]Ӿ. ۵ıָ [J]. йѧ, 1998, 18(5): 433- 436.
ոڣ1998-09-15
