Chin.JGeochem.(2012)31:195- 203DOL: 10.1007/511631-012-0568-0Combinatorial analysis on spatial informationstatistics for the karst water environment inGuiyang, ChinaWANG Zhongmeil' , ZHU Lijun' , YANG Ruidong'. YANG Shengyuan', DING Jianping', and YANG Genlan'' College of Resources and Envionmental Engineering of Gutzhou Universiry, Giuiyang 50003, China2 Key Laboralory of Karsi Enmironment and Geohazard Preverion (Guizhou Universiry,. Ministry of Educarion, Gulyang 50003,. China3 Gwizhou lnsiute of Geological Environment Monion, Guiyang 50001.0 China●Coresponding author, E-mail: re.zmwang@gzu.edu.cnReceived September 19, 2010; accepted October 20. 2010◎Science Press and Institute of Gcochemistry. CAS and Springer-Verlag Berlin Heidelberg 2012Abstract The karst groundwater system is extremely vulnerable and easily contaninaled by human activities. Toundersland the spatial distribution of contaminants in the groundwater of karst urban areas and contributors to thecontamination, this paper employs the spatial informnation statistics analysis theory and method to analyze the karstgroundwater environment in Guiyang City. Based on the karsl groundwater quality data detected in 61 detectionpoints of the research area in the last three years, we made Kriging evaluation isoline map with some ions in the karstgroundwater, such as so. , Fe'*, Mn2+ and F, analyzed and evaluated the spatial distribution, extension and varia-tion of four types of ions on the basis of this isoline map. The results of the analysis show that the anomaly areas ofsoz",Fe2*, Mn2*.F and other ions are mainly located in Baba' a0. Mawangmiao and Sanqiao in northwestern Gui-yang City as well as in its downtown area by reasons of the original non-point source pollution and the contamina-tion caused by human activities (indutial and domestic pollution).Key words karst water environment; spaial information saisics; analysis1 Introductionshortage. Among its limited water resources, ground-water is the main source of water supply with theGuizhou Province is located in the heart of theamount of karst water over 90%. Therefore, we shouldkarst area of Southwest China. It covers a karst area ofprotect the limited water resource for local people's129600 km', accounting for 73.4% of the total area.health condition and sustainable development of Gui-The groundwalter system of Guiyang City is a typicalyang City. This paper attempts to analyze and evaluatekarst water system which is generally acknowledgedthe spatial distribution, extension and variation of theto be extremely vulnerable to the contaminationSO4", Fe", Mn' and F pollution sources in the karstcaused by human activities (Liang Yunchao et al..groundwater of Guiyang City.2008). With the development of Guiyang's economy,coal mining, industrial waste and municipal wastewa-bution randomness of evaluation parameters is theter poison the karst groundwater system. Especiallymajor factor that causes the uncertainty of waterthe pollution caused by coal mining makes the con-evaluation models (Sun Nazheng, 1989). The issue ofcentration values of so2, Fe'*, Mn2+ and F in thehow to quantify the randomness of various spatial dis-karst groundwater exceed the national standards.tribution parameters in groundwater environmentMoreover, the industrial waste discharged from theevaluation so as to ensure the reliability of utilizingprocesses of desulfurization and defluorination bythe evaluation models becomes crucial. Previous sta-cement plants, aluminium plants, tire factories, ther-tistical methods based on traditional simulation mod-mal powder plants and gas companies in Guiyang Cityels中國煤化工lics can no longerwill contributea lot to SO4 , Fe"*, Mn2* and F pollu-provithe spatial varia-tion of the karst groundwater. Meanwhile, municipaltionCNMHGdly use single de-MHsewage will cause F pollution in the groundwater. Onterministic models or complete stochastic models tothe other hand, Guiyang is a city of severe wateraccurately describe the spatial distribution uncertaintywww.gyig.ac.cn ww w.springerink .com勾Springer.196Chin.J.Geochem.(2012)31:195- 203of various parameters in water environmental assess-2.2 Evaluation model of spatial information statis-ment. Nevertheless, based on regionalized variableticstheory, spatial informnation statistics with variablefunction as its main tool is the science to study naturalThe model applied to analyzing the spatial varia-phenomena with randomness and structural character-tion of parameters in water environmental assessmentistics in their spatial distribution (Wang Renduo andshould help provide the structural information aboutHu Guangdao, 1989). As a result of applying this newspatial distribution parameters in water environmentalstatistical method, the deficiency of classical statisticsassessment, and practically display the spatial correla-in analysis on randomness of spatial distribution pa-tion of various evaluated parameters as well as therameters in water environmental assessment can bereciprocal transformational and synergic relationshipremedied, the spatial structural information and thebetween their randomness and structural characteris-spatial optimum estimation of the distributed parame-tics within the relevant spatial coverage. In actualters can be gained, and the analysis and evaluation ofpractice, we usually provide corresponding theoreticalthe spatial distribution uncertainty of various parame-models for discrele data of variable functions, in orderters in water environmental assessment tends to beto produce a quantitative analysis on the spatial varia-more rational. This paper thus adopts the spatial in-tion of parameters in water environmental assessment.formnation statistics theory and method to analyze andThe frequently used models are Spherical Model,evaluate the karst groundwater environment in Gui-Gaussian Model and Exponential Model (Zhangyang City.Zhengtong and Liu Shuchun, 1999; Jourmel and Hui-jbregts, 1982). Gaussian Model has good continuity2 Statistical model of spatial informationbut poor stability. Exponential Model can not expressfinite range. Spherical Model is a model with good2.1 Mathematical hypothesis of spatial informationcontinuity and stability as well as finite range expres-sion. Therefore, according to the ftting analysis ofstatisticsdata, Spherical Model is more suitable for analyzingthe spatial variation of karst groundwater in the re-search area. Spherical Model is presented as follows:ronmental assessment, on one hand, being affected bythe randomness and other uncertain factors of the mi-croscopic geometric structure of environmental media,啊)=c.+c SpHa)0<網(wǎng)sa1)shows randomness in their spatial distribution, on theother hand, being controlled by various macroscopiclaws of spatial distributed parameters, shows a certainregulanity in spatial distribution, namely structuralSph(a)=2.1-0<同≤acharacleristics. To objectively analyze the spatial2a2a'(2)variation of parameters in water environmental as-sessment with both randomness and structural charac-where, Co, c, a are parameters of water environmentalteristics is the foundation t0 create a reasonable spatialassessment, and they are Nugget constant, Sill anddistribution prediction model of environmental pol-lutants. To analyze the spatial variation uncertainty ofRang.is the Variogram. 川is the distancethe parameters by combining randomness with cer-between the regionalized vaniable information values.tainty, we can portray the representative variables ofRang (a) quantitatively shows the autocorrelationrandomness in spatial distribution parameters as re-average influential range of various evaluated pagionalized variables. The impossibility to obtain arameters in different spatial directions (Zhang Zhengseries of sample values of the evaluation parameters atet al, 2002). By taking into account the relative poany point of the water environmental assessment arcations of the spatial value points of environmentcreates difficulties for the quantiative analysis of spa-evaluation parametcrs, sill (Co+c) is a quantitative de-tial distribution randomness and structure property ofscription of the relative discrete degrec, and a limitvarious evaluation parameters (Liu Guodong and Dingvalue of the random variation and structural variationJing, 1996). Therefore, a necessary mathematical hy-of the spatial distributed parameters. Nugget constantpothesis is needed, namely, stationary hypothesisis the reflection of a random variation factor. It pro-(Ding Jing and Deng Yuren, 1988; Kundzewicz,vides, on one hand, the discontinuity variation and1995). With these basic hypothetical conditions, weexpe中國煤化工fommation on thecan create a fundamental mathematical model for spa-spati;various evaluatedtial variation simulation of the randomness and struc-paran:Y;CNMHGale.ontheotherture property of distribution parameters in enviroD-hand, the Integratea vanauon Intormation on the ran-mental assessmenLdomness and structural characteristics of spatial dis-Chin.J.Geochem.(2012)31:195 -203tributed parameters on the scale smaller than theDaye, Anshun, Guanling and Yangliujing formations.measurement scale.The sampling depth is 104.04 255.5 m with drillholes and pumping as the monitoring types. The3 Spatial variation of karst water environ-groundwater samples were collected with prescribedmentcollection criteria, added with protective agents,sealed and submitted to the laboratory of the Bureau3.1 Overview of the study area and sample collec-of Geology and Mineral Exploration and Develop-tionment of Guizhou Province for testing. Full. analysismethod was employed to determine SO2 , Fe'+ and F.3.1.1 Overview of the study area and karst water en.Mn^* was determined and tested by a specific analysisvironmental backgroundmethod.-130The study area includes Guiyang City and itssuburban area, east to Longli, west to Yeyatang, southDulayigSanjing Riverto Huaxi and north to Dulaying. The specific location10Dashandkongnis shown in Fig. 1. Located in the watershed area ofBbu'aothe Yangtze River Basin and the Pearl River Basin, theios -3012 1205study area is at 26°11' to 27*22' noth latitude and1202512034 1204106*07' to 107917' east longitude. The study area,11 ixg。1201/ Wudangmarked by uneven terrain, is a basin-shaped depres-sion, with its average altitude of 1218 m. The altitudeNanming Riner 13267of most parts of the area is from 1000 to 1400 m. InGuiyang. Cit1206,the subtropical humid temperate climate, the study309area has its annual average temperature of around12072，12075 o“2024 201200Longi .15.3C，and the annual average preciptation of1197-1248 mm. Its dense population is distributed in2047 2052the center of the city, mainly in both Nanming and心*n "y Caninging100Yunyan districts, with a total pollution of 18077 per-。2034ache1004son/sq-km in these two districts according to the sta-Ergerhitistics data. And the specific population distribution is110091101)also shown in Fig. 1. The exposed strata (outcrop) aresampling poinrelatively complete, which are from the Cambrian toHuxi Rincr11020可Donedly pulhtedthe Quatemary, except for the Cretaceous, mainly de-0123 4kmposited as carbonate rock and clastic rock. The stratawere mostly deposited in the Triassic, secondly in theFig 1. Disribuio of tbe sampling points.Permian, with the main outcrop category of shallowoceanic carbonate platform (i.e. dolomite and lime-3.2 Spatial variation of evaluation factorsstone), and the continental clastic strata after the laterTriassic. In the study area, folds are clear, faults andfissures are well developed. Particularly in the urbanThough there are many items tested in qualityarea and northern Guiyang, the faults and fissuresanalysis of karst groundwater in the study area, thecrisscross and the underground karst fissures andheavy metal ions Cd, Cr, and As are all within thetubes cluster together in a dense network, which causestipulated standards according to the data analysis ofthe zones to be full of karst groundwater, and ex-the study area in three years (2006 -2008). Therefore,tremely vulnerable to pollution at the same time.our types of ions which are easy to cause karstgroundwater。pollution and exceed the standards,3.1.2 Sample collectionnamely, SO4, Fe*, Mn*, and F, were selected toconduct the analysis of water eavironment change.The study area of quality testing and control, nowTbe testing data of karst groundwater quality in thecovering 768 km', has 61 groundwater quality detec-study就ea from 2006 to 2008, provided by the Geo-tion points. For the specific distribution of the points,logical Environmental Monitoring Station of Guizhouplease sce Fig. 1. Samples were collected respectively, A comparativein dry season (March) and wet scason (uly) for waterstudy中國煤化工our seced ionsquality analysis. The aquifers of the monitoring pointsin botwas conducted toare composed of limestone of the Permian Maokouanaly2CNMH. Gat ofkaitFormation, and limestone and dolomite of the Triassicgroundwater environment.Chin.J.Geochem.(2012)31:195- -20399increase in the dry seasons.all appear in the upper reaches of the Nanming Riverthat may cause water pollution of the river.3.2.2 Fe'+The Kriging evaluation isoline map of Fe+ in300both wet and dry seasons (Figs. 4 and 5) indicates thatDulariduethe anomaly belt of Fe* is mainly distributed inBaba' a0, Mawangmiao and Sanqiao in northwesternGuiyang City.037 。Yoliane River00 E200-ulayingJanJiang Hioer0w.:50 k22 120005150LomYuliang Rier100129Ermrn13-2 90,。]Sampling網(wǎng)imt50-0= JRiver201011 of 20002lsoline of 20021248回Jtolime of 200，101550ErgerhalFig. 5. Evaluation isolie map of Fe2* in the dry seasons of threeconsecuive years from 2006 lo 2008.L。Jsuplion pelnt511020-1soll of 200In the dry seasons of the three years from 2006 to01232008, the high concentration anomaly belt of Fe'+ ap-5(20025peared in Baba' ao, Mawangmiao and Sanqiao inFig 4. Evaluation isolie map of Fe2+ in the wet seasons of threenorthwestern Guiyang City, with the highest concen-consecutive years from 2006 to 2008.tration of 0.23 mg/L from detection point 12025,which was similar to that in the wet seasons. TheIn the wet seasons of 2006, the high concentra-highest Fe concentrations in the dry seasons wereion anomaly belt of Fe'+ appeared in Baba'ao inhigher than those in the wet seasons, with the concen-tration variation value of 0.05 mg/L, which was thetration of 0.05 mg/L from detection point 12025. Insame as the highest Fe* concentration in the wet sea-the wet seasons of 2007 and 2008, the high concentra-sons of 2006. The reason for that can be mainly asso-ion anomaly belt of Fe+ appeared in Sanqiao inciated with precipitation and human activities.westem Guiyang City, with the highest concentrationof 0.18 mg/L. The reasons for the anomaly may be:3.2.3 Mn?+(1) original non-point source pollution. There existcoal measure strata deposited in the Permian LongtanThe Kriging evaluation is oline map of Mn2+ inFormation in orthwesterm Guiyang City. With theboth wet and dry seasons Figs. 6 and 7) indicates thatexploitation of coal underground, a large amount ofthe anomaly belt of Mn2+ is mainly distributed inFe'+ enters into the underground aquifer in northwest-Baba'ao and Mawangmiao in northwesterm Guiyanger Guiyang and contaminate the groundwater; and8s well as the center of the city.(2) industrial point source polution. The nitrogen ferIn the wet seasons of the hree years from 2006tilizer plant, phosphate fertilizer plants, pickle facto-to 2008, the high concentation anomaly belt of Mn2+ries, machinery plants, agricultural machinery facto-main中國煤化工,Mawangmiao inries, waste recycling plants, bean products factories,norththe highest Mn2+pharmaceutical factories, hospitals and steel plants inconcegYHCN M H Gi detecticn pointnorthwesterm Guiyang all can cause Fe'+ contamina-12025. In tne wet season ot 2U07, the highest Mn2+tion. The anomaly zones of high Fes* concentrationsconcentrations monitored in Xintianzhai and Wudang200ChinJ.Geochem.(2012)31:195- 203in northeasterm Guiyang from detection points 12042and 12014 ranged from 0.08 to 0.14 mg/L. In the wet00season of 2008, the highest Mn2+ concentrationsDulayingSanjin Rivermonitored in Longdongbao in southeasterm Guiyangfrom detection point 12055 were 0.1 mg/L. According50-3ios 3012to the groundwater quality standards, the localgroundwater can be listed from Grade II to Grade IV.The main reasons for the water pollution are: (1)07‘Yullang Riveroriginal non-point source pollution. There exist coal00-Perming River. 106measure strata deposited in the Permian LongtanFormation in northwester Guiyang City. With theexploitation of coal underground, a large amount of50- 512082LomgliMn*+ enters into the underground aquifer in north-westermn Guiyang and contaminate the groundwater;and (2) industrial point source pollution.01102007Ergexhai↑30Sanftarg River二Sapn PoImt13009haxst River2 Isoline of 206io18lsoline of 2007250-joms 3012 130050123k2lsoline of 2001204250Fig 7. Evaluation isoline map of Mn2t in the dry seasons of threeTulieng River20consecutive years from 2006 to 2008.Jantine Rier 120103.2.4 F12072， 120752bLongll .The Kriging evaluation isoline map of F in bothwet and dry seasons (Figs. 8 and 9) indicates that theanomaly belt of F is mainly distributed in northwest-Kiooheu0em Guiyang City.Erguhal110135000個Huani fiver2 Jol of 200( DU1qyiheui012342olite of 20072loine of 200815025得中bin，0.1.Fig 6. Evaluation isoline mp of Mn2* in the wet seasons of threeconsecutive yeans from 2006 to 2008.0oSaraing Rirver 1307In the dry seasons of the three years from 2006 to2008, the general distribution of the high concentra-tion anomaly belt of Mn'+ was different from that in50 f1280loel!the wet seasons. It is mainly distributed in Baba'aoand Mawangmiao in northwesterm Guiyang as well asthe center of the city, with the highest Mn^+ concentra-tion of 0.54 mg/L from monitoring point 3026 in theseErgethaithree consecutive years. However, the highest Mn2+. 200concentration in the dry seasons was the same as that/110150-1in the wet seasons. It is because of the fact that thishuati Rirer10011me of 20area is the main industrial and domestic pollution areaZnollne of 200中國煤化工ntmin Guiyang City. The groundwater quality in this areais relatively poor, falling in Grade III and the amountMYHCNM H G2∞0of Mn2+ exceeds the standards. Thbus, it is the areaFig 8. Evaluation isoline map ofF in the wet scasons of three con-with relatively poor groundwater quality in Guiyang.secutive years from 20006 to 2008.Chin.J.Geochem.(2012)31:195- 203201-10S04, Fet, Mn?+ and F is mainly due to original30point source pollution and human activities (includingSnjang Biverindustrial pollution and domestic contamination), and-0.1the spatial distribution and variation tendency of theseofour pollution sources are similar to each other. By thecombinatorial analysis chart produced with the char-acteristics of each one, we can study the variation ofthese four pollutants.20Kering Rinr 12067Based on the analyses of the combinatorial chart1280(Figs. 10 and 11) of SO4*, Fe'+, Mn2+ and F, thevariation characteristics of these four pollutants can be150-12018Legllshown in the following:In the wet seasons, the anomaly belts of thesefour pollutants are basically overlapped, with the main100 )10100distribution area in Baba' ao, Mawangmiao and San-Erethiqiao in northwestern Guiyang City. The high SO4~concentration belt appears in Xintianzhai in north-@Sapling pointeastern Guiyang together with slight F anomaly. FromRverHhopri Rier2006 to 2008, these four pollutants formed an anom-C 1lwolime of 200012342 Isoliue of 200aly area with a significant gradient change in Baba'ao,口Isoline of 200and formed a small anomaly area with noticeable200 .250variation in the downtown area and Xintianzhai.Anomaly areas between various pollutants show slightdifferentiation, and have a tendency of urban-towardFig. 9. Evaluation isoline map of F in the dry seasons of three com-extension.secutive years from 2006 to 2008.The distribution of the anomaly belts of thesefour pollutants in the dry seasons is similar to that inIn the wet seasons of the three years from 2006the wet seasons, the anomaly areas are basically over-to 2008, the high concentration anomaly belt of F islapped, which are mainly distributed in Baba'ao,mainly distributed in Baba'ao in northwester Gui-Mawangmiao and Sanqiao in northwestem Guiyangyang City, with the highest F concentration of 1.44City. The anomaly belts in the dry seasons appearmg/L from detection point 12025 in these three con-more centralized than those in the wet seasons.secutive years. The main reasons for the anomaly maybe the industrial point source pollution and domesticdifferentiation. The distribution area of these fourcontamination. The industrial wastewater dischargedpollutants in the dry seasons is larger than that in theby the aluminium plants and tire factories in north-wet seasons.The above analyses show that the overall varia-westerm Guiyang City together with domestic sewagetion tendency of these four pollutants (SO42, Fe+t,seriously pollutes local karst groundwater.In the dry seasons of the three years from 2006 toMn2+ and F) is that they are mainly distributed in2008, the high concentration anomaly belt of F is dis-Baba' ao, Mawangmiao and Sanqiao in northwestermtributed in Baba'ao, northwestem Guiyang City, withGuiyang City in both wet and dry seasons, and thethe highest F concentration of 1.6 mg/L from detec-anomaly areas of them are of more centralized dis-tion point 12025 in these three consecutive years,tribution. The reason for the phenomena is that thewhich is similar to that in the wet seasons. The highestindustrial point source pollution is the main pollutionF concentrations in the dry seasons were higher thansource in the dry seasons, whereas, in the wet sea-those in the wet seasons, with the concentration varia-sons, the range of pollution is expanded due to thetion value of 0.16 mg/L, which may be caused by pre-influence of sewage irrigation pollution and rain-water leaching, which make the distribution mannercipitation and human activities.of anomaly belts turm from centralization to decen-tralization.4 Combinatorial analyses on spatial infor-To sum up, the combinatorial chart indicates thatmation statistics for karst waterthe anomaly belts of s04", Fe*, Mn2+ and F aremainl中國煤化工igmiao and San-qiao.ly belts of theseThe above spatial distribution analyses of evalua-HCNMHGsfromthepollution factors indicate that the pollution of caused bytion of sewage irrigation.202ChinJ.Geochem.(2012)31:195- -203的:(293.0Fig. 10. Evaluation isoline combinatorial map of SO2 , Fe", Mn2* and F in the wet sceasons of three consecuive years from 2006 to 2008.蘭20Fig. 11. Evaluation iolie combinalorial map of so*". Fe*. Mn* and F in the dry seasons of tree cnsecutve years from 200 to 208.85 Conclusions and suggestionsCity. The fact that a larger distribution area of pollu-tion and a lower degree of contamination appear in theThe uncertainty of spatial distribution parameterswet seasons rather than in the dry seasons, mainly re-in groundwater assessment mainly results from thesulting from original non-point source pollution andanisotropy of natural water-bearing media, distortionthe contamination caused by human activities (in-of sampling and testing as well as testing errors. Thiscluding both industrial and domestic pollution ).uncertainty of spatial distribution has its particularityof both randomness and structure. Thus, with variableAcknowledgementsThis research project is fi-function theoretical model, integrated variation indexnancially supported by the Natural Science Founda-and various spatial optimum estimation, the spatialtion of Guizhou Province [Grant No. J(2009)2029],information statistics analysis can provide a more ra-Leading Academic Discipline Program, 211 Projecttional evaluation on the uncertainty of spatial distribu-for Guizhou University (the 3* phase), Young Scien-tion parameters in groundwater assessment than clas-tists Project of Natural Science Foundation ofsical statistics methods.Guiz中國煤化工)072), Young ScirThe anomaly belts of so2, Fe+. Mn2+ and F areentist1 lege of Resourcesmainly distributed in Baba' ao, Mawangmiao and San-andYHC N M H G Guizhou Univer-qiao and the downtown area in northwestem Guiyangsity (Grant NO. chI UYUC).ChinJ.Geochem.(2012)31:195 -203203ReferencesLiu Guodong and Ding Jing (1996) A review and prospet of unertaintymethods employed in the water environment 小Advances in Envi-ronmental Sciences 4, 46-51 (in Chinese with Engish abstract),Ding Jing and Deng Yuren (1988) Siochastie Hydrology [M]. ChengduSun Neczheng (1989) Mathematical Modeling of Groundwater PolutionUniversity of Science and Technology Publishing House, Chengdu (in(M]. pp.256 259. Geolngical Publishing House. Beijing.Chinese).Wang Renduo and Hu Guangdao (1989) Linear Geosaristcs IM]. Geologi-Journel A.G and Hijbregrs C.H. (uanslaed by Hou Jingui, Huangcal Publising House. BejigJingxian et al.) (1982) Mining Geostaristics (M]. pp.30 169. Metal.lungy Industy Press. Beijing (in Chinese).ciple and method for the spatial variaion simulaing of parameiers inKundzewicr zW. (1995) New Uncernainry Concepts in Hydrology andthe asessmnent of water environmental U Geotechnical InvesigaionWarer Resore |M]. Cambridge Univesity Pres, London.and Sunveying, 178, 17-43.Liang Yunchao, Liu Congqiang. and Satake H. et al. (2008)87 a and8*sZhang Zhengtong and Liu Shuchun (1999) Spatial statistical analysis onVariations of C- and sO.2 in groundwater and surface water of Gui-randomncs of spatal disributed parameters in groundwater envi-yang area, China [1]. Advances in Earth Sciences. 2 151-159 (inronmental assesment [I]. Aca Scieniae Circunsanice. 19,Chinese widh English abstract).410- 414.中國煤化工MHCNMHG