Siyabonga ngokuvakashela i-Nature.com.Inguqulo yesiphequluli oyisebenzisayo inokusekelwa okulinganiselwe kwe-CSS.Ukuze uthole ulwazi olungcono kakhulu, sincoma ukuthi usebenzise isiphequluli esibuyekeziwe (noma uvale imodi yokuhambisana ku-Internet Explorer).Okwamanje, ukuze siqinisekise ukwesekwa okuqhubekayo, sizobonisa isayithi ngaphandle kwezitayela ne-JavaScript.
Ukunukubezeka komhlabathi kuyinkinga enkulu ebangelwa imisebenzi yabantu.Ukusatshalaliswa kwendawo kwezinto ezingaba ubuthi (PTEs) kuyahlukahluka ezindaweni eziningi zasemadolobheni nasezindaweni eziseduze nedolobha.Ngakho-ke, kunzima ukubikezela ngokwendawo okuqukethwe kwama-PTEs enhlabathini enjalo.Isamba samasampula angu-115 atholwe ku-Frydek Mistek e-Czech Republic.), i-magnesium enqunyiwe) i-calcium nickMg (i-potassium nickM) I-spectrometry yokuphuma kwe-plasma ehlanganisiwe ngendlela eguquguqukayo.I-variable yokusabela i-Ni kanye nezibikezelo ziyi-Ca, Mg, kanye ne-K.I-matrix yokuhlobana phakathi kokuguquguquka kwempendulo kanye nokuguquguquka kokubikezela kubonisa ukuhlobana okwanelisayo phakathi kwezakhi.Imiphumela yokubikezela ibonise ukuthi Ukusekela Umshini Wokuhlela I-Vector (SVMR) wenze kahle, nakuba impande yayo elinganiselwe isho iphutha lesikwele/i-RM. .946 mg/kg) beziphakeme kunezinye izindlela ezisetshenzisiwe.Amamodeli axutshiwe e-Empirical Bayesian Kriging-Multiple Linear Regression (EBK-MLR) asebenza kabi, njengoba kufakazelwa ama-coefficients wokunquma angaphansi kuka-0.1.I-Empirical Bayesian Kriging-Support Vector Machine Regression (mg5kg) enemodeli ehamba phambili engu-MAK/SERM7 kanye nemodeli engcono kakhulu ye-SVMR7 (77.368 mg/kg) amanani kanye ne-coefficient ephezulu yokunquma (R2 = 0.637).Ukukhishwa kwesu lokumodela le-EBK-SVMR kubonwa ngeso lengqondo kusetshenziswa imephu ezihlelayo.Ama-neurons aqoqiwe endizeni yemodeli ye-hybrid CakMg-EBK-SVMR abonisa amaphethini amaningi ombala abonisa imiphumela ye-SVMR eqagela i-Nipering yedolobha. kuyindlela esebenzayo yokubikezela ukugxila kuka-Ni emhlabathini wasemadolobheni naseduze kwedolobha.
I-Nickel (Ni) ibhekwa njengesakhamzimba esincane ezitshalweni ngoba inomthelela ekulungiseni i-nitrogen emkhathini (N) kanye ne-urea metabolism, kokubili okudingekayo ukuze imbewu ihlume.Ngaphezu komnikelo wayo ekuhlumeni kwembewu, i-Ni ingasebenza njengesivimbeli sesikhunta nesamagciwane futhi ikhuthaze ukukhula kwezitshalo. ama-fertilizer asekelwe ku-ckel ukuze kuthuthukiswe i-nitrogen fixation2.Ukusetshenziswa okuqhubekayo komanyolo owenziwe nge-nickel ukuze kunothisa umhlabathi futhi kwandiswe ikhono le-legumes ukulungisa i-nitrogen enhlabathini ngokuqhubekayo kwandisa ukugcwala kwe-nickel emhlabathini. i-trient yokukhula kwezitshalo1.Ngokusho kwe-Liu3, i-Ni itholwe iyisici se-17 esibalulekile esidingekayo ekuthuthukisweni nasekukhuleni kwezitshalo.Ngaphezu kwendima ye-nickel ekuthuthukisweni nasekukhuleni kwezitshalo, abantu bayayidinga ukuze benze imisebenzi ehlukahlukene.I-Electroplating, ukukhiqizwa kwama-alloys asekelwe ku-nickel, kanye nokukhiqizwa kwemishini yokuthungela kanye ne-automotive nickel sector embonini ye-automotive nickel. -ama-alloys nama-athikili ane-electroplated asetshenziswe kakhulu ku-kitchenware, izinsiza ze-ballroom, izimpahla zembonini yokudla, ugesi, izintambo kanye nekhebula, izinjini zamajethi, izifakelo zokuhlinzwa, izindwangu, nokwakhiwa kwemikhumbi5. ukuqhuma kwe-canic, uhlaza, ukusha kwehlathi, nezinqubo zokuma komhlaba;kodwa-ke, imithombo ye-anthropogenic ihlanganisa amabhethri e-nickel/cadmium embonini yensimbi, i-electroplating, i-arc welding, idizili kanye namafutha kaphethiloli, kanye nokukhipha komoya okuphuma ekushisweni kwamalahle nokushiswa kwemfucuza nokushiswa kodaka Ukunqwabelana kwe-Nickel7,8.Ngokuvumelana no-Freedman no-Hutchinson9 no-Manyiwa etchinson9.10, imithombo eyinhloko yokungcoliswa kwenhlabathi engaphezulu endaweni eseduze neseduze ikakhulukazi izincibilikisi nezimayini ezenziwe nge-nickel-copper.Inhlabathi ephezulu ezungeze indawo yokuhluza i-nickel-copper yase-Sudbury e-Canada yayinezinga eliphakeme kakhulu lokungcoliswa kwe-nickel ku-26,000 mg/kg11. 11.Ngokuvumelana ne-Alms et al.12, inani le-nickel elikhiphekayo le-HNO3 endaweni ephezulu yokulinywa yesifunda (ukukhiqizwa kwe-nickel eRussia) lalisuka ku-6.25 liye ku-136.88 mg/kg, elihambisana nenani elingu-30.43 mg/kg kanye nokuhlushwa okuyisisekelo okungu-25 mg/kg faka noma ungcolise inhlabathi.Imiphumela engase ibe khona ye-nickel kubantu ingase iholele kumdlavuza ngokusebenzisa i-mutagenesis, ukulimala kwechromosomal, isizukulwane se-Z-DNA, ukulungiswa kwe-DNA excision evinjiwe, noma izinqubo ze-epigenetic13.Ekuhlolweni kwezilwane, i-nickel itholakale inamandla okubangela izinhlobonhlobo zezimila, futhi i-carcinogenic nickel complexes ingase ikhulise izimila ezinjalo.
Ukuhlolwa kokungcoliswa komhlabathi kuye kwachuma ezikhathini zamuva nje ngenxa yezinkinga eziningi eziphathelene nempilo ezivela ebudlelwaneni bezitshalo nenhlabathi, ubudlelwano bezinto eziphilayo zomhlabathi nomhlabathi, ukucekelwa phansi kwemvelo, nokuhlola umthelela wendawo. i-ive soil mapping (PSM).According to Minasny and McBratney16, predictive soil mapping (DSM) has proven to be a prominent subdiscipline of soil science.Lagacherie and McBratney, 2006 define DSM as “the creation and filling of spatial soil information systems through the use of in situal and nonBrat-non-Brand information systems”. t al.I-17 iveza ukuthi i-DSM yesimanje noma i-PSM iyindlela ephumelela kakhulu yokubikezela noma yokwenza imephu ukusatshalaliswa kwendawo kwama-PTE, izinhlobo zenhlabathi kanye nezakhiwo zomhlabathi.I-Geostatistics kanye ne-Machine Learning Algorithms (MLA) izindlela zokumodela ze-DSM ezidala amamephu edijithali ngosizo lwamakhompyutha asebenzisa idatha ebalulekile nencane.
I-Deuts18 ne-Olea19 zichaza i-geostatistics ngokuthi "iqoqo lamasu ezinombolo abhekana nokumelwa kwezimfanelo zendawo, ikakhulukazi zisebenzisa amamodeli we-stochastic, njengokuthi ukuhlaziywa kochungechunge lwesikhathi kuyichaza kanjani idatha yesikhashana."Ngokuyinhloko, i-geostatistics ihilela ukuhlolwa kwama-variograms, okuvumela Ukulinganisa nokuchaza ukuncika kwamanani wendawo kusuka kudathasethi ngayinye20.Gumiaux et al.20 iphinde ibonise ukuthi ukuhlolwa kwama-variograms ku-geostatistics kusekelwe ezimisweni ezintathu, okuhlanganisa (a) ukwenza ikhompuyutha isikali sokuhlobana kwedatha, (b) ukuhlonza nokwenza ikhompuyutha i-anisotropy ngokungafani kwedathasethi kanye (c) ngaphezu kokuthi Ngaphezu kokucabangela iphutha elingokwemvelo ledatha yokulinganisa ehlukanisiwe nemiphumela esetshenziswayo endaweni, le mithelela ye-gepoo esetshenziswayo ibuye ibe yimiphumela yendawo. izibalo, ezihlanganisa i-kriging evamile, i-co-kriging, i-kriging evamile, i-empirical Bayesian kriging, indlela ye-kriging elula nezinye izindlela zokuhumusha ezaziwayo ukuze kumephu noma ukubikezela i-PTE, izici zomhlabathi, nezinhlobo zenhlabathi.
I-Machine Learning Algorithms (MLA) iyindlela entsha uma kuqhathaniswa esebenzisa amakilasi edatha amakhulu angawolayini, akhuthazwa ama-algorithms ngokuyinhloko asetshenziselwa ukumbiwa kwedatha, ukuhlonza amaphethini kudatha, futhi asetshenziswa ngokuphindaphindiwe ekuhlukaniseni emikhakheni yesayensi efana nesayensi yenhlabathi nemisebenzi yokubuyisela. Amaphepha amaningi ocwaningo ancike kumamodeli we-MLA ukuze abikezele i-PTE emhlabathini, njenge-Tan et al.22 (amahlathi angahleliwe okulinganisa kwensimbi esindayo emhlabathini wezolimo), Sakizadeh et al.23 (ukumodela kusetshenziswa imishini ye-vector yokusekela kanye namanethiwekhi e-neural okwenziwa) ukungcoliswa komhlabathi).Ngaphezu kwalokho, i-Vega et al.24 (INQALI yokulinganisa ukugcinwa kwensimbi esindayo kanye ne-adsorption enhlabathini) Sun et al.25 (ukusetshenziswa kwe-cubist ukusatshalaliswa kwe-Cd enhlabathini) namanye ama-algorithms afana nomakhelwane oseduze kuka-k, ukuhlehla okuthuthukile okujwayelekile, kanye nokuhlehla okuthuthukile Izihlahla nazo zasebenzisa i-MLA ukubikezela i-PTE enhlabathini.
Ukusetshenziswa kwama-algorithms e-DSM ekubikezeleni noma ekudwebeni kubhekana nezinselele ezimbalwa.Ababhali abaningi bakholelwa ukuthi i-MLA iphakeme kune-geostatistics futhi ngokuphambene nalokho.Nakuba eyodwa ingcono kunomunye, ukuhlanganiswa kwakho kokubili kuthuthukisa izinga lokunemba kwemephu noma ukubikezela ku-DSM15.I-Woodcock ne-Gopal26 Finke27;U-Pontius no-Cheuk28 kanye no-Grunwald29 baphawula ngokushiyeka kanye namaphutha athile kumephu yomhlabathi ebikezelwe.Ososayensi bomhlabathi bazamile amasu anhlobonhlobo ukuze bathuthukise ukusebenza kahle, ukunemba, kanye nokubikezela kwemephu ye-DSM kanye nokubikezela.Inhlanganisela yokungaqiniseki nokuqinisekisa ingenye yezici eziningi ezihlukene ezididiyelwe ku-DSM ukuze kuncishiswe ukusebenza kahle.15 iveza ukuthi ukuziphatha kokuqinisekisa nokungaqiniseki okwethulwa ngokudalwa kwemephu nokubikezela kufanele kuqinisekiswe ngokuzimele ukuze kuthuthukiswe ikhwalithi yemephu.Imikhawulo ye-DSM ibangelwa ikhwalithi yenhlabathi ehlakazekile ngokwendawo, ebandakanya ingxenye yokungaqiniseki;kodwa-ke, ukuntuleka kwesiqiniseko ku-DSM kungase kuvele emithonjeni eminingi yephutha, okuyiphutha le-covariate, iphutha lemodeli, iphutha lendawo, kanye nephutha lokuhlaziya 31.Ukungalungi kwe-Modelling okufakwe ku-MLA kanye nezinqubo ze-geostatistical kuhlotshaniswa nokuntula ukuqonda, ekugcineni okuholela ekweqinisweni kwenqubo yangempela32.Kungakhathaliseki ukuthi imodeli ye-maccurated, imodeli ye-maccurated, i-accurated imodeli ye-maccurated imodeli ye-maccurated izibikezelo, noma ukungenelela33.Muva nje, sekuvele umkhuba omusha we-DSM okhuthaza ukuhlanganiswa kwe-geostatistics kanye ne-MLA ekudwebeni imephu nokubikezela.Ososayensi abaningi benhlabathi nababhali, njengoSergeev et al.34;I-Subbotina et al.35;Tarasov et al.36 kanye noTarasov et al.Abangu-37 basebenzise ikhwalithi enembile ye-geostatistics nokufunda komshini ukuze bakhiqize amamodeli ayingxube athuthukisa ukusebenza kahle kokubikezela nokubhala imephu.ikhwalithi.some yalezi hybrid noma amamodeli we-algorithm ahlanganisiwe yinethiwekhi ye-neraural nettual desiging (MLP-RK), i-anc-nkrk) 37 I-Multilayeral Network Kriging-Multilayer Perceptron (Ann-K-MLP) 37 noCor Iging ne-gaussian inqubo yenqubo regression38.
Ngokuka-Sergeev et al., ukuhlanganisa amasu okumodela ahlukahlukene kunamandla okuqeda amaphutha futhi kwandise ukusebenza kahle kwemodeli eyingxube ewumphumela kunokuba kuthuthukiswe imodeli yayo eyodwa.Kulo mongo, leli phepha elisha liphikisa ngokuthi kuyadingeka ukusebenzisa i-algorithm ehlanganisiwe ye-geostatistics kanye ne-MLA ukuze kudalwe amamodeli e-hybrid angcono kakhulu ukuze abikezele ukuchuma kwe-Ni ezindaweni zasemadolobheni nasezindaweni ezizungezile ze-KEB imodeli futhi uyixube namamodeli we-Support Vector Machine (SVM) kanye namamodeli we-Multiple Linear Regression (MLR).UkuHybridization kwe-EBK nanoma iyiphi i-MLA akwaziwa.Amamodeli axubile amaningi abonwayo ayinhlanganisela ye-criging evamile, esalayo, ehlehlayo, kanye ne-MLA.EBK indlela yokuhlanganisa i-geostatistical interpolation esebenzisa inkambu ye-spatially defined localized parameters ne-spatially defined localized parameters endaweni, okuvumela ukuhlukahluka kwendawo39.EBK iye yasetshenziswa ocwaningweni oluhlukahlukene, okuhlanganisa ukuhlaziya ukusatshalaliswa kwekhabhoni yemvelo enhlabathini yasemapulazini40, ukuhlola ukunukubezeka komhlabathi41 kanye nokwenza imephu yezakhiwo zomhlabathi42.
Ngakolunye uhlangothi, i-Self-Organising Graph (i-SeOM) i-algorithm yokufunda esetshenziswe ezihlokweni ezihlukahlukene ezifana ne-Li et al.43, Wang et al.44, uHossain Bhuiyan et al.45 kanye no-Kebonye et al.46 Thola izimfanelo zendawo kanye nokuqoqwa kwezinto.Wang et al.44 iveza ukuthi i-SeOM iyindlela yokufunda enamandla eyaziwa ngekhono layo lokuqoqa nokucabanga izinkinga ezingahlangene nomugqa.Ngokungafani nezinye izindlela zokuqaphela iphethini ezifana nokuhlaziywa kwengxenye eyinhloko, ukuhlanganisa okungaqondakali, ukuhlanganisa okulandelanayo, kanye nokwenza izinqumo ngemibandela eminingi, i-SeOM ingcono ekuhleleni nasekuboneni amaphethini e-PTE.Ngokuvumelana ne-War.I-44, i-SeOM ingakwazi ukuqoqa ngokwendawo ukusatshalaliswa kwama-neurons ahlobene futhi inikeze ukubonwa kwedatha yokulungiswa okuphezulu.I-SeOM izobuka ngeso lengqondo idatha yokubikezela ye-Ni ukuze ithole imodeli engcono kakhulu yokubonisa imiphumela yokuchazwa okuqondile.
Leli phepha lihlose ukukhiqiza imodeli yemephu eqinile enembayo ephelele yokubikezela okuqukethwe kwe-nickel ezindaweni ezisemadolobheni kanye naseduze kwedolobha. Sicabanga ukuthi ukwethembeka kwemodeli exubile kuncike kakhulu ethonyeni lamanye amamodeli axhumene nemodeli yesisekelo.ngakho-ke, sizozama ukuphendula imibuzo yocwaningo engase iveze amamodeli axubile.Nokho, inembe kangakanani imodeli ekubikezeleni isici esiqondiwe?Futhi, liyini izinga lokuhlola ukusebenza kahle okusekelwe ekuqinisekiseni nasekuhloleni ukunemba?Ngakho-ke, imigomo ethile yalolu cwaningo bekuwu (a) ukudala imodeli yengxube ehlanganisiwe ye-SVMR noma i-MLR isebenzisa i-EBK (c) ukuqhathanisa umphumela wemodeli ye-pro engcono kakhulu, u-upose imodeli eyisisekelo ye-EBK (c) uqhathanise imodeli ye-pro inhlabathi yasemadolobheni noma eseduze nedolobha , kanye (d) nokusetshenziswa kwe-SeOM ukuze kwakheke imephu enokulungiswa okuphezulu kokuhlukahluka kwendawo ye-nickel.
Ucwaningo lwenziwa e-Czech Republic, ikakhulukazi esifundeni sase-Frydek Mistek esifundeni sase-Moravia-Silesian (bheka uMdwebo 1).Indawo yocwaningo inzima kakhulu futhi iyingxenye enkulu yesifunda sase-Moravia-Silesian Beskidy, esiyingxenye yonqenqema olungaphandle lwezintaba ze-Carpathian. Indawo yocwaningo iphakathi kuka-4 no-0 ′19 ° ′ no-0 ′ 49 ° ukuphakama kuphakathi kuka-225 no-327 m;Nokho, uhlelo lwe-Koppen lokuhlukanisa isimo sezulu sesifunda lulinganiselwe njenge-Cfb = isimo sezulu esishisayo solwandle, Kukhona imvula eningi ngisho nasezinyangeni ezomile.Amazinga okushisa ayahluka kancane unyaka wonke phakathi -5 °C no-24 °C, akuvamile ukuba abe ngaphansi -14 °C noma ngaphezulu kuka-30 °C noma ngaphezulu kuka-30 °C ngonyaka. Indawo yonke ingamakhilomitha-skwele angu-1,208, kanye no-39.38% womhlaba olinywayo kanye no-49.36% wamahlathi. Ngakolunye uhlangothi, indawo esetshenziswe kulolu cwaningo ingamakhilomitha-skwele angu-889.8. E-Ostrava nasezindaweni ezizungezile, imboni yensimbi nemisebenzi yensimbi isebenza kakhulu. sion) kanye nezinsimbi ze-alloy (i-nickel ikhulisa amandla engxubevange kuyilapho igcina i-ductility yayo enhle nokuqina), kanye nezolimo ezijulile ezifana nokufakwa kukamanyolo we-phosphate nokukhiqizwa kwemfuyo kuyimithombo engaba khona yocwaningo ye-nickel esifundeni (isb., ukwengeza i-nickel kumawundlu ukuze kwandiswe izinga lokukhula kwamawundlu kanye nezinkomo ezidla kancane) . izinqubo.Izakhiwo zomhlabathi zihlukaniseka kalula nombala wenhlabathi, ukwakheka, kanye nokuqukethwe kwe-carbonate.Ukwakheka kwenhlabathi kuphakathi nokucolekile, kususelwa ezintweni eziwumzali.Ziyi-colluvial, i-alluvial noma i-aeolian ngokwemvelo.Ezinye izindawo zenhlabathi zibonakala zinamabala ongaphezulu kanye nenhlabathi engaphansi, ngokuvamile enokhonkolo kanye ne-bleaching.Nokho, ama-cambisols nama-stagnosols yizinhlobo ezingu-4 ezivame kakhulu endaweni engu-48 kusukela endaweni engu-48ng kusukela endaweni engu-48 kusukela ku-48ng. 5 m, ama-cambisol abusa i-Czech Republic49.
Imephu yendawo yokufunda [Imephu yendawo yocwaningo yakhiwe kusetshenziswa i-ArcGIS Desktop (ESRI, Inc, inguqulo 10.7, URL: https://desktop.arcgis.com).]
Kutholwe amasampula enhlabathi engaphezulu ayi-115 enhlabathini yasemadolobheni kanye nangaseceleni kwedolobha esifundeni saseFrydek Mistek. Isampula yesampula esetshenzisiwe kwakuyigridi evamile enamasampula omhlabathi aqhelelene ngo-2 × 2 km, futhi inhlabathi engaphezulu yayikalwa ngokujula kuka-0 kuya ku-20 cm kusetshenziswa idivayisi ye-GPS ephathwa ngesandla (Leica Zeno 5 bag, i-Sampletory bag bag, i-Sampletory bag, i-Sampletory bag, i-GPS, i-lab, i-lab, i-Ziplobele ifakwe kahle ku-GPS). .Amasampula omisiwe emoyeni ukuze akhiqize amasampula agayiwe, ahlanjululwe ngohlelo lwemishini (Fritsch disc mill), futhi ahlungwa (usayizi wesisefo 2 mm).Beka igramu elingu-1 lamasampula omhlabathi omisiwe, afakwe i-homogenized futhi ahlungiwe emabhodleleni e-teflon abhalwe ngokucacile. Emkhunjini ngamunye we-Teflon, khipha u-5 ml we-H 37 othomathikhi we-HH 37 ml kanye ne-3 ml ye-HTC othomathikhi engu-3. eyodwa ku-asidi ngayinye), vala kancane bese uvumela amasampula ukuthi ame ubusuku bonke ukuze aphendule (uhlelo lwe-aqua regia) .Beka i-supernatant epuleti lensimbi elishisayo (izinga lokushisa: 100 W kanye no-160 °C) amahora angu-2 ukuze kube lula inqubo yokugaya amasampula, bese upholisa. i-supernatant ibe ishubhu le-PVC elingu-50 ml elinamanzi ahlanjululwe.Ukwengeza, i-1 ml yesisombululo sokuhlanjululwa ihlanjululwe ngo-9 ml wamanzi ahlanzekile futhi ihlungiwe ku-12 ml tube elungiselelwe i-PTE pseudo-concentration.Ukugxila kwama-PTEs (As, Cd, Cr, Cu, Mn, Ni, Cabled by Mgn ICP-Coducts) ma Optical Emission Spectroscopy) (Thermo Fisher Scientific, USA) ngokwezindlela ezijwayelekile kanye nesivumelwano.Qinisekisa Izinqubo Zokuqinisekisa Ikhwalithi Nokulawula (QA/QC) (SRM NIST 2711a Montana II Soil) .Ama-PTE anemikhawulo yokutholwa engaphansi kwesigamu ayengafakwa kulolu cwaningo.Umkhawulo wokuthola we-PTE osetshenziswe kulolu cwaningo lwekhwalithi, i-0u000 yokulawulwa kwekhwalithi ngayinye yayiyi-4. kuqinisekiswa ngokuhlaziya izilinganiso zereferensi.Ukuqinisekisa ukuthi amaphutha ancishisiwe, ukuhlaziya kabili kwenziwa.
I-Empirical Bayesian Kriging (EBK) ingenye yezindlela eziningi zokuhumusha ze-geostatistical ezisetshenziswa ekumodeleni emikhakheni ehlukahlukene njengesayensi yenhlabathi.Ngokungafani nezinye izindlela zokuhumusha ze-kriging, i-EBK ihlukile ezindleleni zendabuko ze-kriging ngokucabangela iphutha elilinganiselwe imodeli ye-semivariogram. indlela yokungaqiniseki nohlelo oluhlotshaniswa nalokhu kuhlelwa kwe-semivariogram eyenza ingxenye eyinkimbinkimbi kakhulu yendlela eyanele ye-kriging.Inqubo yokuhumusha ye-EBK ilandela imibandela emithathu ephakanyiswe ngu-Krivoruchko50, (a) imodeli ilinganisela i-semivariogram kusuka kudathasethi yokufakwayo (b) inani elisha elibikezelwe elisuselwa kudathasethi eqageliwe evela kudathasethi ngayinye eyi-compuvari ekhiqiza imodeli yokugcina eyi-compuvari evela kudathasethi ngayinye ekhiqiziwe. idathasethi ehlanganisiwe.Umthetho wezibalo wase-Bayesia unikezwa njengengemuva
Lapho \(Prob\left(A\right)\) imele okwangaphambili, \(Prob\left(B\right)\) amathuba asemaceleni ashaywa indiva ezikhathini eziningi, \(Prob (B,A)\ ) .Isibalo se-semivariogram sisekelwe emthethweni we-Bayes, obonisa ukuthambekela kokubheka idathasethi ye-observation, inani le-semiovari elinqunywa kusuka ku-Bay semivariogram. okusho ukuthi mangakanani amathuba okudala isethi yedatha yokubuka kusuka ku-semivariogram.
Umshini wokusekelayo uyi-algorithm yokufunda yomshini ekhiqiza i-hyperplane ehlukanisayo efanele ukuze ihlukanise amakilasi afanayo kodwa angazimele ngokomugqa.I-Vapnik51 idale i-algorithm yokuhlukanisa inhloso, kodwa isanda kusetshenziswa ukuxazulula izinkinga eziqondiswe ekuhlehleni. Ngokuka-Li et al.52, i-SVM ingenye yezindlela ezinhle kakhulu zokuhlukanisa izigaba zomshini futhi isetshenziswe ukuhlanganiswa kwe-SVMupport eyodwa. ion - SVMR) isetshenziswe kulokhu kuhlaziya.UCherkassky kanye no-Mulier53 baphayona i-SVMR njengokuhlehla okusekelwe ku-kernel, ukubala kwakho okwenziwe kusetshenziswa imodeli yokuhlehla ngomugqa enemisebenzi yamazwe amaningi.55, i-epsilon (ε) -SVMR isebenzisa idathasethi eqeqeshiwe ukuze ithole imodeli yokumelela njengomsebenzi ongazweli we-epsilon osetshenziswa ukwenza imephu idatha ngokuzimela ngokukhetha okungcono kakhulu kwe-epsilon kusukela ekuqeqesheni idatha ehlotshaniswayo.Iphutha lebanga elisethiwe alinakwa kusukela enanini langempela, futhi uma iphutha likhulu kuno-ε(ε), imodeli yedatha yokuqeqesha eyinkimbinkimbi iphinde inciphise imodeli ye-vector eyinkimbinkimbi. s.Isibalo esihlongozwe yi-Vapnik51 siboniswe ngezansi.
lapho u-b emele umkhawulo wesikala, \(K\left({x}_{,}{ x}_{k}\right)\) imele umsebenzi we-kernel, \(\alpha\) imele isiphindaphindi se-Lagrange, N imelela idathasethi yezinombolo, \({x}_{k}\) imele okokufaka kwedatha, futhi \.i-OnM i-SV esebenza ngokhiye osetshenziswayo, futhi \.i-OnM i-SV esebenza ngokhiye osetshenziswayo umsebenzi wesisekelo se-radial (RBF).I-RBF kernel isetshenziswa ukuze kutholwe imodeli ye-SVMR elungile, ebaluleke kakhulu ukuze kutholwe inhlawulo ecashile kakhulu yokusetha isici C kanye ne-kernel parameter gamma (γ) yedatha yokuqeqeshwa ye-PTE. Okokuqala, sihlole isethi yokuqeqesha sabe sesihlola ukusebenza kwemodeli kusethi yokuqinisekisa.Ipharamitha yokuqondisa esetshenziswa i-sigm ye-sigima.
Imodeli yokuhlehla kwemigqa eminingi (MLR) imodeli yokuhlehla emelela ubuhlobo phakathi kokuhluka kwempendulo kanye nenani lokuguquguquka kokubikezela ngokusebenzisa amapharamitha ahlanganiswe ngomugqa abalwe kusetshenziswa indlela yesikwele esincane.Ku-MLR, imodeli yesikwele esincane iwumsebenzi wokubikezela wezakhiwo zenhlabathi ngemva kokukhethwa kweziguquguquko ezichazayo.Kuyadingeka ukusebenzisa impendulo ukuze kusungulwe ukuguquguquka okusetshenziswayo njengomugqa wokuchaza njengomugqa wencazelo P.P. tory variables.Isibalo se-MLR sithi
lapho u-y ewukushintshashintsha kwempendulo, \(a\) i-intercept, n iyinani lezibikezelo, \({b}_{1}\) ukuhlehla okuyingxenye kwama-coefficients, \({x}_{ i}\) imele isibikezelo noma i-variable echazayo, futhi \({\varepsilon }_{i}\) imele iphutha eliphindwe kabili, eyaziwa ngokuthi imodeli.
Amamodeli axutshiwe atholwe ngokuhlanganisa i-EBK nge-SVMR ne-MLR. Lokhu kwenziwa ngokukhipha amanani abikezelwe kusukela ekuhumusheni kwe-EBK. Amanani abikezelwe atholwe ku-Ca, K, ne-Mg ehlanganisiwe atholakala ngenqubo yokuhlanganisa ukuze kutholwe okuguquguqukayo okusha, okufana ne-CaK, i-CaMg, ne-KMg, kanye ne-KMg. okuguquguqukayo okutholiwe i-Ca, K, Mg, CaK, CaMg, KMg kanye ne-CaKMg.Lokhu okuguquguqukayo kwaba izibikezelo zethu, okusiza ukubikezela ukugxila kwe-nickel enhlabathini yasemadolobheni kanye ne-peri-urban.I-algorithm ye-SVMR yenziwa kuzibikezelo ukuze kutholwe imodeli exubile ye-Empirical Bayesian Kriging-Support Vector Machine (EBK_lySVM) iphinde ibe yi-piemical algorithm ukuthola imodeli ye-Empirical Empiri). I-Bayesian Kriging-Multiple Linear Regression (EBK_MLR).Imvamisa, okuguquguqukayo okuthi Ca, K, Mg, CaK, CaMg, KMg, kanye ne-CaKMg kusetshenziswa njengama-covariates njengezibikezelo zokuqukethwe kwe-Ni emhlabathini wasemadolobheni kanye nangasedolobheni.Imodeli eyamukeleka kakhulu etholiwe (EBK_SVM noma i-EBK_ML noma i-EBK_R) iboniswa kusetshenziswa i-Fib-R. igebe 2.
Ukusebenzisa i-SeOM sekuphenduke ithuluzi elidumile lokuhlela, lokuhlola, kanye nokubikezela idatha emkhakheni wezezimali, ukunakekelwa kwezempilo, imboni, izibalo, isayensi yomhlabathi, nokunye.I-SeOM idalwe kusetshenziswa amanethiwekhi e-neural okwenziwa nezindlela zokufunda ezingagadiwe zokuhlela, ukuhlola, nokubikezela.Kulolu cwaningo, i-SeOM yasetshenziselwa ukubona ngeso lengqondo ukugxilwa kwe-Ni ngokususelwa kunqubo ye-Niom yedolobha kanye nokubikezela kwedatha ye-peri-urban. ukulinganisa kusetshenziswa njengokuguquguquka kwevekhtha ye-input-dimensional43,56.Melssen et al.57 chaza ukuxhunywa kwe-vector yokufaka kunethiwekhi ye-neural ngokusebenzisa isendlalelo esisodwa sokufaka ku-vector ephumayo ene-vector eyodwa enesisindo.Okukhiphayo okukhiqizwa yi-SeOM imephu enezinhlangothi ezimbili ehlanganisa ama-neurons ahlukene noma ama-node alukwe abe amamephu ane-hexagonal, ayindilinga, noma ayisikwele we-topological ngokuya ngokusondela kwawo.Ukuqhathanisa osayizi bemephu ngokusekelwe ku-mepographic quant kanye nephutha le-OME (iphutha le-QE) kanye ne-mepographic. 86 kanye no-0.904, ngokulandelana, kukhethwa, okuyiyunithi yemephu engu-55 (5 × 11).Isakhiwo se-neuron sinqunywa ngokwenani lamanodi ku-equation ye-empirical.
Inani ledatha elisetshenziswe kulolu cwaningo amasampula angu-115. Indlela engahleliwe yasetshenziswa ukuze kuhlukaniswe idatha ibe idatha yokuhlola (25% ukuze kuqinisekiswe) namasethi edatha yokuqeqesha (75% ukuze kulinganiswe).Idathasethi yokuqeqeshwa isetshenziselwa ukukhiqiza imodeli yokuhlehla (ukulinganisa), futhi idathasethi yokuhlola isetshenziselwa ukuqinisekisa ikhono lokwenza okuvamile58.Lokhu kwenziwa ukuze kuhlolwe ukufaneleka kwawo wonke amamodeli wenhlabathi asetshenziswayo ahlukahlukene. -inqubo yokuqinisekisa, ephindwa izikhathi ezinhlanu.Okuguquguqukayo okukhiqizwa ukuhumusha kwe-EBK kusetshenziswa njengezibikezelo noma iziguquguquko ezichazayo ukuze kubikezelwe ukuguquguquka okuqondiwe (PTE).Ukumodela kusingathwa ku-RStudio kusetshenziswa ilabhulali yamaphakheji(Kohonen), umtapo wezincwadi, umtapo wolwazi(modelr), umtapo wolwazi("e1071″), umtapo wolwazi("plyr"),umtapo wezincwadi"("ibhulabhulabhu)"("ibhulabhu ye-"Metric") ("ibhulabhu ye-carospect)" ("ibhulabhu ye-"Metric").
Kusetshenziswe amapharamitha okuqinisekisa ahlukahlukene ukuze kunqunywe imodeli engcono kakhulu efanelekile ukubikezela ukugxila kwe-nickel enhlabathini kanye nokuhlola ukunemba kwemodeli nokuqinisekiswa kwayo.Amamodeli we-Hybridization ahlolwe kusetshenziswa iphutha eliphelele eliphelele (MAE), iphutha le-root mean square (RMSE), kanye ne-R-squared noma i-coefficient determination (R2) .I-R2 inkokhiso yempendulo ngokuhluka kwe-RM imelela i-progress. ubukhulu be-ance ezilinganisweni ezizimele buchaza amandla okubikezela emodeli, kuyilapho i-MAE inquma inani langempela lomthamo.Inani le-R2 kufanele libe phezulu ukuze kuhlolwe imodeli yengxube engcono kakhulu usebenzisa imingcele yokuqinisekisa, inani eliseduze liba ngu-1, liphakeme ukunemba.Ngokusho kuka-Li et al.59, inani lokunquma elingu-R2 elingu-0.75 noma ngaphezulu libhekwa njengesibikezeli esihle;ukusuka ku-0.5 ukuya ku-0.75 kuwukusebenza kwemodeli okwamukelekayo, futhi ngaphansi kuka-0.5 ukusebenza kwemodeli okungamukelekile.Lapho ukhetha imodeli kusetshenziswa izindlela zokuhlola imibandela yokuqinisekisa ye-RMSE kanye ne-MAE, amanani aphansi atholiwe ayenele futhi athathwa njengokukhetha okungcono kakhulu.Isibalo esilandelayo sichaza indlela yokuqinisekisa.
lapho u-n emelela usayizi yenani eliboniwe\({Y}_{i}\) limelela impendulo elinganisiwe, futhi \({\widehat{Y}}_{i}\) imelela inani lokuphendula elibikezelwe, ngakho-ke, ekubonweni kokuqala kuka-i.
Izincazelo zezibalo zokuhlukahluka kwezibikezelo nezimpendulo zethulwe kuThebula 1, okubonisa incazelo, ukuchezuka okujwayelekile (SD), i-coefficient of variation (CV), ubuncane, ubukhulu, i-kurtosis, kanye nokutsheka.Amanani aphansi nomkhawulo wezinto ahlelwe ngendlela enciphayo ye-Mg < Ca < K < Ni kanye ne- Ca < Mg < K < Ni, ngokulandelanayo. Impendulo esuka ku-6 endaweni yocwaningo isuka ku-4 kuya ku-4. .39 mg/kg.Ukuqhathaniswa kwe-Ni nesilinganiso somhlaba (29 mg/kg) kanye nesilinganiso saseYurophu (37 mg/kg) kubonise ukuthi ingqikithi yejiyomethri ebaliwe yendawo yocwaningo yayingaphakathi kwebanga elibekezeleleka.Noma kunjalo, njengoba kuboniswa i-Kabata-Pendias11, ukuqhathaniswa kokugxilisa kwe-nickel (Ni) ocwaningweni lwamanje kukhombisa ukuthi ukugxilisana kwe-concentration ye-nickel yamanje ye-nickel yenhlabathi yezolimo e-Sweden yisilinganiso esiphezulu senhlabathi yezolimo. k enhlabathini yasemadolobheni kanye naseduze kwedolobha ocwaningweni lwamanje (Ni 16.15 mg/kg) ibiphezulu kunomkhawulo ovumelekile wama-60 (10.2 mg/kg) we-Ni enhlabathini yasemadolobheni yase-Polish ebikwe ngu-Różański et al.Ngaphezu kwalokho, i-Bretzel ne-Calderisi61 baqophe inani eliphansi kakhulu le-Nimgkg/i-6m yocwaningo ku-Turban yamanje kutholwe ukugxiliswa kwe-Nimgkg/2m ku-Urban yamanje (1.7). ukugxiliswa kwe-nickel ephansi (12.34 mg/kg) enhlabathini yasemadolobheni e-Hong Kong, ephansi kunokugxilisa i-nickel yamanje kulolu cwaningo.U-Birke et al63 ubike isilinganiso esimaphakathi se-Nickel esingu-17.6 mg/kg endaweni yezimayini endala kanye nendawo yezimboni zasemadolobheni e-Saxony-Anhalt, eJalimane, eyayiyi-1.45 mg/kg ephakeme kunokuqukethwe kwe-ni.C kwe-5. inhlabathi kwezinye izindawo zasemadolobheni nasezindaweni ezingaphansi kwedolobha zendawo yocwaningo ingase ihlotshaniswe kakhulu nemboni yensimbi nensimbi kanye nemboni yensimbi.Lokhu kuhambisana nocwaningo lukaKhodadust et al.64 ukuthi imboni yensimbi nokusebenza kwensimbi kuyimithombo eyinhloko yokungcoliswa kwe-nickel enhlabathini.Nokho, izibikezelo nazo zazisuka ku-538.70 mg/kg kuya ku-69,161.80 mg/kg ku-Ca, 497.51 mg/kg kuya ku-3535.68 mg/kg ku-K, kanye no-6kg/000mg. t al.Abangu-65 baphenye ingqikithi ye-Mg ne-K yenhlabathi emaphakathi ne-Serbia.Bathole ukuthi ukugxila okuphelele (410 mg/kg kanye no-400 mg/kg, ngokulandelana) kwakungaphansi kokugxilisa kwe-Mg no-K kocwaningo lwamanje.Akubonakali, empumalanga ye-Poland, i-Orzechowski ne-Smolczynski66 ibonise, ingqikithi ye-M1 ye-Mg ne-Cass ihlolwe ingqikithi ye-M1 kanye ne-Cass ye-Cas 590 mg/kg) kanye no-K (810 mg/kg) Okuqukethwe emhlabathini ongaphezulu kuphansi kunento eyodwa kulolu cwaningo.Ucwaningo lwakamuva olwenziwa ngu-Pongrac et al.I-67 ibonise ukuthi ingqikithi ye-Ca ehlaziywe enhlabathini engu-3 ehlukene e-Scotland, UK (inhlabathi yaseMylnefield, inhlabathi yaseBalruddery kanye nomhlabathi waseHartwood) ibonise okuqukethwe kwe-Ca okuphezulu kulolu cwaningo.
Ngenxa yokugxilisa okulinganiselwe okulinganiselwe kwama-elementi ayisampula, ukusatshalaliswa kwesethi yedatha yama-elementi kubonisa ukutsheka okuhlukile.Ukutsheka nokurtosis kwezinto kusuka ku-1.53 ​​kuye ku-7.24 kanye no-2.49 kuya ku-54.16, ngokulandelana.Zonke izici zibaliwe zinokugwegwa kanye namazinga e-kurtosis ngenhla +1, ngakho-ke ukusabalalisa kwedatha kubonisa ukuthi i-CV iqondise kahle. wezakhi futhi zibonisa ukuthi i-K, Mg, kanye ne-Ni ibonisa ukuhlukahluka okusesilinganisweni, kuyilapho i-Ca inokuhlukahluka okuphezulu kakhulu.Ama-CV ka-K, Ni kanye noMg achaza ukusatshalaliswa kwawo okufanayo.Ngaphezu kwalokho, ukusatshalaliswa kwe-Ca akufani futhi imithombo yangaphandle ingase ithinte izinga layo lokunothisa.
Ukuhlotshaniswa kwezinto eziguquguqukayo zokubikezela nezinto zokuphendula kubonise ukuhlobana okwanelisayo phakathi kwezakhi (bheka Umfanekiso 3) .Ukuhlobana kubonise ukuthi i-CaK ibonise ukuhlobana okusesilinganisweni nenani le-r = 0.53, njengoba kwenza i-CaNi. Nakuba u-Ca no-K bebonisa ukuhlangana okuthobekile komunye nomunye, abacwaningi abafana no-Kingston et al.68 kanye ne-Santo69 basikisela ukuthi amazinga abo enhlabathini ahambisana ngokuphambene.Nokho, i-Ca ne-Mg iphikisana ne-K, kodwa i-CaK ihlobana kahle.Lokhu kungase kube ngenxa yokusetshenziswa komanyolo onjenge-potassium carbonate, ephakeme ngo-56% ku-potassium.I-Potassium yayihlotshaniswa ngokusesilinganisweni ne-magnesium (KM 63 eduze, i-potassium sulfate iyizici ezimbili ezihlobene ne-Infa). I-te, i-potassium magnesium nitrate, ne-potash isetshenziswa enhlabathini ukuze kwandiswe amazinga okuntula kwayo. I-Nickel ihlotshaniswa ngokusesilinganisweni ne-Ca, K kanye ne-Mg ngamavelu angu-r = 0.52, 0.63 kanye no-0.55, ngokulandelana. kokubili i-magnesium ne-calcium kunciphisa imiphumela enobuthi ye-nickel emhlabathini.
I-matrix yokuhlobana yezakhi ezibonisa ubudlelwano phakathi kwezibikezelo nezimpendulo (Qaphela: lesi sibalo sihlanganisa i-scatterplot phakathi kwama-elementi, amaleveli okubaluleka asekelwe ku-p <0,001).
Umfanekiso wesi-4 ubonisa ukusatshalaliswa kwendawo kwezinto.Ngokuvumelana no-Burgos et al70, ukusetshenziswa kokusatshalaliswa kwendawo kuyindlela esetshenziselwa ukulinganisa nokugqamisa izindawo ezishisayo ezindaweni ezingcolile.Amazinga okunothisa we-Ca ku-Fig. 4 angabonakala engxenyeni esenyakatho-ntshonalanga yemephu yokusabalalisa kwendawo.Isibalo sibonisa ukunothisa kwe-calcium okusesilinganisweni ngenxa yemithombo ye-calcium emaphakathi ngenxa yokucetshiswa kwe-Cap. ukusetshenziswa kwe-quicklime (i-calcium oxide) ukunciphisa ubumuncu bomhlaba kanye nokusetshenziswa kwayo ezigayweni zensimbi njengomoya-mpilo we-alkaline enqubweni yokwenza insimbi.Ngakolunye uhlangothi, abanye abalimi bakhetha ukusebenzisa i-calcium hydroxide enhlabathini ene-asidi ukwenza i-pH ingathathi hlangothi, ephinde ikhulise okuqukethwe kwe-calcium enhlabathini71.I-Potassium iphinde ibonise izindawo ezishisayo enyakatho-ntshonalanga nasempumalanga yemephu. s.Lokhu kuyahambisana nezinye izifundo, ezifana noMadaras noLipavský72, Madaras et al.73, Pulkrabová et al.74, Asare et al.75, owabona ukuthi ukuqiniswa kwenhlabathi nokwelashwa nge-KCl ne-NPK kubangele okuqukethwe okuphezulu kwe-K emhlabathini.Ukucebisa kwe-Spatial Potassium enyakatho-ntshonalanga yemephu yokusabalalisa kungase kube ngenxa yokusetshenziswa komanyolo owenziwe nge-potassium njenge-potassium chloride, i-potassium sulfate, i-potassium nitrate, i-potash, ne-potash ukwandisa okuqukethwe kwe-potassium emhlabathini ompofu.Zádorová et al.76 kanye noTlustoš et al.77 iveze ukuthi ukusetshenziswa komanyolo osuselwe ku-K kwandisa okuqukethwe kwe-K emhlabathini futhi kuzokwandisa kakhulu izakhi zomhlabathi ngokuhamba kwesikhathi, ikakhulukazi i-K ne-Mg ebonisa indawo eshisayo emhlabathini.Izindawo ezishisayo ezimaphakathi enyakatho-ntshonalanga yemephu naseningizimu-mpumalanga yemephu.Ukulungiswa kwe-Colloidal enhlabathini kunciphisa ukugcwala kwe-magnesium emhlabathini ophuzi obangela ukuntuleka kwe-magnesium emhlabathini ophuzi. i-potassium magnesium sulfate, i-magnesium sulfate, ne-Kieserite, iphatha ukuntula (izitshalo zibonakala zinsomi, zibomvu, noma zinsundu, okubonisa ukuntula kwe-magnesium) enhlabathini enezinga le-pH elivamile6. Ukunqwabelana kwe-nickel ezindaweni ezisemadolobheni nasezindaweni eziseduze nedolobha kungase kubangelwe imisebenzi ye-anthropogenic yokukhiqiza insimbi ye-nickel7.
Ukusatshalaliswa kwendawo kwezinto [imephu yokusabalalisa yendawo idalwe kusetshenziswa i-ArcGIS Desktop (ESRI, Inc, Version 10.7, URL: https://desktop.arcgis.com).]
Imiphumela yenkomba yokusebenza yemodeli yezakhi ezisetshenziswe kulolu cwaningo iboniswa kuThebula 2. Ngakolunye uhlangothi, i-RMSE kanye ne-MAE ye-Ni zombili zisondele ku-zero (0.86 RMSE, -0.08 MAE). Ngakolunye uhlangothi, kokubili amanani e-RMSE kanye ne-MAE ka-K ayamukeleka.Imiphumela ye-RMSE ne-MAE ibe mikhulu nge-calcium ne-magnesium esethiwe futhi imiphumela ye-MAERM emikhulu kanye ne-SERM yedatha enkulu. lolu cwaningo olusebenzisa i-EBK ukubikezela u-Ni lutholakale lungcono kunemiphumela kaJohn et al.54 sisebenzisa i-synergistic kriging ukubikezela ukugxila kuka-S enhlabathini kusetshenziswa idatha efanayo eqoqiwe. Imiphumela ye-EBK esiyifundile ihlobana naleyo ka-Fabijaczyk et al.41, Yan et al.79, Beguin et al.80, Adhikary et al.81 kanye noJohane et al.82, ikakhulukazi u-K no-Ni.
Ukusebenza kwezindlela ezingazodwana zokubikezela okuqukethwe kwe-nickel enhlabathini yasemadolobheni naseduze kwedolobha kuye kwahlaziywa kusetshenziswa ukusebenza kwamamodeli (Ithebula 3).Ukuqinisekiswa kwemodeli nokuhlola ukunemba kuqinisekisile ukuthi isibikezelo se-Ca_Mg_K esihlanganiswe nemodeli ye-EBK SVMR iveze ukusebenza okungcono kakhulu.Iphutha lemodeli ye-Calibration Ca_Mg_Mso i-square model Ca_Mg_K-SER (iphutha le-SVM-EBK) imodeli ye-square (SVM_KR) imodeli yesikwele (iphutha le-SVM_EBK-EBK) I-MAE) ibingu-0.637 (R2), 95.479 mg/kg (RMSE) kanye no-77.368 mg/kg (MAE) Ca_Mg_K-SVMR ibingu-0.663 (R2), 235.974 mg/kg (RMSE) kanye no-166.946 mg/kg (lesseds, itholwe i-Ca_M2 value_None R2). 63 mg/kg R2) kanye ne-Ca_Mg-EBK_SVMR (0.643 = R2);imiphumela yabo ye-RMSE ne-MAE ibiphezulu kunaleyo ye-Ca_Mg_K-EBK_SVMR (R2 0.637) (bona Ithebula 3). Ngaphezu kwalokho, i-RMSE kanye ne-MAE ye-Ca_Mg-EBK_SVMR (RMSE = 1664.64 kanye ne-MAE = 1031.49) imodeli kanye ne-13g_5 enkulu kunaleyo eyi-13g_4. K-EBK_SVMR. Ngokunjalo, imodeli ye-RMSE ne-MAE ye-Ca_Mg-K SVMR (RMSE = 235.974 kanye ne-MAE = 166.946) imodeli ingu-2.5 no-2.2 mikhulu kunaleyo ye-Ca_Mg_K-EBK_SVMR RMSE ne-MAE ibonisa ukuthi idatha ilinganiswe kahle kangakanani ne-RMSE ngokulandelana. I-RSME kanye ne-MAE yabonwa.Ngokuka-Kebonye et al.46 kanye noJohane et al.54, uma i-RMSE ne-MAE isondela kuqanda, kuba ngcono imiphumela.I-SVMR kanye ne-EBK_SVMR inamanani aphezulu e-RSME kanye ne-MAE. Kwaqashelwa ukuthi izilinganiso ze-RSME zaziphezulu ngokuqhubekayo kunamanani e-MAE, okubonisa ukuba khona kwabangaphandle.Ngokusho kwe-Leates, i-McCabes isho ukuthi i-McCabes8 ingaphezu kwephutha le-McCabe3 Okutuswayo njengenkomba yokuba khona kwama-outliers.Lokhu kusho ukuthi uma isethi yedatha ihluka kakhulu, amanani e-MAE nawe-RMSE aphezulu.Ukunemba kokuhlolwa kokuqinisekisa okuphambene kwemodeli exubile ye-Ca_Mg_K-EBK_SVMR yokubikezela okuqukethwe kwe-Ni emhlabathini wasemadolobheni namadolobhana kwakungu-63.70%.Ngokuvumelana ne-Li.59, leli zinga lokunemba liyimodeli yezinga lokusebenza elamukelekayo.Imiphumela yamanje iqhathaniswa nocwaningo lwangaphambili lukaTarasov et al.36 imodeli yayo eyingxube edale i-MLPRK (Multilayer Perceptron Residual Kriging), ehlobene nenkomba yokuhlola ukunemba kwe-EBK_SVMR ebikwe ocwaningweni lwamanje, i-RMSE (210) kanye ne-MAE (167.5) yayiphezulu kunemiphumela yethu ocwaningweni lwamanje (RMSE 95.479, MAE6 68 yocwaningo lwamanje) 77. lokho kukaTarasov et al.36 (0.544), kusobala ukuthi i-coefficient of determination (R2) iphezulu kule modeli exubile.Imajini yephutha (RMSE kanye ne-MAE) (EBK SVMR) yemodeli exubile iphansi ngokuphindwe kabili.Ngokufanayo, u-Sergeev et al.34 barekhode u-0.28 (R2) kumodeli ye-hybrid ethuthukisiwe kumodeli ye-hybrid ethuthukisiwe (i-Multilarecord) isifundo se-Record6 yamanje (R2).Izinga lokunemba lokubikezela lale modeli (EBK SVMR) lingu-63.7%, kuyilapho ukunemba kokubikezela okutholwe u-Sergeev et al.I-34 ingu-28%.Imephu yokugcina (Fig. 5) idalwe kusetshenziswa imodeli ye-EBK_SVMR kanye ne-Ca_Mg_K njengesibikezelo sezindawo ezishisayo kanye ne-nickel emaphakathi ukuya kweye-nickel kuyo yonke indawo yocwaningo.Lokhu kusho ukuthi ukugxiliswa kwe-nickel endaweni yocwaningo ngokuyinhloko kumaphakathi, nokugxila okuphezulu kwezinye izindawo ezithile.
Imephu yokuqagela yokugcina imelelwa kusetshenziswa imodeli ye-hybrid engu-EBK_SVMR futhi kusetshenziswa i-Ca_Mg_K njengesibikezelo.[Imephu yokusabalalisa yendawo idalwe kusetshenziswa i-RStudio (inguqulo 1.4.1717: https://www.rstudio.com/).]
Kwethulwa kuMfanekiso 6 ukugxiliswa kwe-PTE njengendiza yokuqamba ehlanganisa ama-neurons angawodwana.Ayikho indiza yengxenye ebonise iphethini yombala efanayo njengoba kukhonjisiwe.Nokho, inani elifanele lama-neurons ngemephu edwetshiwe ngu-55.I-SeOM ikhiqizwa kusetshenziswa imibala eyahlukene, futhi uma amaphethini wombala afana kakhulu, kulapho kuqhathaniswa khona nakakhulu izici zamasampula, amaphethini wombala ofanayo (ngokwe-MCa, amaphethini wombala owodwa) ngokuhambisana nesici sombala esisodwa se-KCa (i-K) ama-neuron aphezulu nama-neuron amaningi aphansi.Ngakho, i-CaK ne-CaMg babelana ngokufana okuthile nama-neurons ane-oda eliphezulu kakhulu kanye namaphethini ombala ophansi ukuya komaphakathi.Womabili amamodeli abikezela ukugxiliswa kwe-Ni enhlabathini ngokubonisa imibala ephakathi nendawo ukuya phezulu yemibala efana nobomvu, osawolintshi kanye nophuzi.Imodeli ye-KMg ibonisa amaphethini amaningi ombala ophakeme asekelwe ekulinganisweni okunembile kanye nephethini yokusabalalisa ephansi kuya kwemaphakathi ibonise ukunemba kombala kwengxenye yombala, imodeli yokusabalalisa okuphakeme. iphethini yombala ophezulu ekhombisa ukugcwala okungaba khona kwe-nickel enhlabathini (bheka Umfanekiso 4).Indiza yengxenye ye-CakMg ibonisa iphethini yombala ohlukahlukene ukusuka kophansi kuye phezulu ngokwesilinganiso sombala esinembile.Ngaphezu kwalokho, ukubikezela kwemodeli kokuqukethwe kwe-nickel (i-CakMg) kufana nokusatshalaliswa kwendawo kwe-nickel eboniswe kumfanekiso we-Figure 5. umhlabathi.Umfanekiso wesi-7 ubonisa indlela yekhonta eqoqweni le-k-izindlela kumephu, ihlukaniswe yaba amaqoqo amathathu ngokusekelwe enanini elibikezelwe kumodeli ngayinye.Indlela yekhonta imelela inani eliphelele lamaqoqo.Kumasampula omhlabathi angu-115 aqoqiwe, isigaba 1 sathola amasampula omhlabathi amaningi, 74.Iqoqo lesi-2 lathola amasampula angu-38 we-simponent, inhlanganisela ye-simponent eyisikhombisa yathola amasampula angu-33. kuhlanganiswe ukuze kuchazwe kahle iqoqo.Ngenxa yezinqubo eziningi ze-anthropogenic nezemvelo ezithinta ukwakheka kwenhlabathi, kunzima ukuba namaphethini eqoqo ahlukaniswe kahle kumephu ye-SeOM esabalalisiwe78.
Okukhipha ingxenye yendiza ngokuhlukahluka ngakunye kwe-Empirical Bayesian Kriging Support Vector Machine (EBK_SVM_SeOM).[Amamephu we-SeOM adalwe kusetshenziswa i-RStudio (inguqulo 1.4.1717: https://www.rstudio.com/).]
Izingxenye ezihlukene zokuhlukanisa iqoqo [amamephu we-SeOM adalwe kusetshenziswa i-RStudio (inguqulo 1.4.1717: https://www.rstudio.com/).]
Ucwaningo lwamanje lubonisa ngokusobala amasu okumodela okugxilwa kwe-nickel enhlabathini yasemadolobheni kanye naseduze kwedolobha. Ucwaningo luhlole amasu okumodela ahlukene, ahlanganisa izakhi nezindlela zokumodela, ukuze kutholwe indlela engcono kakhulu yokubikezela ukugxila kwe-nickel enhlabathini.Izici ze-SeOM zokwakheka kwepulani yendawo yendlela yokumodela ibonise iphethini yombala ophezulu ukusuka kokuphansi ukuya phezulu esikalini sombala we-spatidic esinembile sokusabalalisa umhlabathi, iqinisekisa kanjani isikali sokusabalalisa sombala we-spatidic. Ukusatshalaliswa okulinganayo kwezingxenye ezikhonjiswe yi-EBK_SVMR (bona Umfanekiso 5). Imiphumela ibonisa ukuthi imodeli yokuhlehla komshini we-vector (Ca Mg K-SVMR) ibikezela ukugcwala kwe-Ni enhlabathini njengemodeli eyodwa, kodwa amapharamitha wokuqinisekisa nokunemba abonisa amaphutha aphezulu kakhulu ngokwe-RMSE kanye ne-MAE. Ngakolunye uhlangothi imodeli ye-EB ephansi futhi imodeli ye-EBR iphansi futhi imodeli ye-KML iphansi. ye-coefficient of determination (R2).Imiphumela emihle itholwe kusetshenziswa i-EBK SVMR nezinto ezihlanganisiwe (CaKMg) ezinamaphutha aphansi e-RMSE kanye ne-MAE ngokunemba okungu-63.7%.Kuvele ukuthi ukuhlanganisa i-algorithm ye-EBK ne-algorithm yokufunda komshini kungakhiqiza i-algorithm ye-hybrid engabikezela ukugcwala kwe-PTEs endaweni njengoba i-Mg ibikezela endaweni yocwaningo ibikezela ukugxilisa ingqondo kwe-Mg ukuze ibikezele inhlabathi. ion of Ni enhlabathini.Lokhu kusho ukuthi ukusetshenziswa okuqhubekayo komanyolo osuselwa ku-nickel kanye nokungcoliswa kwezimboni komhlaba yimboni yensimbi kunomkhuba wokwandisa ukugcwala kwe-nickel enhlabathini.Lolu cwaningo lwembula ukuthi imodeli ye-EBK inganciphisa izinga lephutha futhi ithuthukise ukunemba kwemodeli yokusabalalisa indawo yenhlabathi emadolobheni noma ezindaweni eziseduze kwedolobha, i-weBKT ibikezela umhlabathi ojwayelekile, i-weBKT nokubikezela kwenhlabathi ye-weBKT. ;ngaphezu kwalokho, siphakamisa ukusebenzisa i-EBK ukuze sihlanganise ngama-algorithms okufunda omshini ahlukahlukene.Ukugxila kwe-Ni kwabikezelwa kusetshenziswa izakhi njengama-covariate;kodwa-ke, ukusebenzisa ama-covariate amaningi kungathuthukisa kakhulu ukusebenza kwemodeli, okungabhekwa njengomkhawulo womsebenzi wamanje.Omunye umkhawulo walolu cwaningo ukuthi inani lamasethi edatha liyi-115.Ngakho-ke, uma idatha eyengeziwe inikezwa, ukusebenza kwendlela ehlongozwayo yokwenza inhlanganisela elungiselelwe kahle ingathuthukiswa.
I-PlantProbs.net.Nickel Ezitshalweni Nasemhlabathini https://plantprobs.net/plant/nutrientImbalances/sodium.html (Ifinyelelwe zingama-28 kuMbasa 2021).
I-Kasprzak, i-KS Nickel ithuthuka ku-toxicology yemvelo yesimanje.surroundings.toxicology.11, 145–183 (1987).
I-Cempel, M. & Nikel, G. Nickel: Ukubuyekezwa kwemithombo yayo kanye ne-toxicology yemvelo.Polish J. Environment.Stud.15, 375-382 (2006).
Freedman, B. & Hutchinson, TC Okokufaka okungcolile okuvela emkhathini nokunqwabelana enhlabathini nasezitshalweni eduze kwesincibilikisi se-nickel-copper e-Sudbury, Ontario, Canada.can.J.I-Bot.58(1), 108-132.https://doi.org/10.1139/b80-014 (1980).
Manyiwa, T. et al.Izinsimbi ezinzima enhlabathini, izitshalo kanye nezingozi ezihlobene nezilwane ezidla amadlelo eduze nemayini ye-nickel ye-Selebi-Phikwe e-Botswana.surroundings.Geochemistry.Health https://doi.org/10.1007/s10653-021-00918-x (2021).
Cabata-Pendias.Kabata-Pendias A. 2011. Trace elements in earth and… - Google Scholar https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=Kabata-Pendias+A.+2011.+Trace+ Elements+in+soils+and+plants+New+Yos+C2%RNY tnG= (Ifinyelelwe zingama-24 Nov 2020).
Almås, A., Singh, B., Agriculture, TS-NJ of & 1995, undefined.Imiphumela yemboni ye-nickel yaseRussia ekugxilweni kwensimbi enzima enhlabathini yezolimo notshani e-Soer-Varanger, Norway.agris.fao.org.
U-Nielsen, GD et al.Ukumuncwa kwe-Nickel nokugcinwa emanzini okuphuza kuhlobene nokudla kanye nokuzwela kwe-nickel.toxicology.application.Pharmacodynamics.154, 67-75 (1999).
I-Costa, M. & Klein, CB Nickel carcinogenesis, mutation, epigenetics noma selection.surroundings.Health Perspective.107, 2 (1999).
I-Ajman, i-PC;I-Ajado, SK;Borůvka, L.;Bini, JKM;Sarkody, VYO;Cobonye, ​​NM;Ukuhlaziywa kwethrendi kwezinto ezingaba ubuthi: ukubuyekezwa kwe-bibliometric.I-Environmental Geochemistry and Health.Springer Science & Business Media BV 2020.https://doi.org/10.1007/s10653-020-00742-9.
Minasny, B. & McBratney, AB Digital Soil Mapping: Umlando Omfishane kanye Nezinye Izifundo.Geoderma 264, 301–311.https://doi.org/10.1016/j.geoderma.2015.07.017 (2016).
McBratney, AB, Mendonça Santos, ML & Minasny, B. Kumephu yenhlabathi yedijithali.Geoderma 117(1-2), 3-52.https://doi.org/10.1016/S0016-7061(03)00223-4 (2003).
Deutsch.CV Geostatistical Reservoir Modeling,… – Google Scholar https://scholar.google.com/scholar?hl=zu&as_sdt=0%2C5&q=CV+Deutsch%2C+2002%2C+Geostatistical+Reservoir+Modeling%2C +Oxford+University+Press7Aprilc2C2C+Page. 21).