Mahalo iā ʻoe no ka kipa ʻana iā Nature.com. ʻO ka polokalamu kele pūnaewele āu e hoʻohana nei he kākoʻo liʻiliʻi no CSS. No ka ʻike maikaʻi loa, paipai mākou e hoʻohana ʻoe i kahi polokalamu kele hou (a i ʻole e hoʻopau i ke ʻano hoʻohālikelike i Internet Explorer).
He pilikia nui ka pollution lepo ma muli o nā hana a ke kanaka. He ʻokoʻa ka mahele ʻana o nā mea ʻona (PTEs) i ka hapa nui o nā kūlanakauhale a me nā wahi āpau. ly coupled plasma emission spectrometry. ʻO ka mea hoʻololi pane ʻo Ni a ʻo nā mea wānana Ca, Mg, a me K. ʻO ka matrix correlation ma waena o ka pane pane a me ka variable predictor e hōʻike ana i kahi pilina maikaʻi ma waena o nā mea. ʻoi aku ka kiʻekiʻe ma mua o nā ʻano hana ʻē aʻe i hoʻohana ʻia. ʻO nā hiʻohiʻona hui ʻia no Empirical Bayesian Kriging-Multiple Linear Regression (EBK-MLR) hana maikaʻi ʻole, e like me ka hōʻike ʻana e nā coefficients o ka hoʻoholo ʻana ma lalo o 0.1. ʻO ka Empirical Bayesian Kriging-Support Vector Machine Regression (EBK-SVMR) ke kumu hoʻohālike maikaʻi loa, me ka haʻahaʻa haʻahaʻa o ka RMSE95 mg / 4kg/kg/kg. nā waiwai a me ka helu kiʻekiʻe o ka hoʻoholo (R2 = 0.637). ʻIke ʻia ka hoʻopuka ʻana o ka ʻenehana hoʻohālike EBK-SVMR me ka hoʻohana ʻana i kahi palapala ʻāina hoʻonohonoho ponoʻī. .
Manaʻo ʻia ʻo Nickel (Ni) he micronutrient no nā mea kanu no ka mea kōkua ia i ka hoʻoponopono ʻana i ka hau (N) a me ka metabolism urea, pono ia mau mea ʻelua no ka germination o nā hua. izers to optimize nitrogen fixation2.Continued application of nickel-based fertilizers to enrich the soil and increase the ability of legumes to fix nitrogen in the soil hoonui mau i ka nickel concentration in the soil.Although nickel is micronutrient for plants, its excessive intake in the soil can do more harm than good.The toxicity of the nickel up essential growth in soil1. e like me Liu3, ua ʻike ʻia ʻo Ni ʻo ia ka 17th mea nui e pono ai no ka hoʻomohala ʻana a me ka ulu ʻana o nā mea kanu. Ua hoʻohana nui ʻia i nā lako kīhini, nā lumi ballroom, nā lako ʻoihana meaʻai, uila, uea a me ke kaula, jet turbines, surgical implants, textiles, and shipbuilding5.Ni-rich level in soils (ie, surface soil) have attributed to both anthropogenic and natural resources, but priorly, Ni is a natural source than anthropogenic4,6.nae, anthropogenic kumu i loko o ka nickel / cadmium pākuʻi i loko o ka oihana kila, electroplating, arc kuʻihao, diesel a me ka wahie aila, a me ka lewa emissions mai ka lanahu kuni a me ka lepo a me ka sludge incineration Nickel accumulation7,8. Wahi a Freedman a me Hutchinson9 a me Manyiwa et al.10, ʻo nā kumu nui o ka pollution topsoil ma ka ʻaoʻao kokoke a pili i ka nui o ka nickel-copper-based smelters a me nā mines. ʻO ka ʻāina kiʻekiʻe a puni ka Sudbury nickel-copper refinery ma Kanada ka kiʻekiʻe kiʻekiʻe o ka nickel contamination ma 26,000 mg / kg11. e pili ana iā Alms et al.12, ka nui o HNO3-extractable nickel i loko o ka 'āina kiʻekiʻe arable 'āina (nickel hana ma Rusia) mai 6.25 a hiki i 136.88 mg / kg, e pili ana i ka mean o 30.43 mg / kg a me ka baseline kuʻina o 25 mg / kg. E like me ka kabata 11, ka palapala noi o ka fertilizersurban soils-phosphorus lepo i ka wā o ka fertilizersurban fertilizer lepo. Hiki i nā kau hua ke hoʻokomo a hoʻohaumia paha i ka lepo. Hiki i nā hopena o ka nickel i loko o ke kanaka ke alakaʻi i ka maʻi kanesa ma muli o ka mutagenesis, ka pōʻino chromosomal, Z-DNA generation, blocked DNA excision repair, a i ʻole epigenetic process13. I nā hoʻokolohua holoholona, ua ʻike ʻia ka nickel e hiki ke hoʻoulu i nā ʻano maʻi ʻokoʻa, a me nā paʻakikī nickel carcinogenic hiki ke hoʻokuʻu ʻia nā tumora.
Ua ulu aʻe nā loiloi hoʻohaumia ʻāina i kēia mau manawa ma muli o ka nui o nā pilikia e pili ana i ke olakino e kū mai ana mai ka pilina lepo-mea kanu, ka lepo a me ka pilina o ka lepo, ka hoʻohaʻahaʻa ʻana i ka ecological, a me ka loiloi hopena o ke kaiapuni. (PSM). Wahi a Minasny a me McBratney16, ua hōʻike ʻia ka palapala ʻāina wānana (DSM) he subdiscipline koʻikoʻi o ka ʻepekema lepo.Lagacherie a me McBratney, 2006 wehewehe ʻo DSM ma ke ʻano "ka hana ʻana a me ka hoʻopiha ʻana i nā ʻōnaehana ʻike lepo spatial ma o ka hoʻohana ʻana i ka in situ a me ka laboratory observational methods and spatial spatial system.17 ka wehewehe ʻana ʻo ka DSM a i ʻole PSM o kēia wā ʻo ia ka ʻenehana maikaʻi loa no ka wānana a i ʻole ka palapala ʻāina i ka māhele ākea o nā PTE, nā ʻano lepo a me nā waiwai lepo. ʻO Geostatistics a me Machine Learning Algorithms (MLA) nā ʻano hana hoʻohālike DSM e hana ana i nā palapala ʻāina i helu ʻia me ke kōkua o nā kamepiula e hoʻohana ana i ka ʻikepili koʻikoʻi a liʻiliʻi.
Ua wehewehe ʻo Deutsch18 a me Olea19 i ka geostatistics "ʻo ka hōʻiliʻili o nā ʻenehana helu e pili ana i ka hōʻike ʻana i nā ʻano kikoʻī, e hoʻohana nui ana i nā hiʻohiʻona stochastic, e like me ke ʻano o ka hōʻike ʻana i ka ʻikepili manawa."ʻO ka mea nui, pili ka geostatistics i ka loiloi o nā variograms, e ʻae ai i ka helu ʻana a wehewehe i nā hilinaʻi o nā waiwai spatial mai kēlā me kēia dataset20.Gumiaux et al.20 hōʻike hou i ka loiloi o variograms ma geostatistics ma luna o ekolu kumu, me (a) helu ana i ka unahi o ka ikepili correlation, (b) ike a me ka helu anisotropy i ka dataset disparity a me (c) ma waho aʻe o ka lawe 'ana i ka inherent hewa o ke ana ikepili i hookaawaleia mai ka local effects, the areaBuilding are also used on these estimated. me ka kriging maʻamau, co-kriging, kriging maʻamau, empirical Bayesian kriging, simple kriging method a me nā ʻenehana interpolation kaulana ʻē aʻe e palapala ʻāina a wānana paha i ka PTE, nā ʻano lepo, a me nā ʻano lepo.
Machine Learning Algorithms (MLA) he ʻano hana hou e hoʻohana ana i nā papa ʻikepili nui ʻole linear, hoʻohana ʻia e nā algorithms i hoʻohana mua ʻia no ka ʻimi ʻana i ka ʻikepili, ka ʻike ʻana i nā ʻano o ka ʻikepili, a hoʻohana pinepine ʻia i ka hoʻokaʻawale ʻana i nā kula ʻepekema e like me ka ʻepekema lepo a me nā hana hoʻihoʻi.22 (nā ululāʻau maʻamau no ka manaʻo metala kaumaha ma nā ʻāina mahiʻai), Sakizadeh et al.23 (hoʻohālike me ka hoʻohana ʻana i nā mīkini vector kākoʻo a me nā pūnaewele neural artificial) haumia lepo ). Eia kekahi, ʻo Vega et al.24 (CART no ka hoʻohālike ʻana i ka paʻa metala kaumaha a me ka adsorption i ka lepo) Sun et al.25 (ʻo ka hoʻohana ʻana i ka cubist ka māhele ʻana o Cd i loko o ka lepo) a me nā algorithm ʻē aʻe e like me k-kokoke loa, hoʻonui i ka hoʻihoʻi hou ʻana, a me ka hoʻoulu hou ʻana Ua hoʻohana pū ʻia nā lāʻau iā MLA e wānana i ka PTE i ka lepo.
ʻO ka hoʻohanaʻana i nā algorithms DSM i ka wānana a iʻole ka palapala'āina e kū ana i nā pilikia. Manaʻo nā mea kākau heʻoi aku ka maikaʻi o ka MLA ma mua o ka geostatistics a me ka hope.Manaʻo ʻo Pontius lāua ʻo Cheuk28 a me Grunwald29 e pili ana i nā hemahema a me kekahi mau hewa i ka wānana ʻana i ka palapala ʻāina.15 e wehewehe pono e hoʻopaʻa kūʻokoʻa ʻia ka ʻano hōʻoia a me ka maopopo ʻole i hoʻokomo ʻia e ka hana ʻana i ka palapala ʻāina a me ka wānana i mea e hoʻomaikaʻi ai i ka maikaʻi o ka palapala ʻāina.eia nae, ka hemahema o ka maopopo i ka DSM hiki ke ala mai na kumu he nui o ka hewa, 'o ia ka covariate hewa, kŘkohu kŘkohu, wahi hewa, a me analytical Hapa 31. Modelling inaccuracies induced i loko o MLA a me geostatistical kaʻina e pili ana me ka nele o ka hoomaopopo ana, ultimately alakai ana i ka oversimplification o ka maoli process32. No ka mea, o ke ano o ke kumu hoohalike, inaccuracies hiki ke wānana i ke kumu hoʻohālike. lation33. I kēia mau lā, ua puka mai kahi ʻano DSM hou e paipai ana i ka hoʻohui ʻana o nā geostatistics a me MLA ma ka palapala ʻāina a me ka wānana.34;Subbotina et al.35;Tarasov et al.36 a me Tarasov et al.Ua hoʻohana ʻo 37 i ka maikaʻi kūpono o ka geostatistics a me ke aʻo ʻana i ka mīkini e hana i nā hiʻohiʻona hybrid e hoʻomaikaʻi i ka pono o ka wānana a me ka palapala ʻāina.maikaʻi. ʻO kekahi o kēia mau hiʻohiʻona algorithm hybrid a i hui pū ʻia ʻo Artificial Neural Network Kriging (ANN-RK), Multilayer Perceptron Residual Kriging (MLP-RK), Generalized Regression Neural Network Residual Kriging (GR- NNRK)36, Artificial Neural Network Kriging-Multilayer Perceptron (ANN-K-MLP)37 a me ka hoʻoponopono hou ʻana o ka Co-MLP.
Wahi a Sergeev et al., ʻo ka hui pū ʻana i nā ʻano hana hoʻohālike like ʻole i hiki ke hoʻopau i nā hemahema a hoʻonui i ka pono o ka hopena hybrid ma mua o ka hoʻomohala ʻana i kāna kumu hoʻohālike hoʻokahi. (SVM) a me Multiple Linear Regression (MLR) models.Hybridization o EBK me kekahi MLA ʻaʻole i ʻike ʻia. ʻO nā hiʻohiʻona hui like ʻole i ʻike ʻia he hui pū ʻana o nā mea maʻamau, koena, regression kriging, a me MLA.EBK he ʻano interpolation geostatistical e hoʻohana ana i kahi kaʻina spatially stochastic, i wehewehe ʻia ma ke ʻano he ʻano ʻokoʻa ʻokoʻa ma luna o ke kahua kūʻokoʻa. Ua ho'ohana 'ia 'o EBK ma nā ha'awina like 'ole, e like me ka nānā 'ana i ka hā'awi 'ana i ke kalapona organik ma nā 'āina mahi'ai40, ka nānā 'ana i ka pollution lepo41 a me ka palapala 'ana i nā waiwai lepo42.
Ma kekahi ʻaoʻao, ʻo Self-Organizing Graph (SeOM) kahi algorithm aʻo i hoʻohana ʻia ma nā ʻatikala like ʻole e like me Li et al.43, Wang et al.44, Hossain Bhuiyan et al.45 a me Kebonye et al.46 E hoʻoholo i nā ʻano kikoʻī a me ka hui pū ʻana o nā mea.Wang et al.44 outline ʻo SeOM kahi ʻenehana aʻo ikaika i ʻike ʻia no kona hiki ke hui a noʻonoʻo i nā pilikia laina ʻole.44, hiki iā SeOM ke hoʻohui i ka mahele o nā neurons pili a hāʻawi i ka ʻike ʻike kiʻekiʻe.
Ke manaʻo nei kēia pepa e hana i kahi ʻano hoʻohālike paʻa me ka pololei kūpono no ka wānana ʻana i ka ʻike nickel ma nā ʻāina kūlanakauhale a me nā ʻāina āpau.No laila, e ho'āʻo mākou e pane i nā nīnau noiʻi e hāʻawi mai i nā ʻano like ʻole. Akā naʻe, pehea ka pololei o ke kumu hoʻohālike i ka wānana ʻana i ka mea i manaʻo ʻia? Eia kekahi, he aha ke kiʻekiʻe o ka loiloi kūpono e pili ana i ka hōʻoia a me ka loiloi pololei? nā ʻāina pili-kūlanakauhale, a me (d) ka noi ʻana o SeOM e hana i kahi palapala ʻāina hoʻonā kiʻekiʻe o ka hoʻololi ʻana i nā spatial nickel.
Ke hanaʻia nei ke aʻoʻana ma Czech Republic, ma ka mokuʻo Frydek Mistek ma ka'āina Moravia-Silesian (e nānā i ka Figure 1). ʻO ka palapala honua o ka wahi aʻoʻana heʻeleʻele loa ia a he hapa nui ia o ka'āpana Moravia-Silesian Beskidy,ʻo ia kekahi hapa o ka palena waho o nā mauna Carpathian. Aia ka wahi aʻo ma waena o 49 ′ 0 ° N a me 49 ° ′. aia ma waena o 225 a 327 m;Eia naʻe, ua helu ʻia ka ʻōnaehana helu Koppen no ke kūlana climatic o ka ʻāina ʻo Cfb = temperate oceanic climate, He nui ka ua i loko o nā mahina maloʻo. Ua ʻano liʻiliʻi nā mahana ma waena o ka makahiki ma waena o -5 °C a me 24 °C, ʻaʻole hiki ke hāʻule ma lalo o −14 °C a i ʻole ma luna o 30 °C, ʻoiai ʻo ka awelika o ka makahiki he 52 mm. wahi he 1,208 square kilometres, me 39.38% o ka aina mahiai a me 49.36% o ka ululāʻau uhi. Ma kekahi lima, o ka wahi i hoʻohana 'ia i loko o keia haʻawina mea e pili ana i 889.8 square kilometres. Ma a puni Ostrava, ka oihana kila a me ka hana metala ua loa active. ʻO ka ikaika o ka huila me ka mālama ʻana i kona ductility maikaʻi a me ka paʻakikī), a me ka mahiʻai koʻikoʻi e like me ka hoʻohana ʻana i ka phosphate fertilizer a me ka hana holoholona he mau kumu noiʻi kūpono o ka nickel ma ka ʻāina (e like me ka hoʻohui ʻana i ka nickel i nā keiki hipa e hoʻonui i ka ulu ʻana o nā keiki hipa a me nā pipi hānai haʻahaʻa). Hiki ke hoʻoheheʻe ʻia mai ka waihoʻoluʻu o ka lepo, ka hale, a me ka ʻona. ʻO nā mbisol ka lanakila ma ka Repubalika Czech49.
Palapala ʻāina aʻo [Ua hana ʻia ka palapala ʻāina aʻo me ka ArcGIS Desktop (ESRI, Inc, version 10.7, URL: https://desktop.arcgis.com).]
Ua loaʻa mai he 115 mau laʻana topsoil mai ke kūlanakauhale a me nā ʻāina āpau i ka moku ʻo Frydek Mistek. ʻO ke kumu hoʻohālike i hoʻohana ʻia he grid maʻamau me nā ʻāpana lepo i hoʻokaʻawale ʻia 2 × 2 km kaawale, a ua ana ʻia ka topsoil ma kahi hohonu o 0 a 20 cm me ka hoʻohana ʻana i kahi mea GPS paʻa lima (Leica Zeno 5 GPS). -dryed e hana pulverized samples, pulverized by a mechanical system (Fritsch disc mill), and sieved (sieve size 2 mm) .Place 1 gram of dryed, homogenized and sieved soil samples in clear labeled teflon bottles. s e kū i ka pō no ka hopena (aqua regia papahana) .E kau i ka supernatant ma kahi wela wela (ka wela: 100 W a me 160 °C) no 2 h no ka hoʻomaʻamaʻa ʻana i ke kaʻina hana o ka hoʻoheheʻe ʻana o nā laʻana, a laila cool.Transfer ka supernatant i ka 50 ml volumetric flask a hoʻoheheʻe i 50 ml me ka wai deionized deionized o ka wai ml. Eia hou, 1 ml o ka dilution solution ua hoʻoheheʻeʻia me 9 ml o ka wai deionized a kānanaʻia i loko o ka 12 ml paipu i hoʻomākaukauʻia no ka PTE pseudo-concentration. Ua hoʻoholoʻia nā manaʻo o PTEs (As, Cd, Cr, Cu, Mn, Ni, Pb, Zn, Ca, Mg, K) e ICP-OES (In Plaductive Optical). fic, USA) e like me nā kaʻina hana maʻamau a me ka ʻaelike.Ensure Quality Assurance and Control (QA/QC) procedures (SRM NIST 2711a Montana II Soil).PTEs me nā palena ʻike ma lalo o ka hapalua ua kāpae ʻia mai kēia haʻawina. ʻO ka palena ʻike o ka PTE i hoʻohana ʻia i kēia haʻawina ʻo 0.0004. (ʻoe). ized, ua hana ʻia kahi loiloi pālua.
ʻO ka Empirical Bayesian Kriging (EBK) kekahi o nā ʻenehana interpolation geostatistical i hoʻohana ʻia i ka hoʻohālike ʻana i nā ʻano ʻano like ʻole e like me ka ʻepekema lepo. ʻO ke kaʻina hana interpolation o EBK e hahai ana i nā pae hoʻohālike ʻekolu i manaʻo ʻia e Krivoruchko50, (a) koho ke kumu hoʻohālike i ka semivariogram mai ka papa helu hoʻokomo (b) ka waiwai wānana hou no kēlā me kēia wahi hoʻokomo e pili ana i ka semivariogram i hoʻokumu ʻia. hope
Ma kahi o \(Prob\left(A\right)\) e hoike ana i ka mua, \(Prob\left(B\right)\) i malama ole ia ka probability marginal i ka hapanui o na hihia, \(Prob (B,A)\ ) . Hoʻokumu ʻia ka helu semivariogram ma luna o ka rula Bayes, e hōʻike ana i ka propensity o ka nānā ʻana i nā ʻikepili i hiki ke hana ʻia mai ka semivariograms a laila e hoʻoholo ai i ka waiwai o ka semivariograms. e hana i ka ʻikepili o ka nānā ʻana mai ka semivariogram.
ʻO ka mīkini vector kākoʻo he algorithm aʻo mīkini e hoʻopuka i kahi hyperplane hoʻokaʻawale maikaʻi loa e ʻike i nā papa kūʻokoʻa like ʻole akā ʻaʻole linearly. Ua hoʻohana ʻia ma kēia loiloi. ʻO Cherkassky lāua ʻo Mulier53 i paionia SVMR ma ke ʻano he kernel-based regression, ka helu ʻana o ia mea i hana ʻia me ka hoʻohana ʻana i ke ʻano hoʻohālikelike linear me nā hana spatial multi-country. Ua hōʻike ʻo John et al54 e hoʻohana ana ka hoʻohālike SVMR i ka hyperplane linear regression, e hana ana i nā pilina nonlinear e hiki ai iā Vosland al.55, epsilon (ε)-SVMR ke hoʻohana nei i ka ʻikepili i aʻo ʻia no ka loaʻa ʻana o kahi kumu hoʻohālike e like me ka hana epsilon-insensitive i hoʻopili ʻia e palapala i ka ʻikepili me ka maikaʻi epsilon bias mai ka hoʻomaʻamaʻa ʻana i ka ʻikepili i hoʻopili ʻia. hōʻike ʻia ma lalo nei e Vapnik51.
kahi e hōʻike ai ka b i ka paepae scalar, \(K\hema({x}_{,}{ x}_{k}\ʻākau)\) hōʻike i ka hana kernel, \(\alpha\) hōʻike i ka Lagrange multiplier, N Hōʻike i kahi papa helu helu, \({x}_{k}\) hōʻike i ka hoʻokomo ʻikepili, a ʻo \(y\) ka hana ʻikepili i hoʻohana ʻia. Hoʻohana ʻia ka kernel RBF no ka hoʻoholo ʻana i ka hiʻohiʻona SVMR maikaʻi loa, he mea koʻikoʻi ia e loaʻa ai ka helu hoʻopaʻi maʻalahi loa C a me ka kernel parameter gamma (γ) no ka ʻikepili aʻo PTE. ʻO ka mua, ua loiloi mākou i ka hoʻonohonoho hoʻomaʻamaʻa a laila hoʻāʻo i ka hana hoʻohālike ma ka hoʻonohonoho hōʻoia.
ʻO ke kumu hoʻohālikelike laina nui (MLR) he ʻano hoʻohālikelike e hōʻike ana i ka pilina ma waena o ka mea hoʻololi pane a me ka helu o nā mea hoʻololi wānana ma o ka hoʻohana ʻana i nā ʻāpana linear pooled i helu ʻia me ka hoʻohana ʻana i ke ʻano square square liʻiliʻi. ʻO ka hoohalike LR
ma kahi o y ka hoololi pane, \(a\) ka hoololi, n ka helu o na mea wanana, \({b}_{1}\) ka hoopau hapa o na coefficients, \({x}_{ i}\) hoike mai i ka mea wanana a wehewehe wehewehe paha, a o \({\varepsilon }_{i}\) ka hewa i ke koena.
Ua loaʻa nā ʻano hoʻohālike i hui ʻia e ka sandwiching EBK me SVMR a me MLR. Hana ʻia kēia ma ka unuhi ʻana i nā waiwai wānana mai ka interpolation EBK. ʻO nā waiwai wānana i loaʻa mai ka interpolated Ca, K, a me Mg e loaʻa ma o ke kaʻina hui e loaʻa ai nā ʻano hou, e like me CaK, CaMg, a me KMg. , K, Mg, CaK, CaMg, KMg a me CaKMg. Ua lilo kēia mau mea hoʻololi i kā mākou wānana, e kōkua ana i ka wānana ʻana i ka nickel concentrations ma nā kūlanakauhale a me nā ʻāina peri-urban. gression (EBK_MLR). ʻO ka mea maʻamau, ua hoʻohana ʻia nā mea hoʻololi Ca, K, Mg, CaK, CaMg, KMg, a me CaKMg ma ke ʻano he covariates ma ke ʻano he wānana o ka ʻike Ni i loko o ke kūlanakauhale a me nā ʻāina pili.
Ua lilo ka hoʻohana ʻana iā SeOM i mea hana kaulana no ka hoʻonohonoho ʻana, loiloi, a me ka wānana ʻana i ka ʻikepili ma ka ʻāpana kālā, mālama ola kino, ʻoihana, ʻikepili, ʻepekema lepo, a me nā mea hou aku. vector variables43,56.Melssen et al.57 wehewehe i ka pilina o ka mea hoʻokomo i loko o ka neural network ma o ka hoʻokahi papa komo i ka mea hoʻopuka vector me ka hoʻokahi paona vector.ʻO ka huahana i hanaʻia e SeOM he palapala 'elua-dimensional i loaʻa i nā neurons likeʻole a iʻole nā nodes i ulanaʻia i loko o nā palapala'āina hexagonal, circular, a square topological paha e like me ko lākou pili. Ua koho ʻia ʻo 4, ʻo ia hoʻi, he 55-map ʻāpana (5 × 11).
ʻO ka helu o nā ʻikepili i hoʻohana ʻia ma kēia noiʻi ʻana he 115 mau hōʻailona. Hoʻohana ʻia nā mea hoʻololi i hana ʻia e ka interpolation EBK ma ke ʻano he wānana a wehewehe wehewehe paha no ka wānana ʻana i ka variable target (PTE). Hoʻohana ʻia ka hoʻohālike ma RStudio me ka hoʻohana ʻana i ka waihona waihona (Kohonen), hale waihona (caret), hale waihona (modelr), hale waihona ("e1071"), hale waihona ("plyr")), hale waihona puke ("caTools prospects")
Ua hoʻohana ʻia nā ʻāpana hōʻoia like ʻole no ka hoʻoholo ʻana i ke kumu hoʻohālike maikaʻi loa i kūpono no ka wānana ʻana i ka neʻe ʻana o ka nickel i ka lepo a no ka loiloi ʻana i ka pololei o ke kumu hoʻohālike a me kāna hōʻoia ʻana. Ua loiloi ʻia nā hiʻohiʻona Hybridization me ka mean absolute error (MAE), root mean square error (RMSE), a me R-squared or coefficient determination (R2). ʻO nā ana kūʻokoʻa e wehewehe i ka mana wānana o ke kumu hoʻohālike, ʻoiai ʻo MAE e hoʻoholo i ka waiwai quantitative maoli. Pono ke kiʻekiʻe o ka waiwai R2 e loiloi i ke kumu hoʻohālike maikaʻi loa me ka hoʻohana ʻana i nā ʻāpana hōʻoia, ʻoi aku ka pili o ka waiwai i ka 1, ʻoi aku ka kiʻekiʻe o ka pololei. Wahi a Li et al.59, ua manaʻo ʻia ka waiwai hōʻailona R2 o 0.75 a ʻoi aku paha he wānana maikaʻi;Mai ka 0.5 a hiki i ka 0.75 ka hana hoʻohālike i ʻae ʻia, a ma lalo o 0.5 ʻaʻole i ʻae ʻia ka hana hoʻohālikelike.
kahi n e hōʻike ana i ka nui o ka waiwai i ʻike ʻia\({Y}_{i}\) i ka pane i ana ʻia, a ʻo \({\widehat{Y}}_{i}\) pū kekahi i ka waiwai pane wānana, no laila, no nā ʻike mua i.
Hōʻike ʻia nā wehewehe ʻikepili o ka wānana a me nā ʻano pane pane ma ka Papa 1, e hōʻike ana i ka mean, deviation maʻamau (SD), coefficient of variation (CV), liʻiliʻi, kiʻekiʻe, kurtosis, a me skewness. ʻO ka palena liʻiliʻi a me ka palena kiʻekiʻe o nā mea i ka hoʻemi ʻana o ka Mg < Ca < K < Ni a me Ca < Mg < K
ʻO ka pilina o nā mea wānana me nā mea pane i hōʻike i kahi pilina kūpono ma waena o nā mea (e nānā i ka Figure 3).Manaʻo ʻo 68 a me Santo69 i ka like ʻole o ko lākou pae i ka lepo. Akā naʻe, ʻo Ca a me Mg ka mea kū'ē iā K, akā pili maikaʻi ʻo CaK. ʻO kēia paha ma muli o ka hoʻohana ʻana i nā mea kanu e like me ka potassium carbonate, ʻo ia ka 56% kiʻekiʻe i ka pālolo. Hoʻopili ʻia ka nickel me Ca, K a me Mg me nā waiwai r = 0.52, 0.63 a me 0.55. ʻO nā pilina e pili ana i ka calcium, magnesium, a me nā PTE e like me nickel he paʻakikī, akā naʻe, hoʻemi ka magnesium i ka calcium absorption.
ʻO ka matrix correlation no nā mea e hōʻike ana i ka pilina ma waena o nā mea wānana a me nā pane (E hoʻomaopopo: aia kēia kiʻi i kahi scatterplot ma waena o nā mea, nā pae koʻikoʻi e pili ana i ka p <0,001).
Hōʻike ka Figure 4 i ka māhele ākea o nā mea. Wahi a Burgos et al70, ʻo ka hoʻohana ʻana i ka puʻunaue spatial he ʻenehana i hoʻohana ʻia e helu a hōʻike i nā wahi wela ma nā wahi haumia. Hiki ke ʻike ʻia nā pae hoʻonui o Ca ma Fig. ʻO ka lime wikiwiki (calcium oxide) e hoʻemi i ka acidity o ka lepo a me kona hoʻohana ʻana i nā wili kila e like me ka oxygen alkaline i ke kaʻina hana. s. Ua kūlike kēia me nā haʻawina ʻē aʻe, e like me Madaras a me Lipavský72, Madaras et al.73, Pulkrabová et al.74, Asare et al.75, nāna i ʻike i ka hoʻopaʻa ʻana o ka lepo a me ka mālama ʻana me KCl a me NPK i hopena i ka ʻike K kiʻekiʻe i ka lepo.ʻO ka hoʻonui ʻia ʻana o ka Potassium Spatial ma ke komohana ʻākau o ka palapala hoʻohele ma muli o ka hoʻohana ʻana i nā mea hoʻomoʻa pālolo e like me ka potassium chloride, potassium sulfate, potassium nitrate, potash, a me ka potash e hoʻonui ai i ka pāpaʻi o nā lepo ʻilihune.Zádorová et al.76 a me Tlustoš et al.77 ua ho'ākāka 'ia ka ho'ohana 'ana i nā mea ho'omomona K i ho'onui i ka K ma ka lepo a e ho'onui loa i ka waiwai o ka lepo i ka wā lō'ihi, 'o ia ho'i, 'o K a me Mg e hō'ike ana i kahi wela o ka lepo. 'O nā wahi wela ma ke komohana akau o ka palapala 'āina a me ka hikina hema o ka palapala 'āina. , e like me ka potassium magnesium sulfate, magnesium sulfate, a me Kieserite, mālama i nā hemahema (ʻike ʻia nā mea kanu i ka poni, ʻulaʻula, a ʻeleʻele paha, e hōʻike ana i ka hemahema o ka magnesium) ma nā lepo me kahi pae pH maʻamau6. ʻO ka hōʻiliʻili ʻana o ka nickel ma ke kūlanakauhale a me ke kaona o nā ʻāina āpau ma muli paha o nā hana anthropogenic e like me ka mahiʻai a me ke koʻikoʻi o ka nickel i ka hana kila stainless78.
ʻO ka hoʻohele ākea o nā mea [ua hoʻokumu ʻia ka palapala ʻāina hoʻohele ma ka hoʻohana ʻana i ka ArcGIS Desktop (ESRI, Inc, Version 10.7, URL: https://desktop.arcgis.com).]
Hōʻike ʻia nā hualoaʻa hōʻike kumu hoʻohālike no nā mea i hoʻohana ʻia i loko o kēia haʻawina ma ka Papa 2. Ma kekahi ʻaoʻao, ua kokoke nā RMSE a me MAE o Ni i ka ʻole (0.86 RMSE, -0.08 MAE). Ma kekahi ʻaoʻao, ua ʻae ʻia nā waiwai ʻelua o RMSE a me MAE o K. ʻoi aku ka maikaʻi ma mua o nā hopena o John et al.54 me ka hoʻohana ʻana i ka synergistic kriging e wānana i nā manaʻo S ma ka lepo me ka hoʻohana ʻana i ka ʻikepili i hōʻiliʻili like.41, Yan et al.79, Beguin et al.80, Adhikary et al.81 a me Ioane et al.82, oi aku o K a me Ni.
Ua loiloiʻia ka hana o nāʻano hoʻokahi no ka wānanaʻana i ka nickel ma nā kūlanakauhale a me nā'āina āpau e hoʻohana ana i ka hana o nā hiʻohiʻona (Table 3). Hōʻoia ka hōʻoia a me ka loiloi pololei i hōʻoia i ka Ca_Mg_K predictor i hui pūʻia me ka EBK SVMR model i ka hana maikaʻi loa. 37 (R2), 95.479 mg/kg (RMSE) a me 77.368 mg/kg (MAE) Ca_Mg_K-SVMR he 0.663 (R2), 235.974 mg/kg (RMSE) a me 166.946 mg/kg (MAE) . Ca_Mg-EBK_SVMR (0.643 = R2);ʻOi aku ka kiʻekiʻe o kā lākou mau hopena RMSE a me MAE ma mua o nā Ca_Mg_K-EBK_SVMR (R2 0.637) (e nānā i ka Papa 3). Eia kekahi, ʻo ka RMSE a me ka MAE o ka Ca_Mg-EBK_SVMR (RMSE = 1664.64 a me MAE = 1031.49) ke kumu hoʻohālike ʻo 17.5 a me 1.49, ʻoi aku ka nui o ka Ca_Mg-EBK. .Pēlā hoʻi, ʻo ka RMSE a me ka MAE o ka Ca_Mg-K SVMR (RMSE = 235.974 a me MAE = 166.946) he 2.5 a me 2.2 ka nui ma mua o ka Ca_Mg_K-EBK_SVMR RMSE a me MAE. Kebonye et al.46 a me Ioane et al.54, ʻoi aku ka pili o ka RMSE a me ka MAE i ka ʻole, ʻoi aku ka maikaʻi o nā hopena. SVMR a me EBK_SVMR ua kiʻekiʻe aʻe nā helu RSME a me MAE. 'o ia ho'i, 'o ka 'oi aku o ka heterogeneous o ka dataset, 'o ia ke ki'eki'e o ka MAE a me ka RMSE waiwai. 'O ka pololei o ka loiloi cross-validation o ka Ca_Mg_K-EBK_SVMR kŘkohu huikau no ka wanana ana o Ni ma loko o ke kaona a me ka lepo o ka aina he 63.70%. E like me Li et al.59, ʻo kēia pae o ka pololei he helu hana hoʻohālike e ʻae ʻia. Hoʻohālikelike ʻia nā hopena o kēia manawa me kahi noiʻi mua e Tarasov et al36 nona ke kŘkohu hybrid i hana i ka MLPRK (Multilayer Perceptron Residual Kriging), pili i ka EBK_SVMR pololei helu helu helu helu helu 'ike i loko o ka haʻawina o keia manawa, RMSE (210) a me ka MAE (167.5) ua oi aku mamua o ko kakou mau hualoaʻa ma ka haʻawina o kēia manawa (RMSE 95.479, MAE 77.368). al.36 (0.544), ua maopopo i ka coefficient o ka hoʻoholo 'ana (R2) i oi aku i loko o keia huikau kükohu.O ka palena o ka hewa (RMSE a me MAE) (EBK SVMR) no ka hui 'ana i elua manawa emi. Likelike, Sergeev et al.34 i hoʻopaʻa 0.28 (R2) no ka hoʻomohala hybrid kŘkohu (Multilayer Kriging2. Ke hoʻopaʻa nei i kēia manawa). ʻO 63.7% ka wānana pololei o kēia kükohu (EBK SVMR), aʻo ka wānana pololei i loaʻa iā Sergeev et al.ʻO 34 ka 28%. ʻO ka palapala 'āina hope (Fig. 5) i hana ʻia me ka hoʻohana ʻana i ke kumu hoʻohālike EBK_SVMR a me Ca_Mg_K ma ke ʻano he wānana e hōʻike ana i nā wanana o nā wahi wela a me ka haʻahaʻa a hiki i ka nickel ma luna o ka wahi haʻawina holoʻokoʻa.
Hōʻike ʻia ka palapala wānana hope loa me ka hoʻohana ʻana i ke kumu hoʻohālike hybrid EBK_SVMR a me ka hoʻohana ʻana iā Ca_Mg_K ma ke ʻano he wānana.
Hōʻike ʻia ma ka Figure 6 he mau ʻano PTE ma ke ʻano he papa hana i loko o nā neurons pākahi. nā neurons a me nā neurons haʻahaʻa loa. No laila, kaʻana like ʻo CaK a me CaMg i kekahi mau mea like me nā neurons kiʻekiʻe loa a me nā ʻano kala haʻahaʻa haʻahaʻa. hōʻike ʻia ke kumu hoʻohālike i kahi ʻano kala kiʻekiʻe e hōʻike ana i ka neʻe ʻana o ka nickel i loko o ka lepo (e nānā i ke kiʻi 4). Hōʻike ka mokulele ʻo CakMg i kahi ʻano kala like ʻole mai ka haʻahaʻa a hiki i ke kiʻekiʻe e like me ke kala pololei. soils.Figure 7 depicts the contour method in the k-means grouping on the map, shared in 3 clusters based on the predicted value in each model.The contour method represents the best number of clusters.Of 115 soil samples ohi, category 1 loaa ka loa o ka lepo laana, 74. Cluster 2 loaa 33 mau laana. ka wehewehe ʻana i ka puʻupuʻu pololei.Ma muli o ka nui o nā kaʻina hana anthropogenic a kūlohelohe e pili ana i ka hoʻokumu ʻana i ka lepo, paʻakikī ke hoʻokaʻawale pono i nā kumulāʻau puʻupuʻu ma kahi palapala SeOM78 i puʻunaue ʻia.
Hoʻopuka ʻia nā mea mokulele e kēlā me kēia Empirical Bayesian Kriging Support Vector Machine (EBK_SVM_SeOM).
ʻO nā ʻāpana hoʻohālikelike hui like ʻole [Ua hana ʻia nā palapala SeOM me ka hoʻohana ʻana iā RStudio (version 1.4.1717: https://www.rstudio.com/).]
Hōʻike maopopo ka haʻawina i kēia manawa i nā ʻano hana hoʻohālike no ka neʻe ʻana o ka nickel ma ke kūlanakauhale a me nā ʻāina peri-urban. ʻO ka māhele ākea o nā mea i hōʻike ʻia e EBK_SVMR (e nānā i ke kiʻi 5). Hōʻike nā hopena i ke ʻano o ka mīkini hoʻihoʻi ʻana o ka mīkini vector kākoʻo (Ca Mg K-SVMR) wānana i ka neʻe ʻana o Ni i ka lepo ma ke ʻano hoʻokahi kumu hoʻohālike, akā ʻo ka hōʻoia ʻana a me ka pololei o ka loiloi loiloi e hōʻike ana i nā hewa kiʻekiʻe loa ma ke ʻano o RMSE a me MAE. R2). Loaʻa nā hopena maikaʻi me ka hoʻohana ʻana i ka EBK SVMR a me nā mea i hui pū ʻia (CaKMg) me nā hewa haʻahaʻa RMSE a me MAE me ka pololei o 63,7%. ʻO nā mea kanu a me nā haumia ʻoihana o ka lepo e ka ʻoihana kila e hoʻonui i ka neʻe ʻana o ka nickel i ka lepo. Ua hōʻike ʻia kēia haʻawina e hiki i ke kumu hoʻohālike EBK ke hōʻemi i ka pae o ka hewa a hoʻomaikaʻi i ka pololei o ke kumu hoʻohālike o ka hāʻawi ʻana i ka lepo i nā ʻāina kūloko a i ʻole nā kūlanakauhale.Eia kekahi, ke manaʻo nei mākou e hoʻohana i ka EBK e hoʻohui me nā algorithm aʻo mīkini like ʻole.akā naʻe, ʻo ka hoʻohana ʻana i nā covariates hou e hoʻomaikaʻi nui i ka hana o ke kumu hoʻohālike, hiki ke manaʻo ʻia he palena o ka hana o kēia manawa.
PlantProbs.net.Nickel i loko o nā mea kanu a me ka lepo https://plantprobs.net/plant/nutrientImbalances/sodium.html (Loaʻa iā 28 ʻApelila 2021).
Ke holomua nei ʻo Kasprzak, KS Nickel i nā toxicology kaiapuni hou.surroundings.toxicology.11, 145–183 (1987).
Cempel, M. & Nikel, G. Nickel: He loiloi o kona mau kumu a me ka toxicology kaiapuni.Polish J. Environment.Stud.15, 375–382 (2006).
ʻO Freedman, B. & Hutchinson, TC Nā mea hoʻohaumia mai ka lewa a me ka hōʻiliʻili ʻana i ka lepo a me nā mea kanu kokoke i kahi mea hoʻoheheʻe nickel-copper ma Sudbury, Ontario, Canada.can.J.Bot.58(1), 108-132.https://doi.org/10.1139/b80-014 (1980).
Manyiwa, T. et al. Nā metala kaumaha i loko o ka lepo, nā mea kanu a me nā pilikia e pili ana i ka hānai ʻana i nā ruminants kokoke i ka Selebi-Phikwe copper-nickel mine ma Botswana.surroundings.Geochemistry.Health https://doi.org/10.1007/s10653-021-00918-x (2021).
Cabata-Pendias.Kabata-Pendias A. 2011. Trace elements in soil and… – Google Scholar https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=Kabata-Pendias+A.+2011.+Trace+ Elements+in+soils+and+NY+plants.9 CRC+Press&btnG= (Ua loaʻa iā 24 Nov 2020).
Almås, A., Singh, B., Agriculture, TS-NJ of & 1995, undefined. Nā hopena o ka ʻoihana nickel Russian ma nā ʻano metala kaumaha ma nā ʻāina mahiʻai a me nā mauʻu ma Soer-Varanger, Norway.agris.fao.org.
ʻO Nielsen, GD et al.Nickel absorption a me ka paʻa ʻana i ka wai inu e pili ana i ka ʻai a me ka nickel sensitivity.toxicology.application.Pharmacodynamics.154, 67-75 (1999).
Costa, M. & Klein, CB Nickel carcinogenesis, mutation, epigenetics or selection.surroundings.Health Perspective.107, 2 (1999).
Ajman, PC;Ajado, SK;Borůvka, L.;Bini, JKM;Sarkody, VYO;Kobonye, NM;Ka nānā 'ana o nā mea 'awa'awa paha: he loiloi bibliometric.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: He Moʻolelo Pōkole a me Kekahi Haʻawina. Geoderma 264, 301–311.https://doi.org/10.1016/j.geoderma.2015.07.017 (2016).
McBratney, AB, Mendonça Santos, ML & Minasny, B. Ma ka palapala ʻāina kikohoʻe.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=en&as_sdt=0%2C5&q=CV+Deutsch%2C+2002%2C+Geostatistical+Reservoir+Modeling%2C +Oxford=+University+Css. ed 28 ʻApelila 2021).
Ka manawa hoʻouna: Iulai-22-2022