{"id":1326,"date":"2015-12-10T14:57:11","date_gmt":"2015-12-10T14:57:11","guid":{"rendered":"http:\/\/www.vtei.cz\/?p=1326"},"modified":"2024-07-16T10:13:02","modified_gmt":"2024-07-16T09:13:02","slug":"the-accuracy-of-flood-extent-simulation-by-an-alternative-tool-aizm","status":"publish","type":"post","link":"https:\/\/www.vtei.cz\/en\/2015\/12\/the-accuracy-of-flood-extent-simulation-by-an-alternative-tool-aizm\/","title":{"rendered":"The accuracy of flood extent simulation by an alternative tool \u2013 AIZM"},"content":{"rendered":"

<\/i> This article is available in Czech only. For translation or more information on this topic, please contact author.<\/h4>\n

Souhrn<\/h2>\n

Relativn\u011b p\u0159esn\u00e9 ur\u010den\u00ed rozsahu povod\u0148ov\u00fdch rozliv\u016f umo\u017e\u0148uje matematick\u00e9 \u010di fyzik\u00e1ln\u00ed modelov\u00e1n\u00ed, kter\u00e9 se pou\u017e\u00edv\u00e1 a\u00a0rozv\u00edj\u00ed po\u00a0dlouhou \u0159adu let. Ruku v\u00a0ruce s\u00a0r\u016fstem p\u0159esnosti, plynouc\u00ed z\u00a0p\u0159esn\u011bj\u0161\u00edch m\u011b\u0159ic\u00edch za\u0159\u00edzen\u00ed, vzr\u016fstaj\u00ed rovn\u011b\u017e n\u00e1roky na\u00a0v\u00fdpo\u010detn\u00ed kapacity a\u00a0\u010das. Existuje v\u0161ak \u0159ada \u00fasek\u016f vodn\u00edch tok\u016f bez namodelovan\u00fdch inundac\u00ed, pop\u0159.\u00a0by bylo \u017e\u00e1douc\u00ed modelovat \u00faseky vodn\u00edch tok\u016f, pro n\u011b\u017e by bylo mo\u017eno odvodit rozsah zaplaven\u00e9ho \u00fazem\u00ed, a\u00a0to na\u00a0z\u00e1klad\u011b zn\u00e1m\u00e9 \u00farovn\u011b hladiny. Alternativn\u00edm p\u0159\u00edstupem k\u00a0vizualizaci inundac\u00ed je model AIZM (Alternativn\u00ed indikativn\u00ed z\u00e1plavov\u00fd model). Jedn\u00e1 se o\u00a0n\u00e1stroj vytvo\u0159en\u00fd z\u0159et\u011bzen\u00edm n\u011bkolika d\u00edl\u010d\u00edch submodel\u016f v\u00a0prost\u0159ed\u00ed ArcGIS ModelBuilderu um\u00edst\u011bn\u00fdch v\u00a0ArcToolboxu. AIZM je schopen p\u0159i minimu po\u017eadovan\u00fdch vstupn\u00edch dat a\u00a0v\u00a0kr\u00e1tk\u00e9m \u010dase vygenerovat indikativn\u00ed rozsah z\u00e1plavy a\u00a0pole hloubek ve\u00a0form\u011b rastru. Je v\u0161ak pot\u0159eba br\u00e1t v\u00a0potaz jist\u00e9 nedostatky, kter\u00fdmi je tento model zat\u00ed\u017een.<\/p>\n

P\u0159edlo\u017een\u00fd \u010dl\u00e1nek shrnuje prvn\u00ed \u010d\u00e1st v\u00fdsledk\u016f anal\u00fdzy vlivu r\u016fzn\u00fdch charakteristik \u00fasek\u016f t\u0159ic\u00edtky r\u016fzn\u00fdch vodn\u00edch tok\u016f (nap\u0159.\u00a0\u0161\u00ed\u0159ky koryta, pod\u00e9ln\u00e9ho sklonu \u010di antropogenn\u00edho ovlivn\u011bn\u00ed pr\u016fb\u011bhu koryta) na\u00a0m\u00edru shody \u010di neshody v\u00fdstup\u016f AIZM s\u00a0referen\u010dn\u00edmi v\u00fdsledky vypo\u010d\u00edtan\u00fdmi v\u00a0r\u00e1mci tvorby map povod\u0148ov\u00e9ho nebezpe\u010d\u00ed a\u00a0rizik s\u00a0vyu\u017eit\u00edm 1D \u010di 2D numerick\u00e9ho modelov\u00e1n\u00ed. D\u00edky tomu je mo\u017eno identifikovat typy vodn\u00edch tok\u016f, na\u00a0nich\u017e poskytuje AIZM dobr\u00e9 v\u00fdsledky, resp. ur\u010dit ty charakteristiky, kter\u00e9 nejv\u00edce ovliv\u0148uj\u00ed p\u0159esnost v\u00fdstupu.<\/p>\n

\u00davod<\/h2>\n

Technologie geografick\u00fdch informa\u010dn\u00edch syst\u00e9m\u016f (GIS) jsou v\u00a0r\u00e1mci hydrologick\u00e9ho modelov\u00e1n\u00ed pou\u017e\u00edv\u00e1ny pro zpracov\u00e1n\u00ed velik\u00fdch objem\u016f prostorov\u00fdch dat. V\u00fdhodou GIS je, \u017ee umo\u017e\u0148uj\u00ed z\u00edskat, resp. odhadnout \u0159adu charakteristik \u00fazem\u00ed vymezen\u00fdch povod\u00ed. D\u00edky dobr\u00e9 anal\u00fdze a\u00a0popisu prostorov\u00e9 variability jednotliv\u00fdch ukazatel\u016f nab\u00edz\u00ed GIS mo\u017enost zlep\u0161en\u00ed kalibrace model\u016f [1]. V\u00a0podm\u00ednk\u00e1ch \u010cesk\u00e9 republiky se GIS syst\u00e9my nap\u0159.\u00a0dob\u0159e osv\u011bd\u010duj\u00ed p\u0159i stanovov\u00e1n\u00ed parametr\u016f hydrologick\u00fdch model\u016f [2]. Z\u00e1kladn\u00ed koncept pou\u017e\u00edv\u00e1n\u00ed GIS technologi\u00ed v\u00a0r\u00e1mci hydrologick\u00e9ho modelov\u00e1n\u00ed m\u016f\u017ee b\u00fdt charakterizov\u00e1n t\u0159emi kroky: 1. p\u0159\u00edpravou vstupn\u00edch dat (tj.\u00a0pre\u00ad\u2011processingem) a\u00a0n\u00e1sledn\u00fdm exportem do\u00a0hydrologick\u00e9ho modelu, 2. modelov\u00e1n\u00edm sam\u00fdm, 3. fin\u00e1ln\u00edm exportem zp\u011bt do\u00a0prost\u0159ed\u00ed GIS s\u00a0n\u00e1sledn\u00fdm zpracov\u00e1n\u00edm, vizualizac\u00ed a\u00a0interpretac\u00ed [3, 4].<\/p>\n

P\u0159\u00edstupy k\u00a0vz\u00e1jemn\u00e9mu propojen\u00ed GIS n\u00e1stroj\u016f s\u00a0hydrologick\u00fdmi modely mohou b\u00fdt r\u016fzn\u00e9: 1. p\u0159\u00edstup\u00a0\u2013 j\u00e1dro GIS je vyu\u017eito v\u00a0modelu, 2. p\u0159\u00edstup\u00a0\u2013 model je zakomponov\u00e1n p\u0159\u00edmo do\u00a0GISu formou programovac\u00edho jazyka Python \u010di ModelBuilderu, 3. p\u0159\u00edstup\u00a0\u2013 GIS a\u00a0model jsou odd\u011bleny a\u00a0GIS slou\u017e\u00ed pro p\u0159\u00edpravu dat a\u00a0zpracov\u00e1n\u00ed v\u00fdsledk\u016f, 4. p\u0159\u00edstup\u00a0\u2013 \u00fazk\u00e9 programov\u00e9 prov\u00e1z\u00e1n\u00ed GIS, hydrologick\u00e9ho modelu a\u00a0statistick\u00e9ho softwaru [5]. Dal\u0161\u00ed modifikace uveden\u00fdch p\u0159\u00edstup\u016f jsou zalo\u017eeny na\u00a0propojen\u00ed s\u00a0webov\u00fdmi aplikacemi [6].<\/p>\n

Jedn\u00edm z\u00a0alternativn\u00edch p\u0159\u00edstup\u016f k\u00a0ur\u010den\u00ed a\u00a0anal\u00fdze rozliv\u016f je model AIZM, kter\u00fd byl vyvinut na\u00a0P\u0159\u00edrodov\u011bdeck\u00e9 fakult\u011b Univerzity Komensk\u00e9ho v\u00a0Bratislav\u011b [7]. Byl vyvinut jako n\u00e1stroj k\u00a0analyzov\u00e1n\u00ed relativn\u00edch v\u00fd\u0161ek nad ter\u00e9nem v\u00a0okol\u00ed vodn\u00edch tok\u016f. Mo\u017enost aproximace \u00farovn\u011b hladiny v\u00a0oblastech, kde nejsou k\u00a0dispozici v\u00fdstupy hydraulick\u00fdch model\u016f, se jev\u00ed jako u\u017eite\u010dn\u00e1 a\u00a0\u017e\u00e1douc\u00ed. Zde je v\u0161ak pot\u0159eba zd\u016fraznit, \u017ee p\u0159edstavovan\u00fd model v\u00a0\u017e\u00e1dn\u00e9m p\u0159\u00edpad\u011b neaspiruje na\u00a0to, aby nahrazoval v\u00fdstupy zaveden\u00fdch sofistikovan\u00fdch postup\u016f hydrologick\u00e9ho modelov\u00e1n\u00ed. V\u00a0danou chv\u00edli se jedn\u00e1 o\u00a0n\u00e1stroj, kter\u00fd je schopen na\u00a0z\u00e1klad\u011b zad\u00e1n\u00ed z\u00e1kladn\u00edch vstupn\u00edch \u00fadaj\u016f (jsou zm\u00edn\u011bny n\u00ed\u017ee) v\u00a0kr\u00e1tk\u00e9m \u010dase simulovat rozsah inundace. V\u00fdsledky modelu AIZM je pak nutn\u00e9 n\u00e1le\u017eit\u011b interpretovat odborn\u00edkem se znalost\u00ed stanoven\u00ed rozliv\u016f a\u00a0hloubek vody v\u00a0inunda\u010dn\u00edm \u00fazem\u00ed s\u00a0v\u011bdom\u00edm ur\u010dit\u00fdch nevyhnuteln\u00fdch nedostatk\u016f.<\/p>\n

Model AIZM pracuje na\u00a0platform\u011b ESRI ArcGIS for Desktop Advanced 10.x s\u00a0extenzemi 3D Analyst a\u00a0Spatial Analyst. Ur\u010dit\u00e9 modifikace modelu formou z\u00e1sahu do\u00a0jeho struktury jsou v\u00a0p\u0159\u00edpad\u011b ni\u017e\u0161\u00edch verz\u00ed mo\u017en\u00e9, ale s\u00a0nutnost\u00ed instalace dal\u0161\u00edch n\u00e1stroj\u016f, nap\u0159.\u00a0sady ArcHydro.<\/p>\n

ArcGIS nab\u00edz\u00ed grafick\u00e9 v\u00fdvojov\u00e9 prost\u0159ed\u00ed ModelBuilder, v\u00a0n\u011bm\u017e je mo\u017en\u00e9 \u0159et\u011bzit za\u00a0sebe r\u016fzn\u00e9 GIS n\u00e1stroje a\u00a0objekty. V\u00a0ModelBuilderu tedy lze sestavovat komplexn\u00ed n\u00e1stroje a\u00a0\u010d\u00e1ste\u010dn\u011b automatizovan\u00e9 pracovn\u00ed postupy. Vytvo\u0159en\u00e9 modely mohou b\u00fdt ulo\u017eeny v\u00a0toolboxech, nebo ve\u00a0form\u011b toolset\u016f. Model AIZM se skl\u00e1d\u00e1 ze t\u0159\u00ed z\u00e1kladn\u00edch \u010d\u00e1st\u00ed a\u00a0n\u00e1stroje na\u00a0vygenerov\u00e1n\u00ed p\u0159\u00ed\u010dn\u00fdch profil\u016f. Jednotliv\u00e9 \u010d\u00e1sti jsou samostatn\u00fdmi toolboxy, kter\u00e9 jsou sestaveny z\u0159et\u011bzen\u00edm z\u00e1kladn\u00edch n\u00e1stroj\u016f ArcGIS. Ve\u00a0smyslu v\u00fd\u0161e uveden\u00e9ho rozd\u011blen\u00ed [5] je model AIZM zalo\u017een\u00fd na\u00a0integraci modelu do\u00a0syst\u00e9mu GIS. Tyto \u0159et\u011bzce funguj\u00ed jako samostatn\u00e9 n\u00e1stroje toolboxu s\u00a0vlastn\u00edmi vstupy, v\u00fdstupy a\u00a0parametry. Prvn\u00ed toolset pracuje s\u00a0povrchem TINu (Triangulated Interpolated Network), druh\u00fd zpracov\u00e1v\u00e1 povrch ve\u00a0form\u011b rastru a\u00a0t\u0159et\u00ed obsahuje analytick\u00e9 n\u00e1stroje zpracov\u00e1vaj\u00edc\u00ed data z\u00a0TINu i\u00a0rastru.<\/p>\n

Mo\u017enosti pr\u00e1ce s\u00a0modelem AIZM jsou r\u016fzn\u00e9. Jednak z\u00a0hlediska p\u0159\u00ed\u010dn\u00fdch profil\u016f, kdy jsou bu\u010f do\u00a0v\u00fdpo\u010dtu zahrnuty p\u0159\u00ed\u010dn\u00e9 profily generovan\u00e9 automaticky pomoc\u00ed vstupn\u00edho digit\u00e1ln\u00edho modelu ter\u00e9nu, nebo je mo\u017en\u00e9 zad\u00e1vat je tak\u00e9 ru\u010dn\u011b. Jednak tak\u00e9 z\u00a0hlediska zp\u016fsobu generov\u00e1n\u00ed rozlivu s\u00a0povolen\u00edm \u010di zak\u00e1z\u00e1n\u00edm pr\u016fsak\u016f podzemn\u00edch vod mimo samotn\u00fd rozliv.<\/p>\n

K\u00a0proveden\u00ed v\u0161ech v\u00fdpo\u010dt\u016f vy\u017eaduje AIZM vstupn\u00ed data (viz\u00a0obr.\u00a01<\/em>). Konkr\u00e9tn\u011b se jedn\u00e1 o\u00a0linii vodn\u00edho toku, centr\u00e1ln\u00ed body le\u017e\u00edc\u00ed p\u0159esn\u011b uprost\u0159ed v\u00fdpo\u010detn\u00ed masky, tabulku hloubek vody (pro ka\u017ed\u00fd rozliv jednu) v\u00a0\u00farovni koryta, resp. centr\u00e1ln\u00edho bodu, d\u00e1le data dan\u00e9ho digit\u00e1ln\u00edho modelu ter\u00e9nu (rastr nebo povrch TIN) a\u00a0tak\u00e9 sadu p\u0159\u00ed\u010dn\u00fdch profil\u016f, je\u017e jsou generov\u00e1ny v\u00a0r\u00e1mci b\u011bhu modelu AIZM, nebo mohou b\u00fdt zad\u00e1ny ru\u010dn\u011b.<\/p>\n\"caletka-1\"<\/a>\n

Obr.\u00a01. Sch\u00e9ma v\u00fdpo\u010dtu rozsahu rozlivu modelem AIZM\u00a0\u2013 uprost\u0159ed st\u0159edov\u00fd bod, na\u00a0okraji vymezeno \u00fazem\u00ed (v\u00fdpo\u010detn\u00ed maska) o\u00a0rozm\u011bru 2\u2009000\u2009\u00d7\u20092\u2009000\u2008m, na\u00a0linii
\nvodn\u00edho toku vyzna\u010deny p\u0159\u00ed\u010dn\u00e9 profily, plocha rozlivu stanoven\u00e9ho modelem AIZM zn\u00e1zorn\u011bna \u0161ed\u00fdm polygonem
\nFig. 1. Scheme of flood extent computation by the AIZM model\u00a0\u2013 the central point\u00a0in the centre, the mask of 2\u2009000\u2009\u00d7\u20092\u2009000\u2008m is marked out at the edges, stream line drawn\u00a0together with crossectional profiles, the flooded area computed by the AIZM model\u00a0is represented by the grey polygon<\/h6>\n

V\u00fdpo\u010det modelu AIZM prob\u00edh\u00e1 od\u00a0centr\u00e1ln\u00edho bodu, kolem n\u011bho\u017e se vygeneruje \u010dtvercov\u00e1 maska o\u00a0rozm\u011bru 2\u2009000\u2009\u00d7\u20092\u2009000\u2008m (rozm\u011br je mo\u017eno i\u00a0zmen\u0161it, zv\u011bt\u0161en\u00ed by vy\u017eadovalo z\u00e1sah do\u00a0algoritmu v\u00fdpo\u010dtu). Ve\u0161ker\u00e9 v\u00fdpo\u010dty se pak odehr\u00e1vaj\u00ed v\u00a0plo\u0161e masky. Je\u00ad\u2011li vstupn\u00ed digit\u00e1ln\u00ed model ter\u00e9nu v\u00a0TIN (v\u00a0p\u0159\u00edpadech, kdy sada bod\u016f pro interpolaci digit\u00e1ln\u00edho modelu ter\u00e9nu je nepravideln\u011b uspo\u0159\u00e1dan\u00e1), je v\u00a0plo\u0161e masky p\u0159eveden do\u00a0rastru a\u00a0op\u011btovn\u011b interpolov\u00e1n do\u00a0podoby rastru s\u00a0velikost\u00ed bu\u0148ky 2\u2008m (nastaveno implicitn\u011b).<\/p>\n

Linie vodn\u00edho toku je rovn\u011b\u017e o\u0159ez\u00e1na podle rozm\u011bru masky a\u00a0pot\u00e9 je rozd\u011blena na\u00a0\u00faseky tak, \u017ee vzd\u00e1lenost dvou sousedn\u00edch okrajov\u00fdch bod\u016f t\u011bchto \u00fasek\u016f je v\u017edy maxim\u00e1ln\u011b 20\u2008m. N\u00e1sleduje vykreslen\u00ed p\u0159\u00ed\u010dn\u00fdch profil\u016f zalo\u017een\u00e9 na\u00a0metod\u011b Thiessenov\u00fdch polygon\u016f. Zde je nutno zm\u00ednit skute\u010dnost, \u017ee automatick\u00e9 vykreslov\u00e1n\u00ed p\u0159\u00ed\u010dn\u00fdch profil\u016f p\u0159edstavuje jeden z\u00a0nejv\u011bt\u0161\u00edch probl\u00e9m\u016f modelu AIZM, co\u017e je diskutov\u00e1no d\u00e1le.<\/p>\n

V\u00fdpo\u010det rozsahu inundace vych\u00e1z\u00ed ze zn\u00e1m\u00fdch hloubek vody pro danou masku. Na\u00a0z\u00e1klad\u011b t\u011bchto zn\u00e1m\u00fdch hloubek z\u00a0tabulky se ur\u010d\u00ed \u00farove\u0148 hladiny vody pro dan\u00fd p\u0159\u00ed\u010dn\u00fd profil. Hodnota relativn\u00ed hloubky plat\u00ed v\u017edy pro pr\u016fse\u010d\u00edk p\u0159\u00ed\u010dn\u00e9ho profilu s\u00a0lini\u00ed vodn\u00edho toku. Protnut\u00ed linie p\u0159\u00ed\u010dn\u00e9ho profilu s\u00a0modelovan\u00fdm reli\u00e9fem definuje rozsah rozlivu ve\u00a0sm\u011bru p\u0159\u00ed\u010dn\u00e9ho profilu. Rozliv mezi sousedn\u00edmi profily je odvozen na\u00a0z\u00e1klad\u011b povrchu TIN, vznikl\u00e9ho p\u0159eveden\u00edm lini\u00ed p\u0159\u00ed\u010dn\u00fdch profil\u016f na\u00a0body (vertexy), a\u00a0na\u00a0z\u00e1klad\u011b digit\u00e1ln\u00edho modelu ter\u00e9nu, kter\u00fd je v\u00a0r\u00e1mci b\u011bhu modelu p\u0159eveden tak\u00e9 na\u00a0povrch TIN. Ode\u010dten\u00edm \u00farovn\u00ed hladin v\u00a0rozlivu a\u00a0digit\u00e1ln\u00edho modelu ter\u00e9nu vznik\u00e1 rastr hloubek jako jeden z\u00a0v\u00fdstup\u016f. Opakov\u00e1n\u00ed v\u00fdpo\u010dtu prob\u00edh\u00e1 podle po\u010dtu hodnot hloubek ve\u00a0vstupn\u00ed tabulce (ka\u017ed\u00e9mu rozlivu odpov\u00edd\u00e1 jedna hodnota). V\u00fdstupy modelu jsou ukl\u00e1d\u00e1ny ve\u00a0form\u011b rastru i\u00a0vektoru.<\/p>\n

Metodika a data<\/h2>\n

Pro ur\u010den\u00ed p\u0159esnosti v\u00fdstup\u016f modelu AIZM bylo vybr\u00e1no celkem t\u0159icet r\u016fzn\u00fdch \u00fasek\u016f vodn\u00edch tok\u016f \u010cesk\u00e9 republiky v\u00a0oblastech s\u00a0v\u00fdznamn\u00fdm povod\u0148ov\u00fdm rizikem na\u00a0z\u00e1klad\u011b sm\u011brnice 2007\/60\/ES o\u00a0vyhodnocen\u00ed a\u00a0zvl\u00e1d\u00e1n\u00ed povod\u0148ov\u00fdch rizik. Pro tyto \u00faseky jsou k\u00a0dispozici hloubky ve\u00a0form\u011b rastr\u016f a\u00a0rozlivy ve\u00a0form\u011b vektorov\u00fdch polygon\u016f pro povod\u0148ov\u00e9 pr\u016ftoky Q5, Q20, Q100 a\u00a0Q500. Tato data jsou pova\u017eov\u00e1na za\u00a0referen\u010dn\u00ed a\u00a0n\u00e1sledn\u00e9 anal\u00fdzy vych\u00e1zej\u00ed z\u00a0jejich porovn\u00e1n\u00ed s\u00a0v\u00fdstupy n\u00e1stroje AIZM.<\/p>\n

Vstupn\u00ed \u00fadaje AIZM<\/h3>\n

Jako linie, resp. osa vodn\u00edch tok\u016f poslou\u017eila vrstva z\u00a0datab\u00e1ze DIBAVOD (Digit\u00e1ln\u00ed b\u00e1ze vodohospod\u00e1\u0159sk\u00fdch dat) spravovan\u00e9 V\u00daV TGM,\u00a0v.v.i.<\/p>\n

Digit\u00e1ln\u00ed model ter\u00e9nu byl interpolov\u00e1n s\u00a0vyu\u017eit\u00edm p\u0159\u00edslu\u0161n\u00fdch mapov\u00fdch list\u016f (klad list\u016f je toto\u017en\u00fd s\u00a0kladem list\u016f St\u00e1tn\u00ed mapy 1 : 5\u2009000\u00a0\u2013 odvozen\u00e9) sady DMR 4G od\u00a0\u010c\u00daZK, kter\u00e1 byla po\u0159\u00edzena leteck\u00fdm laserov\u00fdm skenov\u00e1n\u00edm. Je tvo\u0159ena s\u00edt\u00ed bod\u016f rovnom\u011brn\u011b rozm\u00edst\u011bn\u00fdch po\u00a0plo\u0161e \u010cesk\u00e9 republiky ve\u00a0vzd\u00e1lenosti 5\u2008m. Jednotliv\u00e9 body p\u0159itom le\u017e\u00ed ve\u00a0st\u0159edu ka\u017ed\u00e9ho gridu, jemu\u017e tak\u00e9 p\u0159\u00edslu\u0161\u00ed hodnota nadmo\u0159sk\u00e9 v\u00fd\u0161ky dan\u00e9ho bodu.<\/p>\n

Pro zvolen\u00e9 centr\u00e1ln\u00ed body v\u0161ech \u00fasek\u016f byly z\u00a0referen\u010dn\u00edch model\u016f zji\u0161t\u011bny hloubky pro rozlivy s\u00a0dobou opakov\u00e1n\u00ed 5, 20 a\u00a0100\u00a0let (tyto hloubky vstupuj\u00ed do\u00a0tabulky hloubek pro dan\u00fd \u00fasek). Pot\u0159ebn\u00e9 vrstvy byly poskytnuty Ministerstvem \u017eivotn\u00edho prost\u0159ed\u00ed.<\/p>\n

Porovn\u00e1n\u00ed rozliv\u016f<\/h3>\n

Rozlivy vypo\u010dten\u00e9 modelem AIZM byly porovn\u00e1ny s\u00a0rozlivy stanoven\u00fdmi referen\u010dn\u00edmi modely. Simulov\u00e1n\u00ed a\u00a0rozlivy na\u00a0v\u0161ech \u00fasec\u00edch jsou rozd\u011bleny v\u017edy na\u00a0dvacet segment\u016f podle sm\u011bru toku, resp. s\u00a0ohledem na\u00a0tvar z\u00e1plavy (obr.\u00a02<\/em>). Ka\u017ed\u00e9mu segmentu pak jednozna\u010dn\u011b odpov\u00edd\u00e1 plocha rozlivu modelovan\u00e9ho Am a\u00a0referen\u010dn\u00edho Ar. Vz\u00e1jemn\u00e9 porovn\u00e1n\u00ed kvantitativn\u00edch charakteristik tok\u016f r\u016fzn\u00fdch parametr\u016f (nap\u0159.\u00a0Rty\u0148ka\u00a0\u2013 Oh\u0159e) nen\u00ed smyslupln\u00e9, a\u00a0proto je m\u00edra shody vyj\u00e1d\u0159ena relativn\u00ed m\u00edrou, \u010dili pom\u011brem.<\/p>\n\"caletka-vzorec-1\"<\/a>\n

Ze vztahu (1) je z\u0159ejm\u00e9, \u017ee pro S\u00a0> 1 je rozliv modelovan\u00fd AIZM v\u011bt\u0161\u00ed oproti modelu referen\u010dn\u00edmu. Pro S\u00a0< 1 je tomu naopak.<\/p>\n\"caletka-2\"<\/a>\n

Obr.\u00a02. Plocha rozlivu na\u00a0\u00faseku Cidliny rozd\u011blen\u00e1 do\u00a0dvaceti segment\u016f s\u00a0vyzna\u010den\u00edm sm\u011bru toku
\nFig. 2. The flood extent on the Cidlina River cut into twenty segments with the flow direction<\/h6>\n

Porovn\u00e1n\u00ed rozliv\u016f je provedeno v\u00a0kontextu kvantitativn\u00edch i\u00a0kvalitativn\u00edch charakteristik vodn\u00edch tok\u016f, resp. v\u0161ech vymezen\u00fdch segment\u016f. Z\u00a0kvantitativn\u00edch charakteristik se konkr\u00e9tn\u011b jedn\u00e1 o\u00a0\u0161\u00ed\u0159ku koryta vypo\u010d\u00edtanou s\u00a0vyu\u017eit\u00edm b\u0159ehov\u00fdch hran z\u00a0datab\u00e1ze DIBAVOD a\u00a0pod\u00e9ln\u00fd sklon dna odvozen\u00fd z\u00a0digit\u00e1ln\u00edho modelu ter\u00e9nu. Sledovan\u00fdmi kvalitativn\u00edmi charakteristikami jsou antropogenn\u00ed ovlivn\u011bn\u00ed sm\u011bru toku odvozen\u00e1 z\u00a0leteck\u00fdch sn\u00edmk\u016f a\u00a0dostupn\u00fdch historick\u00fdch map, pr\u016fto\u010dnost p\u0159\u00edb\u0159e\u017en\u00ed z\u00f3ny ve\u00a0smyslu p\u0159\u00edtomnosti p\u0159ek\u00e1\u017eek (z\u00e1stavba, souvislej\u0161\u00ed vegetace), charakter b\u0159eh\u016f a\u00a0svah\u016f bl\u00ed\u017e\u00edc\u00ed se p\u0159irozen\u00e9mu a\u00a0p\u0159\u00edtomnost vodohospod\u00e1\u0159sk\u00fdch objekt\u016f (nap\u0159.\u00a0jez\u016f \u010di stup\u0148\u016f). Metodika hodnocen\u00ed je nast\u00edn\u011bna v\u00a0tabulce 1<\/em>.<\/p>\n

Tabulka 1. Klasifikace kvalitativn\u00edch charakteristik \u00fasek\u016f vodn\u00edch tok\u016f
\nTable 1. Classification of qualitative characteristics for river sections<\/h5>\n\"caletka-tabulka-1\"<\/a>\n

Statistick\u00e9 zpracov\u00e1n\u00ed za\u00a0\u00fa\u010delem zji\u0161t\u011bn\u00ed popisn\u00fdch statistik a\u00a0proveden\u00ed n\u00e1sledn\u00fdch anal\u00fdz bylo uskute\u010dn\u011bno s\u00a0vyu\u017eit\u00edm softwaru PAST, Statistica a\u00a0s\u00a0pomoc\u00ed programovac\u00edho jazyka R.<\/p>\n

Prvotn\u00ed pr\u016fzkum dat prost\u0159ednictv\u00edm popisn\u00fdch statistik uk\u00e1zal, \u017ee mezi t\u0159iceti vymezen\u00fdmi \u00faseky vodn\u00edch tok\u016f panuje zna\u010dn\u00e1 variabilita hodnot pom\u011br\u016f S.
\nPatrn\u00e9 je to z\u00a0krabicov\u00fdch graf\u016f sestrojen\u00fdch pro ka\u017ed\u00fd \u00fasek vodn\u00edho toku tvo\u0159en\u00fd v\u017edy dvaceti d\u00edl\u010d\u00edmi segmenty (viz\u00a0obr.\u00a03<\/em>). Bylo tak\u00e9 zji\u0161\u0165ov\u00e1no, zda se v\u00fdsledky li\u0161\u00ed s\u00a0ohledem na\u00a0vzd\u00e1lenost od\u00a0st\u0159edov\u00e9ho bodu. Jin\u00fdmi slovy, bylo ov\u011b\u0159ov\u00e1no, zda se nejvzd\u00e1len\u011bj\u0161\u00ed segmenty po\u00a0sm\u011bru toku pod\u00edlej\u00ed na\u00a0rozptylu hodnot pom\u011bru S. Proto byly anal\u00fdzy provedeny pro \u00faseky se v\u0161emi dvaceti segmenty a\u00a0pro \u00faseky s\u00a0odebran\u00fdmi prvn\u00edmi a\u00a0posledn\u00edmi p\u011bti segmenty (po\u00a0sm\u011bru toku). V\u00a0tabulce 2<\/em> je uvedeno percentu\u00e1ln\u011b zastoupen\u00ed hodnot S. Je evidentn\u00ed, \u017ee pro v\u0161echny rozlivy je v\u00fdrazn\u011b vy\u0161\u0161\u00ed pod\u00edl S\u00a0> 1 vyjad\u0159uj\u00edc\u00ed, \u017ee rozliv modelovan\u00fd pomoc\u00ed AIZM je v\u011bt\u0161\u00ed oproti referen\u010dn\u00edmu a\u00a0tak\u00e9 \u017ee odstran\u011bn\u00edm zm\u00edn\u011bn\u00fdch segment\u016f se tento pod\u00edl je\u0161t\u011b zvy\u0161uje.<\/p>\n

Tabulka 2. Procentu\u00e1ln\u00ed zastoupen\u00ed hodnot pom\u011bru S\u00a0> 1 a\u00a0S\u00a0< 1 pro celou sadu dat a\u00a0data zkr\u00e1cen\u00e1 p\u0159i rozlivu Q5, Q20 a\u00a0Q100
\nTable 2. Percentage of ratio S\u00a0> 1 and S\u00a0< 1 in both whole and shortened dataset for flood extents Q5, Q20 and Q100<\/h5>\n\"caletka-tabulka-2\"<\/a>\n

Zna\u010dn\u00e1 variabilita jednotliv\u00fdch \u00fasek\u016f m\u016f\u017ee souviset s\u00a0n\u011bkter\u00fdmi problematick\u00fdmi aspekty interpolace digit\u00e1ln\u00edho modelu ter\u00e9nu. V\u00a0\u00farovni koryta se mohou objevit \u201efale\u0161n\u00e9\u201c elevace, kter\u00e9 se v\u00fdpo\u010dtem p\u0159enesou a\u00a0projev\u00ed v\u00a0nadhodnocen\u00e9m rozsahu rozlivu. Stejn\u011b tak se mohou objevit i\u00a0p\u0159\u00edpady depres\u00ed, kter\u00e9 analogicky vedou k\u00a0podhodnocen\u00e9mu rozlivu. Tento probl\u00e9m v\u0161ak\u00a0nast\u00e1v\u00e1 m\u00e9n\u011b \u010dasto. Dal\u0161\u00ed chyby b\u00fdvaj\u00ed spojeny s\u00a0meandruj\u00edc\u00edmi \u00faseky tok\u016f. Na\u00a0konvexn\u00edch b\u0159ez\u00edch se linie p\u0159\u00ed\u010dn\u00fdch profil\u016f radi\u00e1ln\u011b rozb\u00edhaj\u00ed, tedy roste jejich vzd\u00e1lenost. Plocha inundace pak m\u016f\u017ee b\u00fdt bu\u010f nadhodnocena, nebo se v\u00a0n\u011bkter\u00fdch p\u0159\u00edpadech mohou objevit i\u00a0fale\u0161n\u00e9 p\u0159etr\u017eky. Tomuto probl\u00e9mu bude v\u011bnov\u00e1n n\u00e1sleduj\u00edc\u00ed v\u00fdzkum sm\u011b\u0159uj\u00edc\u00ed k\u00a0upraven\u00ed algoritm\u016f tak, aby k\u00a0t\u011bmto chyb\u00e1m nedoch\u00e1zelo.<\/p>\n\"caletka-3\"<\/a>\n

Obr.\u00a03. Uk\u00e1zka boxplot\u016f hodnot pom\u011br\u016f S\u00a0na\u00a0\u00faseku Oh\u0159e p\u0159i rozlivech Q5, Q20 a\u00a0Q100; data cel\u00e1 (A), zkr\u00e1cen\u00e1 (B)
\nFig. 3. Example of boxplots of ratios S\u00a0on the Oh\u0159e River section for flood extents Q5, Q20 and Q100; whole dataset (A), shortened (B)<\/h6>\n

Vztah mezi vzorky pom\u011br\u016f S\u00a0a\u00a0ostatn\u00edmi ji\u017e zm\u00edn\u011bn\u00fdmi kvantitativn\u00edmi charakteristikami je vyj\u00e1d\u0159en hodnotami Pearsonova a\u00a0Spearmanova korela\u010dn\u00edho koeficientu. Spearman\u016fv koeficient se vyzna\u010duje t\u00edm, \u017ee nen\u00ed tolik ovliv\u0148ov\u00e1n odlehl\u00fdmi hodnotami, jako tomu je u\u00a0Pearsonova koeficientu. Z\u00a0uk\u00e1zky vypo\u010d\u00edtan\u00fdch hodnot v\u00a0tabulce 3<\/em> je z\u0159ejm\u00e9, \u017ee Pearson\u016fv korela\u010dn\u00ed koeficient ukazuje na\u00a0vztah mezi pr\u016fm\u011brn\u00fdm pod\u00e9ln\u00fdm sklonem segmentu a\u00a0odpov\u00eddaj\u00edc\u00edm pom\u011brem S. Obdobn\u00e9 v\u00fdsledky vych\u00e1zej\u00ed pro v\u0161echny zbyl\u00e9 rozlivy pro vzorek dat cel\u00fd i\u00a0zkr\u00e1cen\u00fd.<\/p>\n

Tabulka 3. Hodnoty Spearmanova a\u00a0Pearsonova korela\u010dn\u00edho koeficientu pro \u0161\u00ed\u0159ku koryta, pod\u00e9ln\u00fd profil a\u00a0pom\u011br S\u00a0rozlivu Q5, Q20, Q100
\nTable 3. Values of Spearman\u2019s\u00a0and Pearson\u2019s\u00a0correlation coefficient for width of channel, longitudinal profile and the ratio S\u00a0of flood extent Q<\/span>5,<\/span> Q<\/span>20,<\/span> Q<\/span>100<\/span><\/h5>\n\"caletka-tabulka-3\"<\/a>\n

Soubor v\u0161ech prom\u011bnn\u00fdch byl podroben anal\u00fdze hlavn\u00edch koordin\u00e1t PCoA\u00a0(z\u00a0angl. Principal Coordinates Analysis), kter\u00e1 byla publikov\u00e1na v\u00a0r.\u00a01966 [8]. Umo\u017e\u0148uje odhalit vliv jednotliv\u00fdch prom\u011bnn\u00fdch na\u00a0v\u00fdsledky modelu. Anal\u00fdza je vhodn\u00e1 i\u00a0v\u00a0t\u011bch p\u0159\u00edpadech, kdy zkouman\u00e9 charakteristiky zahrnuj\u00ed v\u00edcestavov\u00e9 kvalitativn\u00ed znaky \u010di sm\u00ed\u0161en\u00e1 data. Grafy na\u00a0obr.\u00a04<\/em> ukazuj\u00ed v\u00fdsledky anal\u00fdzy hlavn\u00edch koordin\u00e1t proveden\u00e9 pomoc\u00ed programovac\u00edho jazyka R.\u00a0Z\u00a0v\u00fdsledk\u016f z\u00edskan\u00fdch pro celou sadu dat se zd\u00e1, \u017ee spolu souvisej\u00ed jednak charakteristiky ovlivn\u011bn\u00ed b\u0159eh\u016f a\u00a0sm\u011bru koryta, jednak charakteristiky pr\u016fto\u010dnost\u00ed b\u0159eh\u016f.<\/p>\n\"caletka-4\"<\/a>\n

Obr.\u00a04. V\u00fdstupy anal\u00fdzy hlavn\u00edch koordin\u00e1t pro cel\u00fd (A) i\u00a0zkr\u00e1cen\u00fd <\/span>(B) vzorek segment\u016f; 1 = \u0161\u00ed\u0159ka koryta, 2 = pod\u00e9ln\u00fd sklon, 3 = ovlivn\u011bn\u00ed sm\u011bru toku, 4 = p\u0159\u00edtomnost vodohospod\u00e1\u0159sk\u00fdch objekt\u016f, 5 = ovlivn\u011bn\u00ed lev\u00e9ho b\u0159ehu, 6 = ovlivn\u011bn\u00ed lev\u00e9ho b\u0159ehu, 7 = pr\u016fto\u010dnost p\u0159\u00edb\u0159e\u017en\u00ed z\u00f3ny na\u00a0prav\u00e9m b\u0159ehu, 8 = pr\u016fto\u010dnost p\u0159\u00edb\u0159e\u017en\u00ed z\u00f3ny na\u00a0prav\u00e9m b\u0159ehu
\nFig. 4. Outputs of the principal coordinates analysis for the entire (A) and shortened (B) sample of the river segments; 1 = channel width, 2 = longitudinal profile, 3 = influencing of the channel\u2019s\u00a0direction, 4 = presence of waterworks, 5 = influencing of the left bank, 6 = influencing of the left bank, 7 = flowage of the litoral zone on the right bank, 8 = flowage of the litoral zone on the right bank<\/h6>\n

Proto\u017ee v\u0161ak sm\u011b\u0159uj\u00ed pod\u00e9l nulov\u00fdch os, nen\u00ed jejich vliv na\u00a0v\u00fdsledky modelu v\u00fdznamn\u00fd. Naproti tomu charakteristika pod\u00e9ln\u00e9ho sklonu od\u00a0ostatn\u00edch pom\u011brn\u011b vybo\u010duje, co\u017e vypov\u00edd\u00e1 o\u00a0jej\u00edm v\u011bt\u0161\u00edm vlivu a\u00a0velikosti pom\u011bru S, co\u017e koresponduje i\u00a0s\u00a0vypo\u010d\u00edtan\u00fdmi korela\u010dn\u00edmi koeficienty. Podobn\u011b je tomu rovn\u011b\u017e u\u00a0dat zkr\u00e1cen\u00fdch.<\/p>\n

Dal\u0161\u00ed p\u0159edstavu o\u00a0struktu\u0159e dat a\u00a0chov\u00e1n\u00ed modelu v\u00a0r\u016fzn\u00fdch \u00fasec\u00edch z\u0159ejm\u011b poskytne precizn\u00ed vyhodnocen\u00ed proveden\u00e9 shlukov\u00e9 anal\u00fdzy. Byla vyu\u017eita Wardova shlukovac\u00ed metoda zalo\u017een\u00e1 na\u00a0anal\u00fdze rozptylu. Vyzna\u010duje se t\u00edm, \u017ee nen\u00ed zat\u00ed\u017eena efektem \u0159et\u011bzen\u00ed [9]. Hodnota kofenetick\u00e9ho koeficientu CC je pro celou sadu dat 0,716 a\u00a0pro zkr\u00e1cen\u00e1 data 0,826. Tyto hodnoty zna\u010d\u00ed, \u017ee pou\u017eitou shlukovac\u00ed metodu je mo\u017eno pova\u017eovat za\u00a0vhodnou.<\/p>\n

Z\u00e1v\u011br<\/h2>\n

Pro t\u0159ic\u00edtku vybran\u00fdch \u00fasek\u016f vodn\u00edch tok\u016f na\u00a0\u00fazem\u00ed \u010cesk\u00e9 republiky bylo sledov\u00e1no, zda se modelovan\u00e9 rozlivy podle AIZM shoduj\u00ed s\u00a0v\u00fdstupy z\u00a0referen\u010dn\u00edch model\u016f. Automatizovan\u00fdm v\u00fdpo\u010dtem v\u00a0prost\u0159ed\u00ed ArcGIS byly ur\u010deny hodnoty kvantitativn\u00edch charakteristik jednotliv\u00fdch segment\u016f \u00fasek\u016f vodn\u00edch tok\u016f. Posouzen\u00edm leteck\u00fdch sn\u00edmk\u016f a\u00a0historick\u00fdch map byly ur\u010deny tak\u00e9 kategorie charakteristik kvalitativn\u00edch. Tyto charakteristiky spole\u010dn\u011b s\u00a0pom\u011brn\u00fdmi hodnotami modelov\u00fdch a\u00a0referen\u010dn\u00edch rozliv\u016f pro\u0161ly anal\u00fdzou hlavn\u00edch koordin\u00e1t za\u00a0\u00fa\u010delem zji\u0161t\u011bn\u00ed vlivu jednotliv\u00fdch charakteristik na\u00a0v\u00fdsledky modelu. Z\u00a0anal\u00fdzy vyplynulo, \u017ee jak pro data \u00fapln\u00e1, tak i\u00a0pro data zkr\u00e1cen\u00e1 je nejdominantn\u011bj\u0161\u00ed charakteristikou zejm\u00e9na pod\u00e9ln\u00fd profil linie vodn\u00edho toku, m\u00e9n\u011b tak\u00e9 \u0161\u00ed\u0159ka koryta. Toto zji\u0161t\u011bn\u00ed koresponduje i\u00a0s\u00a0v\u00fdsledky v\u00fdpo\u010dt\u016f korela\u010dn\u00edch koeficient\u016f. Zkr\u00e1cen\u00edm \u00fasek\u016f tok\u016f vzd\u00e1len\u011bj\u0161\u00edch od\u00a0st\u0159edov\u00e9ho bodu s\u00a0ode\u010dtenou hloubkou se rozptyly hodnot pom\u011bru S,\u00a0resp. chyba stanoven\u00e9ho rozlivu zmen\u0161ily jen u\u00a0n\u011bkter\u00fdch z\u00a0vybran\u00fdch \u00fasek\u016f. K\u00a0p\u0159esn\u011bj\u0161\u00ed interpretaci t\u00e9to skute\u010dnosti v\u0161ak bude nutn\u00e9 dokon\u010dit dal\u0161\u00ed anal\u00fdzy.<\/p>\n

Uveden\u00e9 v\u00fdsledky p\u0159edstavuj\u00ed pouze d\u00edl\u010d\u00ed \u010d\u00e1st rozs\u00e1hlej\u0161\u00ed anal\u00fdzy. V\u00a0n\u00e1sleduj\u00edc\u00edch anal\u00fdz\u00e1ch bude dokon\u010deno d\u016fkladn\u00e9 vyhodnocen\u00ed v\u00fdsledk\u016f shlukov\u00e9 anal\u00fdzy proveden\u00e9 Wardovou metodou a\u00a0v\u00fdsledky modelu AIZM vypo\u010d\u00edtan\u00e9 na\u00a0podkladu DMR 4G budou porovn\u00e1ny s\u00a0v\u00fdpo\u010dty na\u00a0podkladu DMR 5G.<\/p>\n

C\u00edlem anal\u00fdz do\u00a0budoucna je na\u00a0z\u00e1klad\u011b objektivn\u00edch v\u00fdsledk\u016f statistick\u00fdch anal\u00fdz identifikovat typy tok\u016f (s\u00a0ur\u010dit\u00fdmi charakteristikami), pro n\u011b\u017e je u\u017eit\u00ed modelu AIZM vhodn\u00e9, \u010di nikoli. Pokud to bude mo\u017en\u00e9, budou tak\u00e9 testov\u00e1ny modifikovan\u00e9 verze modelu AIZM, zejm\u00e9na ve\u00a0vztahu k\u00a0interpolaci digit\u00e1ln\u00edho modelu ter\u00e9nu.<\/p>\n

Pod\u011bkov\u00e1n\u00ed<\/h3>\n

Na\u00a0tomto m\u00edst\u011b bych r\u00e1d pod\u011bkoval p\u0159edev\u0161\u00edm Mgr.\u00a0Martinu Kolimu za\u00a0spolupr\u00e1ci. D\u00e1le bych velmi r\u00e1d pod\u011bkoval V\u00fdzkumn\u00e9mu \u00fastavu vodohospod\u00e1\u0159sk\u00e9mu TGM.,\u00a0v.v.i., za\u00a0poskytnut\u00ed pot\u0159ebn\u00fdch podkladov\u00fdch dat digit\u00e1ln\u00edho modelu reli\u00e9fu a\u00a0N\u00ad\u2011let\u00fdch rozliv\u016f, bez nich\u017e by tato pr\u00e1ce nemohla vzniknout.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

The estimation of the extent of inundations is enabled by methods of physical and mathematical modelling which have been developing for many years. Increasing accuracy, related to precise measurements, demands higher computing capacity and more time. <\/p>\n","protected":false},"author":8,"featured_media":1128,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[86,88],"tags":[137,136],"coauthors":[181],"class_list":["post-1326","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-hydraulics-hydrology-and-hydrogeology","category-informatics-cartography-in-water-management","tag-flood-extent","tag-inundation"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts\/1326","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/comments?post=1326"}],"version-history":[{"count":2,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts\/1326\/revisions"}],"predecessor-version":[{"id":30297,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts\/1326\/revisions\/30297"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/media\/1128"}],"wp:attachment":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/media?parent=1326"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/categories?post=1326"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/tags?post=1326"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/coauthors?post=1326"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}