{"id":4993,"date":"2018-08-10T08:31:05","date_gmt":"2018-08-10T07:31:05","guid":{"rendered":"https:\/\/www.vtei.cz\/?p=4993"},"modified":"2024-07-16T14:31:45","modified_gmt":"2024-07-16T13:31:45","slug":"discharge-prediction-by-swat-model-in-case-of-olesna-water-reservoir","status":"publish","type":"post","link":"https:\/\/www.vtei.cz\/en\/2018\/08\/discharge-prediction-by-swat-model-in-case-of-olesna-water-reservoir\/","title":{"rendered":"Discharge prediction by SWAT model in case of Ole\u0161n\u00e1 water reservoir"},"content":{"rendered":"

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

 <\/p>\n

Souhrn<\/h2>\n

P\u0159\u00edsp\u011bvek je zam\u011b\u0159en na posouzen\u00ed v\u00fdsledk\u016f simulace pr\u016ftok\u016f komplexn\u00edm modelem SWAT, kter\u00fd je obt\u00ed\u017en\u00fd pro u\u017e\u00edv\u00e1n\u00ed z\u00a0hlediska pot\u0159eby vstupn\u00edch dat o\u00a0p\u016fd\u00e1ch a\u00a0managementu povod\u00ed. Na p\u0159\u00edkladu povod\u00ed VD Ole\u0161n\u00e1, kter\u00e9 pat\u0159\u00ed mezi men\u0161\u00ed povod\u00ed, ne\u017e je b\u011b\u017en\u00e9 pro pou\u017e\u00edv\u00e1n\u00ed SWATu, byla provedena kalibrace vybran\u00fdch parametr\u016f p\u0159i srovn\u00e1n\u00ed s\u00a0m\u011b\u0159en\u00fdmi daty a\u00a0n\u00e1sledn\u011b validace kalibrovan\u00e9ho modelu na valida\u010dn\u00ed pozorovan\u00e9 \u0159ad\u011b. Model m\u00e1 v\u00a0tomto p\u0159\u00edpad\u011b probl\u00e9m s\u00a0predikc\u00ed denn\u00edch pr\u016ftok\u016f spojen\u00fd s\u00a0v\u011bt\u0161\u00edmi sr\u00e1\u017ekov\u00fdmi ud\u00e1lostmi, m\u011bs\u00ed\u010dn\u00ed pr\u016ftoky pak predikuje velmi dob\u0159e. Model je velmi vhodn\u00fd p\u0159edev\u0161\u00edm pro posouzen\u00ed dopadu r\u016fzn\u00fdch sc\u00e9n\u00e1\u0159\u016f (klimatick\u00e9 zm\u011bny, vyu\u017e\u00edv\u00e1n\u00ed krajiny v\u00a0povod\u00ed apod.) v\u00a0dlouhodob\u00e9m \u010dasov\u00e9m m\u011b\u0159\u00edtku. Dal\u0161\u00ed rozvoj modelu je pl\u00e1nov\u00e1n pro kalibraci parametr\u016f modelu pro predikci transportu \u017eivin z\u00a0povod\u00ed do n\u00e1dr\u017ee.<\/p>\n\"\"<\/a>\n

\u00davod<\/h2>\n

Matematick\u00e9 modely se v\u00a0hydrologii uplat\u0148uj\u00ed ji\u017e \u0159adu let ruku v\u00a0ruce s\u00a0rostouc\u00edm v\u00fdkonem v\u00fdpo\u010detn\u00ed techniky a\u00a0dostupnost\u00ed vstupn\u00edch dat. Umo\u017e\u0148uj\u00ed realizaci virtu\u00e1ln\u00edch experiment\u016f, kter\u00e9 lze empiricky prov\u00e9st bu\u010f jen velmi obt\u00ed\u017en\u011b, nebo v\u016fbec. V\u00a0praxi se vyu\u017e\u00edvaj\u00ed ke studiu hydrologick\u00fdch proces\u016f a\u00a0odhad\u016fm dopadu r\u016fzn\u00fdch sc\u00e9n\u00e1\u0159\u016f na hydrologick\u00e9 procesy. Z\u00e1kladn\u00edm p\u0159edpokladem je, aby zvolen\u00fd model dostate\u010dn\u011b realisticky simuloval pot\u0159ebn\u00e9 procesy pro minimalizaci nejistoty v\u00fdsledk\u016f. P\u0159edev\u0161\u00edm v\u00a0sou\u010dasn\u00e9 dob\u011b, ve kter\u00e9 pot\u0159ebujeme pl\u00e1novat \u00fa\u010dinn\u00e1 opat\u0159en\u00ed v\u00a0krajin\u011b k\u00a0adaptaci na prob\u00edhaj\u00edc\u00ed klimatick\u00e9 zm\u011bny, m\u016f\u017ee b\u00fdt role t\u011bchto model\u016f pro posouzen\u00ed efektivity adapta\u010dn\u00edch opat\u0159en\u00ed nezastupiteln\u00e1. P\u0159edev\u0161\u00edm v\u00a0simulac\u00edch proces\u016f v\u00a0zem\u011bd\u011blsky vyu\u017e\u00edvan\u00fdch povod\u00edch je \u0161iroce uplat\u0148ov\u00e1n model SWAT (akronym pro soil and water assessment tool), jeho\u017e obliba ve sv\u011bt\u011b neust\u00e1le roste.<\/p>\n

Tento p\u0159\u00edsp\u011bvek m\u00e1 za c\u00edl p\u0159edstavit model SWAT, jeho\u017e u\u017eit\u00ed na \u00fazem\u00ed \u010cesk\u00e9 republiky je velmi vz\u00e1cn\u00e9, p\u0159esto\u017ee m\u00e1 velk\u00fd potenci\u00e1l. Postupn\u011b je stru\u010dn\u011b pops\u00e1n samotn\u00fd model a\u00a0n\u00e1sledn\u011b je na p\u0159\u00edkladu povod\u00ed vodn\u00ed n\u00e1dr\u017ee Ole\u0161n\u00e1 (33\u2008km2<\/sup>), kter\u00e1 slou\u017e\u00ed jako zdroj vody pro pr\u016fmysl a\u00a0zna\u010dn\u011b trp\u00ed p\u0159\u00edsunem sediment\u016f a\u00a0\u017eivin, provedena kalibrace a\u00a0validace simulac\u00ed pr\u016ftok\u016f pro denn\u00ed \u010dasov\u00fd krok a\u00a0n\u00e1sledn\u00e9 srovn\u00e1n\u00ed s\u00a0m\u011bs\u00ed\u010dn\u00edm \u010dasov\u00fdm krokem.<\/p>\n

Materi\u00e1l a\u00a0metodika<\/h2>\n

Model SWAT<\/h3>\n

Model SWAT je vytv\u00e1\u0159en p\u0159edev\u0161\u00edm z\u00e1sluhou Jeffa Arnolda na US Department of Agriculture Agricultural Research Service (USDA-ARS). SWAT je kontinu\u00e1ln\u00ed model v\u00a0m\u011b\u0159\u00edtku povod\u00ed, kter\u00fd pracuje s\u00a0denn\u00edm \u010dasov\u00fdm krokem a\u00a0je vytvo\u0159en pro predikci vlivu managementu povod\u00ed na hydrologii, sedimenty a\u00a0chemick\u00e9 l\u00e1tky spojen\u00e9 se zem\u011bd\u011blskou \u010dinnost\u00ed, jako jsou \u017eiviny, pesticidy, t\u011b\u017ek\u00e9 kovy a\u00a0dal\u0161\u00ed.<\/p>\n

SWAT vy\u017eaduje pom\u011brn\u011b podrobn\u00e9 \u00fadaje o\u00a0p\u016fd\u011b, informace o\u00a0vyu\u017e\u00edv\u00e1n\u00ed p\u016fdy\/hospoda\u0159en\u00ed a\u00a0v\u00fd\u0161kopisn\u00e1 data. Mezi z\u00e1kladn\u00ed vstupn\u00ed parametry pro p\u016fdu pat\u0159\u00ed hloubka, zrnitostn\u00ed slo\u017een\u00ed, obsah organick\u00e9 hmoty a\u00a0hydropedologick\u00e9 vlastnosti, a\u00a0to p\u0159edev\u0161\u00edm hydrologick\u00e1 skupina a\u00a0nasycen\u00e1 hydraulick\u00e1 vodivost. Parametry lze p\u016fd\u00e1m p\u0159id\u011blit zvl\u00e1\u0161\u0165 pro jednotliv\u00e9 horizonty. D\u016fle\u017eit\u00e9 parametry charakterizuj\u00edc\u00ed typ vyu\u017eit\u00ed p\u016fdy jsou \u010d\u00edsla odtokov\u00fdch k\u0159ivek pro jednotliv\u00e9 hydrologick\u00e9 skupiny p\u016fd. Pro urbanizovan\u00e9 plochy je nav\u00edc podstatn\u00fd parametr pod\u00edlu nepropustn\u00e9ho povrchu. Pro zem\u011bd\u011blsky vyu\u017e\u00edvan\u00e9 plochy je mo\u017enost parametrizace podrobn\u00e9ho managementu zahrnuj\u00edc\u00edho rotaci plodin, agrotechnick\u00e9 lh\u016fty set\u00ed, sklizn\u011b, orby, hnojen\u00ed, pastvy apod. V\u00fd\u0161kopisn\u00e1 data slou\u017e\u00ed k\u00a0odvozen\u00ed sklonitosti \u00fazem\u00ed, kter\u00e1 je rozd\u011blena do mno\u017estv\u00ed t\u0159\u00edd podle u\u017eivatele. D\u00e1le lze p\u0159idat dal\u0161\u00ed v\u00fdznamn\u00e9 \u010dinnosti hospoda\u0159en\u00ed nap\u0159. bodov\u00e9 zdroje, odb\u011bry vody, zavla\u017eov\u00e1n\u00ed a\u00a0dal\u0161\u00ed. Nejd\u016fle\u017eit\u011bj\u0161\u00ed vstupn\u00ed klimatologick\u00e1 data jsou denn\u00ed sr\u00e1\u017ekov\u00e9 \u00fahrny a\u00a0minim\u00e1ln\u00ed a\u00a0maxim\u00e1ln\u00ed teplota vzduchu. Pokud jsou k\u00a0dispozici podrobn\u011bj\u0161\u00ed sr\u00e1\u017ekov\u00e9 \u00fahrny, lze je vyu\u017e\u00edt. Dal\u0161\u00ed vstupn\u00ed klimatologick\u00e1 data jsou \u00fahrny glob\u00e1ln\u00edho z\u00e1\u0159en\u00ed, relativn\u00ed vlhkost vzduchu a\u00a0pr\u016fm\u011brn\u00e1 rychlost v\u011btru. Ty jsou pot\u0159eba pro metody v\u00fdpo\u010dtu evapotranspirace, kter\u00e9 jsou n\u00e1ro\u010dn\u011bj\u0161\u00ed na data. Hodnoty v\u0161ech t\u011bchto veli\u010din lze simulovat pomoc\u00ed stochastick\u00e9ho gener\u00e1toru po\u010das\u00ed.<\/p>\n\"\"<\/a>\n

Obr. 1. Diagram postupu v\u00fdpo\u010dtu hydrologick\u00e9 bilance v HRU\/subpovod\u00edch
\nFig. 1. Flowchart of hydrological budget calculation for HRU\/subbasins<\/h6>\n

Proces v\u00fdpo\u010dtu lze rozd\u011blit na \u010d\u00e1st prob\u00edhaj\u00edc\u00ed na \u00farovni krajiny a\u00a0\u010d\u00e1st p\u0159edstavuj\u00edc\u00ed procesy v\u00a0r\u00e1mci hlavn\u00edch vodn\u00edch tok\u016f. Pro v\u00fdpo\u010det vodn\u00ed bilance a\u00a0n\u00e1sledn\u011b odnos\u016f jednotliv\u00fdch l\u00e1tek na \u00farovni krajiny je plocha povod\u00ed rozd\u011blena do d\u00edl\u010d\u00edch subpovod\u00ed a\u00a0d\u00e1le do hydrologick\u00fdch odpov\u011bdn\u00edch jednotek, tzv. HRU (Hydrologic Response Unit), tvo\u0159\u00edc\u00ed kvazihomogenn\u00ed plochy s\u00a0unik\u00e1tn\u00ed kombinac\u00ed typu p\u016fdy, krajinn\u00e9ho pokryvu a\u00a0t\u0159\u00eddy sklonitosti, kter\u00e9 zachycuj\u00ed p\u0159edev\u0161\u00edm rozd\u00edly v\u00a0hodnot\u00e1ch evapotranspirace, stejn\u011b jako v\u00a0hodnot\u00e1ch odtoku pro jednotliv\u00e9 HRU a\u00a0rovn\u011b\u017e pro jednotliv\u00e9 plodiny. Z\u00e1kladn\u00ed procesy zahrnuty v\u00a0hydrologick\u00e9 bilanci povod\u00ed jsou obsa\u017eeny v\u00a0rovnici (1). K\u00a0jednotliv\u00fdm proces\u016fm existuje v\u00a0n\u011bkter\u00fdch p\u0159\u00edpadech nab\u00eddka v\u00edcero matematick\u00fdch popis\u016f hlavn\u011b v\u00a0z\u00e1vislosti na n\u00e1ro\u010dnosti vstupn\u00edch dat. Nap\u0159\u00edklad u\u00a0v\u00fdpo\u010dtu infiltrace sr\u00e1\u017eek lze pou\u017e\u00edt metodu odtokov\u00fdch k\u0159ivek, pro kterou posta\u010duj\u00ed denn\u00ed sr\u00e1\u017ekov\u00e9 \u00fahrny, nebo metodu Green&Ampt, kter\u00e1 vy\u017eaduje sr\u00e1\u017eky v\u00a0podrobn\u011bj\u0161\u00edch \u010dasov\u00fdch intervalech. Podobn\u011b u\u00a0v\u00fdpo\u010dtu potenci\u00e1ln\u00ed evapotranspirace lze vyu\u017e\u00edt metody n\u00e1ro\u010dn\u011bj\u0161\u00ed na meteorologick\u00e1 data, konkr\u00e9tn\u011b Penmann-Monteith nebo Pristley-Taylor, ale z\u00e1rove\u0148 metodu Hargreaves, kter\u00e1 vy\u017eaduje pouze \u00fadaje o\u00a0teplot\u00e1ch. Odtok a\u00a0koncentrace l\u00e1tek jsou potom modelov\u00e1ny odd\u011blen\u011b pro ka\u017edou HRU zvl\u00e1\u0161\u0165 a\u00a0n\u00e1sledn\u011b slou\u010deny do celkov\u00e9 hodnoty odtoku a\u00a0koncentrace z\u00a0plochy subpovod\u00ed podle diagramu na obr. 1<\/em>.<\/p>\n\"\"<\/a>\n\n\n\n\n\n\n\n\n\n
Kde<\/td>\nSWt<\/sub><\/em><\/td>\nje<\/td>\nfin\u00e1ln\u00ed obsah vody v\u00a0p\u016fd\u011b,<\/td>\n<\/tr>\n
<\/td>\nSW0<\/sub><\/em><\/td>\n<\/td>\nobsah vody v\u00a0p\u016fd\u011b na po\u010d\u00e1tku,<\/td>\n<\/tr>\n
<\/td>\nRden<\/sub><\/em><\/td>\n<\/td>\ndenn\u00ed \u00fahrn sr\u00e1\u017eek,<\/td>\n<\/tr>\n
<\/td>\nQpov<\/sub><\/em><\/td>\n<\/td>\npovrchov\u00fd odtok,<\/td>\n<\/tr>\n
<\/td>\nE<\/em><\/td>\n<\/td>\nevapotranspirace,<\/td>\n<\/tr>\n
<\/td>\nw<\/em><\/td>\n<\/td>\ninfiltrace do vad\u00f3zn\u00ed z\u00f3ny,<\/td>\n<\/tr>\n
<\/td>\nQpod<\/sub><\/em><\/td>\n<\/td>\nz\u00e1kladn\u00ed odtok.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Procesy prob\u00edhaj\u00edc\u00ed na \u00farovni hlavn\u00edch vodn\u00edch tok\u016f p\u0159edstavuj\u00ed transformaci vstupuj\u00edc\u00edho odtoku vody, sediment\u016f, \u017eivin a\u00a0dal\u0161\u00edch l\u00e1tek. Hlavn\u00ed toky jsou zde schematizov\u00e1ny do homogenn\u00edch kan\u00e1l\u016f s\u00a0lichob\u011b\u017en\u00edkov\u00fdm profilem. Rychlost proud\u011bn\u00ed vody je po\u010d\u00edt\u00e1na p\u0159es manningovu rovnici a\u00a0zm\u011bna pomoc\u00ed metod Variable storage nebo Muskingum. Nap\u0159\u00edklad zm\u011bna \u017eivin je po\u010d\u00edt\u00e1na podle modelu QUAL2E.<\/p>\n

V\u00fdstupy z\u00a0modelu jsou na \u00farovni jednotliv\u00fdch HRU, subpovod\u00ed a\u00a0hlavn\u00edch vodn\u00edch tok\u016f v\u00a0denn\u00edm, m\u011bs\u00ed\u010dn\u00edm nebo ro\u010dn\u00edm \u010dasov\u00e9m kroku. Pro HRU jsou ve v\u00fdsledc\u00edch hlavn\u011b hodnoty vodn\u00ed bilance (povrchov\u00fd, podpovrchov\u00fd a\u00a0z\u00e1kladn\u00ed odtok, obsah vody v\u00a0p\u016fd\u011b, potenci\u00e1ln\u00ed a\u00a0aktu\u00e1ln\u00ed evapotranspirace atd.), d\u00e1le nap\u0159. mno\u017estv\u00ed biomasy, odnos sediment\u016f, bilance \u017eivin apod. Pro subpovod\u00ed se jedn\u00e1 o\u00a0velmi podobn\u00e9 v\u00fdsledky jako na \u00farovni HRU, kde doch\u00e1z\u00ed k\u00a0jejich sou\u010dtu. Na \u00farovni vodn\u00edch tok\u016f jsou v\u00fdsledkem vstupy ze subpovod\u00ed, co se t\u00fd\u010de vody, sediment\u016f a\u00a0dal\u0161\u00edch l\u00e1tek a\u00a0jejich transformace vodn\u00edm tokem.<\/p>\n

Vy\u010derp\u00e1vaj\u00edc\u00ed p\u0159ehled o\u00a0v\u0161ech simulovan\u00fdch procesech a\u00a0jejich matematick\u00e9m popisu, doporu\u010den\u00e9m pou\u017eit\u00ed apod. pod\u00e1v\u00e1 podrobn\u00e1 dokumentace k\u00a0modelu [1].<\/p>\n

Probl\u00e9m u\u017eit\u00ed modelu SWAT v\u00a0podm\u00ednk\u00e1ch \u010cesk\u00e9 republiky spo\u010d\u00edv\u00e1 hlavn\u011b v\u00a0nedostupnosti vstupn\u00edch dat o\u00a0p\u016fd\u00e1ch, co\u017e se t\u00fdk\u00e1 p\u0159edev\u0161\u00edm textury p\u016fdy, obsahu organick\u00e9 hmoty a\u00a0jejich hydropedologick\u00fdch vlastnost\u00ed a\u00a0d\u00e1le zem\u011bd\u011blsk\u00e9ho managementu \u00fazem\u00ed, hlavn\u011b osevn\u00ed postupy a\u00a0u\u017e\u00edvan\u00e1 agrotechnika. Datab\u00e1ze dostupn\u00e9 ke SWATu jsou tvo\u0159eny p\u0159edev\u0161\u00edm pro \u00fazem\u00ed USA.<\/p>\n

Tabulka 1. Zastoupen\u00ed sklonitostn\u00edch t\u0159\u00edd, p\u016fd a\u00a0vyu\u017eit\u00ed \u00fazem\u00ed ve vstupn\u00edch datech
\nTable 1. Composition of slope classes, soils and land use in input data<\/h5>\n\"\"<\/a>\n

Tvorba modelu<\/h3>\n

Jako vstupn\u00ed v\u00fd\u0161kopisn\u00e1 data byl pou\u017eit digit\u00e1ln\u00ed model reli\u00e9fu 4. generace (DMR\u00a04G) poskytovan\u00fd \u010c\u00daZK. Podle sklonitosti bylo povod\u00ed rozd\u011bleno na t\u0159\u00eddy 0\u20137\u00b0, 7\u201317\u00b0 a\u00a0nad 17\u00b0. Data o\u00a0p\u016fd\u00e1ch byla odvozena p\u0159edev\u0161\u00edm z\u00a0komplexn\u00edho pr\u016fzkumu zem\u011bd\u011blsk\u00fdch p\u016fd proveden\u00e9ho v\u00a060. letech minul\u00e9ho stolet\u00ed. Rozlo\u017een\u00ed p\u016fdn\u00edch subtyp\u016f bylo p\u0159evzato z\u00a0p\u016fdn\u00edch map a\u00a0vlastnosti t\u00fdkaj\u00edc\u00ed se hloubek horizont\u016f, zrnitosti, obsahu organick\u00e9 hmoty byly p\u0159evzaty z\u00a0v\u00fdb\u011brov\u00fdch sond vztahuj\u00edc\u00edch se k\u00a0dan\u00e9mu hospod\u00e1\u0159sk\u00e9mu obvodu [2] a\u00a0hodnoty pro lesn\u00ed p\u016fdy byly p\u0159evzaty z\u00a0typologick\u00fdch p\u016fdn\u00edch sond od \u00daHUL. Hydropedologick\u00e9 vlastnosti, jako je nasycen\u00e1 hydraulick\u00e1 vodivost, byly odhadnuty podle pedotransferov\u00fdch funkc\u00ed [3]. Typ krajinn\u00e9ho pokryvu byl odvozen interpretac\u00ed ortofota pro dosa\u017een\u00ed lep\u0161\u00edho rozli\u0161en\u00ed, ne\u017e kter\u00e9 poskytuje jinak hojn\u011b u\u017e\u00edvan\u00e1 datab\u00e1ze Corine Land Cover, a\u00a0management zem\u011bd\u011blsk\u00fdch ploch byl zji\u0161t\u011bn na z\u00e1klad\u011b konzultace s\u00a0agronomem podniku Beskyd Agro a.\u00a0s., kter\u00fd hospoda\u0159\u00ed na zna\u010dn\u00e9 plo\u0161e v\u00a0povod\u00ed. Vstupn\u00ed denn\u00ed meteorologick\u00e1 data poch\u00e1z\u00ed z\u00a0nejbli\u017e\u0161\u00edch m\u011b\u0159ic\u00edch stanic v\u00a0r\u00e1mci s\u00edt\u011b \u010cHM\u00da. Hodnoty min. a\u00a0max. teplot, \u00fahrn\u016f z\u00e1\u0159en\u00ed, relativn\u00ed vzdu\u0161n\u00e9 vlhkosti a\u00a0rychlosti v\u011btru poch\u00e1z\u00ed ze stanice v\u00a0Mo\u0161nov\u011b. Hodnoty \u00fahrn\u016f sr\u00e1\u017eek poch\u00e1z\u00ed ze \u010dty\u0159 sr\u00e1\u017ekom\u011brn\u00fdch stanic.<\/p>\n\"\"<\/a>\n

Obr. 2. P\u016fdn\u00ed typy, vyu\u017eit\u00ed \u00fazem\u00ed a \u010dlen\u011bn\u00ed na subpovod\u00ed
\nFig. 2. Soil types, land use and subbasins delineation<\/h6>\n

Samotn\u00e1 diskretizace povod\u00ed na subpovod\u00ed a\u00a0HRU prob\u011bhla v\u00a0prost\u0159ed\u00ed programu Quantum GIS s\u00a0vyu\u017eit\u00edm z\u00e1suvn\u00e9ho modulu QSWAT [4]. Celkov\u011b bylo povod\u00ed rozd\u011bleno na 21 subpovod\u00ed a\u00a01 102 HRU.<\/p>\n

Jak je patrn\u00e9 z\u00a0tabulky 1<\/em>, vyu\u017eit\u00ed \u00fazem\u00ed povod\u00ed je spojeno p\u0159edev\u0161\u00edm se zem\u011bd\u011blsk\u00fdmi plochami a\u00a0lesy. Orn\u00e1 p\u016fda je situov\u00e1na p\u0159edev\u0161\u00edm na luvizem\u00edch, hn\u011bdozem\u00edch a\u00a0pseudoglej\u00edch. Mezi nejv\u00edce p\u011bstovan\u00e9 plodiny zde pat\u0159\u00ed \u0159epka, kuku\u0159ice, ozim\u00e1 p\u0161enice a\u00a0ozim\u00fd je\u010dmen. Na kambizem\u00edch pak p\u0159evl\u00e1d\u00e1 vyu\u017eit\u00ed jako travn\u00ed porosty, nej\u010dast\u011bji louky a\u00a0v\u00a0men\u0161\u00ed m\u00ed\u0159e pastviny a\u00a0ve vy\u0161\u0161\u00edch poloh\u00e1ch lesy. Plo\u0161n\u00e9 rozlo\u017een\u00ed zobrazuj\u00ed mapy na obr. 2<\/em>.<\/p>\n

Kalibrace modelu<\/h3>\n

Pro kalibraci vybran\u00fdch parametr\u016f byly v\u00fdsledky simulac\u00ed pr\u016ftok\u016f porovn\u00e1ny s\u00a0m\u011b\u0159en\u00fdmi \u0159adami z\u00a0vodom\u011brn\u00e9 stanice v\u00a0obci Palkovice (poskytuje \u010cHM\u00da) a\u00a0z\u00a0bilan\u010dn\u00edch p\u0159\u00edtok\u016f do VD Ole\u0161n\u00e1 (poskytuje st\u00e1tn\u00ed podnik Povod\u00ed Odry) v\u00a0denn\u00edch a\u00a0m\u011bs\u00ed\u010dn\u00edch pr\u016fm\u011brn\u00fdch hodnot\u00e1ch. U\u017eit\u00ed t\u011bchto dvou typ\u016f dat vn\u00e1\u0161\u00ed do kalibrace parametr\u016f odli\u0161nou nejistou vzhledem k\u00a0rozd\u00edln\u00e9mu p\u016fvodu (pr\u016ftoky \u010cHM\u00da podle vodn\u00edch stav\u016f p\u0159epo\u010dteny p\u0159es m\u011brn\u00e9 k\u0159ivky, Povod\u00ed Odry pak podle bilance mezi odtokem a\u00a0kol\u00eds\u00e1n\u00edm vody v\u00a0n\u00e1dr\u017ei). Pro kalibraci modelu bylo zvoleno obdob\u00ed mezi roky 2007 a\u00a02011 a\u00a0pro validaci modelu, tedy ov\u011b\u0159en\u00ed v\u00fdkonnosti modelu na \u010dasov\u00e9 \u0159ad\u011b meteorologick\u00fdch dat, obdob\u00ed 2012 a\u017e 2015. Zvolen\u00e9 vstupn\u00ed parametry byly adjustov\u00e1ny automaticky algoritmem SUFI-2 (Sequential Uncertainty Fitting) implementovan\u00fdm v\u00a0programu SWAT-CUP\u00a0[5]. Zm\u011bna parametr\u016f byla umo\u017en\u011bna bu\u010f \u00faplnou zm\u011bnou hodnoty, co\u017e se t\u00fdk\u00e1 hlavn\u011b konstant, nebo relativn\u00ed zm\u011bnou. Hodnoty zm\u011bn jsou v\u00a0r\u00e1mci algoritmu vzorkov\u00e1ny metodou latinsk\u00fdch \u010dtverc\u016f. Jako objektivn\u00ed funkce byl zvolen nejroz\u0161\u00ed\u0159en\u011bj\u0161\u00ed vzorec v\u00a0hydrologii Nash-Sutcliffe efficiency NSE (rovnice\u00a0(2)) pro porovn\u00e1n\u00ed v\u00fdsledk\u016f simulac\u00ed a\u00a0pozorovan\u00fdch dat.<\/p>\n\"\"<\/a>\n\n\n\n\n
Kde<\/td>\nQsim<\/sub><\/em><\/td>\nje<\/td>\nsimulovan\u00fd pr\u016ftok,<\/td>\n<\/tr>\n
<\/td>\nQobs<\/sub>\u00a0<\/em><\/td>\n<\/td>\npozorovan\u00fd pr\u016ftok.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

V\u00fdsledky a\u00a0diskuse<\/h2>\n

C\u00edlem studie bylo na z\u00e1klad\u011b m\u011b\u0159en\u00fdch pr\u016ftokov\u00fdch \u0159ad prov\u00e9st kalibraci a\u00a0validaci modelu, a\u00a0t\u00edm prov\u011b\u0159it funk\u010dnost modelu SWAT na p\u0159\u00edkladu povod\u00ed VN Ole\u0161n\u00e1.<\/p>\n

Byla provedena automatick\u00e1 kalibrace vybran\u00fdch d\u016fle\u017eit\u00fdch parametr\u016f, na n\u011b\u017e jsou v\u00fdstupy model\u016f citliv\u00e9. Jejich hodnoty ukazuje tabulka 2<\/em>. Velk\u00fd vliv zde maj\u00ed parametry spojen\u00e9 se z\u00e1kladn\u00edm odtokem s\u00a0hodnotami odpov\u00eddaj\u00edc\u00edmi povod\u00edm s\u00a0rychlou odezvou, co\u017e odpov\u00edd\u00e1 p\u0159edpokladu na z\u00e1klad\u011b p\u0159evl\u00e1daj\u00edc\u00edho karpatsk\u00e9ho fly\u0161e v\u00a0geologick\u00e9m podlo\u017e\u00ed [6]. Pot\u0159eba sn\u00ed\u017eit hodnoty odtokov\u00fdch k\u0159ivek m\u016f\u017ee b\u00fdt \u010d\u00e1ste\u010dn\u011b zp\u016fsobena p\u0159\u00edtomnost\u00ed plo\u0161n\u00fdch odvod\u0148ovac\u00edch staveb, kter\u00e9 sni\u017euj\u00ed pravd\u011bpodobnost vzniku povrchov\u00e9ho odtoku. P\u0159\u00edtomnost takov\u00fdchto staveb lze do modelu zahrnout, nicm\u00e9n\u011b nejsou k\u00a0dispozici p\u0159\u00edli\u0161 podrobn\u00e9 informace o\u00a0jejich um\u00edst\u011bn\u00ed a\u00a0parametrech.<\/p>\n

Tabulka 2. Kalibrovan\u00e9 hodnoty parametr\u016f modelu
\nTable 2. Calibrated model parameters values<\/h5>\n\"\"<\/a>\n

V\u00a0hodnocen\u00ed modelu p\u0159i kalibraci model simuluje pr\u016ftoky v\u00a0denn\u00edm kroku s\u00a0v\u011bt\u0161\u00ed \u00fasp\u011b\u0161nost\u00ed po stanici v\u00a0Palkovic\u00edch, p\u0159i\u010dem\u017e v\u00a0m\u011bs\u00ed\u010dn\u00edm kroku se tento rozd\u00edl smaz\u00e1v\u00e1 (tabulka 3<\/em>). Hodnota NSE je pom\u011brn\u011b citliv\u00e1 na shodu p\u0159i velk\u00fdch pr\u016ftoc\u00edch, kter\u00e9 se v\u00a0denn\u00edm kroku p\u0159i vstupu denn\u00edch \u00fahrn\u016f sr\u00e1\u017eek obt\u00ed\u017en\u011b predikuj\u00ed, p\u0159edev\u0161\u00edm pr\u00e1v\u011b na men\u0161\u00edch povod\u00edch. Nejv\u00edc je tento probl\u00e9m patrn\u00fd pro kv\u011bten roku 2010\u2008s\u00a0v\u00fdskytem p\u0159ibli\u017en\u011b p\u011btilet\u00e9ho pr\u016ftoku, kde doch\u00e1z\u00ed k\u00a0nejv\u011bt\u0161\u00edmu podhodnocov\u00e1n\u00ed, a\u00a0to dokonce o\u00a0v\u00edce ne\u017e 50\u00a0%. V\u00a0denn\u00edm kroku je ve SWATu pou\u017eita metoda odtokov\u00fdch k\u0159ivek, kter\u00e1 nem\u016f\u017ee bez zohledn\u011bn\u00ed intenzity sr\u00e1\u017eek mnohdy adekv\u00e1tn\u011b vypo\u010d\u00edst v\u00fd\u0161ku infiltrace [7]. Toto negativum se p\u0159est\u00e1v\u00e1 projevovat s\u00a0rostouc\u00ed plochou povod\u00ed.<\/p>\n

Tabulka 3. Hodnoty NSE pro obdob\u00ed kalibrace a\u00a0validace
\nTable 3. NSE values for calibration and validation<\/h5>\n\"\"<\/a>\n

I\u00a0po shlazen\u00ed do m\u011bs\u00ed\u010dn\u00edho kroku z\u016fst\u00e1v\u00e1 \u010d\u00e1st neshody, kter\u00e1 souvis\u00ed p\u0159edev\u0161\u00edm se zimn\u00edmi a\u00a0jarn\u00edmi m\u011bs\u00edci (obr. 3<\/em>). Model zde tedy h\u016f\u0159e p\u0159edpov\u00edd\u00e1 procesy spojen\u00e9 s\u00a0tvorbou sn\u011bhov\u00e9 pokr\u00fdvky a\u00a0n\u00e1sledn\u00fdm odt\u00e1v\u00e1n\u00edm. P\u0159esto se citlivost zm\u011bny parametr\u016f spojen\u00fdch s\u00a0procesy ohledn\u011b sn\u011bhu na zm\u011bny hodnot NSE neuk\u00e1zala jako v\u00fdznamn\u00e1. V\u011bt\u0161\u00ed vliv na NSE m\u00e1 srovn\u00e1n\u00ed vysok\u00fdch pr\u016ftok\u016f v\u00a0letn\u00edch m\u011bs\u00edc\u00edch, kter\u00e9 dosahuj\u00ed vy\u0161\u0161\u00edch kulmina\u010dn\u00edch hodnot. V\u00a0r\u00e1mci doporu\u010den\u00fdch krit\u00e9ri\u00ed pro hodnocen\u00ed na z\u00e1klad\u011b NSE lze konstatovat, \u017ee model produkuje velmi dobr\u00e9 v\u00fdsledky [8]. Pravd\u011bpodobn\u011b bude pot\u0159eba pro objektivn\u00ed funkci zvolit jin\u00fd vztah ne\u017e tradi\u010dn\u00ed NSE, kter\u00fd je citliv\u011bj\u0161\u00ed i\u00a0k\u00a0ni\u017e\u0161\u00edm hodnot\u00e1m pr\u016ftok\u016f, nap\u0159. Kling-Gupta efficiency (KGE) [9], co\u017e je hodnota vych\u00e1zej\u00edc\u00ed z\u00a0rozkladu NSE, nebo hledat optimum nap\u0159\u00ed\u010d v\u00edce m\u00edrami v\u00fdkonnosti. Nejkorektn\u011bj\u0161\u00edm \u0159e\u0161en\u00edm je pak u\u017eit\u00ed tzv. ekvifin\u00e1ln\u00edch model\u016f, tedy nepou\u017e\u00edvat model s\u00a0jednou konkr\u00e9tn\u00ed sadou parametr\u016f, ale se v\u0161emi, kter\u00e9 produkuj\u00ed rozumn\u00e9 v\u00fdsledky.<\/p>\n\"\"<\/a>\n

Obr. 3. Porovn\u00e1n\u00ed simulovan\u00fdch a pozorovan\u00fdch m\u011bs\u00ed\u010dn\u00edch pr\u016ftok\u016f
\nFig. 3. Comparison of simulated and observed monthly flows<\/h6>\n

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

Model SWAT, kter\u00fd je vyv\u00edjen v\u00a0USA, ji\u017e z\u00edskal popularitu nap\u0159\u00ed\u010d sv\u011btem, p\u0159esto je v\u00a0\u010cesk\u00e9 republice jeho u\u017eit\u00ed ojedin\u011bl\u00e9. Jeho u\u017eit\u00ed je limitov\u00e1no \u010dasto \u0161patnou dostupnost\u00ed vstupn\u00edch dat o\u00a0vlastnostech p\u016fd a\u00a0managementu \u00fazem\u00ed. Na p\u0159\u00edkladu povod\u00ed k\u00a0VD Ole\u0161n\u00e1 byl proveden sb\u011br pot\u0159ebn\u00fdch dat na z\u00e1klad\u011b dostupn\u00fdch zdroj\u016f a\u00a0bylo provedeno ov\u011b\u0159en\u00ed v\u00fdkonnosti modelu v\u00a0p\u0159edpov\u011bdi pr\u016ftok\u016f po kalibraci modelu. V\u00fdstupy v\u00a0denn\u00edm kroku se uk\u00e1zaly jako problematick\u00e9 p\u0159edev\u0161\u00edm p\u0159i intenzivn\u011bj\u0161\u00edch sr\u00e1\u017ekov\u00fdch ud\u00e1lostech, u\u00a0kter\u00fdch nelze adekv\u00e1tn\u011b vypo\u010d\u00edst proces infiltrace, kter\u00fd v\u00a0p\u0159\u00edpad\u011b men\u0161\u00edch povod\u00ed nab\u00fdv\u00e1 na v\u00fdznamnosti. V\u00a0r\u00e1mci validace oproti kalibraci pak v\u00fdkonnost je\u0161t\u011b kles\u00e1. V\u00fdsledky simulac\u00ed v\u00a0m\u011bs\u00ed\u010dn\u00edm kroku vych\u00e1zej\u00ed velmi dob\u0159e konzistentn\u011b i\u00a0v\u00a0r\u00e1mci validace.<\/p>\n

Model bude d\u00e1le kalibrov\u00e1n hlavn\u011b pro zp\u0159esn\u011bn\u00ed proces\u016f spojen\u00fdch se sn\u011bhovou pokr\u00fdvkou adjustac\u00ed relevantn\u00edch parametr\u016f tak, aby i\u00a0v\u00a0denn\u00edm kroku adekv\u00e1tn\u011b predikoval hydrologickou bilanci. N\u00e1sledn\u011b bude provedena kalibrace a\u00a0validace proces\u016f spojen\u00fdch se \u017eivinami podle m\u011b\u0159en\u00fdch dat koncentrac\u00ed dusi\u010dnanov\u00e9ho dus\u00edku a\u00a0celkov\u00e9ho fosforu, pouze v\u00a0m\u011bs\u00ed\u010dn\u00edm kroku (mal\u00e1 frekvence m\u011b\u0159en\u00ed), co\u017e p\u0159edstavuje dal\u0161\u00ed v\u00fdzvu, hlavn\u011b pro\u00a0dal\u0161\u00ed zp\u0159es\u0148ov\u00e1n\u00ed managementu povod\u00ed. Takov\u00fdto model d\u00e1le poslou\u017e\u00ed k\u00a0progn\u00f3ze dopadu potenci\u00e1ln\u00edch klimatick\u00fdch zm\u011bn a\u00a0efektivity adapta\u010dn\u00edch opat\u0159en\u00ed.<\/p>\n

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

P\u0159\u00edsp\u011bvek vznikl za podpory intern\u00edho projektu Ostravsk\u00e9 univerzity SGS05\/P\u0159F\/2017-2018 Pozn\u00e1n\u00ed v\u00fdvoje a\u00a0sou\u010dasn\u00e9ho stavu krajiny Z\u00e1padn\u00edch Karpat a\u00a0v\u00fdchodo-sudetsk\u00fdch poho\u0159\u00ed s\u00a0ohledem na \u010dinnost \u010dlov\u011bka a\u00a0sou\u010dasn\u00fdch p\u0159\u00edrodn\u00edch hazard\u016f.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

Mathematical models have been used in hydrology for decision making support for many years. They allow the realization of virtual experiments that can be done empirically either very difficult or not at all. The SWAT model, developed in the US, has gained world wide popularity but its application in the Czech Republic is quite rare yet.<\/p>\n","protected":false},"author":8,"featured_media":4941,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[86],"tags":[1290,1288,1289,1287],"coauthors":[1251],"class_list":["post-4993","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-hydraulics-hydrology-and-hydrogeology","tag-calibration-and-validation","tag-input-data","tag-olesna-reservoir","tag-swat-model"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts\/4993","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=4993"}],"version-history":[{"count":2,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts\/4993\/revisions"}],"predecessor-version":[{"id":30468,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts\/4993\/revisions\/30468"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/media\/4941"}],"wp:attachment":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/media?parent=4993"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/categories?post=4993"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/tags?post=4993"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/coauthors?post=4993"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}