J. Hydrol. Hydromech., 5, 22, 2, 57 7 ANALYSIS OF FLOOD EVENTS ON SMALL RIVER CATCHMENTS USING THE KINFIL MODEL P. KOVÁR, P. CUDLÍN, M. HERMAN, F. ZEMEK, M. KORYTÁR? This paper deals with some ways of carrying out an analysis of a flood event using the KINFIL hydrological model on small catchments where both land use and management play a significant role, and where these human activities can influence design discharges. The combination of GIS techniques with the KINFIL model, which is conceived on physically based infiltration approach and on a kinematic wave transformation of direct runoff, provides a tool for analysing historical rainfall-runoff events, for assessing design discharges, and for simulating some hypothetical flood scenarios. KINFIL is a complex model using the correspondence of Curve Number (CN) with soil parameters and the correspondence of kinematic wave transformation with the physiographical parameters of the Všeminka catchment in Eastern Moravia (Czech Republic), which was used in the tests. Two versions of the KINFIL model (KINFIL, KINFIL2) were implemented. The infiltration part of the model is the same in both versions. KINFIL assumes a more schematic geometrization of the catchment topography, distributing the catchment area to a V-shaped form in which a main channel collects direct runoff from both side planes or segments. This is not fully in accord with the topography of the sub-catchments. KINFIL 2 is a more sophisticated version, where the topography is GIS-organized, taking fully into account the river network and its corresponding subcatchments division. The latter version is geographically (and also physically) better based, and the results of the simulation of the July 997 flood waves in the Vseminka experimental catchment fit better with the observed waves. All the topographical and morphological data were analysed and prepared for the KINFIL model (particularly for the KINFIL 2 version), using GIS facilities. Thus the KINFIL 2 version can be applied in future for design discharge assessment when simulating scenarios of various land uses expressing the model parameters.? P. Kovár, M. Korytár, Czech University of Agriculture Prague, Forestry Faculty, Kamýcká 29, 65 2 Praha 6, Czech Republic. P. Cudlín, M. Herman, F. Zemek, Institute of Landscape Ecology, Academy of Sciences, Na Sádkách 7, 37 5 Ceské Budejovice, Czech Republic. 57
P. Kovár, P. Cudlín, M. Herman, F. Zemek, M. Korytár Introduction KEY WORDS: Infiltration Approach, Kinematic Wave, GIS, Rainfall- Runoff Models, Water Balance Models. P. Kovár, P. Cudlín, M. Herman, F. Zemek, M. Korytár: ANALÝZA POVODNOVÝCH UDÁLOSTÍ NA MALÝCH POVODÍCH S VYUŽITÍM MODELU KINFIL. Vodohosp. Cas., 5, 22, 2; 3 lit., 5 obr., 6 tab. Príspevek analyzuje povodnové prípady implementace hydrologického modelu KINFIL, používaného na malých povodích, kde hospodárské využití pozemku a antropogenní vlivy hrají podstatnou roli. Zámerem príspevku je informovat o možnostech využití GIS pri fragmentaci malých povodí za úcelem zpresnení vstupních dat pro hydrologický model KINFIL. Kombinace GIS a KINFIL, který je fyzikálne založen na teorii infiltrace a transformace prímého odtoku kinematickou vlnou, poskytuje nástroj pro analýzu jak historických srážko-odtokových prípadu, tak hypotetických scénárových simulací. Model KINFIL využívá dríve odvozených vztahu mezi hodnotami císel odtokových krivek CN a koncepcních pudních parametru (nasycené hydraulické vodivosti a sorptivity) spolu s parametry transformace na testovaném povodí Všeminky na východní Morave. Byly testovány dve modelové verze: KINFIL a KINFIL 2. Infiltracní cást modelu je v obou verzích stejná. Verze KINFIL je založena na schematické geometrizaci, kde povodí je V-tvaru a do hlavního toku je sváden prímý odtok zobou stran zparalelne usporádaných desek a každá z techto desek je sériove (kaskádove) clenena podle sklonu svahu. Takto definované desky ale nemusí plne respektovat clenení na subpovodí. Naproti tomu nový prístup KINFIL 2, který za úcelem fragmentace povodí využívá prostredku GIS, plne respektuje prubeh rícní síte povodí a její clenení na subpovodí. Základem pro získání požadovaných parametru modelu KINFIL 2 je digitální model reliéfu terénu (DEM). Tato verze je geograficky lépe založená a výsledky simulovaných prutoku povodne z cervence 997 na experimentálním povodí Všeminky se lépe shodují s prutoky merenými. KLÍCOVÁ SLOVA : infiltracní prístup, kinematická vlna, GIS, modely hydrologické bilance. Recent trends in sudden weather changes in Central Europe have clearly led to a rise in the frequency of hydrological extremes, i. e., floods and droughts. Catchment management, including land use, also plays a significant role in rainfall-runoff relationships. Implementation of hydrological models gives an opportunity to make a better analysis of flood situations, with particular reference to direct runoff processes in the context of changes caused by human 58
Analysis of flood events on small river catchments using the KINFIL model Fig.. Land use on the Vseminka catchment. Obr.. Využití pozemku na povodí Všeminky. 59
P. Kovár, P. Cudlín, M. Herman, F. Zemek, M. Korytár activities. Simulation of design discharges on small catchments under the influence of these activities is a key task for hydrometeorological institutes. However, in the case of small catchments, the reliability of this data varies and one possible way to improve it is by applying hydrological models. One such model simulating direct runoff from ungauged catchments is the KINFIL model [K o v a r, P e c h, 996], which can be utilized by experienced hydrologists. The authors have developed the KINFIL model in two versions; KINFIL is simpler, describing a runoff pattern schematically with the use of simple geometrical units (rectangles and segments). KINFIL 2 fully respects the hierarchy of subcatchment systems, but these systems are also geometrized just by rectangles (see Figs 2 and 3). The aim of this paper is to compare the two models, to show their applicability and to assess which is more efficient for implementation in flood reconstruction and subsequently for forecasting or design purposes. Fig. 2. Runoff pattern for KINFIL schematization. Obr. 2. Odtokové schéma pro model KINFIL. 6
Analysis of flood events on small river catchments using the KINFIL model Fig. 3. Runoff pattern for KINFIL 2 schematization. Obr. 3. Odtokové schéma pro model KINFIL 2. Outlines for schematization: Zásady schematizace. Borders between cascades are waterdivides. Borders within one cascade were derived from slopes respecting land use. All planes were transferred to the same area rectangulars. According to average width and real subcatchment area, the plane theoretical length was computed. All these operations were carried out using GIS ARC/INFO. Numbering:? xx right part? 2xx left part? xx - x7x cascade number? xx - xx3 plane number in cascade (downslope) 6
P. Kovár, P. Cudlín, M. Herman, F. Zemek, M. Korytár Model structure The present version of KINFIL uses the Curve Number method [US SCS, 986], but suppresses its weak physical background by substituting physically based infiltration theory for the common empirical CN approach. The correspondences between CN values and soil parameters, such as saturated hydraulic conductivity (K S ) and sorptivity (S f ), were derived through a correlation technique of these parameters with design rainfalls for the territory of the Czech Republic. These correspondences were used for a further simulation of some historical rainfall-runoff events, implementing the KINFIL model namely on typical mountainous catchments. The infiltration part of the model is based on the Morel-Seytoux equations [M o r e l - S e y t o u x, V e r d i n, 98], based on the Green/Ampt concept distinguishing pre- and post-ponding infiltration from constant or variable rainfall. The second basic component of the KINFIL model is a simulation of runoff. In the present version of the model, this process is based on a kinematic flow approximation. For the numerical solution, the explicit Lax-Wendroff finite difference scheme was implemented in the model. Three simulation components, a cascade of planes, converging or diverging segments and channel reaches were used to simulate the topography of a catchment. A more detailed model structure is described elsewhere [K o v á r, 992]. A recent innovation in the geometrization of a catchment is to respect consistently the hierarchy of sub-catchments in a flow direction. The traditional approach to the geometrization of topographical surfaces of a catchment cascade of planes, convergent or divergent segments, and to channel reaches, is represented by the KINFIL model version. A newer and better physically-based approach which fully respects hydromorphological division into the river network (keeping flow direction as the first prerequisite) is introduced by KINFIL2. This version assumes that individual small sub-catchments are substituted by a system of serial/parallel cascades of planes arranged according to flow direction. However, this system should not go into too great topographic detail, but puts emphasis on slopes and roughness conditions. This outline was confirmed by the KINFIL model sensitivity analysis [K o v á r, 994]. Differences between the approaches of KINFIL and KINFIL 2 are shown in Figs 4 and 5. Model implementation and results The Všeminka experimental catchment (located in the Zlín district) has a heterogeneous land use pattern, and was used for the KINFIL implementation. The shape of the Všeminka catchment is longitudinal, the upper sides of the 62
Analysis of flood events on small river catchments using the KINFIL model Fig. 4. Comparison of ordinates of observed and reconstructed discharges (KINFIL, KINFIL 2) on the Vseminka catchment FLOOD WAVE (6 July 997). Obr. 4. Porovnání prutokových poradnic merených a vypoctených vln (KINFIL, KINFIL 2) na povodí Všeminky. 63
P. Kovár, P. Cudlín, M. Herman, F. Zemek, M. Korytár Fig. 5. Comparison of ordinates of observed and reconstructed discharges (KINFIL, KINFIL 2) on the Vseminka catchment FLOOD WAVE 2 (8 23 July 997) Obr. 5. Porovnání prutokových poradnic merených a vypoctených vln (KINFIL, KINFIL 2) na povodí Všeminky. 64
Analysis of flood events on small river catchments using the KINFIL model central valley are forested down from the water divide, and land use in the other parts is diversified. The physiographic characteristics of the catchment are given in Table, and the species and age structure of the forest, together with the hydrological soil groups, are presented in Tab. 2. These groups were selected according to a widely used method concerning both agricultural and forested lands [US SCS, 986; R i t z e m a, 994]. The land use in the catchment is presented in Fig. ; a more detailed catchment description is given elsewhere [C u d l í n et al, 998]. T a b l e. Basic physiographic characteristics of Všeminka catchment. T a b u l k a. Základní fyzickogeografické charakteristiky povodí Všeminky. Catchment characteristics Catchment area 2.5 km 2 Minimum catchment altitude 27 m Length of river 9.2 km 2 Average catchment altitude 4 m Average river slope 3.6 % Forestation 48.2 % Average catchment slope 9.4 % Catchment perimeter 23.6 km Maximum catchment altitude 62 m Length of riparian vegetation 3.8 km Land use. Arable land 2 ha (9.3 %) 2. Permanent grassland, meadows 52 ha (24.2 %) 3. Public gardens, orchards 7 ha (5.5 %) 4. Urbanized area with vegetation 9 ha (4.3 %) 5. Wild greenery, riparian vegetation 82 ha (8.5 %) 6. Forest 36 ha (48.2 %) The Všeminka catchment was heavily impacted by the significant regional rainfall that led to two flood events in July 997 that affected almost the Central Europe and, in particular, the whole of Moravia and Eastern Bohemia. Besides maps and other available materials [C u d l í n et al, 998], the following data used were for the KINFIL implementation: Hourly rainfall data from the Slušovice raingauge recorder Hourly runoff data from water-stage recorder at the Slušovice outlet Land use data, topographical and physiographical characteristics of the catchment (as shown in Tab. ) Hydrological and soil parameters from maps and soil analysis (as shown in Table 2) Initial soil moisture conditions (SMD, API) were studied using the WBCM- 5 [K o v á r et al., 2] Water Balance Model (as shown in Tab. 3). For our analysis of floods on the Všeminka catchment, flood wave (6. 7.. 7. 97) and flood wave 2 (8. 7. 23. 7. 97)) were selected. Tab. 3 gives basic information about this historical two-fold wave event. 65
P. Kovár, P. Cudlín, M. Herman, F. Zemek, M. Korytár T a b l e 2. Species and age structure of forest; hydrological soil groups Všeminka. T a b u l k a 2. Druhová a veková struktura lesa; hydrologické skupiny pud Všeminka. Species and age structure of forest [ha] Age [years] 4? 4 Coniferous forest 77 75 565 Deciduous forest 7 27 295 Hydrological soil groups [ha] Class of slope [ha] Land use [ha] Class [%] [ ] A B C D Class A B C D I. I. 53 47 92 II. 5 3 II. 46 24 92 III. 5 3 6 2 22 III. 4 66 42 IV. 2 6 27 5 IV. 3 58 V.? 2? 884 5 V. 5 8 5 T a b l e 3. Basic information on floods in July on the Všeminka catchment. T a b u l k a 3. Základní informace o povodni v cervenci 997 na povodí Všeminky. Basic data Flood wave Flood wave 2 Beginning of causal rainfall 6/7/97 6. hr 8/7/97 2. hr End of causal rainfall 8/7/97 22. hr 22/7/97. hr Total depth of causal rainfall 23.3 mm 7.4 mm Total depth of effective rainfall 6.4 mm 53. mm Interception 2. mm 2. mm Cumulative infiltration (incl. surf. retention) 59.9 mm 5.4 mm Soil moisture deficit in upper layer (depth =,3 m).4 /. mm 6.4 /. mm Total saturation content of upper layer 2. mm 2 mm Antecedent precipitation index (beginning/end) 9.8 / 52.7 mm.2 / 7. Average coefficient of satur. hydraulic conductivity K S.22 mm.hr -.22 mm.hr - Average storage suction factor S f 8.5 mm 8.5 mm Time to ponding 2. hr.98 hr Ponding rainfall 4. mm.hr -.6 mm.hr - Peak flow.4 m 3.s - 7.33 m 3.s - Both KINFIL model versions ( and 2) were implemented with the topographic schematization given in Tabs 4 and 5, resp. Figs 2 and 3, illustrating the different approach described above (in the section on Model Structure). The results are given in the form of graphs including ordinates of rainfall, observed and computed (reconstructed) runoff, using both versions of KINFIL (, 2), in Figs 4 and 5. ANUDEM software was used to interpolate the digital elevation model (DEM) from the altitude contour lines (map scale : ). PCI Geomatica software efficiently facilitated determination of the drainage pattern, re-arrangement of the sub-catchment into the cascade of planes, and derivation of the runoff characteristics (slopes, land use, roughness, etc.) particularly for KINFIL 2. 66
Analysis of flood events on small river catchments using the KINFIL model T a b l e 4. KINFIL schematization of the Všeminka catchment (according to Fig.2). T a b u l k a 4. KINFIL schematizace povodí Všeminky (podle obr. 2). Cascade no. Area [km 2 ] Plane no. Area [km 2 ] Average width [km] Length [km] Slope [ ] DP 4.4 DP DP2 DP3.68.47.26 2..8.7.6.75.6.5 DP2 2.4 DP2.8.9.65.2 DP22.96.8.5 DP3.626 DP3.84.8.6.5 DP32.786.75.45 DP4.96 DP4.96.2.8. DP5.69 DP5.69.375.45.5 SP.3 SP.8.45.5.4 (TH = 5 ) SP2.5.5.2 DL.5 DL.45.3.8 DL2.5.5.7.6 DL2.8 DL2.42.7.6.6 DL22.39.65.5 DL3 2.32 DL3 DL32 DL33.725.725.87.45.5.5.6.65.55.45 DL4.9 DL4.595.7.6.85 DL42.595.7.5 DL5 2. DL5.9.6.4.5 DL52.2.8.3 DL6.9 DL6.63.9.7. SL (DIVERG).22 (TH = 3 ) DL62 SL SL2 Discussion and conclusions.27.68.54.25.9.3.9.3.5.3 The aim of this paper was to present the KINFIL model for reconstructing significant rainfall-runoff events and for improving the effective assistance of GIS tools in the KINFIL model for catchment parameter analysis and subsequently for process modelling. A clear visualization of the spatial parameter variability and drainage pattern can also be performed well by interconnecting the hydrological model with ARC/INFO. The quality of the fit of the observed and computed discharge pairs by the KINFIL and KINFIL 2 models is not absolutely convincing, however the Slušovice rain gauge lies outside the catchment and does not provide highly reliable rainfall data. Also, interconnection of the KINFIL short-term event model with the WBCM-5 water balance model is beneficial, namely concerning initial soil moisture conditions (SMD) and antecedent precipitation index values (API). This can easily be done 67
P. Kovár, P. Cudlín, M. Herman, F. Zemek, M. Korytár T a b l e 5. KINFIL 2 schematization of the Všeminka catchment (according to Fig.3). T a b u l k a 5. KINFIL 2 schematizace povodí Všeminky (podle obr. 3). Cascade no. Area [km 2 ] Plane no. Area [km 2 ] Average width [km] Length [km] Slope [ ] Arable Grass Forest Urbanized DP 2.59 DP 2 3.766 DP 3.839 DP 4.57 DP 5.886 DP 6.43 DP 7.58 DL 2.947 DL 22 3.84 DL 23.77 DL 24.68 DL 25.666 DL 26.73 2 3 2 22 3 32 4 42 5 52 53 6 62 7 72 2 22 23 22 222 23 232 233 24 242 243 25 252 26 262.378.982.23 3.3.663.379.46.93.377.2.499.77.286.45.433.47.539.5.258.996.78.64.968.9.42.563.23.32.534.959.24 2.45 2.3.54.67.98.72.8 2.7.5.38.2.54.5.56.4.9.52.32.7.85.7.22.2.5.8.39.2.4.3.9.42.9.33.78.42.7.7.33.46.6.73.34.63.4.267.24.4.222.23.247.64.25.5.56.64.33.86.49.96.78.249.97.74.222.9.25.23.54.223.69.53.53.45.5.76 3 3 4 55 5 78 2 64 2 5 6 3 3 24 6 24 68 9 5 2 49 4 6 56 6 4 2 53 4 2 4 58 2 6 2 46 2 5 69 6 5 6 5 89 45 2 87 45 96 33 99 37 9 29 33 9 32 3 36 96 37 34 88 36 79 47 37 95 28 8 89 6 26 22 68 3 6 7 2 42 3 2 5 45 3 48 42 68 even in an on-line regime in a daily step. Tab. 6 shows the quality of the fit of the KINFIL and KINFIL 2 results expressed in the form of a statistical analysis. The pair-values of the observed and computed discharge ordinates were evaluated by the coefficient of determination RE (i. e., efficiency coefficient) and the coefficient of variation PE (WMO, 992). For the best fit RE?. and PE? (both RE and PE are dimensionless). The observed and computed total runoff volumes were assessed 68
Analysis of flood events on small river catchments using the KINFIL model by the relative ratio expressed in percentage error TVOL [%]. The comparison of the peaks is expressed similarly by the peak errors PEAK [%]. Tab. 6 shows that implementation of the better physically based KINFIL 2 provides more satisfactory results. T a b l e 6. Goodness of fit, the KINFIL, 2 results. T a b u l k a 6. Hodnocení shody merených a vypoctených hydrogramu modelem KINFIL, 2. KINFIL KINFIL 2 VSEMINKA catchment, wave no. WAVE (6.. 7. 97) RE [ ] PE [ ] TVOL [%] PEAK [%] RE [ ] PE [ ] TVOL [%] PEAK [%].55.48 28.3.3.67.4 25.2 5. WAVE 2 (8. 25. 7. 97).48.43 4. 3.5.54.39 24.6 3.5 Although it was established, after summarizing experiences from the flood of 997 in Central Europe, that the huge depth and intensity of the rainfall were the main factors in the formation of the flood situation, the decrease in arable land in favour of permanent grassland or mixed forest had distinctly reduced direct runoff. Our scenario model analyses provide evidence of this. Good and diversified land use (with a high percentage of forest and permanent grassland) can have a positive influence on direct runoff formation to increase infiltration and thus the retention capacity of a catchment. The KINFIL modelling technique has demonstrated good applicability and its successful use is confirmed. Acknowledgement. The authors fully acknowledge the financial support from the Grant No. 75/3/4399-CEZ8 Possibilities to increase an ecological stability, retention and accumulation capacities of water in the landscape. List of symbols CN - curve number [- ], SMD - soil moisture deficit [mm], API 3 - antecedent precipitation index 3 days before event [mm], S f - storage suction factor [mm], K S - saturated hydraulic conductivity [mm.hr - ], RE - determination coefficient [- ], PE - variation coefficient [- ], TVOL - runoff volume error [%], PEAK - maximum discharge error [%]. 69
P. Kovár, P. Cudlín, M. Herman, F. Zemek, M. Korytár REFERENCES [] CHOW, W. T. - MAIDMENT, D. R. - MAYS, L. W. : Applied hydrology. USA : McGraw-Hill, 998, 572 pp. [2] CUDLÍN, P. - KOVÁR, P. - ZEMEK, F. - ŠERÁ, B. - HERMAN, M. - KANTOR, P.: Príciny snížení vodohospodárské funkce lesa v krajine. [Zpráva o rešení úkolu VaV 6/2/98, DÚ /7.] Ceské Budejovice : UEK, 998, 54 str. [3] HLADNÝ, J. (koordinátor): Vyhodnocení povodnové situace v cervenci 997. [Souhrnná zpráva.] MŽP CR, 998, 63 str. [4] INSTITUTE OF HYDROLOGY: Flood Studies Report No. 49, Methods of Flood Estimation, Wallingford-Oxon, U.K., 978. [5] KOVÁR, P.: Možnosti urcování návrhových prutoku na malých povodích modelem KINFIL. Vodohosp. Cas., 4, 992, 2, str. 97-22. [6] KOVÁR, P.: Využití hydrologických modelu pro urcování maximálních prutoku na malých povodích. [Doktorská disertace.] Praha : LF CZU, 994. [7] KOVAR, P. - SVITAK, M.: A Possibility of Modelling Direct Runoff on Small Catchments using GIS. Zeitschrift für Kulturtechnik und Landentwicklung, 35, 994, pp. 374-38. [8] KOVAR, P. - PECH, P. : Modelling Floods with Respect of Changes in Land Use on Small Catchments. Proceedings INTERPRAEVENT, 96, 996, Band 4, pp. 9-, GaPa, SRN. [9] KOVÁR, P. - CUDLÍN, P. - KORYTÁR, M. - ZEMEK, F. - HERMAN, M.: Comparative study of water balance on the experimental catchments Všeminka and Drevnice. Plant Production (Rostlinná výroba), 47, 2, 6, pp. 26-266. [] MOREL-SEYTOUX, H. J. - VERDIN, J. P.: Extension of the Soil Conservation Service. Rainfall-Runoff Methodology for Ungauged Watersheds - Colorado, USA, 98. [] RITZEMA, H. P.: Drainage Principles and Applications. ILRI Publication 6, 2 nd edition, Wageningen, The Netherlands, 994, 94 pp. [2] US SCS: Urban hydrology for small watersheds. U. S. Soil Conservation. Technical Release, 55, 986, 3, Service, Washington. [3] WMO: Simulated real-time intercomparison of hydrological models. WMO No. 779, Operational Hydrology, 992, No. 38, Geneva. Received 2 September 2 Review accepted 26 November 2 ANALÝZA POVODNOVÝCH UDÁLOSTÍ NA MALÝCH POVODÍCH S VYUŽITÍM MODELU KINFIL Pavel K o v á r, Pavel C u d l í n, Michal H e r m a n, František Z e m e k, Michal K o r y t á r Zámerem príspevku je informovat o možnostech využití GIS pri fragmentaci malých povodí za úcelem zpresnení vstupních dat pro hydrologický model KINFIL. Posoudili jsme vlivy tradicní a nove vytvorené fragmentace povodí na verohodnost modelu overenou skutecnými mereními. Kombinace GIS a KINFIL, který je fyzikálne založen na teorii infiltrace a transformace prímého odtoku kinematickou vlnou, poskytuje nástroj pro analýzu jak historických srážko-odtokových prípadu, tak hypotetických scénárových simulací. 7
Analysis of flood events on small river catchments using the KINFIL model Úloha srovnání dvou prístupu fragmentace povodí byla aplikována na experimentálním povodí Všeminky na východní Morave. Byly tedy testovány dve modelové verze: KINFIL a KINFIL 2. Infiltracní cást modelu je v obou verzích stejná. Starší prístup segmentace KINFIL je založen na schematické geometrizaci plochy povodí. Predpokládá, že povodí je V-tvaru, do hlavního toku je sváden prímý odtok zobou stran, z nekolika vedle sebe ležících desek povodí. Takto definované desky ale nemusí plne respektovat clenení na subpovodí. Naproti tomu nový prístup KINFIL 2, který za úcelem fragmentace povodí využívá prostredku GIS, plne respektuje prubeh rícní síte povodí a její clenení na subpovodí. Základem pro získání požadovaných parametru k naplnení KINFIL je digitální model reliéfu terénu (DEM). Z nej jsou postupne vypocteny následující charakteristiky:. soustredený odtok vody zde nabývá každý pixel povodí hodnoty rovné poctu pixelu do nej stékajících ; 2. potenciální odtoková sít odvozena ze soustredeného odtoku; 3. výpocet subpovodí ve forme kaskády ploch a výpocty základních charakteristik (svažitost, kategorie využití území, drsnost povrchu, prekážky, atd.) nutných k naplnení modelu KINFIL 2. Pro všechny výše popsané procedury byl použit software PCI Geomatica. Model KINFIL je založen na kombinaci teorie infiltrace a transformace prímého odtoku kinematickou vlnou ; osvedcil se na rade experimentálních povodí pri rekonstrukci historických povodnových prípadu. Používá fyzikálne-geometrické, hydraulické a klimatické parametry povodí a je prednostne urcen pro stanovení návrhových prutoku pro ruzné scénárové situace. Pro analýzu povodne na Všemince byly použity dve povodnové vlny: první v dobe 6. 7. 7. 97 a druhá v dobe 8. 7. 23. 7. 97. Verze modelu KINFIL 2 více odpovídá skutecným geografickým, ale i fyzikálním pomerum povodí a také výsledky simulace dvou povodnových vln z cervence 997 se více približují skutecne namereným hodnotám. I když je vhodné otestovat model pro další povodnové vlny a nejlépe na dalších povodích, ukazuje se, že verze KINFIL 2, podporená topografickou analýzou provedenou prostredky GIS, je vhodným prostredkem pro ohodnocení odtokových pomeru povodí pri simulaci ruzných scénáru využití povodí. Seznam symbolu CN - císlo odtokové krivky [-], SMD - deficit pudní vlhkosti [mm], API 3 - index predchozí srážky 3 dní pred událostí [mm], S f - retencní soucinitel sacího tlaku [mm], K S - soucinitel nasycené hydraulické vodivosti [mm.hr - ], RE - koeficient determinace [-], PE - koeficient variace [-], TVOL - relativní chyba objemu odtoku [%], PEAK - relativní chyba v kumulacním prutoku [%]. 7