Vízügyi Közlemények, 1984 (66. évfolyam)

3. füzet - Kovács György: Az átlagos lefolyás meghatározása a folytonos mezők elvének alkalmazásával

382 Kovács György The determination of average runoff values by use of continuum approach by dr. Gy. KOVÁCS By analyzing the multiannual average runoff values based on observations (in 4 cases on estimates) of 74 gauging stations it has been proved that the use of the continuum approach was justified because these parameters are the cumulative representations of surface characteristics fluctuating in a random way over the subcatchment located upstream of the station in question. Based on this concept the areal variability of the specific runoff values could be defined. In order to quantify the increasing uncertainties emerging parallel to a decrease in the area, the expected variance was determined as a function of the catchment area. To help the solution of practical problems the base-values of multiannual average runoff of 12 catchments covering practically the entire area of Transdanubia - except The Little Plains and Mezőföld characterized by extremely small slopes and zero natural runoff - could be determined. These specific values were usable for planning because the expected mean discharges could be obtained in any cross-section by multiplying the base-value of specific runoff and the area pertaining to the catchment in question. By completing detailed calculations it was possible to supply the planners by numerical results about uncertainties arising from the variability of this surface of random nature. For any cross-section to which a catchment larger than 1000 km

2 belonged the mean discharge calculated according to this method could be accepted for desing without any precondition. In this range the variance was 7 to 8 percent compared to the mean of the actual data. It is not probable that this deviation will exceed 15 percent, an error still accepted in hydrological investigations. If a smaller catchment was analyzed the expected mean did not change but uncertainty has increased: the variance was 12 percent, 16 percent and beyond 30 percent when the area was decreased to 400 km

2, 100-400 km

2 or below 100 km

2, respectively. Extreme deviations may amount to double variances and in such cases the maximum error committed in the above mentioned categories could increase to 24, 32 or 60 percent and in catchments with an extension of around 25 km

2 it may jump to 100 percent. It is obvious that in catchments smaller than 400...500 km

2 the magnitude of the possible errors may exceed the limits still thought as acceptable in hydrology. Therefore, it is indispensable to set up a model for the assessment of deterministic relationships between the most characteristic parameters of an area precipitation, sloping conditions, the ratio of land use categories, etc. and the differences of actual and average runoff with regard on the dynamic nature of the catchment surface. The analysis of numerous available data allowed to estimate not only the expected value of multiannual specific runoff but also the temporal variablitiy of consecutive individual annual values. It has been proved that the distribution of annual discharges could be fairly well approximated by a two-parameter gamma-function. Parameter к of this function took the values 5,12 and 7 in the northern part of Transdanubia, in the middle zone and in the south, respectively. An additional result of our investigations was that the areal variations inside a zone did not show up substantial differences. As a consequence, the generation of artificial time-series for the simulation of the actual conditions could be executed quite easily by using the above mentioned average parameters and the values of multiannual runoff that are characteristic for the particular catchment. Although, multiannual average runoff values and their yearly variations are important information for a planner - this has become obvious from the aforementioned investigations - the determination of a series of different other parameters is also needed if comlpete information is required about a cross-section without having regular measurements. By determinig the effect of the variability of the ground surface on variations in runoff and by explaining the phenomenon on the basis of the concept of continuous fields it might be possible to set up similar procedures for the calculation of other hydrological characteristics as well. * * *