Академическое издательство «Гео»

Keywords: streamflow formation, stochastic modeling, forecasting the inflow into the reservoir, ensemble forecasting.



GEOGRAPHY AND NATURAL RESOURCES, 2016, № 2, p. 175-182

Forecasting the water inflow into the Krasnoyarsk and Sayano-Shushenskoe reservoirs in the second quarter of the year

D.A. Burakov*, I.N. Gordeev**, A.V. Ignatov***, O.E. Petkun*, L.A. Putintsev**, and A.A. Chekmarev***

* daburakov@yandex.ru, petkun.oksana@yandex.ru
** biggin@sibmail.com, leo77789@mail.ru
*** ignatov@irigs.irk.ru, cvv32@yandex.ru

We consider the various methods of constructing models intended to forecast the average water inflow, in the second quarter of the year, into two reservoirs on the Yenisei river. To solve modeling problems used a new computer technology implemented in the specialized “Stochastic Modeling” software package. Independent data were employed to verify the variants of the models for the formation of variability in quarterly inflow as generated based on different algorithms. A more sophisticated and robust model for forecasting the inflow was constructed as an ensemble of partial models. Based on aggregate results of modeling, we suggest the method of constructing a forecast of the average (for the second quarter) lateral inflow into the Krasnoyarsk reservoir and the inflow into the Sayano-Shushenskoe reservoir by use of observational data accumulated by Srednesibirskoe UGMS (Weather Control and Environmental Monitoring Service), based on an ensemble of partial models. It is established that such an operation reduces the probability of forecasting errors implying an arbitrary selection of models. We constructed forecasts of the aforementioned characteristics using real-time data for 2015. It is stated that the solution of the forecasting problem can be facilitated by using additional information.

Keywords: streamflow formation, stochastic modeling, forecasting the inflow into the reservoir, ensemble forecasting.
главная

www.izdatgeo.ru