General review of the study

The objective
The main objective of this study is to investigate possible links between teleconnection patterns with Mongolian rivers runoff and to select more predictable predictors and explore the method of the long range runoff forecasting for selected rivers of Mongolia for the different users by using teleconnection patterns of the large-scale interaction of ocean and atmosphere.

Used data and method
The data used in this study consist of monthly flows for May– September and spring and summer flood runoff in 37 major rivers of Mongolia, recorded at 37 stations by the Institute of Meteorology and Hydrology since 1950 and some possible data on the water equivalent of snow mantle for selected river basin such as Baruunturuun, Tes, Khalkh gol and Tsagaanturuut .
Statistical values of snow coverage of Khovd river between the period 1972 to 1982 calculated from Meteor-30 and 31 satellite channel’s with 1.2 km resolution. These data are supplemented by the snow coverage map of February and March 2003-2005 by NOAA KLM satellites.
Indexes of eleven prominent teleconnection patterns for 1950-2004 identified in the Northern Hemisphere extratropics, El Nino-Southern Oscillation (ENSO) index and sea surface temperature (SST) anomalies averaged over the 4 regions are extracte from Database of the Climate Prediction Center of USA [http://www.cpc.ncep.noaa.gov].

For any forecasting problem, there are two frequently used forecasting approaches: the linear regression and the discriminant prediction. The regression approach, with single or multiple predictors, uses the best-fit relationship to estimate the predictand, while the discriminant approach estimates the probabilities that the predictand will fall into a set of prescribed categories. We are used a bi-linear regression method which is the oldest, and by far the most widely used technique for the runoff forecasting. The accuracy of the relation is given by the correlation coefficient.

Besides this traditional methods and approaches such as empiric, graphic, analytic and comparative methods in data analysis have been used. The analytical and cartography procedures of the study employed for image processing and mapping software ILWIS and Arc View

The novelty
However, Mongolia is located far from any ocean and surrounded by the high central Asian mountain ranges this study provides evidence of a statistically-significant link between some teleconnection patterns of large scale interaction of ocean and atmosphere to the runoff of some region`s rivers of Mongolia. It is one of the main novelty of this study. Consequently, for some rivers of Mongolia long range forecasting equations were drawn for July-August monthly discharge and summer flood with lead time 2-3 months.

The results of the study have shown the possibility of the prediction of the river spring flood by using remote sensing data with sufficient accuracy and lead time 1-1.5 month. Also one of the particularity of this study is a creation of small database on snow coverage area data derived from satellite images for the river catchments dominating snow melt flood.