7. Results of streamflow modeling
Continuous runoff modeling with daily temporal resolution was carried
out for the Suntar river basin for the period 1957-1964, using input
meteorological data from four meteorological stations (Suntar-Khayata,
Nizhnyaya Baza, Vostochnaya and Agayakan); and for the period 1966-2012
using the data from two stations (Vostochnaya and Agayakan). Water
balance components distribution for these periods are presented in Table
5, and the comparison of observed and calculated streamflow hydrographs
– in Fig. 5, Fig. 6.
Calculated mean annual precipitation for the Suntar river basin is 344
mm for the 1957-1964 period. Estimated streamflow is 199 mm, which is
10% higher than the observed value (180 mm). Estimated ET from the
whole basin equaled 143 mm.
Maximum precipitation and streamflow annual values for the entire
simulation period reached 486 and 348 mm in 1959, while minimum values
were 259 mm in 1958 and 136 mm in 1963 for the Suntar River basin. The
coefficient of variation of annual streamflow is 0.30.
Here we compare the water balance distribution with other research basin
in the region, the KWBS station, the Kontakovy creek watershed (area
21.3 km2, average altitude 1070 m). The value of ET
for the KWBS is assessed within the range from 114 to 137 mm (Lebedeva
et al., 2017; Zhuravin, 2004). Mean annual precipitation and streamflow
reached 420 and 280 mm for the period 1948–1997.
The average and median Nash-Sutcliffe efficiency (NS) for the Suntar
River amounted to 0.75 in 1957-1964. The same value for the period of
1966-2012 is lower (average 0.58, median 0.67 with maximum and minimum
values of 0.88 and -0.90, respectively). We attribute this decrease of
efficiency to the lack of meteorological data in the second period.
Overall, despite some overestimation of streamflow, the calculated
streamflow hydrographs match the observed ones quite satisfactorily,
both in phases and absolute discharge values. Overestimation of
simulated streamflow during spring freshet may be associated with the
spread of underchannel taliks. In spring dry alluvial deposits in the
river channels are filled with snowmelt water and delay the start of
freshet and decrease its magnitude (Grave et al., 1964). Such phenomena
are also described by Mikhaylov (2013).
We also compared simulated and observed maximum discharges. The maximum
simulated and observed discharges were 1200 and 1659 respectively during
1957-1964 and 1905 and 1910 during 1966-2012 for the Suntar River basin
(Table).
Based on the simulation results, the contribution of each runoff
formation complex (RFC) into total streamflow of the Suntar River was
evaluated (Table 6, Fig. 7). Goltsy complex that covers only 7% of the
basin provides 20% of the total streamflow, and the runoff coefficient
reaches 0.91. Tundra is the largest contributor to the runoff formation
at the Suntar river catchment – 49% of the total runoff, with a runoff
coefficient of 0.74. The total streamflow from the taiga and sparse
forest landscapes, which take 56% of the territory, is about 31%. The
contribution of the goltsy landscape increases in dry years and may
reach up to 28% (for example, in 1963 the total annual streamflow was
only 130 mm, while the streamflow from the goltsy complex was simulated
as 513 mm).