Geoscience Reference
In-Depth Information
Figure B3.3.2
Variation in average degree-day factor, F, over the melt season used in discharge pre-
dictions in three large basins: the Dischma in Switzerland (43.3 km
2
, 1668-3146m elevation range);
the Dinwoody in Wyoming, USA (228 km
2
, 1981-4202m elevation range); and the Durance in France
(2170 km
2
, 786-4105m elevation range) (after Rango, 1995, with kind permission of Water Resource
Publications).
Where a catchment covers a wide range of elevations, then it is usual to divide the area up
into a number of elevation bands. Rango (1995) reports using bands of 500m in the Snowmelt
Runoff Model (SRM). Air temperatures may then be adjusted for elevation using a simple lapse
rate multiplied by the elevation difference from the temperature station so that for band k
T
k
=
T
s
+
(
E
k
−
E
s
)
/
100 (B3.3.4)
where
is the lapse rate [K per 1000 m],
T
s
is the temperature at the recording station,
E
k
is the
mean elevation of band
k
, and
E
s
is the elevation of the recording station. A typical lapse rate
used would be of the order of 6.5 K/1000m. A further important adjustment is to account for
the change in snow-covered area within each elevation band. This can be done either by using
standardised depletion curves for the changing average water equivalent of the pack in each
elevation band (Figure B3.3.3) or by updating the current snow-covered area using remote
sensing (Rango, 1995).
The SRM Snowmelt Runoff Model is a particularly interesting implementation of the degree-
day method because of the way in which it also makes use of satellite imaging to determine
snow-covered areas in predicting snowmelt runoff in large catchments (see Rango (1995) for
more detail).
In some cases, shorter than daily time steps are required for runoff prediction. Rango and
Martinec (1995) report a test of a modification to the degree-day method for use in the SRM
model that incorporates a radiation component such that:
M
=
F
∗
max(0
, T
−
T
F
)
+
R
n
/
m
(B3.3.5)
where
M
is now an hourly melt water equivalent,
F
∗
is now a new degree-day coefficient,
R
n
is the net radiation and
m
is the latent heat of melting. This modification is clearly more
demanding in terms of data but can significantly improve predictions of snowmelt rates (see
also Hock, 2003). Net radiation, in particular, may not be measured directly and may need
