af
al
International Archives of the Photogrammetry, Remote Sensing and
zone basis relations were used in distributed hydrological
models such as the snowmelt runoff model (Martinec et al..
1994). The snow depletion curves are watershed specific. in
that they represent the characteristic response of the watershed
to snowmelt, which is a function of the land cover and
elevation. characteristics within the basin. Hall and Martinec
(1985) proposed an approximation to compute snow cover
depletion curves and estimate daily snow cover variation from
Landsat MSS scenes. This procedure seems to fit well for some
alpine catchments when the depletion curves are computed by
làking the entire basin as a single unit. When the snowmelt
depletion curves are derived on an elevation zone basis, the
above approximation may not be suitable for some catchments.
This is evident in particular for the lower elevation zones of the
catchment, where the snow cover disappears rapidly.
The show covered arca on cach day during the snowmelt period
was estimated for cach elevation zone, using either second
order or third order polynomial {it employing a threshold to the
snow cover data of cach elevation zone, estimated from remote
sensing satellite data. Interestingly the equations developed
gave à high coefTicient of determination. The daily snow cover
was computed using the above equations for each elevation
zone and expressed as a percentage of the total area in order to
be input into the model. Figure 9 presents an example of the
modified depletion curves for three elevation zones of
Cordevole river basin indicating the distribution of daily snow
cover areal extent on each day of melt period. However, these
snow depletion. values do not reflect short-term changes
resulting from snowfalls during the snowmelt period. Weekly
satellite images with high spatial resolution may be more useful
for precise estimation of daily snow cover and the resulting
depletion curves.
second nregr pitynedicul cgg2bpns
1 ud Cm EAS EN ARTs EIA LDA RD aS ToU un
a, end E 454 188 (t i ‘
how ; SILKE soup agi
: A 1 y à PU (d i i = F pod
11 >
T ^.
- EN
e.
n
= ^".
M. |
*
^w
to bn Zone-I
^
ET
b M. A:
i ».
~
Ki ^. |
Zune-2 ~~
^» =,
frm C. .
^. 7
^ a Se
N a.
\ "
3 V N
p
y i
u 0 Xe j
3
^ s ko Nas al odas ^
\ > É
TS axe As ere Nn
ü Y, 41 f st nx 126
Figure 9 Modified depletion curves tor three elevation zones of
Cordevele river basin
7. HYDROLOGICAL MODEL
The main aim of hydrological modeling is to provide a forecast
of the future performance of a hydrological system. The
snowmelt runoff model should be able to simulate the
contribution owing to the snow depletion in the basin to the
total river discharge, day by day, during the melting scason.
snowmelt runoff simulation models generally consist of a
snowmelt model and a transformation modcl. À snow model is
1225
Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
defined as a mathematical description of melt related processes
which gives the flux of melt water at the bottom of the
snowpack as output. The snowmelt and transformation model
can be lumped or distributed in nature. Lumped models usc one
scl of parameter values to define the physical and hydrological
characteristics of a watershed. Distributed models attempt (o
account for the spatial variability by dividing the basin into
sub-areas and computing snowmelt runoff for each sub-arca
independently, with a set of parameters corresponding to cach
of the sub-arcas. The distributed representation used in most
snowmelt runoff simulation models is the separation of the
watershed into distinct clevation zonc. These clevation zoncs
are appropriate for alpine regions, where snowmelt and
temperature are strongly related to elevation.
During the inter-comparison of various snowmelt runoff
models conducted by the World Meteorological Organization
(1986), of the models tested, the snowmelt runofT model
(Martinec et al... 1983) exploited the increasing. availability of
snow cover mapping from satellites. This model is a
deterministic distributed temperature index model that takes
into consideration the precipitation and air temperature, along
with all other predetermined catchment parameters. An
advantage of this model is that it is easy to usc operationally
because of a limited amount of data is required for thc forecast.
usually precipitation and temperature
The Snow Runoff Model (SRM) considered for three elevation
zones tn the present study reads as:
{ : i T
(1 ii" I : ut E AJ Ic ^ iH L / "A
Mo gii
r pr uq aiid
Fi Fan = AT Hy ^ + € rue del
Es SON
i FE
cde fa EA S ( p] |
Sh d |
(| — À à 4 (J Ka
where is the average daily discharge (m' s). C, is the runoff
coefficient, with C, referring to snowmelt and C,, to rain. a, is
the degree day factor (em "C" das!) T, is the number of
degree days above the base of 0" C, AT, is thc adjustment by
temperature lapse rate tor different altitudes of meteorological
stations, S, is the ratio of snow-covered arca to the total area
P, is the precipitation contributing to the runoff, K, is thc
rcccssion coefficient derived from historical discharges, n is the
index referring to thc sequence of days, A is the arca of the
basin (m^), a, b. c refer to elevation zones 1, 2, 3 respectively,
and 1027/86400 covcrts cm m? day! to m s^
Using the above model daily discharges have been computed
for the two basins for the snowmelt period of April to July. The
pre-determined morphological parameters and hydrological
paramcters and daily snow cover area estimated from three
elevation zones from satellite remote sensing data has been
utilized in computing the discharges. Figure 10 presents the
distribution of measured and simulated discharges for La Vizza
basin for the period of 122 days starting from 1 April 1984.
Visual comparison of thc discharges indicates there is a goad
correlation between measured and simulated discharges,
7.1 Model performance
To analyze the performance of the model, linear regression
analyses has been made and correlation between. measured and
simulated discharges has been determined. The correlation
