215
Modeling Snowmelt Runoff under Climate Change Scenarios
modeling is a crucial element in any attempt to predict runoff from snow-dominated
areas (Hock 2003).
10.2.1 coMPaRiSon of teMPeRatuRe index and eneRgy
Balance Modeling aPPRoacheS
Melt models developed to simulate the accumulation and melt of snowpack can be
broadly categorized as either temperature index also known as degree-day models or energy balance models. All the operational runoff models use one of these
two approaches for modeling snowmelt. According to Rango and Martinec
(1994), most of the operational runoff models reported in the literature employ
the degree-day approach. These include the Streamflow Synthesis and Reservoir
Regulation (SSARR) model (U.S. Army Corps of Engineers 1975), tank model
(Sugawara et al. 1984), University of British Columbia watershed (UBC) model
(Quick and Pipes 1977), SRM (Martinec and Rango 1986), Hydrologiska Byråns
Vattenbalansavdelning model (Bergstrom 1975), empirical regressive model (Turcan
1981), and HyMet model (Tangborn 1984). Some employ the energy-based approach,
including the energy and mass-balance model (ISNOBAL; Marks et al. 1999), energy
balance model for snow and soil (SYNTHERM; Jordan 1991), simultaneous heat and
water model (Flerchinger and Saxton 1989), Système Hydrologique Européen model
(Morris 1982), and Utah energy balance model (Tarboton et al. 1995).
The degree-day or temperature index approach has been in use for over 75 years
(Collins 1934); Hock (2003) reported that Finsterwalder and Schunk (1887) first used
empirical relationships between air temperatures and melt rates for an Alpine glacier
as the predecessor of temperature index models. Since then, many researchers have
employed the simplicity of the degree-day melt model in simulating snowmelt and
also for operational SRMs.
The degree-day approach involves computing the daily snowmelt depth by multiplying the number of degree-days by the degree-day factor (Kustas et al. 1994). This
modeling approach can be used over large areas with limited data input requirements
yet can provide realistic simulations of discharge (Brubaker et al. 1996). However,
the degree-day method only works under conditions where the energy input into the
snow cover can be easily predicted by the temperature or where there is a well-defined
relationship between the energy input into the snow and the air temperature (Garen
and Marks 2005). It has been shown to work poorly under conditions lacking a good
relationship between the temperature and the energy input into the snowpack such as
rain on snow. Hock (2003) also pointed out that the basic degree-day approach does
not account for topographical effects such as slope, aspect, and shading, which are
common with complex mountains. Nevertheless, due to their good performance, low
data requirements, and simplicity, Hock (2003) argued that temperature index models will retain their leading position in snowmelt modeling in the future.
In contrast, the energy balance melt approach, as illustrated in Figure 10.1, is a
more physically based type of model, enabling it to account directly for many of the
physical processes that affect snowmelt (Kustas et al. 1994). Incorporating the physical processes involved in snowmelt increases the data input requirements needed to
run these models. Previous studies emphasized that there is the scarcity of input data
Modeling Snowmelt Runoff under Climate Change Scenarios
modeling is a crucial element in any attempt to predict runoff from snow-dominated
areas (Hock 2003).
10.2.1 coMPaRiSon of teMPeRatuRe index and eneRgy
Balance Modeling aPPRoacheS
Melt models developed to simulate the accumulation and melt of snowpack can be
broadly categorized as either temperature index also known as degree-day models or energy balance models. All the operational runoff models use one of these
two approaches for modeling snowmelt. According to Rango and Martinec
(1994), most of the operational runoff models reported in the literature employ
the degree-day approach. These include the Streamflow Synthesis and Reservoir
Regulation (SSARR) model (U.S. Army Corps of Engineers 1975), tank model
(Sugawara et al. 1984), University of British Columbia watershed (UBC) model
(Quick and Pipes 1977), SRM (Martinec and Rango 1986), Hydrologiska Byråns
Vattenbalansavdelning model (Bergstrom 1975), empirical regressive model (Turcan
1981), and HyMet model (Tangborn 1984). Some employ the energy-based approach,
including the energy and mass-balance model (ISNOBAL; Marks et al. 1999), energy
balance model for snow and soil (SYNTHERM; Jordan 1991), simultaneous heat and
water model (Flerchinger and Saxton 1989), Système Hydrologique Européen model
(Morris 1982), and Utah energy balance model (Tarboton et al. 1995).
The degree-day or temperature index approach has been in use for over 75 years
(Collins 1934); Hock (2003) reported that Finsterwalder and Schunk (1887) first used
empirical relationships between air temperatures and melt rates for an Alpine glacier
as the predecessor of temperature index models. Since then, many researchers have
employed the simplicity of the degree-day melt model in simulating snowmelt and
also for operational SRMs.
The degree-day approach involves computing the daily snowmelt depth by multiplying the number of degree-days by the degree-day factor (Kustas et al. 1994). This
modeling approach can be used over large areas with limited data input requirements
yet can provide realistic simulations of discharge (Brubaker et al. 1996). However,
the degree-day method only works under conditions where the energy input into the
snow cover can be easily predicted by the temperature or where there is a well-defined
relationship between the energy input into the snow and the air temperature (Garen
and Marks 2005). It has been shown to work poorly under conditions lacking a good
relationship between the temperature and the energy input into the snowpack such as
rain on snow. Hock (2003) also pointed out that the basic degree-day approach does
not account for topographical effects such as slope, aspect, and shading, which are
common with complex mountains. Nevertheless, due to their good performance, low
data requirements, and simplicity, Hock (2003) argued that temperature index models will retain their leading position in snowmelt modeling in the future.
In contrast, the energy balance melt approach, as illustrated in Figure 10.1, is a
more physically based type of model, enabling it to account directly for many of the
physical processes that affect snowmelt (Kustas et al. 1994). Incorporating the physical processes involved in snowmelt increases the data input requirements needed to
run these models. Previous studies emphasized that there is the scarcity of input data