Data and experiments

Abegg and Froesch (1994) demonstrated that for Switzerland a continuous snowpack cannot be guaranteed anymore below the critical altitude of 1200 m a.s.l. Therefore, we selected two high-altitude stations representing locations above and below this threshold. Data for the winter periods of Disentis (1190 m a.s.l., 8°51'E; 46°42'N) and Santis (2490m a.s.l., 9°20'E; 47°15'N) were compiled from the Swiss Meteorological Service (Figure 19.1; Bantle 1989). For both stations, a snow-poor (1988-1989) and a snow-rich winter (1998-1999) were selected: at Disentis, from December 1, 1988 to March 31, 1989 (hereafter called D88-89) and from October 20, 1998 to April 30, 1999 (D98-99) and for Santis, from September 1, 1988 to July 31, 1989 (S88-89) and from September 1, 1998 to August 31, 1999 (S98-99).

In order to infer the impact of climate change at these two stations, three temperature change scenarios are devised (Table 19.1): (1) a 2 x CO2 simulation done with the ECHAM1/LSG-AOGCM-model downscaled statistically over Switzerland (Gyalistras etal. 1998), (2) a scenario allowing an increase of Tmin only, inferred from an analysis of temperature changes in the twentieth where EL is the liquid and EF the solid part of the soil moisture. TfK is the freezing point.

Precipitation is considered as solid if Tair is less than that of the triple point of water. Liquid precipitation on a snowpack induces snowmelt. Melted snow (MS/dt < 0 in Equation (19.4)) goes directly into the soil as liquid moisture, latent heat is released, and energy is transferred to the surface. The surface energy budget is computed

Figure 19.1 The situation of the two study sites in Switzerland
Table 19.1 Temperature increase scenarios used in this study based on the observed values (in °C). DJF stands for December to February, MAM for March to May, JJA for June to August, SON for September to November. Data taken from Gyalistras et al. (1998) and Weber et al. (1997)




















century in the mountain regions of central Europe (Weber etal. 1997), and (3) a scenario of Weber etal. (1997) considering an increase of Tmax separately. These three scenarios involve seasonal temperature changes. The two latter scenarios are projected in the future to simulate a possible temperature change for the period 2080-2100. To infer the minimum and maximum scenarios, anomalies of the observed air temperatures for both years and both stations are calculated with respect to a baseline period (1978-1998). The resulting positive (Tmax) and negative (Tmin) anomalies were scaled by a factor so as to obtain the required average seasonal increase rate (Figure 19.2). All other parameters are left unchanged.

For the estimation of the snow duration, a threshold value of 2.5 cm for Disentis and 5.0 cm for Santis are applied. These thresholds were deduced from several tests about the constancy of snow cover duration. Only those days were counted where the observed and simulated snow depth exceeded these thresholds and did not fall back to zero again until the end of the investigated winter seasons.

The model's internal parameters and initial conditions of soil moisture and surface temperature were tuned for each station and year separately in comparison with the observed snow depth (Table 19.2). The following were fixed for all simulations: the surface albedo without snow to 0.285, the soil porosity to 45%, the soil drainage relaxation coefficient to 4 days, the soil hydrological depth to 500 mm, the surface emissivity to 1, the initial snow depth to 0, the surface roughness, z0>sfc, to 0.001, the snow roughness height, z0>snow, to 0.5 x 10-4 m and the snow masking depth to 0.07 m. There was no initial frozen soil moisture content, WF, and the liquid soil water, WL, was fixed at 45 kg m-3.

The energy fluxes are analyzed according to the diagnostic formulation of Oke (1987) to calculate the contribution of each process to melt a volume of snow:

Here, AQm is the latent heat storage change due to melting or freezing and AQs represents the convergence or divergence of sensible heat fluxes within a pack; this term includes internal energy gains or losses due to variations of radiation and heat conduction. The radiation fluxes are analyzed in terms of daily and monthly

-Ctrl -First -Second -Third

A r




VA / n\


\ i

i JZ/~ v—


T Min


Figure 19.2 Illustration of the temperature increase scenarios. The first scenario is an increase in the mean, the second, in the minimum, and the third, in the maximum temperature

Table 19.2 Model parameters and initial conditions at D88-89, D98-99 and S88-89, S98-99. The asymptotic value of snow albedo stands for the value of the albedo after snow has aged. The duration of the snow-ageing process is determined by the snow-ageing coefficient. The extinction factor of the snowpack determines the quantity of radiation, which penetrates the pack

D88-89 D98-99 S88-89 S98-99 Units

Psnowif Msnow < 10 kg m 2


Psnow if Msnow > 10 kgm-2






Asymptotic snow a


Snow-aging coefficient


Radiation extinction factor


of the snowpack of the snowpack

60 220 220 kgm-3

225 270 220 kgm-3

10 5 6 day

900 800 700

IT 600

S 500

1 400

W 300 200 100 0

46925803697025 4 5 6 8 9 0 2 3 4 5 2 3 4 2222233333

t-^fi^cn(MLooot-^fcDcn(MLooo 7 8 9 0 2 3 4 6 7 8 9 1 2 3

Figure 19.3 Observed (obs) and simulated (sim) snowpacks at Disentis and at Santis for the 1988-1989 and 1998-1999 periods. Note the different scales used

46925803697025 4 5 6 8 9 0 2 3 4 5 2 3 4 2222233333

t-^fi^cn(MLooot-^fcDcn(MLooo 7 8 9 0 2 3 4 6 7 8 9 1 2 3

Days of year

Figure 19.3 Observed (obs) and simulated (sim) snowpacks at Disentis and at Santis for the 1988-1989 and 1998-1999 periods. Note the different scales used averages concentrating on the melting period starting at the maximum snow accumulation.


0 0

Post a comment