Snowcover variability

Snow cover has been established as an important component of the land-atmosphere climate system. Snow cover experiences the largest fluctuations, both spatially and temporally, of all varying surface conditions. GCMs need to correctly simulate the i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | i i i ( r i i i i i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | i i i ( r i i i i

I i i ' i i i i i i I i i i i t i i i i I i I i i i i i i i i i I i i i i i t i i i

30 40 50 60 70 80 90

Latitude

Figure 4.14. Surface air temperature change in 15 coupled GCMs that have doubled atmospheric CO2 concentrations. All models show "polar amplification" and enhanced warming in the Arctic compared to tropical latitudes. Bony et al. (2006). (Plate 4.14.)

I i i ' i i i i i i I i i i i t i i i i I i I i i i i i i i i i I i i i i i t i i i

30 40 50 60 70 80 90

Latitude

Figure 4.14. Surface air temperature change in 15 coupled GCMs that have doubled atmospheric CO2 concentrations. All models show "polar amplification" and enhanced warming in the Arctic compared to tropical latitudes. Bony et al. (2006). (Plate 4.14.)

natural variability of snow cover from seasonal time scales out to interannual or even decadal time scales. GCMs also need to correctly parameterize the influence of snow cover on the climate. Several multi-GCM studies have evaluated how well models simulate snow-cover extent compared with remotely sensed data sets for the Northern Hemisphere. The first, Foster et al. (1996), (see Fig. 4.15), compared the snow cover and snow mass in seven GCMs with NOAA's visible remotely sensed snow-cover data set, the U.S. Air Force snow depth climatology (SDC), and microwave remotely sensed data sets for snow mass.

In the Northern Hemisphere, snow-cover extent is observed to be at a maximum in February (to about 40° N) and at a minimum in August, while Northern Hemisphere snow mass is observed to be at a maximum in March and at a minimum in August. It should be noted that although snow-cover extent historically peaked in February, as reported in Foster et al. (1996), January is currently the month with the observed peak in snow-cover extent, as reported in Frei et al. (2003). Also, snow cover is observed to retreat more quickly in the spring than it advances in the fall. Models are poorer at simulating snow cover during the transitional seasons of the spring and fall compared to the more stable winter and summer seasons; Frei et. al. (2003) found that model simulated snow cover tends to advance too slowly in the fall and to retreat too slowly in the spring, when compared with observations. Why observed

Eurasia Average Monthly Snow Cover Eurasia Average Monthly Snow Cover

Eurasia Average Monthly Snow Cover Eurasia Average Monthly Snow Cover

0123456789 10 1112 0123456789 10 11 12

Months Months

0123456789 10 1112 0123456789 10 11 12

Months Months

Figure 4.15. Snow cover and mass intercomparison for Eurasia from NOAA, SMMR and seven different GCMS. Taken from Foster et al. (1996).

snow cover melts and retreats faster in the spring than it advances in the fall and why models have greater difficulty simulating snow cover as observed during the transitional seasons is still not understood.

Comparisons between GCM simulated and observed snow cover indicate general differences at greater than the 95% significance level for every month of the year. The different GCMs show a wide range of values for snow cover and snow mass and even differ in respect to which months the maximum and minimum in snow cover and snow mass occur (see Fig. 4.15). Most of the errors in the GCM output probably result from inaccuracies in modeling surface temperatures, which stem from errors in the surface energy balance, and the parameterization schemes of precipitation, which often produce excessive or deficient precipitation. Properly simulating snow mass and snow cover involves improvements in model hydrodynamics, surface energy exchange, and other physical model parameterizations, all of which make the problem quite complex.

In the second snow-cover variability study, Frei and Robinson (1998) conducted an atmospheric model intercomparison project (AMIP)-type comparison of snow extent in 27 different models participating in AMIP-1. All participating GCMs were forced with observed sea surface temperatures from 1979-1988. The authors verified the models' snow extent results against NOAA's visible remotely sensed data set for snow cover. In general, they found that model-generated fall and winter snow extent was underestimated, while spring snow extent was overestimated when compared with the remotely sensed snow-cover extent. GCMs tended to underestimate snow cover more in North America than in Eurasia in the fall and winter while overestimating snow extent more in Eurasia than in North America in the spring. All models underestimated the observed interannual variability of snow cover during fall and winter, with the majority of models displaying less than half the observed range in snow-cover variability. However, some models do better in the spring.

Correlations between interannual snow-cover anomalies simulated in the GCMs with those observed were poor. Spearman rank correlations are generally 0.0 ± 0.25 for fall and winter. The models were even unable to reproduce extreme events; therefore, the authors concluded that snow cover is not forced by sea surface temperatures (SSTs), although this conclusion contradicts the observational study of Ye (2001), which identified SSTs as a forcing for snow-cover variability. Finally the authors examined the dependence of snow-cover variability on model resolution. Models with coarser resolution, horizontally and vertically, tended to have greater root mean square errors and greater interannual variability than higher resolution models.

Frei et al. (2003), (see Fig. 4.16) completed a similar evaluation of snow-cover variability in models participating in AMIP-2. In general, they found that model-simulated snow cover was greatly improved when compared to AMIP-1. Most notable was that the seasonal biases reported for AMIP-I had been corrected. Inter-annual variability was also notably improved in AMIP-2, although still less than that observed. Regional biases remained among the models, especially in Eurasia where biases were systematic.

Finally Frei et al. (2005) evaluated how well GCMs simulated snow water equivalent (SWE) over North America only among the suite of AMIP-2 GCMs. They conclude that the numerical models accurately simulate the seasonal timing of the relative spatial patterns of continental snow depth. However, the models do tend to ablate snow cover too quickly in the spring; in contrast to Foster et al. (1996) who found the models are too sluggish in melting spring snow cover. They also found that the multi-model mean was best at accurately simulating snow-cover climatology when compared with individual model results. Peak monthly SWE varied among the models by 50% of the observed peak value of ~1500 km3, which has important repercussions for the water balance of the North American continent; such large errors need to be improved. They found little correspondence in interannual variability between the models and observed snow mass and therefore conclude that at least in the models, SSTs are not forcing variations in continental-scale snow mass.

Figure 4.16. Observed and modeled mean seasonal snow-cover area for the Northern Hemisphere (NH), North America (NA) and Eurasia (EU). Observed values shown on left-hand side of each panel; model results are numbered. Horizontal lines indicate observed values ±0.05. Taken from Frei et al. (2003).

Not only do GCMs need to simulate the natural variability of snow cover from seasonal to interannual and even decadal time scales to represent the observed climate, but also to predict the role of snow cover in an anthropogenically changed climate accurately. Yet, as will be discussed below, not only do GCMs lack skill in simulating the natural variability of snow cover, the simulated influence of snow cover on the atmosphere also lacks consistency among different GCMs.

Renewable Energy 101

Renewable Energy 101

Renewable energy is energy that is generated from sunlight, rain, tides, geothermal heat and wind. These sources are naturally and constantly replenished, which is why they are deemed as renewable. The usage of renewable energy sources is very important when considering the sustainability of the existing energy usage of the world. While there is currently an abundance of non-renewable energy sources, such as nuclear fuels, these energy sources are depleting. In addition to being a non-renewable supply, the non-renewable energy sources release emissions into the air, which has an adverse effect on the environment.

Get My Free Ebook


Post a comment