Land surface models

Numerous LSMs of differing complexity and architecture are in existence. A major research avenue is model intercomparison. This has been conducted under the Project for Intercomparison of Land Surface Parameterization Schemes (PILPS), within the GEWEX Global Land/Atmosphere System Study. PILPS 2e (Bowling et al., 2003) is a component of this effort aimed at evaluating the performance of uncoupled ("stand alone") LSMs in northern high latitudes. The project was motivated by the need to improve high-latitude land surface representations within NWP models and GCMs.

PILPS 2e compared simulations of land surface processes from 21 models run for the Torne-Kalix (58 000 km2) catchment in northern Scandinavia (Figure 9.4). The location was selected because reasonably detailed and high quality data for a 20-year period (1979-98) were available. Identical atmospheric forcing data (precipitation, air temperature, specific humidity, wind speed, downward shortwave and longwave radiation) were provided to each modeling group. Also provided were land cover classifications, leaf area index and soil characteristics.

A few examples of the models that participated in PILPS 2e are:

• Canadian Land Surface Scheme (CLASS) This model (Verseghy, 1991; Verseghy et al., 1993) was specifically designed with cold land processes in mind and was among the first to incorporate a separate snow layer (and frozen ground). CLASS has been the focus of extensive activity by the Canadian Mackenzie GEWEX study group, which has targeted more realistic treatments of soil, land cover, hydrologic routing and sublimation.

• VIC (Variable Infiltration Capacity Model) VIC was developed with the purpose of producing realistic runoff and streamflow via parameterizations of subgrid soil moisture variability, as well as spatial variability in precipitation

Plate 6 Modeled maximum thaw depth (a) and the date of maximum thaw depth (b) for 1999 over the Arctic terrestrial drainage. Areas with no permafrost in the drainage and the Greenland Ice Sheet are indicated by dark grey shading (from Oelke et al., 2003, by permission of AGU). See color-plates section.

250 225

25 0

DT,max(cm)

1-Dec

350 325

1-Nov

300

1-Oct

275

1-Sep

250

1-Aug

B

225 200

1 -Jul

Figure 9.4 The Torne-Kalix river system modeled under the PILPS 2e project and location of observation stations. Precipitation stations are shown by circles, temperature stations by stars and synoptic stations by diamonds. Abbreviations are A, Abisko Research Station; H, Haparanda; K, Katterjakk; N, Naimakka; P, Pajala. The dotted line indicates the approximate northern limit of trees. A natural bifurcation diverts about 22% of the Torne river runoff into the adjacent Kalix river basin. Dark shading indicates sub-catchments for calibration and validation (from Bowling et al., 2003, by permission of Elsevier).

Figure 9.4 The Torne-Kalix river system modeled under the PILPS 2e project and location of observation stations. Precipitation stations are shown by circles, temperature stations by stars and synoptic stations by diamonds. Abbreviations are A, Abisko Research Station; H, Haparanda; K, Katterjakk; N, Naimakka; P, Pajala. The dotted line indicates the approximate northern limit of trees. A natural bifurcation diverts about 22% of the Torne river runoff into the adjacent Kalix river basin. Dark shading indicates sub-catchments for calibration and validation (from Bowling et al., 2003, by permission of Elsevier).

forcings and land cover. Newer parameterizations have been developed (Cherkauer et al., 2003) to improve representation of important high latitude processes such as sublimation, snow redistribution, soil freezing, and the dynamics of lakes and wetlands.

• CHASM The CHASM (CHAmeleon Surface Model) framework (Desbor-ough, 1999) contains options that allow it to utilize a range of hydrologic parameterizations and surface energy balance configurations. In simplest form, CHASM collapses to the simple "bucket" model of Manabe (1969), while in its most complex mode it has a tiled mosaic structure (e.g. Koster and Suarez, 1992). A tile is a homogeneous land surface type (e.g., boreal forest or tundra). Each grid cell can have numerous tiles, for which the energy balance is calculated separately.

• ECMWF This is a tiled version of the land surface scheme described by Viterbo and Beljaars (1995). It is designed for operational use within the ECMWF forecast model (also the basis for the ERA-40 reanalysis). The tiled version used in PILPS 2e (van den Hurk and Viterbo, 2002) has yielded major improvements in the simulation of snow processes in boreal forest areas.

Among the problems identified under PILPS activities especially critical to model performance in the Arctic are: (1) treatment of snowpack redistribution due to wind, and the accompanying issue of parameterization of sublimation; (2) snowfall interception by tree canopies and the fate of this intercepted snow in boreal forests; (3) the effects of lakes and wetlands on the timing of runoff, evaporation, and the regional energy balance; (4) soil freezing and permafrost.

For example, Pomeroy et al. (1993) and Pomeroy and Essery (1999) demonstrated that incorporation of blowing snow physics is necessary to simulate adequately sublimation losses in prairie snow environments. Similar issues exist in Arctic environments. Other and arguably more first-order problems identified by Slater et al. (2001) as part of the precursor PILPS 2d effort include: (a) difficulties in adequately modeling the albedo and fractional coverage of snow, and associated effects on ablation; (b) the wide variety of model structures and resulting impacts on surface energy budget treatments; (c) the tendency for "colder" models to create a stability-induced cutoff of turbulent heat fluxes, yielding a larger sensitivity to downwelling longwave radiation when surface net radiation is negative.

Figures 9.5 and 9.6 summarize some results from the model intercomparison as means to the decade 1989-98 (the decade 1979-88 was used for model spin-up). Quite striking is the large scatter in predicted March snow water equivalent. Averaged modeled snow water equivalent over the basin ranges from 119 to 268 mm. Models with high latent heat flux and an average downward sensible heat flux (a heat source for the surface) tend to have the lowest snow accumulation. Not surprisingly given the scatter in snow water equivalent, mean annual runoff also differs substantially, from 301 to 481 mm. For some models subsurface runoff dominates, while for others, runoff is solely from the surface. Differences in modeled snow accumulation and surface-subsurface runoff partitioning contribute to large variations in the shapes of mean hydrographs.

100 Bowling Tips

100 Bowling Tips

Playing bowling with your friends can help you decide if it is indeed the hobby that you want to invest your time on today. Aside from that, it can help you get a better feel of the sport. More importantly, when you play with your friends, it would become a more fun activity, which you can look forward to each week.

Get My Free Ebook


Post a comment