CCSM Land Model Working Group Meeting
Breckenridge, June 25, 2003
The CCSM Land Model Working Group met on Wednesday, June 25, 2003. The format consisted of several 15-minute talks that provided updates to ongoing projects and to identify important new areas of research for the working group.
Forrest Hoffman (ORNL) discussed the proposed vectorization of CLM. The proposed vectorization plan results in a single code for scalar and vector platforms. It maintains the scientific functionality of the model, is portable to various machines, and does not significantly degrade performance on existing supported platforms. The working group endorsed this vectorization plan.
Liang Yang (University of Texas) discussed changes to the hydrology of CLM (interception and runoff). He presented simulations using fractional precipitation within a grid cell to modify interception (a scheme proposed by Guiling Wang), a modified TOPMODEL approach to runoff, and a full TOPMODEL that used a global topographic wetness index. The combined changes of interception and the modified TOPMODEL scheme improved the simulation of surface fluxes over the Amazon. The full TOPMODEL only marginally improved the simulation compared to the modified TOPMODEL. Defining the fractional coverage of precipitation in a grid cell still remains an issue.
David Gochis (NCAR) discussed his work related to runoff and infiltration. His scheme allows surface ponding and subsequent re-infiltration as surface runoff moves over land. It also parameterizes runoff with fully dynamical routing, saturated sub-surface routing, and exfiltration to the soil surface.
Marcia Branstetter (ORNL) discussed soil moisture sensitivity in CCSM2/CLM2. Her worked showed that the top soil layers recover quickly from a soil moisture perturbation, but that the deeper layers recover over significantly longer periods. This raises some question about how realistic the slow recovery of the bottom soil layer is.
Menglin Jin (University of Maryland) discussed the surface emissivity used in CLM. Compared to MODIS emissivity, CLM needs a more realistic surface emissivity over arid and semi-arid regions of North Africa. This could be corrected through a new soil emissivity in this region. Over the long term, the emissivity of soil should vary in relation to soil type, soil moisture, and leaf area index.
Robert Dickinson (Georgia Tech) discussed directions for the improvement of CLM. He discussed problems with surface albedo, especially in snow covered areas in spring and in arid/semi-arid regions. How to combine the effects of vegetation and soil on surface albedo remains an issue. He discussed the need for a two-leaf model rather than the single big leaf in the current model. He discussed downscaling of rainfall and solar radiation to improve canopy evapotranspiration. He also discussed soil hydraulic conductivity and the need to represent macropores.
Ian Baker (CSU) discussed whether CLM should be modified to have a prognostic canopy air space. This is needed to add carbon isotopes to the model for biogeochemistry applications, but also aids the simulation of heat and moisture fluxes by eliminating the need for iterative solutions and making the solution more stable. The working group endorsed this change as promising, though its impact on climate simulation and model performance needs to be thoroughly evaluated.
David Noone (Caltech) discussed adding water isotopes to CLM. Water isotopes are important to understand the hydrologic cycle. Results from the NCAR LSM are very promising. He plans to migrate the isotope code to CLM over the next 6-12 months.
Peter Thornton (NCAR) provided an update on his work to add the terrestrial carbon cycle to CLM. He described changes to the representation of the sunlit/shaded canopy in CLM and also to the Vmax parameter that controls photosynthesis/stomatal conductance. These changes are necessary to improve the simulation of gross primary production and also alleviate some of the low transpiration bias in CLM.
Sam Levis (NCAR) provided an update on his work to add dynamic vegetation to CLM. He showed results of the prognostic leaf phenology in the model and its impact on surface air temperature.
Xubin Zeng (University of Arizona) discussed the new parameterization of under-canopy turbulence that reduces a large warm temperature bias in arid/semi-arid regions. This parameterization has been included in the latest version of CLM. He also compared modeled albedo to MODIS albedo and discussed the need for a new surface albedo parameterization.
1. The working group strongly endorsed the plan to vectorize CLM. It is hoped that this vectorization will proceed rapidly so as to not delay scientific development of the model.
2. Model developments that should be considered for the next version of CLM include:
3. Addition of water isotopes to CLM is highly desirable.
G. Bonan. co-chair
S. Running, co-chair
L. Ruby Leung