Towards an Explicit Representation of Root Distributions in Global Models

(RB Jackson)

Workshop products (completed and planned)

Jackson, RB, HA Mooney, & ED  Schulze 1997 A global budget for fine root biomass, surface area, and nutrient contents. Proceedings National Academy of Sciences, USA 94:7362-7366.
Kleidon, A, & M Heimann 1998 A method of determining rooting depth from a terrestrial biosphere model and its impacts on the global water- and carbon cycle. Global Change Biology, in press.
Jackson, RB, HJ Schenk, J Canadell, GD Colello, RE Dickinson, T Dunne, CB Field, P Friedlingstein, M Heimann, K Hibbard, DW Kicklighter, A Kleidon, RP Neilson, WJ Parton, OE Sala, and MT Sykes 1999 Root distributions and vegetation change: consequences for global carbon and water fluxes. In preparation.
Jackson, RB, L Pitelka, eds. 1999 Belowground Ecology and Global Change. Invited feature in Ecological Applications (feature approved; 6 articles to be submitted 6/98). Several of the articles are based on workshop ideas and results.  One of these is the new Neilson/Running BIOMAP model that incorporates the root distributions discussed at the workshop into a new dynamic global vegetation model.

Related Publications

Jackson, RB 1998 The importance of root distributions for hydrology, biogeochemistry, and ecosystem functioning.  In (J Tenhunen, P Kabat, eds., Dahlem Conference) Integrating hydrology, ecosystem dynamics, and biogeochemistry in complex landscapes.  John Wiley and Sons, Chichester.  In press.

Workshop participants

Axel Kleidon            Pep Canadell
Pep Canadell            Xia Li
Greg Colello            Ron Neilson
Robert Dickinson     Bill Parton
Tom Dunne              Osvaldo Sala
Chris Field               Martin Sykes
Pierre Friedlingstein
Martin Heimann
Kathy Hibbard
Dave Kicklighter

The first workshop accomplished three goals: It 1) Outlined global rooting patterns for biomes and plant functional types, 2) Discussed the issues associated with root representations in global models - the issues and conditions for which they will be important, and 3) Outlined a series of future modeling exercises as sensitivity analyses and subsequent joint ecosystem and global simulations.

Needs Identified at the Workshop (Conceptual Issues)
 A) Sensitivity analysis of rooting depth, site-specific and global
 B) Global rooting map(s): 1) biome-specific, 2) physical constraints, 3) climatic controls
 C) Inferential techniques and inverted rooting depths

Needs Identified at the Workshop (Data Inputs)
Fine root biomass by biome and plant functional type
Data on field capacity and wilting point
Frequency of biological vs physical constraints of rooting depth
Relationship between soil carbon and live roots; soil C and horizon characterization
Spatial heterogeneity, horizontal pattern (plant scale; landscape scale)
Relationship of LAI to fine roots (area basis!)
Cost function
Effect of soil texture on rooting depth and total root biomass
Turnover time of roots
Root phenology
Live vs dead roots
Rules for penetration of roots (bulk density, pH, C horizon)
Database of adjacent habitats of differing soil depths

Summary of outlined modeling exercises

Site-specific sensitivity runs
 A detailed sensitivity analysis on a network of well-documented field stations (and representing the major global biomes) will provide baseline information to construct a global rooting depth map and provide a framework to compare model output at variable rooting depths with existing site data. Output variables will include net primary productivity (NPP), gross primary productivity (GPP), actual evapotranspiration (AET), transpiration (Ts), and stream runoff. For models that utilize multiple soil depths, the exponential Gale and Grigal (1987) equation (1-bd) will be used to derive the root biomass distribution, where d= soil depth. The experimental design for the modeling exercise by each group will include 28 potential sites with variable soil depths (Table 1).
 Ideally, a 10y climate database will be developed for each field station. The climate data needed by modeling groups include daily precipitation, and minimum and maximum air temperature. Daily shortwave incident radiation and daylength will be generated from the latitude, longitude,  temperature and precipitation data supplied by the Biome-BGC group. Daily data can be aggregated for those models that require monthly inputs (e.g. TEM, CENTURY, MAPSS, CASA).  Sub-hourly climate data will be degraded for SiB. In addition, relative humidity, wind speed and longwave radiation will be obtained from the Re-Analysis data available on CD-ROM for SiB. Some models are driven by LAI (MAPSS, CASA) which will be supplied by either the averaged seasonal LAI from other models that generate it (BGC, CENTURY) or whichever model provides the most “reasonable” LAI.  Models will initialize the sites as mature ecosystems and go through at least two 10-year climate cycles for spin-up purposes.
Table 1. Detailed information for site-specific analyses

Possible Sites                                                                            Biome   
LTER:  Konza                                                                           Tallgrass Prairie
     CPER (3)                                                                              Shortgrass Prairie (burned, unburned)
     Coweeta, Harvard Forest, Hubbard Brook                          Temperate Forest
     Andrews                                                                               Temperate Coniferous Forest
     Bonanza Creek, Niwot                                                          Boreal Forest, Tundra
Osvaldo's Sites (9)                                                                    Semiarid shortgrass/steppe
La Copita Research Area                                                           Subtropical thorn savanna parkland
Luquillo Forest                                                                          Tropical forest
Nairobi, Lamtô                                                                          Tropical Savanna
Flakalada      Boreal Forest
Mongolia                                                                                    Shrub-steppe, grassland
Monte Verde                                                                              Tropical forest

Rooting Depths (m): 0.3, 0.5, 0.7, 1.0, 1.4, 2.0, 2.4, 2.8, 4.0
Participating Modeling Groups (site-specific data): BIOME3, Biome-BGC, CASA, CENTURY, MAPSS, SiB, TEM

Global simulations
 Carbon models will be initialized with the global Potsdam climate data from the PIK exercise, and GCM’s will run on their own climate for global simulations.  Models will be initialized with global rooting depths provided from a variety of possible sources:  1) Biome-specific rooting depths as provided from the site sensitivity analyses described above; (2) A map derived from physical soil constraints (Scholes, IGBP); or (3) a map derived from inferential (e.g. NDVI) datasets. A subset of models may examine the biogeochemical consequences of shrub encroachment globally.