Search Results

You are looking at 1 - 1 of 1 items for :

  • Author or Editor: Krista A. Dunne x
  • Journal of Hydrometeorology x
  • Refine by Access: Content accessible to me x
Clear All Modify Search
P. C. D. Milly
,
Sergey L. Malyshev
,
Elena Shevliakova
,
Krista A. Dunne
,
Kirsten L. Findell
,
Tom Gleeson
,
Zhi Liang
,
Peter Phillipps
,
Ronald J. Stouffer
, and
Sean Swenson

Abstract

LM3 is a new model of terrestrial water, energy, and carbon, intended for use in global hydrologic analyses and as a component of earth-system and physical-climate models. It is designed to improve upon the performance and to extend the scope of the predecessor Land Dynamics (LaD) and LM3V models by better quantifying the physical controls of climate and biogeochemistry and by relating more directly to components of the global water system that touch human concerns. LM3 includes multilayer representations of temperature, liquid water content, and ice content of both snowpack and macroporous soil–bedrock; topography-based description of saturated area and groundwater discharge; and transport of runoff to the ocean via a global river and lake network. Sensible heat transport by water mass is accounted throughout for a complete energy balance. Carbon and vegetation dynamics and biophysics are represented as in LM3V. In numerical experiments, LM3 avoids some of the limitations of the LaD model and provides qualitatively (though not always quantitatively) reasonable estimates, from a global perspective, of observed spatial and/or temporal variations of vegetation density, albedo, streamflow, water-table depth, permafrost, and lake levels. Amplitude and phase of annual cycle of total water storage are simulated well. Realism of modeled lake levels varies widely. The water table tends to be consistently too shallow in humid regions. Biophysical properties have an artificial stepwise spatial structure, and equilibrium vegetation is sensitive to initial conditions. Explicit resolution of thick (>100 m) unsaturated zones and permafrost is possible, but only at the cost of long (≫300 yr) model spinup times.

Full access