From the earliest attempt at numerical weather prediction up until today's efforts on improving the land surface hydrologic parameterization in General Circulation models (GCMs), it has been recognized that the realistic characterization of atmospheric phenomena requires accurate representation of surficial processes. Lewis F. Richardson who attempted numerical weather prediction using hand calculations around the First World War period includes, in his notes, a parameterization of the surface evapotranspiration and plant stomatal control of the vapor exchange between land and atmosphere. He writes:
``Almost three-quarters of a century later and using high-speed digital computers, the research community is essentially implementing Lewis F. Richardson's original scheme of stomatal resistance to vapor flux between the saturated interior of leaves at temperatureLet the rate of loss of water from a leaf be denoted by T, then
here K is the conductance of the stomatal openings and
is the saturated vapour density at
.'' [ Richardson, 1922].
and
near-surface air humidity 
. At first sight it may appear that not
much scientific progress has been made in the interim years. Transpiration,
turbulence and other processes related to land-atmosphere exchange are
complex phenomena and in fact there has been significant recent advances in
the study of land-atmosphere interaction. The land, biosphere, atmosphere
and ocean systems are coupled across a wide range of space and time scales
such that each discovery leads to a deeper and larger scientific question.
Research inquiry in this area is now performed in both the hydrologic and
the atmospheric science communities.
The purpose of this paper is to summarize the recent (1991-1994) research advances in the study of two-way land-atmosphere interaction as reported in the scientific literature. In the years preceding this period, there are large volumes of papers on this topic some of which form the foundations for recent investigations. Limiting this review to the last four years necessarily overlooks some milestone investigations. The integral influence of the past should be apparent in the hypotheses, methodology and nomenclature of these recent studies reviewed here.
