Geoscience Reference
In-Depth Information
A water budget quantifies systematically the flows and reservoirs of water in the water
cycle based on the principle of the conservation of mass. That principle assumes that water
is neither created nor destroyed in the system. However, GCMs may be subject to para-
metrisation errors which affect water conservation and thus the closure of the water budget.
It is essential that such problems are addressed in order that reliable estimates of future
changes in the water cycle can be made. For example, Collins et al. (
2011
) showed a
budget diagram of long-term means of global water fluxes from the various sub-models in
a climate configuration of the Met Office Unified Model (MetUM), HadGEM2-AO. The
fluxes were shown to be close to equilibrium, with the exception of the atmosphere. This
apparent imbalance has been traced to a lack of water conservation in two areas of the
atmospheric model and has since been corrected (J. Rodriguez, personal communication,
May 2012). In addition, Collins et al. (
2011
) showed that a net freshwater flux field must be
applied to the ocean in order to balance the lack of iceberg-calving processes in the GCM
(see Sect.
2.3.4
), a problem which is common to many climate models.
Global water conservation is usually represented with an equation that involves verti-
cally integrated quantities:
o
W
ot
þr
H
Q
¼
E
P
where the first term on the left-hand side represents the precipitable water tendency, the
second term represents the moisture flux divergence, E is the evapotranspiration from the
surface and P the precipitation at the surface. Excess of evaporation over precipitation is
balanced by the local rate of change of moisture storage and the loss through horizontal
advection. Budget equations like the one above have been used before to estimate quan-
tities like moisture, energy and momentum, deducing the fluxes as a residual, after com-
puting the other variables in the budget equation (e.g., Trenberth et al.
2007
).
Rodr
´
guez et al. (
2011
) used this method to estimate the freshwater fluxes from the
atmosphere and land surface over large-scale ocean basins, in both a range of MetUM
models and in reanalyses. This showed an excess of evaporation in the tropical-subtropical
Atlantic in the models, which is manifested mainly as a lack of water vapour imported
across the African boundary. Models also have difficulties in representing the surface
fluxes over the tropical-subtropical Pacific. The freshwater budget in this region suffers
from too much import of water vapour across the Indonesia boundary and an excessive
export of water vapour to the mid-latitude regions. This surplus of water vapour from the
tropical-subtropical regions makes, in turn, the surface fluxes over the oceans too fresh at
mid-latitudes and can also produce a fresh bias over the Southern Ocean.
The above errors are consistent with an excessively strong hydrological cycle. Water vapour
recycles between 12 and 24 % faster in climate models, compared with estimates from obser-
vations (Rodriguez et al. 2011). The shorter water vapour residence times in the GCMs are the
effect of a combination of an enhanced global precipitation rate and a deficit in the water content of
the atmosphere (e.g., Demory et al.
2012
). These characteristics are related to the tendency for
models to rain more frequently and with less intensity than observed (Stephens et al.
2010
).
The characteristics of the water cycle in various regions in the globe are now being
studied. Regional water budgets on different timescales are being evaluated, and the
capacity of the GCMs to simulate the workings of the water cycle in those areas is being
assessed. Preliminary analysis of the monthly mean water budget components in Central
Africa in a recent climate configuration of the MetUM (J. Rodriguez, personal commu-
nication, May 2012) suggests that, although the seasonal variability of rainfall is well
represented, the moisture convergence is underestimated for most of the year. This makes
