CONCLUSIONS
The field measurement systems described here were
successful in obtaining reliable data for determining
throughfall and stemflow at the three remote rainforest
sites. The use of large automatically recorded tipping
buckets removed the need for frequent sites visits. The
throughfall system design employed in this study has
been shown to work well across the range of rainfall
conditions experienced during the comparison periods.
Trough replicates suggest that the trough length employed
in this study was capable of capturing the variability
in throughfall exhibited beneath the rainforest at our
measurement sites. However, at sites where the canopy
is more variable than in our study it may be necessary
to extend the spatial extent of the trough system to
capture the greater throughfall variability. The stemflow
system design also functioned across a range of rainfall
conditions and reasonable estimates of stemflow were
obtained despite the range of tree species, tree sizes and
bark textures sampled.
The canopy water balance method for estimating cloud
interception was shown to be more preferable than the
alternative gauge calibration method. When using the
canopy water balance method the cloud interception
gauge is needed only to define the occurrence (not the
amount) of cloud interception. Cloud interception gauges
can only be used in isolation if they are calibrated,
however, this calibration is affected by wind speed and
rainfall intensity and therefore, estimates obtained can
be highly uncertain. The strength of the canopy water
balance method is its ability to quantify total precipitation
input to forests that are exposed to significant cloud
interception and it works well when there are significant
differences in Sf C Tf between days with and
without cloud interception. However, at locations where
cloud interception is limited, regressions between Pga and
Sf C Tf may not be significantly different, thus rendering
the approach meaningless from a statistical point of
view (Hafkenscheid, 2000; Bruijnzeel, 2001).
Overall, the systems described and tested in this
paper have performed very well at three remote sites
over comparison periods ranging between 581 and
787 days. Canopy interception rates, based on stemflow
and throughfall measurements from these sites and others
where the same methods have been used can now
be reported with confidence. The application of the technique
described in this paper to other rainforest sites in
North Queensland will be presented in McJannet et al.
(2006).