A similar analysis using a much longer tree-ring reconstruction,
to simulate worse conditions than those experienced in
the historical record, would require seasonal resolution or better
from the tree-ring data. While sub-annual tree-ring chronologies
and reconstructions are feasible, further statistical downscaling is
likely to be the only option to pursue applications such as these.
The method of matching pre-instrumental proxy flow sequences
with corresponding analogue years from the gauge record
(Woodhouse and Lukas, 2006) is one approach to downscaling,
using the monthly flows from the analogue years. Another possibility
would be to combine the strength of paleohydrology, record
length, with the advantages of stochastic hydrology, large data sets
of synthetic flows (Sobajic et al., 2010). Long (e.g., 1000 years) synthetic
time series of weekly or monthly river flows could derived so
that all relevant statistical properties of the historic and tree-ring
time series are preserved, capturing the dominant modes of variability
from seasonal to multi-decadal.