[NARCCAP-discuss] Question: Bias and regional applicability studies/results?

[Ethan Bodnaruk]

Dear NARCCAP staff and users:

With 12 sets of data available through NARCCAP (6 regional models, each run
with 2 different global drivers), I'm wondering if there is already a
consensus about whether or not some of the models are particularly
well-suited (low bias or other performance metric) to certain regions of
the US (NE, SE, etc).  In other words, is there any guidance on which
models to use in research applications?  Or is it still best to analyze
multiple data sets to see the variation between models?

Also, is there any additional analysis available on the UDEL (Univ
Delaware) observed data and regional NCEP runs?  The graphs (
http://www.narccap.ucar.edu/results/ncep-results.html) contain a lot of
info that's hard to tease out visually, and I wonder if anyone has already
summarized or analyzed these further.

My interest is in the effect of climate change on urban forestry (how to
use urban forestry to mitigate climate change, provide a host of ecosystem
services, and how to plain maintenance/planting for a different climate).
I've written code that calculates monthly averages of precip and temp for
the future time period and Reanalysis time period, and plots them as a
boxplot.  This helps visualize climate change and shows predicted monthly
differences in a clear manner (drier August, wetter January, etc.).  I'm
wondering if this is something useful to the community or if something
similar has been done before.  If it is useful, I would probably need to
look more at observed weather data and bias.

Any thoughts?

Cheers,
Ethan Bodnaruk

-- 
Ethan Bodnaruk
M.S. Nuclear Engineering
Ph.D. Student, Ecological Engineering
SUNY School of Environmental Science & Forestry
Syracuse, NY 13210 USA