Transcript No Slide Title

Evaluation of Hydrologically Relevant PCM Precipitation Characteristics over the Continental U.S.
C. M. Zhu, A.W. Wood and D.P. Lettenmaier
Department of Civil & Environmental Engineering, Box 352700, University of Washington, Seattle, WA 98195
2
Introduction
Monthly Cycle
gridded to 1/8 degree, were aggregated to the PCM scale. The precipitation statistics
comparison includes i) spatial variation in the annual and seasonal mean; ii) the monthly
precipitation cycles and cumulative probability distribution; iii) daily precipitation distribution;
iv) the diurnal precipitation cycle. The period 1950-1999 was used as the basis for daily
through annual comparisons; and the period 1995-1999 was used for the hourly analysis.
Annual and Seasonal Comparison
PCM (B06.22)
Observed
PCM - Observed
Annual
Daily Statistics
In eastern US, there is no obvious difference between
PCM and observation in daily precipitation intensity
distribution, but in central US, PCM has much higher
precipitation intensities than observed.
The ability of the DOE PCM ( Parallel Climate Model ) to reproduce selected precipitation
statistics was evaluated by comparing the PCM historical climate runs B06.22 (daily output)
and B06.64 (hourly output) with gridded observations (described in Maurer, et al. (in review))
over a range of temporal scales from sub-daily to annual. The domain for this comparison is
the continental U.S. with a grid cell resolution of PCM scale (T42 horizontal resolution, about
2.8 degree). The observation data, taken from long-term Cooperative Observer stations and
1
3
In western US, PCM tends to underestimate the
longer storm inter-arrival periods, and in eastern
and central US the reverse is true.
•
•
•
•
PCM tends to overestimate the storm duration in
Great Plains, underestimate it in the Pacific
Northwest, West Coast and Northeast.
PCM reproduces the monthly cycle well in the eastern Great Plains,
Midwest and the Ohio River basin.
In the Great Plains and east of the Rocky Mountains, PCM overestimates
precipitation in summer and autumn.
In the Southeast, PCM underestimates precipitation throughout the year,
with the greatest biases in summer and fall.
In Western US, east of the coastal ranges, PCM overestimate the
precipitation in winter, spring and fall.
4
Hourly Features
MAM
JJA
SON
DJF
Summer:
Shown above is the spatial distribution comparison of annual and seasonal long-term
(1950-1999) average precipitation.
 For annual average precipitation, PCM shows significant negative bias in the
southeastern US and in the Pacific Northwest, and a large positive bias centered in
eastern Colorado.
 In spring and autumn, the differences between PCM and observation are smaller than
in summer and winter: PCM has a small negative southeastern US and Pacific
Northwest for both. For spring, in the central and northeastern US, there is a small
positive precipitation bias.
 summer precipitation differences drive the positive bias in the annual precipitation
difference in Colorado, and winter differences are the source of the negative bias in
the Pacific Northwest.
Shown above are the precipitation cumulative probability distributions for January and
July.
PCM precipitation exhibits a strong diurnal variation, with precipitation maximum in
the afternoon, everywhere except the southeastern US coast and Gulf Coast. This variation matches
observations well in the West and Southeast, but particularly overestimates diurnal variation in the
northern Great Plains and Central US.
Winter:
January:
PCM and observed CDFs are similar in the eastern US; however, the
extreme PCM precipitation is much less than observed. For much of the western and
central US, the reverse is true, and generally PCM precipitation is greater than observed.
July:
In the East Coast, the PCM precipitation agrees fairly well with the
observations, but is low for extreme months; whereas in the central US, PCM
simulation is greater than the observed at nearly all percentiles. In the West, results vary.
PCM matches well the lack of diurnal variation in observed winter precipitation.
References
For PCM model information, see: http://www.cgd.ucar.edu/pcm/
Maurer, E.P., A.W. Wood, J.C. Adam, D.P. Lettenmaier, and B. Nijssen, 2001, A Long-Term Hydrologically-Based Data
Set of Land Surface Fluxes and States for the Continental United States, J. Climate (in review). Draft available from:
http://www.ce.washington.edu/~edm/pub_pres.html