Transcript JMA Best Track JMA Best Track

Downscaling Tropical Cyclones
from global re-analysis:
Statistics of multi-decadal variability
of TC activity in E Asia, 1948-2007
VON STORCH Hans and FESER Frauke
1Institute
for Coastal Research, GKSS Research Center, Germany
2clisap-Klimacampus, University of Hamburg, Germany
Estimating the changing risk related to typhoons –
Problems and issues:
Inhomogeinity of observational evidence (higher
accuracy, more details in recent times).
Data about typhoon-related damages are affected by
changing socio-economic regional usage and conditions.
Change may be related to interannual and interdecadal
variability or to systematic climate change.
Representativity of near surface wind
speed measurements
• Causes of inhomogenities:
• Changes in
– Instruments
– Sampling frequencies
– Measuring units
– Environments (e.g. trees, buildings)
– Location
Representativity of near surface wind
speed measurements
1.25 m/s
Dotted –
station
relocation
Representativity of near surface wind speed measurements
Example: Inhomogeneity of analyses of hurricanes –
Erin, in September 2001
Max: 20 m/s
Analysis using all available
surface observations
Landsea et al., 2004
Max: 52 m/s
Analysis using additional
aircraft reports.
Losses from Atlantic
Hurricanes
The increase in damages related
to extreme weather conditions is
massive – but is it because the
weather is getting worse?
Hardly
“Great Miami”, 1926, Florida, Alamaba
– damages of 2005 usage - in 2005
money: 139 b$
Katrina, 2005: 81 b$
Pielke, Jr., R.A., Gratz, J., Landsea, C.W., Collins,
D., Saunders, M., and Musulin, R., 2008.
Normalized Hurricane Damages in the United
States: 1900-2005. Natural Hazards Review
• Any assessment of how weather patterns have
changed in recent decades requires long and
homogeneous time series.
• Local time series representing wind speeds are
usually not homogeneous, even for a few decades
(sensitivity to instrumentation and surrounding).
• Homogeneous description of variability of mesoscale storms in recent decades has also not been
achieved.
• Model-based “reconstructions” may help
Dynamical downscaling,
deriving regionally
disaggregated descriptions
from global re-analysis or
global climate change
scenarios.
Climate = statistics of weather
The genesis of climate
 “downscaling”
Cs = f(Cl, Φs)
with
Cl = larger scale climate
Cs = smaller scale climate
Φs = physiographic detail at smaller scale
von Storch, H., 1999: The global and regional climate system. In: H. von
Storch and G. Flöser: Anthropogenic Climate Change, Springer Verlag,
ISBN 3-540-65033-4, 3-36
Problem for synoptic systems solved by CoastDat@GKSS in
N Europe, using RCM spectrally nudged to NCEP
- retrospective analysis 1958-2007
- good skill with respect to statistics,
but not all details are recovered.
Observations – black;
Hindcast - green
At a platform
in the
S North Sea
Wind speed [m/s]
Sig. wave height [m]
Wind direction [degrees]
(Weisse and Günther. 2007)
Mean wave direction [degrees]
Weisse, R., H. von Storch and F. Feser, 2005: Northeast Atlantic and North Sea storminess as simulated by a regional
climate model 1958-2001 and comparison with observations. J. Climate 18, 465-479
The added value of RCM runs in hindcast mode
1. Enhanced variability of medium spatial scales
2. Better description of medium scale variations.
3. Better description of variability in topographically structured
regions (e.g., coasts)
4. Additional dynamical features (e.g., Polar Lows, typhoons)
5. Spatially complete data on fine grid, needed to run ocean
and wave field models
In the red
marked areas,
the day-to-day
wind description
by CLM
compares better
than NCEP reanalysis to
QUICKSCAT
satellite data
Winterfeldt, 2008
Simulation of
additional
dynamical detail
Formation of
Catalina Eddy in
description of
Californian
Climate (CaRD10)
Kanamitsu, SIO, pers. comm.
Climate simulations with CLM of Polar Lows 1948-2007
driven by the NCEP reanalysis
 spectral nudging
 about 50 km grid resolution

Zahn, M., and H. von Storch, 2008: A longterm climatology of North Atlantic Polar Lows.
Geophys. Res. Lett., 35, L22702, doi:10.1029/2008GL035769
Based on these findings, we
believe that we can use the
concept also for describing
Typhoon stats variability and
possible trends and constructing
scenarios
• We have implemented the dynamical downscaling
approach for E Asian marine weather.
• The key question is – will we master the description of
typhoons?
• Done: Case studies and seasons – formation of typhoons
induced by large scale dynamics and NOT by initial values.
Feser, F., and H. von Storch, 2008: Regional modelling of the western Pacific
typhoon season 2004, Meteor. Z. 17 (3), 1-10. DOI: 10.1127/0941-2948/2008/0282
Feser, F., and H. von Storch, 2008: A dynamical downscaling case study for
typhoons in SE Asia using a regional climate model. Mon. Wea. Rev. 136, 1806-1815
• Presently under examination: Continuous 6-decade
simulations constrained by NCEP global re-analyses.
16 km
50 km
Case study:
Typhoon Winnie, August 1997 simulated with different set-ups
12 TCs
in
Season
2004
only 10 were
found in
CLM
simulation
Following Zhang et al.,
2007. Meteor. Atmos.
Phys.
Simulated typhoons are weaker than found in the „best
track data“ – too high core pressures, too weak winds,
But considerably stronger than in the driving NCEP reanalyses.
The model‘s performance does not improve much by
enhancing the horizontal resolution from about 50 km
to about 16 km.
Experiments with different convection parametrizations
are presently doen with encouraging results.
Complete simulation of 1948-2007
using CLM with 0.5º grid resolution
and NCEP/NCAR reanalysis
Spectral nudging of scales
larger than about 800 km.
All „best tracks“, 1951-2007
E Asian tropical cyclones as given by
CLM downscaling
JMA best track data
40
30
20
10
0
1945
1950
1955
1960
1965
1970
1975
1980
Note: different criteria employed
1985
1990
1995
2000
2005
2010
Interannual variability
36 “best tracks
26 tracks in CLM
16 “best tracks”
16 tracks in CLM
Annual TC count
Bias CLM vs Best Track
15
CLM - Best Track
11 year running mean
10
5
0
-5
-10
-15
1960
1980
2000
Findings so far:
1) CLM-simulated typhoons too
weak.
2) Number and interannual
variability in CLM similar to
„best track“ data set.
3) Long term trends in CLM and
in in „best track“ markedly
different.
4) In CLM, intensification since
about 1980.
5) In „best track“, weakening
since about 1980.
But …
in China
Ren, F., G. Wu, W. Dong, X. Wang, Y. Wang, W. Ai, and W. Li, 2006: Changes in tropical cyclone
precipitation over China. Geophys Res. Lett. 33, L20702, doi:10.1029/2006GL027951
1953
Typhoon Season 1953
Several typhoons of
the Best Track data
show very large
drops in core
pressure –
Are they realistic?
JMA Best Track (black lines) – 23 typhoons,
CLM (colored lines) 28 typhoons
Typhoon 195307
(NINA)
1953-08-08 06:00
1953-08-19 00:00
JMA Best Track
Typhoon 195307
(NINA)
JMA Best
Track
1953-08-08 06:00
1953-08-19 00:00
Largest drop in
core pressure
August 11-12
65 hPa in 6 hours
JMA Best
Track
Typhoon 195313
(TESS)
1953-09-18 00:00
1953-09-27 18:00
Largest drop in BT core
pressure:
93 hPa in 6 hours
The CLM shows much smaller drops in core pressure
- probably too small in most cases.
The most extreme pressure falls described in the “best
track” data set took place over the open ocean, where
no satellite data was available in 1953 – how was it
observed?
Later years show overall less extreme pressure 6-hourly
drops.
Use caution when using earlier “best track” years.
Overall Conclusions
Dynamical Downscaling of NCEP reanalysis with
regional climate models returns typhoon
climatology better than NCEP, even if cyclones
are still too weak.
Results concerning change
- Strong year-to-year variability
- Little decadal variability
- No overall trend in numbers
- Trends in intensities opposite in CLM and in “best
track”.
- “Best track” suffer likely from severe
inhomogeneities in the early years (e.g., 1953)