Weather pattern-based evaluation of the Intermediate Complexity Atmospheric Research Model (ICAR) Johannes Horak 1 , Marlis Hofer 1 , Fabien Maussion 1 , Ethan Gutmann 2 , Alexander Gohm 1 and Mathias W. Rotach 1 1 Department
of Atmospheric and Cryospheric Sciences Universit¨ at Innsbruck, Austria
2 National
Center for Atmospheric Research Boulder, Colorado, USA
12.04.2018 1/25
Motivation Local effects of a changing global climate
Glaciers
Hydrology
Agriculture
Cities 2/25
Motivation
Reliability of the method local variability of precipition well represented by the method?
3/25
Model Intermediate Complexity Atmospheric Research Model (ICAR) (Gutmann et al., 2016) I quantities stored on 3D grid I advected within wind field I microphysics I I
physics based downscaling computationally frugal
Wind field I calculated analytically I based in linear theory I calculated for every forcing time step ⇒ Sequence of steady states 4/25
ICAR - Windfield
U
5/25
ICAR - Windfield
U
from coarse scale forcing
6/25
ICAR - Windfield
U
from high resolution DEM
7/25
ICAR - Windfield ts af
dr up
U
ts af
dr wn do equations: Barstad and Grøn˚ as (2006) 8/25
ICAR - Windfield ts af
dr up
U
ts af
dr wn do equations: Barstad and Grøn˚ as (2006) 9/25
Study Region New Zealand
10/25
Domain - South Island of New Zealand alpine stations coastal stations topography above 1000 m model domain north-westerlies predominant
40°S
42°S
44°S
precipitation data provided by NIWA, NZ MetService, NZ University of Otago, NZ
46°S
48°S 164°E
166°E
168°E
170°E
172°E
174°E
176°E 11/25
Setup ERA-Interim forcing I
∆t = 6 h ∆A ≈ 60 × 83 km2
40°S
also used as to determine added value
downscale to I ∆t = 1 h ∆A ≈ 4 × 4 km2 I model top at ≈ 5.7 km above topography 10 year study period I 01/2006 to 12/2016 Settings I ICAR standard settings I NO tuning to observations
42°S
44°S
46°S
48°S 164°E
166°E
168°E
170°E
172°E
174°E
176°E
12/25
Weather Patterns 1000 hPa isohypses
I I I
12 synoptic weather patterns (Kidson, 2000) daily classification since 1948 by NIWA, NZ defined by 24h mean elevation of 1000 hPa lvl I I
I
example: Trough - pattern on ≈ 12% of days
linked to regional moistening / drying
100 m 50 m 0m -50 m
Trough - pattern 13/25
Weather Patterns 30 25
-42S
20 15
-44S
10
-46S
5
-48S
4E 17
2E 17
0E 17
8E 16
16
6E
0
-40S
pattern - mean (mm/day)
-42S -44S -46S -48S
10.0 7.5 5.0 2.5 0.0 2.5 5.0 7.5 10.0
16 6E 16 8E 17 0E 17 2E 17 4E
mean (mm/day)
-40S
14/25
Weather Patterns
Weather patterns - ideal for investigating ICAR I not part of downscaling method I indicator of physicality Weather Pattern ⇒ local moistening and drying
can ICAR model the measured variability?
15/25
Precipitation Variability at Alpine Station Ivory (z = 1390 m) Taihoro Nukurangi
(niwa.co.nz)
60 40 20
R
N W W H SE H E N E H W
H
H
TN W TS W
SW
0
T
precipitation (mm / day)
measured by
16/25
Precipitation Variability at Alpine Station Ivory (z = 1390 m)
ICAR
60 40 20 0
T SW TN W TS W H HN W W HS E HE NE HW R
precipitation (mm / day)
measured
weather pattern
17/25
Precipitation Variability at Alpine Station Ivory (z = 1390 m)
ICAR
ERAI
60 40 20 0
T SW TN W TS W H HN W W HS E HE NE HW R
precipitation (mm / day)
measured
weather pattern
18/25
Precipitation Variability at Alpine Station Ivory (z = 1390 m)
measured
r 2 = 0.80
2 1 0
T SW TN W TS W H HN W W HS E HE NE HW R
P/Pmean
3
ICAR
weather pattern
19/25
Precipitation Variability at Alpine Station Ivory (z = 1390 m)
2
ERAI
r 2 = 0.001
1 0
T SW TN W TS W H HN W W HS E HE NE HW R
P/Pmean
measured
weather pattern
20/25
Calculate for every station albert_burn
arthur_pass
brewster
cook
ivory_nv
mahanga
mueller_hut
philistine
potts
raikai
christchurch_intl_cliflo
hokitika_aerodrome_cliflo
invercargill_airpor_cliflo
kaikoura_cliflo
wellington_aero_aws_cliflo
4 2 0
4 2 0
4 2
H HE HN W HS E HW NE R SW T TN W TS W W
0
4 2
T TN W TS W W
R SW
H HE HN W HS E HW NE
T TN W TS W W
R SW
H
HE HN W HS E HW NE
T TN W TS W W
R SW
H
HE HN W HS E HW NE
0
21/25
Al b Ar ert B thu urn r Br Pas ew s st Co er ok I v M Mu aha ory ell nga e Ph r Hu ilis t Ch R tine ris aik tch ai H ur Inv okit ch erc ika Ka argi W ikou ll ell ra ing ton
r 2 (1)
Coefficients of Determination ICAR ERAI
1.00
0.75
0.50
0.25
0.00
22/25
Caveats
ICAR underestimates precipitation I ERA-Interim too dry? I strong influence of model top ⇒ further studies needed I
workaround: correction factor per site
Convection parametrizations not tested (yet)
23/25
Summary
Investigated variability of local precipitation due to synoptic weather patterns I I I I
added value of ICAR compared to ERA-Interim local variability well explained by ICAR local variability linked to synoptic situation relevant processes well approximated
⇒ ICAR suited to investigate the local effects of a future climate
24/25
Outlook I I I
extend analysis to gridded precipitation data (e.g. GPM) variability of local temperature does ERA5 explain variability better?
More details in paper later this year I skill scores (MSE and HSS based) I performance indicators for ICAR Updates / Contact: I
[email protected] I or on ResearchGate.net
Thank you! 25/25
Literature I
Barstad, I. and Grøn˚ as, S. (2006). Dynamical structures for southwesterly airflow over southern norway: the role of dissipation. Tellus A, 58(1):2–18. Gutmann, E., Barstad, I., Clark, M., Arnold, J., and Rasmussen, R. (2016). The intermediate complexity atmospheric research model (icar). Journal of Hydrometeorology, 17(3):957–973. Kidson, J. W. (2000). An analysis of new zealand synoptic types and their use in defining weather regimes. International journal of climatology, 20(3):299–316.
25/25
The End
25/25
Appendix
Supplemental data and plots
25/25
Precipitation Variability w. Standard Deviation at Ivory
75 50 25 0
T SW TN W TS W H HN W W HS E HE NE HW R
precipitation (mm / day)
measured
25/25
r (1)
r (1)
0.5
Al Arbert thu Bu Br r Parn ew ss s Coter o MuMahIvorky el an Phler Hga i Ch Rlistinut ris ai e tch ka i InvHokurch i e ti Karcargka W iko ill ell u ing ra ton
Al Arbert thu Bu Br r Parn ew ss s Coter o MuMahIvorky el an Phler Hga i Ch Rlistinut ris ai e tch ka i InvHokurch e iti Karcargka W iko ill ell u ing ra ton
Correlation for permuted weather pattern data icar original data mean r of permuted data of permuted pattern data erai original data mean r of permuted data of permuted pattern data
1.0
0.0
25/25
