NARCCAP Users Meeting April 2011 Results from NCEP-driven RCMs Overview Based on Mearns et al....
-
Upload
charleen-preston -
Category
Documents
-
view
216 -
download
0
Transcript of NARCCAP Users Meeting April 2011 Results from NCEP-driven RCMs Overview Based on Mearns et al....
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Results from NCEP-driven RCMsResults from NCEP-driven RCMs
Overview Based on Mearns et al. (BAMS, 2011)Overview Based on Mearns et al. (BAMS, 2011)
William J. Gutowski, Jr. William J. Gutowski, Jr. Iowa State University Iowa State University
and and The NARCCAP TeamThe NARCCAP Team
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Simulations AnalyzedSimulations Analyzed
• DomainDomain
- Most of North America- Most of North America
• PeriodPeriod
- 1980-2004- 1980-2004
• Boundary ConditionsBoundary Conditions
- NCEP/DOE reanalysis- NCEP/DOE reanalysis
• ResolutionResolution
- 50 km- 50 km
MM5MM5Iowa Iowa
State/State/PNNLPNNL
RegCM3RegCM3UC Santa CruzUC Santa Cruz
ICTPICTP
CRCMCRCMQuebec,Quebec,OuranosOuranos
HADRM3HADRM3Hadley CentreHadley Centre
RSMRSMScrippsScripps
WRFWRFNCAR/NCAR/PNNLPNNL
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Analysis RegionsAnalysis Regions
Comparison with 0.5Comparison with 0.5oo gridded observations from Univ. Delaware gridded observations from Univ. Delaware
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
1. Means & Variability1. Means & Variability
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Temperature Bias - DJFTemperature Bias - DJFTemp.Temp.BiasBias
DJFDJF-10 10
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Temperature BiasTemperature BiasTemp.Temp.BiasBias
JJAJJA-10 10
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Interannual Variance
0.45 2.23
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Temperature BiasTemperature Bias
Interannual Variance
0.45 2.23
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Temperature BiasTemperature Bias
-60 60
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Temperature BiasTemperature Bias
-60 60
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Pattern CorrelationPattern Correlation
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Regional Regional Annual Annual CyclesCycles
-10
35
0
6
-10
35
35
35
- 5
- 5 0
8
0
0
6
6
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Regional Regional Annual Annual CyclesCycles(bias)(bias)
-10
35
0
6
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
TemperatureTemperature--
PrecipitationPrecipitationCorrelationsCorrelations
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
0
3
6
9
12
15
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
mm
/day
RCM3 MM5IECPC CRCMWRFP HRM3Observed (GPCC) Observed (UDEL)Observed (CRUT) Ensemble
1982-83 El Nino
1997-98 El Nino
multi-year drought
Monthly time series of precipitation in coastal California
small spread, high skillSubstantial annual cycle
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Correlation with Observed Precipitation - Coastal California
Model Correlation
HadRM3 0.857
RegCM3 0.916
MM5 0.925
RSM 0.945
CRCM 0.946
WRF 0.918
Ensemble 0.947
Ensemble mean has a higher correlation than any model
All models have high correlations with observed monthly time series of precipitation.
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Monthly Time Series - Deep South
Model Correlation
HadRM3 0.489
RegCM3 0.231
MM5 0.343
RSM 0.649
CRCM 0.649
WRF 0.513
Ensemble 0.640
RSM+CRCM 0.727
A “mini ensemble” of RSM and CRCM performs best in this region.
0
3
6
9
12
15
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
mm
/da
y
RCM3 MM5I ECPCCRCM WRFP HRM3Observed (GPCC) Observed (UDEL) Observed (CRUT)Ensemble
Ensemble (black curve)
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
2. Precipitation Extremes - Daily2. Precipitation Extremes - Daily
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Region AnalyzedRegion Analyzed
Boreal forest
Pacific coast
California coast
Great LakesMaritimes
Upper Mississippi
River
Deep South
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Precipitation Frequency vs. IntensityPrecipitation Frequency vs. Intensity
99.5%
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
99.5%
Precipitation Frequency vs. IntensityPrecipitation Frequency vs. IntensityNARCCAP - JJA
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Precipitation Frequency vs. IntensityNARCCAP - JJA
(OBS = Co-op Stations)
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Precipitation Frequency vs. IntensityPrecipitation Frequency vs. Intensity
99.5%
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Days with Simultaneous Extremes on Days with Simultaneous Extremes on ““NN”” Grid Points Grid Points
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Days with Simultaneous Extremes on Days with Simultaneous Extremes on ““NN”” Grid Points Grid Points
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Composite Structure of Extreme Events - DJF
NARRm/s
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Composite Structure of Extreme Events - DJF
NARR
MM5
GFDL
CCSM + 9 hr
NARCCAPNARCCAP
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Composite Structure of Extreme Events - DJF
MM5
GFDL
CCSMNARR 2-m specific humidity anomaly
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Composite Structure of Extreme Events - DJF
NARCCAPNARCCAP
NARR
GFDL
HadRM3
CCSM + 9 hr2-m specific humidity anomaly
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Thank You!Thank You!
(www.narccap.ucar.edu)(www.narccap.ucar.edu)
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Higher resolution is necessary, but not sufficient, for Higher resolution is necessary, but not sufficient, for simulating short-term (e.g., daily) precipitation simulating short-term (e.g., daily) precipitation extremes.extremes.
Coarser models (and nudged regional models) tend to Coarser models (and nudged regional models) tend to have daily extremes covering a wider area than have daily extremes covering a wider area than observed extremes. observed extremes.
Focusing on environments conducive to extremes Focusing on environments conducive to extremes yields relevant climatic behavior, even in relatively yields relevant climatic behavior, even in relatively coarse models.coarse models. This conclusion rests on the assumption that important This conclusion rests on the assumption that important
small-scale features are not missing (e.g., low-level small-scale features are not missing (e.g., low-level jets).jets).
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Part I: Interannual Variability
• Results shown for 1981-2002Results shown for 1981-2002
• Comparison with 0.5Comparison with 0.5oo gridded gridded precipitation analysis from the University precipitation analysis from the University of Delawareof Delaware
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Precipitation analysis for two regionsPrecipitation analysis for two regions
Coastal California Deep
South
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Monthly Time Series - Deep South
Model Correlation
HadRM3 0.489
RegCM3 0.231
MM5 0.343
RSM 0.649
CRCM 0.649
WRF 0.513
Ensemble 0.640
Two models (RSM and CRCM) perform much better. These models inform the domain interior about the large scale.
0
3
6
9
12
15
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
mm
/da
y
RCM3 MM5I ECPCCRCM WRFP HRM3Observed (GPCC) Observed (UDEL) Observed (CRUT)Ensemble
Ensemble (black curve)
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Full ensembleFull ensemble RSM + Canadian RCMRSM + Canadian RCM
The "mini-ensemble" has better correlation than the full The "mini-ensemble" has better correlation than the full ensemble in the southern and eastern parts of the domain.ensemble in the southern and eastern parts of the domain.
Other measures of forecast skill (such as bias) are not Other measures of forecast skill (such as bias) are not necessarily better.necessarily better.
Correlation of Monthly Time SeriesCorrelation of Monthly Time SeriesCorrelation of Monthly Time SeriesCorrelation of Monthly Time Series
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Ensemble error and spread (January)Ensemble error and spread (January)
BiasBias Ensemble spreadEnsemble spread
There are hints of a spread-skill relation but it is There are hints of a spread-skill relation but it is not consistent.not consistent.
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
ensemble July minus Juneensemble July minus June observed July minus Juneobserved July minus June
The ensemble reproduces the dipole of June-The ensemble reproduces the dipole of June-July precipitation change, but the monsoon does July precipitation change, but the monsoon does not extend as far north as observed.not extend as far north as observed.
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Part 2: Extreme Monthly PrecipitationPart 2: Extreme Monthly Precipitation
• ObservationsObservations Precip: University of Washington VIC retrospective analysisPrecip: University of Washington VIC retrospective analysis
500 hPa Heights: North American Regional Reanalysis500 hPa Heights: North American Regional Reanalysis
• Comparison period: 1982 -1999Comparison period: 1982 -1999 1979-1981 omitted - spinup1979-1981 omitted - spinup UW data end in mid-2000UW data end in mid-2000
• AnalysisAnalysis Cold season (Oct-Mar)Cold season (Oct-Mar) 10 wettest months (top 10%)10 wettest months (top 10%)
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Regions AnalyzedRegions Analyzed
Boreal forest
Pacific coast
California coast
Deep South
Great LakesMaritimes
Upper Mississippi
River
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Frequency – Coastal CAFrequency – Coastal CA
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Ranked Precipitation – Coastal CARanked Precipitation – Coastal CA
Ensemble Ensemble average of average of top 10 top 10 = 9% smaller = 9% smaller than UWthan UW
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Interannual Variability – Coastal CAInterannual Variability – Coastal CA
59 of 60 (98%) 59 of 60 (98%) simulated simulated extremes occur extremes occur in cold seasons in cold seasons with an with an observed observed extreme.extreme.
(random (random chance: 27) chance: 27)
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Composite Composite 500 hPa Height 500 hPa Height Anomalies Anomalies
Top 10 ExtremesTop 10 Extremes
Coastal CACoastal CA
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Frequency – Deep SouthFrequency – Deep South
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Ranked Precipitation – Deep SouthRanked Precipitation – Deep South
Ensemble Ensemble average of average of top 10 top 10 = 22% smaller = 22% smaller than UWthan UW
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Interannual Variability – Deep SouthInterannual Variability – Deep South
27 of 60 (45%) 27 of 60 (45%) simulated simulated extremes occur extremes occur in cold seasons in cold seasons with an with an observed observed extreme.extreme.
(random (random chance: 27) chance: 27)
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
500 hPa Height Anomalies – 500 hPa Height Anomalies – Deep South ExtremeDeep South Extreme
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Summary Summary Monthly PrecipitationMonthly Precipitation
Where there is a substantial periodic cycle:Where there is a substantial periodic cycle:- Models simulate well the interannual variability- Models simulate well the interannual variability- Models simulate well monthly, regional extremes- Models simulate well monthly, regional extremes
Where there is no substantial periodic cycle:Where there is no substantial periodic cycle:- Models simulate poorly the interannual var. & - Models simulate poorly the interannual var. & extremesextremes- Interior nudging improves interannual variability- Interior nudging improves interannual variability- Interior nudging does not help extremes- Interior nudging does not help extremes
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Hydrologic Analysis (Takle et al.)Hydrologic Analysis (Takle et al.)
SWAT model domainSimulation period: last 2 decades of 20C
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Hydrologic Analysis (Takle et al.)Hydrologic Analysis (Takle et al.)
Streamflow Interannual Variability
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Hydrologic Analysis (Takle et al.)Hydrologic Analysis (Takle et al.)
Precipitation Annual Cycle
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Hydrologic Analysis (Takle et al.)Hydrologic Analysis (Takle et al.)
Streamflow Annual Cycle
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Summary Summary MONTHLY PRECIPITATIONMONTHLY PRECIPITATIONWhere there is a substantial periodic cycle:Where there is a substantial periodic cycle:- Models simulate well the interannual variability- Models simulate well the interannual variability- Models simulate well monthly, regional extremes- Models simulate well monthly, regional extremes
Where there is no substantial periodic cycle:Where there is no substantial periodic cycle:- Models simulate poorly the interannual var. & - Models simulate poorly the interannual var. & extremesextremes- Interior nudging improves interannual variability- Interior nudging improves interannual variability-Interior nudging does not help extremesInterior nudging does not help extremes
UPPER MISSISSIPPI STREAMFLOWUPPER MISSISSIPPI STREAMFLOWEnsemble replicates well the interannual variabilityEnsemble replicates well the interannual variabilityAnnual cycle simulated less wellAnnual cycle simulated less well
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
2b. Precipitation Extremes - Monthly2b. Precipitation Extremes - Monthly
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Ranked Precipitation – Coastal CARanked Precipitation – Coastal CA
Ensemble Ensemble average of average of top 10 top 10 = 9% smaller = 9% smaller than UWthan UW
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Ranked Precipitation – Upper MSRanked Precipitation – Upper MS
Ensemble Ensemble average of average of top 10 top 10 = 6% smaller = 6% smaller than UWthan UW
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Ranked Precipitation – Deep SouthRanked Precipitation – Deep South
Ensemble Ensemble average of average of top 10 top 10 = 22% smaller = 22% smaller than UWthan UW
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Regional Regional Annual Annual CyclesCycles
-10
35
0
6
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Temperature BiasTemperature Bias
Interannual Variance
0.45 2.23
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Composite Structure of Extreme Events - DJFComposite Structure of Extreme Events - DJF
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Composite Structure of Extreme Events - DJFComposite Structure of Extreme Events - DJF
NARCCAP Users MeetingNARCCAP Users Meeting April 2011April 2011
Composite Structure of Extreme Events - DJFComposite Structure of Extreme Events - DJF