Use of Regional Models in Impacts Assessments L. O. Mearns NCAR Colloquium on Climate and Health NCAR, Boulder, CO July 22, 2004

“Most GCMs neither incorporate nor provide information on scales smaller than a few hundred kilometers. The effective size or scale of the ecosystem on which climatic impacts actually occur is usually much smaller than this. We are therefore faced with the problem of estimating climate changes on a local scale from the essentially large-scale results of a GCM.” Gates (1985) “One major problem faced in applying GCM projections to regional impact assessments is the coarse spatial scale of the estimates.” Carter et al. (1994)

But, once we have more regional detail, what difference does it make in any given impacts assessment? What is the added value? Do we have more confidence in the more detailed results?

Elevation (meters) NCAR CSM Topography 2.8 deg. by 2.8 deg. RegCM Topography 0.5 deg. by 0.5 deg.

Resolutions Used in Climate Models High resolution global coupled ocean- atmosphere model simulations are not yet feasible (~ km) High resolution global atmospheric model simulations are feasible for time-slice experiments ~ km resolution for years (~ 100 km) Regional model simulations at resolution km are feasible for simulations years (~ 50 km)

Benefits of High Resolution Modeling Improves weather forecasts (e.g., Kalnay et al. 1998), down to to 10 km and improves seasonal climate forecasts, but more work is needed (Mitchell et al., Leung et al., 2002). Improves climate simulations of large scale conditions and provides greater regional detail potentially useful for climate change impact assessments Often improves simulation of extreme events such as precipitation and extreme phenomena (hurricanes).

Regional Climate Modeling Adapted from mesoscale research or weather forecast models. Boundary conditions are provided by large scale analyses or GCMs. At higher spatial resolutions, RCMs capture climate features related to regional forcings such as orography, lakes, complex coastlines, and heterogeneous land use. GCMs at 200 – 250 km resolution provide reasonable large scale conditions for downscaling.

Regional Modeling Strategy Nested regional modeling technique Global model provides: – initial conditions – soil moisture, sea surface temperatures, sea ice – lateral meteorological conditions (temperature, pressure, humidity) every 6-8 hours. –Large scale response to forcing (100s kms) Regional model provides finer scale response (10s kms)

RCM Nesting Technique

Regional Climate Model Schematic RotatedMercatorProjection GLCCVegetation Reanalysis & GCM Initial and Boundary Conditions Hadley & OI Sea Surface Temperatures USGSTopography

Agriculture: Brown et al., 2000 (Great Plains – U.S.) Guereña et al., 2001 (Spain) Mearns et al., 1998, 1999, 2000, 2001, 2003, 2004 (Great Plains, Southeast, and continental US) Carbone et al., 2003 (Southeast US) Doherty et al., 2003 (Southeast US) Tsvetsinskaya et al., 2003 (Southeast U.S.) Easterling et al., 2001, 2003 (Great Plains, Southeast) Thomson et al., 2001 (U.S. Pacific Northwest) Pona et al., (in Mearns, 2001) (Italy) Use of Regional Climate Model Results for Impacts Assessments

Use of RCM Results for Impacts Assessments 2 Water Resources: Leung and Wigmosta, 1999 (US Pacific Northwest) Stone et al., 2001, 2003 (Missouri River Basin) Arnell et al., 2003 (South Africa) Miller et al., 2003 (California) Wood et al., 2004 (Pacific Northwest) Forest Fires: Wotton et al., 1998 (Canada – Boreal Forest) Human Health: New York City Health Project (ongoing)

Examples of RCM Use in Climate and Impacts Studies Precipitation and Hydrology over S. Africa Water Resources in Pacific Northwest Agriculture - Southeast US Human Health – New York European Prudence Program New Program – NARCCAP

Regional Climate Modeling and Hydrological Impacts in Southern Africa Arnell et al., 2003, J. Geophys. Research

Arnell et al., 2003, J. of Geophys. Res.

Climate Simulations of Western U.S. Strong Effect of Terrain Model ability to resolve terrain features is critical Observed snow pack, March, 1998Observed mean annual precipitation Leung et al., 2004, Climatic Change (Jan.)

Elevation (meters) NCAR/DOE Topography 2.8 deg. by 2.8 deg. MM5 Topography 0.5 deg. by 0.5 deg.

Observed and Simulated El Nino Precipitation Anomaly RCM reproduces mesoscale features associated with ENSO events RCM SimulationObservationNCEP Reanalyses

Global and Regional Simulations of Snowpack GCM under-predicted and misplaced snow Regional SimulationGlobal Simulation

Climate Change Signals TemperaturePrecipitation PCM RCM

Extreme Precipitation/Snowpack Changes Lead to significant changes in streamflow affecting hydropower production, irrigation, flood control, and fish protection

Special Issue of Climatic Change (60:1-148) Issues in the Impacts of Climatic Variability and Change on Agriculture 1.Mearns, L. O., Introduction to the Special Issue on the Impacts of Climatic Variability and Change on Agriculture 2.Mearns, L. O., F. Giorgi, C. Shields, and L. McDaniel, Climate Scenarios for the Southeast US based on GCM and Regional Model Simulations. 3.Tsvetsinskaya, E., L. O. Mearns, T. Mavromatis, W. Gao, L. McDaniel, and M. Downton,The Effect of Spatial Scale of Climate Change Scenarios on Simulated Maize, Wheat, and Rice Production in the Southeastern United States. 4.Carbone, G., W. Kiechle, C. Locke, L. O. Mearns, and L. McDaniel, Response of Soybeans and Sorghum to Varying Spatial Scales of Climate Change Scenarios in the Southeastern United States. 5.Doherty, R. M., L. O. Mearns, R. J. Reddy, M. Downton, and L. McDaniel, A Sensitivity Study of the Impacts of Climate Change at Differing Spatial Scales on Cotton Production in the SE USA. 6.Adams, R. M., B. A. McCarl, and L. O. Mearns, The Economic Effects of Spatial Scale of Climate Scenarios: An Example From U. S. Agriculture.

Models Employed Commonwealth Scientific and Industrial Research Organization (CSIRO) GCM – Mark 2 version Spectral general circulation model Rhomboidal 21 truncation (3.2  x 5.6  ); 9 vertical levels Coupled to mixed layer ocean (50 m) 30 years control and doubled CO  runs NCAR RegCM2 50 km grid point spacing, 14 vertical levels Domain covering southeastern U.S. 5 year control run 5 year doubled CO  runs

Domain of RegCM denotes study area + denotes RegCM Grid Point (~ 0.5 o ) X denotes CSIRO Grid Point (3.2 o lat. 5.6 o long)

RegCM Topography (meters) Contour from 100 to 4000 by 100 (x1)

Summer Fall Maximum Temperature Minimum Temperature CSIRORegCMCSIRORegCM 5.00 to to 4.00 to to to to to to to 1.00 Climate Change - Δ Temperature ( o C)

% Change in Corn Yields

Regionalization of the climate change scenarios matters in terms of the economic indicators of the ASM Shows up in aggregate economic welfare (different orders of magnitude); Regional patterns of agricultural production are altered; -more spatial variability with RegCM; -Southern states are more negatively affected by RegCM. Conclusions

The contrast in economic net welfare based on spatial scale of climate scenarios is similar in magnitude to the economic contrast resulting from use of two very different AOGCM simulations in the US National Assessment. Conclusions (Con’t.)

Modeling the Impact of Global Climate and Regional Land Use Change on Regional Climate and Air Quality over the Northeastern United States C. Hogrefe, J.-Y. Ku, K. Civerolo, J. Biswas, B. Lynn, D. Werth, R. Avissar, C. Rosenzweig, R. Goldberg, C. Small, W.D. Solecki, S. Gaffin, T. Holloway, J. Rosenthal, K. Knowlton, and P.L. Kinney This project is supported by the U.S. Environmental Projection Agency under STAR grant R

NY Climate & Health Project: Project Components Model Global Climate Model and Evaluate Land Use Model Regional Climate Model Regional Air Pollution (ozone, PM 2.5 ) Evaluate Health Impacts (heat, air pollution) –For 2020s, 2050s, and 2080s

Model Setup GISS coupled global ocean/atmosphere model driven by IPCC greenhouse gas scenarios (“A2” high CO 2 scenario presented here) MM5 regional climate model takes initial and boundary conditions from GISS GCM MM5 is run on 2 nested domains of 108km and 36km over the U.S. CMAQ is run at 36km to simulate ozone 1996 U.S. Emissions processed by SMOKE and – for some simulations - scaled by IPCC scenarios Simulations periods : June – August June – August

A Model Look Into the 2050’s How will modeled temperature and ozone in the northeastern U.S. change under the “A2” (high CO 2 growth) scenario (assume constant VOC and NO x emissions)? How will CMAQ ozone predictions change when IPCC “A2” projected changes in ozone precursor emissions (VOC+8%, NO x +29.5%) are included in the simulation?

Research questions Health Risk Assessment: –Deaths due to short-term heat exposures –Hospital admissions due to short-term ozone exposures Development of model linkages Scale Intercomparison: as we go from 108 -> 36 -> 4km scale, what difference do we see in impact estimates? How do model results compare at different scales for NY metro region.

Global Climate Model NASA-GISS Land Use / Land Cover SLEUTH, Remote Sensing Regional Climate ClimRAMS MM5 Air Quality MODELS-3 Public Health Risk Assessment reflectance; stomatal resistance; surface roughness heat Ozone PM2.5 meteorological variables: temp., humidity, etc. meteorological variables IPCC A2, B2 Scenarios

Daily Maximum O 3 Predictions July , 1996

MM5 Current Climate

GCM and RCM Projections

Tests with 12 and 4 km Resolution

RCMs and Simulation of Extremes Do they do better?

1993 Midwest Summer Flood USHCN Observations RegCM Record high rainfall (>200 year event) Thousands homeless 48 deaths $15-20 billion in Damage J. Pal

1988 Great North American Drought CRU Observations RegCM Driest/warmest since 1936 $30 billion in Agricultural Damage

Putting spatial resolution in the context of other uncertainties Must consider the other major uncertainties regarding future climate in addition to the issue of spatial scale – what is the relative importance of uncertainty due to spatial scale? These include: –Specifying alternative future emissions of ghgs and aerosols –Modeling the global climate response to the forcings (i.e., differences among GCMs)

PRUDENCE Project Multiple AOGCMs and RCMs over Europe: Simulations of Future Climate

Summary of RegCM3 Results for A2 and B2 scenarios Nested in HADAM3 time-slice RegCM3 – 50 km HadAM3 time slice – 100 km Years – vs –2099 Control run results Changes in Climate Giorgi et al., 2004

Emissions Scenarios CO 2 Emissions (Gt C) CO 2 Concentrations (ppm) A2 B2

Map of Domain & Topography

Winter Precipitation: Reference Simulation DJF CRU DJF RegCMDJF HadAMH

Summer Surface Air Temperature: Reference Simulation JJA CRU JJA RegCMJJA HadAMH

Summer Precipitation: Reference Simulation JJA CRU JJA RegCM JJA HadAMH

Winter Temperature Change: B2 & A2 Scenarios DJF HadAMH: B2 DJF RegCM: B2 DJF RegCM: A2DJF HadAMH: A2 WARM HOT WARM

Summer Temperature Change: B2 & A2 Scenarios JJA HadAMH: B2 JJA RegCM: B2 JJA RegCM: A2JJA HadAMH: A2 WARM HOT WARM

Summer Precipitation Change: B2 & A2 Scenarios DRY WET JJA HadAMH: B2 JJA RegCM: B2 JJA RegCM: A2JJA HadAMH: A2 DRY WET

NARCCAP North American Regional Climate Change Assessment Program Multiple AOGCM and RCM Climate Scenarios Project over North America

Participants Linda O. Mearns, National Center for Atmosheric Research, Ray Arritt, Iowa State, George Boer, CCCma, Daniel Caya, OURANOS, Phil Duffy, LLNL, Filippo Giorgi, Abdus Salam ICTP, William Gutowski, Iowa State, Isaac Held, GFDL, Richard Jones, Hadley Centre, Rene Laprise, UQAM, Ruby Leung, PNNL, Jeremy Pal, ICTP, John Roads, Scripps, Lisa Sloan, UC Santa Cruz, Ron Stouffer, GFDL, Gene Takle, Iowa State, Warren Washington, NCAR, Francis Zwiers, CCCma

Main NARCCAP Goals Exploration of multiple uncertainties in regional model and global climate model regional projections Development of multiple high resolution regional climate scenarios for use in impacts models

NARCCAP domain

A2 Emissions Scenario GFDLCCSM HADAM3 link to European Prudence CGCM current future Provide boundary conditions MM5 Iowa State/ PNNL RegCM3 UC Santa Cruz ICTP CRCM Quebec, Ouranos HADRM3 Hadley Centre RSM Scripps WRF NCAR/ PNNL NARCCAP PLAN

A2 Emissions Scenario GFDL AOGCM CCSM Global Time Slice / RCM Comparison at same resolution (50km) Six RCMS 50 km GFDL Time slice 50 km CAM3 Time slice 50km compare

When to Use High Resolution Consider the importance of regional detail compared to other uncertainties in project High resolution useful when there are high resolution forcings: complex topography, complex coastlines, islands, heterogeneous land- use Consider also statistical downscaling (Wilby) More guidance on web at: www. ipcc-ddc.cru.uea.ac.uk