Title: Model based climate projections for the western US and Utah
1Model based climate projections for the western
US and Utah
Thomas ReichlerDepartment of Atmospheric
Sciences, U. Utahthomas.reichler_at_utah.eduAcknow
ledgmentsJohn Horel Leigh Jones, U. Utah
2The IPCC-AR4 Report
- Projected Patterns of Precipitation Change
- Relative changes in precipitation (in ) for
20902099, relative to 19801999 - Multi-model averages based on the SRES A1B
scenario - Somewhat unclear situation for Utah
3- Background on climate modeling
- Statistical downscaling
- Dynamical downscaling
4What Exactly is a Model?
- Merriam-Webster
- A model is only a representation of reality
5Why modeling climate?
- Climate is far too complex to
- observe it consistently
- reproduce it in a laboratory
- understand it analytically
- Alternative
- devise a mathematical model that simulates the
key processes that govern climate - Model can be used to
- understand what cannot be directly observed
- perform experiments past, present, and future
- gain quantitative understanding for change
6Types of models
Statistical Dynamical Combined
- Statistical relationship established from past
data applied to future data - Reasonable as long as relationships are not
changing
- Explicitly calculate the evolution of the climate
system from first principles - Capture physical causes for relationships
- More reliable, in particular if relationships are
changing -
- Statistical/dynamical downscaling
7Dynamical models
- Simulate the 3D evolution of atmospheric and
oceanic flow and key physical processes that
determine climate - Capture the interactions among its components and
important feedbacks - Examine statistical effects of the motions on
key-climate quantities such as precipitation and
temperature -
- Global models (GCMs)
- global coverage
- large scale 50-500 km
- Regional models (RCMs)
- usually nested into the output of a global model
- small scale 1-50 km
8Model Development
- Model components
- Atmosphere, ocean, sea ice, land ice, land
surface, biosphere, chemistry - Model development requires major resources
- Worldwide 24 models, 12 centers
- 4 US centers NSF (NCAR), NASA (GISS, GSFC), NOAA
(GFDL)
9Model Weather
High-resolution (0.5 x 0.5) atmosphere-only
simulation with the NOAA-GFDL model AM2.1
10Observed and Modeled Climate
- annual mean precipitation, 1979-1999
Observations (CMAP)
NOAA GFDL CM2.1
11Model Uncertainties
- Fundamentally,
- climate models are
- physics-based
- But certain approximations are unavoidable
- Well-known physics approximation due to the
numerical discretization of continuous equations - Empirically-known physics evaporation over water
- Unknown physics precipitation processes
- Unresolved physics due to limited model
resolution - These approximations and fits introduce
uncertainties into the predictions
12SRES Emission Scenarios
- Estimates of future emissions
- From population, energy, economics models
- Translated into greenhouse gas concentrations
- Adds to uncertainty
- A1B
- most commonly used
- CO2 stabilization at
- 700 ppm
- quite optimistic
- A2
- no stabilization
- CO2 exceeds 800 ppm
- more realistic
13How credible are climate models?
- Models are routinely validated by comparing them
against observations for present, recent, or past
climate - Reichler Kim (2008) Multivariate performance
index - Models have become demonstrably more realistic
- Multi-model mean (MMM) outperforms most models
14Model Uncertainties
- A model is always a simplification of the
complexity of nature - Some quotations
- All models are wrong, but some are useful (George
Box) - As simple as possible, but not simpler (Albert
Einstein) - Much of the uncertainties are related to the
coarse spatial model resolution - clouds, precipitation, turbulence,
atmosphere-ocean interaction - prediction of regional change
15Spatial Resolution
16How to Overcome Coarse Resolution?
- GCM at higher spatial resolution
- very expensive
- x2 resolution, x16 resources
- very clean
- Dynamical downscaling
- nest high resolution RCM into coarse resolution
GCM - expensive
- model uncertainties
- North American Regional Climate Change
Assessment Project (NARCCAP) - Statistical downscaling
- cheap
17- Background on climate modeling
- Statistical downscaling
- Dynamical downscaling
18Statistical Downscaling
- For present climate, establish a statistical
relationship betw. coarse model data (
predictor) and fine-scale observations (
predictand) - Correct model deficiencies by applying the
relationship, which was established for todays
climate, to model data for future climate (
downscaled) -
- Problematic assumption relationship is stationary
19High-resolution US Downscaling
- Lawrence Livermore National Laboratory (LLNL),
Bureau of Reclamation, and Santa Clara University
(SCU) - Statistics Wood et al. 2004, Maurer 2007
- US only 1/8 degree (ca. 12x12 km)
- Monthly mean precipitation and temperature,
1950-2099 - From sixteen GCMs (used for IPCC-AR4) and three
scenarios (A2, A1B, B1) - http//gdo-dcp.ucllnl.org/downscaled_cmip3_project
ions/
20Example US Precipitation
- annual climatology (mm/month)
Downscaled model data 1980-1999
PRISM (observations), 1971-2000
21Precipitation Change SRES A1B (A2)
- 20 year averages
- Reference 1990 (1980-1999)
- SRES A1B 2050 (2040-2059)
- SRES A2 2090 (2080-2099)
Winter Nov-Apr
Summer May-Oct
Multi-model means 16 models
22Precipitation Change
Nov-Apr
May-Oct
mm/month
Absolute
Relative
23Precipitation Change
Nov-Apr
May-Oct
mm/month
Absolute
of models with positive change
24Precipitation Change
Nov-Apr
May-Oct
mm/month
Absolute
Relative
25Precipitation Change
Nov-Apr
May-Oct
mm/month
Absolute
of models with positive change
26Seasonal Cycle Changes
- Northern vs. Southern Utah
Longitude Latitude
Northern Utah 114W-109W 39.5W-42N
Southern Utah 114W-109W 37W-39.5N
27Precipitation Change A1B
Northern Utah
Southern Utah
28Precipitation Change A2
Northern Utah
Southern Utah
29Temperature Change SRES A1B (A2)
30Temperature Change
Nov-Apr
May-Oct
C
1980-1999
C
2040-2059
31Temperature Change
Nov-Apr
May-Oct
C
1980-1999
C
Change
32Temperature Change
Nov-Apr
May-Oct
C
1980-1999
1C
C
2040-2059
1C
33Temperature Change
Nov-Apr
May-Oct
C
1980-1999
C
Change
34Temperature Change A1B
Northern Utah
Southern Utah
35Temperature Change A2
Northern Utah
Southern Utah
36- Background on climate modeling
- Statistical downscaling
- Dynamical downscaling
37Dynamical (Downscaling)
- North American Regional Climate Change Assessment
Program (NARCCAP) - www.narccap.ucar.edu
- Not yet completed
- Preliminary results GFDL AM2.1 (M180)
- 0.6x0.5, 50 km
- 1 model and 1 member only
- SRES A2
- 28 year averages 2039-2066 vs. 1969-1996
38Precipitation Change A2
Northern Utah
OBS
Southern Utah
39Precip. High vs. Low Resolution
- GFDL, A2, 2039-2066 minus 1969-1996, 1 member
Nov-Apr
May-Oct
mm/month
High-res
Low-res.
40Summary
41Precipitation Change
inches/month
mm/6 month
Absolute
Relative ()
15.75 14.43 13.12 11.8 10.49 9.18 7.87 6.56
13
6
5
14
42Temperature Change
4.5
C
3.4
2.0
1.6
43Impact on Water Supply
- Water supply and the availability of water is
controlled by - precipitation trends
- variability trends and timing of precipitation
(seasonal cycle, extreme events) - temperature trends (evapotranspiration, snow
fraction, snow melt) - Increasing temperatures will negatively impact
water supply for the region - What will dominate? Increase in temperature or
increase in precipitation.
44Cautionary Note
- These results represent probably the best
available CC information for Utah - However, they still contain large uncertainties
due to - model errors (particularly large for
precipitation) - assumption of stationarity for statistical
downscaling - emission scenario uncertainties
- Nevertheless, almost all models indicate robust
- 5-10 increases in precipitation for Northern
Utah in winter - 5-10 decreases in precipitation for Northern
Utah in summer - increases in temperatures by 3F in winter and
7F in summer, which will contribute to shortage
of water
45Thank You
46(No Transcript)
47For A1B scenario during 21st century, average of
ca. 20 models. globally dT0.7 1.2 1.8 2.3 2.9
C, dP 1 1.7 2.6 3.7 4.6 (2000-2019,2020-2039,
2040-2059, 2080-2099) From Giorgi and Bi
(2005) Updated regional precipitation and
temperature changes for the 21st century from
ensembles of recent AOGCM simulations, GRL.
48Precipitation
May-Oct
Nov-Apr
A1B
A2
49Precipitation
May-Oct
Nov-Apr
A1B
A2
50Temperature
May-Oct
Nov-Apr
A1B
A2
51Temperature
May-Oct
Nov-Apr
A1B
A2
52Precipitation GFDL AM2.1 (M180)
- A2, 2039-2066 minus 1969-1996
Nov-Apr
May-Oct
mm/month
Absolute
Relative