Simulation of Present-Day Climate in the Caribbean using PRECIS RCM

1
Simulation of Present-Day Climate in the
Caribbean using PRECIS RCM
M
Mr. Arnoldo Bezanilla Marlot
2
SIMULATION OF THE PRESENT-DAY CLIMATE IN THE
CARIBBEAN USING PRECIS RCM.
Abel Centella Arnoldo Bezanilla Israel
Borrajero Daniel Martinez Institute of
Meteorology, Cuba
2
Climate Change Impacts on the Caribbean, UWI
Mona, Jamaica, June 15,17-2007
3
SIMULATION OF THE PRESENT-DAY CLIMATE IN THE
CARIBBEAN USING PRECIS RCM.
Abel Centella Arnoldo Bezanilla Israel
Borrajero Daniel Martinez Institute of
Meteorology, Cuba
3
Climate Change Impacts on the Caribbean, UWI
Mona, Jamaica, June 15,17-2007
4
OUTLINE OF PRESENTATION

• What is PRECIS?
• Simulating Present-Day Climate in the Caribbean
  using PRECIS.
• How well does PRECIS represent the Caribbean
  Climate?
• PRECIS and Tropical Cyclones. Preliminary results
• PRECIS-CARIBE Online Access to PRECIS Output
• What have we done, future plans.
• Summary

5
Regional Climate Modeling (I)
The main goal of regional climate models
(RCMs) is to reproduce the main climatic features
in complex terrain, where mesoscale forcing
becomes important (Giorgi and Mearns, 1991) and
coarse-resolution global climate models (GCMs)
are not sufficient for assessing local climate
change (Aldrian et al. 2004).
6
Regional Climate Modeling (II)
An example of a region where
present-generation GCMs are especially lacking in
their ability to represent complex terrain and
land-sea contrasts is the Caribbean zone. In this
region, tropical and extra-tropical systems
interact (Alfonso and Naranjo, 1996) and
frequently produce complex meteorological
conditions. The sea-breeze circulation in islands
and peninsulas favors the development of
convective systems (Riehl, 1979). In such a
complex meteorological situation, a
high-resolution model is necessary for meaningful
regional climate prediction
7
PRECIS (Providing REgional Climates for Impact
Studies)

• A Regional Modelling System derived from Hadley
  Centre GCM
• Providing REgional Climates for Impact Studies
• PC-based regional climate model
• It can be set up and run over any area
• PRECIS is freely available
• It has two resolutions 25 km and 50 km
• PRECIS is a tool for dynamical downscaling

8
Simulating Present-Day Climate in the Caribbean
using PRECIS

• Quasi-observed climate PRECIS forced by ECWRF
  Reanalysis ERA15 (RCM ERA)
• PRECIS driven by HadAM3 control run (RCM CTRL)
• Climate Research Unit database was used as
  observed climate (CRU OBS)
• RCM ERA and RCM CTRL are identical apart from
  their driving data.

9
How well does PRECIS represent the Caribbean
Climate?

• RCM ERA and RCM CTRL were compared with CRU OBS
• RCM ERA and RCM CTRL biases were also compared
• Temperature and Precipitation for land areas only

PRECIS land sea mask
10
TEMPERATURERCM CTRL- CRU OBS
11
PRECIPITATIONRCM CTRL- CRU OBS
12
RCM CTRL- CRU OBSTemperature and Precipitation
seasonal cycle for Cuba

• Warm and dry biases are evident in RCM CTRL
  simulation
• Greater biases occur over Cuba and other islands
• The dry biases are more intense in summer

13
TEMPERATURERCM ERA- CRU OBS
14
PRECIPITATIONRCM ERA- CRU OBS
15
RCM ERA- CRU OBSTemperature and Precipitation
seasonal cycle for Cuba

• Warm and dry biases also exist in RCM ERA
  simulation
• Greater biases occur over Cuba and other islands
• The dry biases are more intense in summer

16
Partial conclusions

• There consistency beetwen CTRL and ERA biases
  suggest the RCM model dont capture very well
  some of the mesoscale process.
• The significant summer dry biases observed in
  both simulations motivate further analysis of
  convective process.
• Comparisons with GCM outputs are also needed to
  assess the improvement that PRECIS could produce.

17
The Frequency of tropical cyclones in the
Caribbean and Mexico as show in Regional Climate
Model simulations

• Cyclone representation is a potential tool for
• Investigate and assess the Model ability to
  represent this type of event.
• Investigate or project future TC behavior in the
  area

18
TCLVs detection methodology

• It is designed to work with daily mean fields
• On a grid of 0.44o (50 km), a point of minimum in
  surface pressure is sought so that the averaged
  pressure over a circumference of 6o (700 km)
  centered in the point is at least 5.5 hPa greater
  than in the point.
• The difference between the maximum and minimum
  values of the wind speed in a neighborhood of 3
  grid points radius (1.3o or 150 km) centered in
  the point of minimum pressure must be at least 40
  km/h (11 m/s).
• The end of the track of each individual vortex
  occurs when in two consecutive days, the points
  of two consecutive TCLV position are located at a
  mutual distance of more than 7o (800 km).
• The above criteria were adjusted in practice by
  trial and error based on a detailed visual
  analysis of selected animated image sequences in
  the output of the model.

19
S
E
W
N
20
Historical North Atlantic and East-Central North
Pacific Tropical Cyclone Tracks, 1851-2005
National Oceanic and Atmospheric Administration,
Tropical Prediction Center/National Hurricane
Center
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
Partial conclusions

• The projected increase in temperature in Eastern
  Pacific near the coast of Mexico and the smaller
  increase projected for Western Tropical Atlantic
  seem to be consistent with a greater projected
  increase of TCLVs for Eastern pacific.

25
The method allow us to identify possible change
in tropical cyclone activity, but

• More experiments are needed
• Add new fields to be used in the detection
  criteria.
• Specify the definition of TCLV based on daily
  means of
• surface parameters using the criteria defined
  by
• Walsh et al. for instant(hourly mean)
  observations in a
• limited run of the model.
• Adjust a theoretical Cyclone Model to obtain
  hourly
• values of main TCLV parameters from the mean
  daily
• fields.
• Validate the definition using the adjustment to
  climatologic
• data.
• Analyze the dependence of TCLVs on projected
• temperature and wind shear projections.

26
PRECIS (Sharing Results)
http//precis.insmet.cu/Precis-Caribe.ht
m PRECIS-CARIBE Online Access to Climate
Change Scenarios in the Caribbean.
27
PRECIS (Sharing Results)
28
PRECIS (Outputs)
Two Kind of format for Data Download NetCDF y
ASCII
29
PRECIS (Outputs)
Examples for small island countries
Grid results over Jamaica
Grid results over Haiti
30
PRECIS (What have we Done)
As part of a very prominent collaborative
initiative between 5C, INSMET and UWI Jamaica and
Barbados

• 15 Years Reanalysis (1979-1993)
• 2 B1 30 Years Baseline Ensembles
• (Control Experiments) 1960-1990 (S)
• 3 A2 30 Years de future scenarios 2070-2100 (S).
• B2 30 Years de future scenarios 2070-2100 (S).

31
PRECIS (At this moment upcoming work)

• ECHAM4 50km
• Control 1960-1990 31yrs
• A2 1990-2100 114 yrs
• B2 1990-2100 114 yrs
• ERA40 with sulphur cycle 50km
• Observed 1957-2002 45yrs
• SRES Control hourly with sulphur cycle 50km
• 1960-1969 10 yrs
• HadCM3 perturbed parameter ensemble members
  (50km)
• 1 High 1 Low 150 Years

32
PRECIS (upcoming work)

• HadCM3 perturbed parameter ensemble members
  (50km)
• 1 High 1 Low 150 Years

33
PRECIS (Possibilities)

• Obtain and Use our own scenarios for climate
  change
• To share the obtained results with all the
  scientific community, and stakeholders.
• The improvement of the Collaboration in the
  Caribbean
• Using the outputs to feed other numeric models
  (Hydrological, Crops models, etc )
• The investigation of the extreme events in the
  future (strength, duration, frequency, season
  shifting, etc )

34
PRECIS (Summary)

• Improvements on the Precis-Caribe Web Site,
  Languages, outputs added for another scenarios
  (HadCM3-B2) and another GCM (ECHAM4 A2 and B2
  scenarios).
• Define whether to make daily data available.
• Add some specific variables, another domain (high
  res Scenario for West Indias).
• Work more Deeply in the analysis and validation
  of the outputs. ERA40 Data.

35
Thanks !!!!
PRECIS Online Access http//precis.insmet.cu/Pre
cis-Caribe.htm Email precis.insmet_at_insmet.cu
Contacts PRECIS Group in Cuba Abel Centella
abel.centella_at_insmet.cu Arnoldo Bezanilla
arnoldo.bezanilla_at_insmet.cu Israel Borrajero
iborajero_at_met.inf.cu