In%20this%20study%20we%20have%20applied%20sea%20surfaces%20temperatures%20(SSTs)%20derived%20from%20remote%20sensing%20and%20Global%20Climate%20Models%20(GCMs)%20projections%20to%20examine%20future%20potential%20distribution%20of%20Chondrus%20crispus,%20Irish%20Moss.%20The%20geographic%20range%20of%20organisms%20is%20related

1
Behind the Map Predicting Marine Species
Habitat Change Using Global Climate Models
Faculty of Science Undergraduate Research
Conference 1st Prize Earth System Sciences
Elizabeth Flanary and Sarah Vereault Department
of Geography, McGill University Supervisor Dr.
Gail Chmura Department of Geography, McGill
University
Data and Methodology SSTs were downloaded from
the Intergovernmental Panel on Climate Change
(http//www.ipcc.ch) for four GCMs (Table 1),
encompassing the years 1961-1999 , and 2079-2099.
The years 1961-1999 were used as the baseline,
and the years 2079-2099 were selected as the
future period because that is when 4C
atmospheric warming is supposed to occur. Only
the A2 scenario, which predicts the highest
increase in emissions, was included to assess
maximum potential change (Nikicenovic, 2000). We
generated change fields (increase or decrease in
SSTs) by subtracting the baseline (average of
1961 -1999) from the average of the future time
period.
Introduction In this study we have applied sea
surfaces temperatures (SSTs) derived from remote
sensing and Global Climate Models (GCMs)
projections to examine future potential
distribution of Chondrus crispus, Irish Moss. The
geographic range of organisms is related to
climate, particularly on continental scales.
Temperature can be a useful first indicator to
habitat suitability (Pearson and Dawson, 2003).
Remote sensing uses satellites to obtain data
that may otherwise be unobtainable on the surface
of the earth. This is a benefit in oceanographic
studies, as the vast extent of oceans are
difficult to sample. The National Oceanic and
Atmospheric Administration Advanced Very
High-Resolution Radiometer (NOAA AVHRR) uses the
thermal infrared channels (? 3.0 1000 ?m) to
remotely sense SST. To address yearly
variability, for example the North Atlantic
Oscillation (NAO), we took the average monthly
temperatures of the NOAA AVHRR temperatures over
the past 12 yr. (The NAO is a cycle of warming
and cooling caused by reversal of pressure that
affects waters and weather of Europe and the east
coast of North America (Ahrens, 2003)). GCM
simulations differ on regional trends, and on
climate variables, but all predict that the
global average surface air temperature will
increase by 4C in the next 100 yr (Cubasch et
al., 2001). The extent of climate change
predicted is dependent on the emissions of
atmospheric greenhouse gases and aerosols,
largely influenced by carbon dioxide (Environment
Canada, 2004). In addition, treatment of surface
conditions, such as ocean currents and
circulation, and wind patterns vary amongst
models. GCMs are often produced on a coarse
resolution grid, and need to be downscaled in
order to be evaluated at the finer resolution
needed for impact studies. Direct interpolation
is one method that has been implemented in
studies that predict species habitat change
(e.g., Bartlein et al., 1997).


Figure 7. GFDL prediction for Feb SST,
2079-2099.
Figure 1. GFDL grid in vector format.
Figure 3. GFDL changefield for February.
Figure 5. February AVHRR SST.
Equation 1. Formula for Inverse Distance
Weighted Interpolation. z(u0) estimated SST at
unknown points z(ui) known data points dij
the distance between each data point and the
unknown point. p power source Lo, and Yeung,
2002.
Table 1. Models used in analysis.


Model Centre Spatial Resolution Spatial Resolution
Model Centre lat x long km
CGCM2 Canadian Centre for Climate Modeling and Analysis Canada 3.75 x 3.75 416 x 281
CCSR/NIES AGCM CCSR OGCM Centre for Climate System Research Japan 5.6 x 5.6 622 x 420
CSIRO Mk2 Commonwealth Scientific and Industrial Research Organization Australia 3.2 x 5.6 355 x 420
GFDL R30 C Geophysical Fluid Dynamics Laboratory United States 2.25 x 3.75 250 x 281
Figure 2. CCSR grid in vector format.
Figure 8. CCSR prediction for Aug SST,
2079-2099.
Figure 4. CCSR changefield for August.
Figure 6. August AVHRR SST.
The data was imported into ERSI ArcGIS 9.0, and a
vector layer was generated which showed
temperature change at spatially referenced points
on a global scale. They were assigned the
Geographic Coordinate System WGS 1984, to
correspond with the NOAA AVHRR data. The data was
trimmed to include only points that occurred in
the NW Atlantic, between 25 and 65N, and 30 and
80W. The visual presentation of the data reveals
the breadth of spatial resolutions (Figs 1 2).
NOAA AVHRR SST data was downloaded from the
Physical Oceanography Distributed Active Archive
Center (http//podaac-www. jpl.nasa.gov) in
pentad format. It had a spatial resolution of 9
km at the equator, and was global in extent. It
was imported into Idrisi 32 v.2, and clipped to
the NW Atlantic study area, then imported into
ArcGIS.
Interpolation is the process by which known data
values and mathematical equations are used to
accurately estimate values at unknown locations.
Inverse distance weighted (IDW) interpolation
supposes that the value of an unknown point is
inversely proportional to a power of its distance
to known points (Equation 1). IDW was used to
interpolate the change fields, using the distance
squared, and a radius of 8 nearest neighbours.
This follows Toblers Law that things that are
closer together are more closely related. This
process produces a smooth surface.
The GCM change fields were added to the NOAA
AVHRR data to create the final output showing
predicted future sea surface temperatures.
Application of GCM SST data Chondrus crispus, or
Irish moss, is a red algae and a source of
carrageenan, commonly used as a thickener and
stabilizer for processed foods such as ice cream
and luncheon meats. It is also used in cosmetics
to soften skin and as an herbal supplement.
Historically there has been a large Irish moss
industry on the South Shore of Massachusetts and
along the Maine coast (http//www. purplesage.org.
uk/profiles/irishmoss.htm).
Conclusion Despite variations in magnitude and
some local cooling effects, all of the GCMs
examined predict that over time the average sea
surface temperature will rise. Changes in
biogeographic ranges due to rising SSTs could
result in the depletion of other species that are
economically valuable in certain areas. As well,
the disruption in habitat and food webs could
have far-reaching ecological consequences.   The
implications from this could have noticeable
effects on marine flora and fauna, local
communities connected to the sea, and the global
community as well.
Discussion When examining and applying the data
and results it is important to keep in mind the
original scale (Table 1). Since the data was
obtained at a very coarse resolution the
temperature predictions are likely not viable at
a local scale. A large portion of the data was
generated through interpolation, and so is not
exactly known. However, at a regional level
predictions from GCMs can be used to evaluate
general trends. Other methods of downscaling
data to a finer resolution include statistical
and dynamical downscaling. However, in a study
using GCMs the results obtained from a dynamical
coupling of a GCM and finer resolution Regional
Climate Model showed no significant difference
from direct interpolation of the GCM (Murphy,
2000). In regard to the distribution of marine
species, temperature is not the only factor that
determines suitable habitat. Salinity, substrate,
nutrient availability, and interaction with other
species were not included in this study, and may
be equally useful in predicting distribution
change.
Figure 11. Future distribution of C. crispus
according to CCCMA.
Figure 13. Change in the distribution of C.
crispus according to CCCMA.
Figure 9. Chondrus crispus current known
distribution.
Figure 10. Distribution of waters -2.1 to 21.3C
annually.
Acknowledgements Funding for this project was
provided by the Climate Change Action Fund
(CCAF), the World Wildlife Fund (WWF), and the
Natural Sciences and Engineering Research Council
(NSERC). Results were generated using the Walter
Hitschfeld Geographic Information Centre
undergraduate labs GIS software and computers.
Thank you to Dr. Gerhard Pohle and Mr. Lou Van
Guelpen of the Atlantic Reference Centre (ARC),
Dr. Gail Chmura, Dr. Jonathan Seaquist, Mr.
Graham MacDonald, and Mr. Tim Horton.
A current thermal distribution layer was created,
which selected areas from the NOAA AVHRR data
that was bounded by the depth and temperature
restrictions. This layer corresponded with the
known distribution, and thus the temperature
bounds are likely a good predictor of habitable
areas for C. crispus. If the two distributions
did not correspond, and areas of acceptable
temperatures were not in the known distribution,
it is likely that other variables have a stronger
influence on species distribution, such as
substrate, salinity, food sources, or predators.
Chondrus crispus occurs on the NW Atlantic coast,
to 20 m water depth and between 40 and 60N, from
New Jersey to Labrador (Lee, 1977). The maximum
and minimum temperature extremes that occur in C.
crispus range are in August and February,
respectfully. From the NOAA AVHRR layer, data was
selected within the depth and latitude
parameters, and the corresponding temperature
extremes were recorded. The temperatures bounding
C. crispus range are 2.1 21.3C.
Figure 12. Future distribution of C. crispus
according to CCSR.
Figure 14. Change in the distribution of C.
crispus according to CCSR.
Works Cited Ahrens, C.D. 2003. Meteorology
Today, 7th Ed. Brooks/Cole-Thompson Learning
Pacific Grove, USA. Bartlein, P. J., Whitlock,
C., and S. L. Shafer. 1997. Future Climate in the
Yellowstone National Park and Its Potential
Impact on Vegetation. Conservation Biology 11
780-792. Cubasch U, Meehl GA, Boer GJ, Stouffer
M, Dix M, Noda A, Senior CA, Raper S, Yap KS
(2001) Projections of future climate change. Pp.
525-582 In Houghton JT, Ding Y, Griggs DJ, Noguer
M, VAN DER Linden PJ, Dai X, Maskell K, Johnson
CA (eds.) Climate change 2001 the scientific
basis, contribution of working group 1 to the
third assessment report of the intergovernmental
panel on climate change. Cambridge University
Press.Environment Canada. The first generation
coupled Global Climate Model (7/6/2004).
Retrieved 9/23/2005, from http//www.cccma.bc.ec.g
c.ca/models/cgcm1.shtml Irish Moss (N.D).
Retrieved 9/25/2005, from http//en.wikipedia.org/
wiki/Irish_moss Irish Moss (N.D.). Retrieved
9/25/2005, from http//www.purplesage.org.uk/profi
les/irishmoss.htm Lee, T. 1977. The seaweed
handbook an illustrated guide to seaweeds from
North Carolina to the Arctic, Mariners Press,
Boston. Lo, C. P., and A. K. W. Yeung. 2002.
Concepts and Techniques of Geographic Information
Systems. Prentice Hall, Upper Saddle River, NJ.
Murphy, James. 2000. Predictions of climate
change over Europe using statistical and
dynamical downscaling techniques. International
Journal of Climatology 20 489 501. Nakicenovic
N, Swart R (Eds.) (2001) Emissions Scenarios
2000, Special Report of the Intergovernmental
Panel on Climate Change. Cambridge University
Press, UK. 570 pp. Pearson, R. G., and T. P.
Dawson. 2003. Predicting the impacts of climate
change on the distribution of species are
bioclimate envelope models useful? Global Ecology
Biology 12 361-371.
To predict where C. crispus will be able to live
in the future, the bounding temperatures of 2.1
and 21.3C were found in layers of future SSTs
generated by the GCMs, in areas where the depth
was also within limits. Only areas that fell into
this range in both February and August were
selected as suitable.
A change image was produced by subtracting the
current thermal distribution layer from the
future thermal distribution layer using raster
calculations. Notably, all four models predict
retraction of the range of C. crispus in southern
New England, where Irish moss is currently
harvested. The degree to which the models predict
loss is variable, and some also predict loss in
northern Labrador, around Prince Edward Island
and northern Nova Scotia.