|
       
| |
Climatic data for the United States reveal changes and variations that may be
significant in redistributing vector-borne and water-borne diseases, as well as direct
climate-induced human morbidity and mortality. Since the turn of the century average daily
temperatures in the conterminous United States have increased by approximately 0.4 degrees
C, with most of this increase occurring during the past 30 years (Karl et al. 1995b).
Recent studies have shown that the hydrologic cycle in the US is changing as indicated by
increases in cloud cover (Karl and Steurer 1990) and precipitation (Groisman and
Easterling 1994) and decreases in pan evaporation (Peterson 1996). Extremes in US
precipitation have been changing with increases in heavy precipitation events and
decreases in lighter precipitation events (Karl et al. 1995a; Karl et al. 1996). Using
data back to 1910, Karl et al. found that the most recent 15 years had the highest values
of Greenhouse Climate Response Index (GCRI) as well as Climate Extremes Index (CEI). It is
becoming increasingly apparent that measurable changes in climate trends are occurring in
the US.
NCDC contains the largest archive of climate and climate-related data in the
world. These data, to be purchased from NCDC, will be used to refine our current
understanding of the relation between infectious disease and weather stress on human
health. Through collaboration across disciplines, existing infectious disease models, such
as USDA's models for tick-borne disease and mosquito-borne dengue fever, will be refined
and new models developed to study water-borne cholera and cryptosporidiosis. Climate data
currently available include the Summary of the Day, Co-Operative data set (TD-3200) which
includes daily values of maximum and minimum temperature and total precipitation for the
entire US Cooperative Observing Network (over 7000 stations). Data are also available from
the Summary of the Day, First Order data set (TD-3210) which includes, in addition to
temperature and precipitation, variables such as humidity, wind speed and direction, and
cloud cover for approximately 400 stations in the US, Caribbean, and Pacific islands. The
Comprehensive Ocean-Atmosphere Data Set (COADS) contains observations of sea-surface
temperature (SSTs), wind, and temperature taken from ships-of-opportunity for the entire
globe. Satellite data (e.g., NOAA Advanced Very High Resolution Radiometer - AVHRR) can be
accessed to provide information on land surface vegetation, ocean temperatures and
inferred circulation, and Expendable Bathythermograph (XBT) data and hydrographic data can
be obtained to determine anomalies for salinity.
Several methods are available to develop climate change scenarios suitable for
use in climate-driven models of disease incidence and spread. These methods include: 1)
direct use of climate model output from a General Circulation Model (GCM) or from regional
climate models nested into a GCM; 2) empirical methods that take existing observed data (
e.g. observations of temperature and precipitation for a given observing station) and
perform a geographical or temporal shift of the data; and 3) the downscaling approach that
combines the use of a GCM simulation and observational data (Robinson and Finkelstein
1989). Each approach has advantages and drawbacks; for example, empirical methods have the
advantage of being tied directly to the existing surface climate, while direct use of the
GCM simulation itself does not provide the spatial resolution necessary to drive impact
models, and assumes statistical distributions under climate change that may not be valid.
Statistical downscaling is an attractive option that can: 1) link a GCM simulation to the
observed statistical distribution for a given observing station using the parameters from
the GCM that are well simulated by the model (Karl et al. 1990); and, 2) develop climate
scenarios using transient GCM simulations (Easterling 1996). Our choice is to purchase
downscaled climate scenarios developed at NCDC using coupled Ocean Atmosphere General
Circulation Models which provide internally consistent scenarios of climate and the
associated ensemble of weather events during the next century. Where appropriate, we also
will purchase statistical simulations developed at NCDC using weather generators and
simple Autoregressive Moving Average (ARMA) models. Any climate scenario, regardless of
the method used, must be considered solely as a plausible outcome of potential climate
change. GCMs are physically-based models of the climate system, so any GCM simulation is
subject to the uncertainties inherent in any non-linear system, particularly one that is
forced by such factors as increased greenhouse gases, and tropospheric aerosols.
| |
|
