Heat-Related Health Effects in North Carolina Counties

Representatives of the North Carolina Department of Health and Human Services are concerned that residents of rural counties in North Carolina might be more susceptible to heat related impacts compared to residents in metropolitan areas. They would like to investigate in more detail the issue about morbidity in rural compared to urban areas and its relationship to climate and weather conditions. In recent research (Rhea et al. 2012) the authors indicate that people involved in outdoor activities such as agricultural, construction or landscaping work, as well as in organized sport activities are most susceptible to heat related illnesses.

To do an impact assessment of climate and weather impacts on applications the first step is to extract Global Climate Model (GCM) data in order to obtain background information about the area of interest. This use case involves extraction of GCM data from several models for the state of North Carolina as well as for several counties using OCGIS. The information for the state of North Carolina will serve as a context information in support of the subsequent analysis of downscaled data, while the data for the counties is for a simple comparison to the subsequently obtained downscaled data for these counties, since the current generation of GCMs does not provide detailed information on a scale smaller than several grid points. The two counties that were chosen initially were: Wake (transitions to coastal plain, contains Raleigh and surrounding area) and Johnston (coastal plain, located to SE of Wake – contains Southfield). In addition, an epidemiologist from the NC Dept. of Health and Human Services recommended Bladen and Robeson county (large native population) – the majority of the county land is used for agriculture; as well as, Scotland and Halifax county (no large cities are located in these counties).

Goal of the use case:  Given the shapefiles for those 6 counties and access to monthly CMIP5 datasets from several models, to write an OCGIS script that extracts temperature and precipitation time series for each county and outputs it in NetCDF and CSV format. These files would then be ready for further analysis.

A US Counties shapefile is available here (use Wake, Johnston, Bladen, Robeson, Scotland and Halifax in North Carolina): https://dl.dropboxusercontent.com/u/867854/us_counties.zip.

Location of staged CMIP5 data (thanks to Joe Barsugli): ftp://ftp.cdc.noaa.gov/Public/jbarsugli/CMIP5_for_Will/

The data are pre-staged on an ftp site and include:

• monthly average of daily average surface air temperature
• monthly average of precipitation rate 

The pre-staged data represent three modeling centers. There is an AOGCM and an Earth System Model from each center, making six models total. The first run of each model was chosen, for the historical and rcp85 scenarios. The historical data is 1860 (usually) - 2005. The rcp85 is 2006 - 2010. An article describing the CMIP5 experiments in more detail is here: http://journals.ametsoc.org/doi/pdf/10.1175/BAMS-D-11-00094.1.

These files were obtained from the ETH Zurich mirror of CMIP5 (https://wiki.c2sm.ethz.ch/Wiki/CMIP5) and have been processed at NOAA PSD from the original ESG files so that they are in a single file for each variable for the length of the experiment.

Reference:

Rhea S., Ising A., Fleischauer A.T., Deyneka L., Vaignah-Batten H. and A. Waller, 2012 – Using near real-time morbidity data to identify heat-related illness prevention strategies in North Carolina. J. Community Health 37: 495-500.