Atlantic Ocean

NC ECONet: Station Density Study

Analysis of Monitoring Station Density of the North Carolina Environment and Climate Observing System (NC ECONet)

A paper by Souleymane Fall
December 2003

 

Background   |   Objectives   |   Data   |   Method

Results   |   Conclusion   |   Acknowledgements   |   Appendix

 

Background

The North Carolina Environment and Climate Observing Network (NC ECONET) combines the automated weather network currently operated by the North Carolina State Climate Office with other stations maintained by the National Weather Service, the Federal Aviation Administration, and the US Natural Resources Conservation Service.

The NC ECONET stations provide real time information for emergency situations such as tornadoes, hurricanes or winter storms. Data for these stations are also provided to government's agencies for energy planning and natural resources management. In addition, data are available to the general public, for a variety of uses. The NC ECONET stations can also play a crucial role in the fight against bioterrorism.

Consequently, it is essential to have a good distribution of these stations throughout North Carolina.

 

Objectives

The objectives of this project are to characterize the distribution of NC ECONET stations with respect to elevation zones and to identify the areas in North Carolina where additional observing stations must be installed, for the best data set.

The following criteria were used:

•  Technical :

    • The distance between two stations should not exceed 50 km.
    • This distance is not uniform: it will vary according to the change in elevation (every 200 meters), given that climatic conditions such as temperature vary with respect to altitude.
    • The network has 8 different types of stations, but only AGNET, ASOS and AWOS stations, which have the most complete set of sensors, will be taken into account. The study will be conducted considering first these three sub-networks; then the same procedure will be followed, but only with AGNET and ASOS stations; finally the same analysis will be performed considering only AGNET stations, which constitute the backbone of the NC ECONET.
  • Geographic : Each county should have at least one station.

 

Data

The data for this project have been obtained from NCSU Libraries and the North Carolina State Climate Office. They consist of shapefiles and a digital elevation model (DEM):

  • econet_nc.shp
  • nc_counties
  • ncdem1000_ft

 

Method

The steps used to achieve the objective are shown in the chart below (Figure 1):

Figure 1 : Flowchart of the methods used in the density analysis

In summary, ArcGIS 8.1 and its extension Spatial Analyst were used.

A subset of the targeted stations is created. Using ArcGIS's Spatial Analyst, the DEM, originally in feet, is converted to meters. A text file specifying elevation zones with intervals of 200 meters is used to reclassify the DEM into elevation zones. The elevation zones, along with the location of stations, are shown in Figure 2.

Figure 2 : Elevation zones and stations used in density analysis: an elevation zone layer (interval 200 meters), with a transparency set at 50%, is displayed on top of a Hillshade layer.

 

A first buffer of 25 km is created around stations without dissolving barriers. This new layer, made transparent at 50%, is displayed on top of the elevations layer, to determine the number of elevation zones in each buffer area. Then the stations are classified with respect to the number of elevation zones in their buffer area. This number of zones will determine the actual buffer distance around each set of stations, based on the following equation:

Buffer Distance = 25 km - (S x 1 km) ........... (1)

Where S is the number of slices (elevation zones in a buffer area). For example, if in a buffer area there are 4 zones, the buffer distance will be: 25 km - 4 km = 21 km.

The different buffer layers are then merged and their boundaries dissolved to get one uniform layer. To identify the counties that are totally or partly covered by stations, this buffer layer is used to clip the NC counties layer. The resulting layer is grouped with the NC counties layer. This new layer represents, for each county:

•  Areas that are covered by NC ECONET stations

•  Areas that are not covered.

The buffer areas are selected, and then the selection is switched: the areas out of the buffer zone are now selected and exported as a new layer.

It's now possible to group the counties into categories (hint: zoom as large as necessary and use the Identify tool). Then using Select by Attributes, each group will be selected, exported as a shapefile, and given a proper symbology.

All new areas are computed using a pre-logic VBA code in the Field Calculator, and then combined with the total area of counties to compute the percentage of covered and non covered areas for each county. The resulting tables are joined with the Counties attribute table, and the counties are classified based on the percentage of coverage.

 

Results

The degree of coverage of counties in North Carolina by AGNET, ASOS, and AWOS sensors is shown in Figure 3. Based on the coverage of these 3 networks, there are 2 counties that are completely covered and 3 counties that are not covered at all (Table 1a and 1b).

 

 

 

Figure 3 : Coverage map when all 3 networks (AWOS, ASOS and AGNET) are included: a) spatial coverage; b) percentage of area covered.

 

 

Figure 4 shows the degree of coverage for AGNET and ASOS stations. These 2 networks are the only sensor suites that record a standard set of hourly measurements including precipitation amounts. Is this analysis, there are zero counties that are completely covered and 10 counties that are not at all covered (Tables 2a and 2b).

 

Figure 4 : Coverage map of AGNET and ASOS stations.
a) spatial coverage; b) percentage of area covered.

 

Coverage by only the AGNET stations is given in Figure 5. The AGNET sensor suite is the only station that measures soil temperature and soil moisture. In this analysis, no county is totally covered and 35 counties are not at all covered (Tables 3a and 3b).

 

Figure 5 : Coverage by AGNET stations.
a) spatial coverage; b) percentage of area covered.

 

 

Conclusion

On the whole, we can distinguish 4 groups of counties:

  • Counties that have at least 1 station and are totally covered
  • Counties that have at least 1 station but are not totally covered
  • Counties that don't have a station and are not covered by stations of neighboring counties
  • Counties that don't have a station but are partially covered by stations of neighboring counties

Western North Carolina, which is a mountainous region with many elevation zones, needs more coverage.

 

Acknowledgements

I gratefully acknowledge the reviewers of this study, Dr Sethu Raman, State Climatologist for North Carolina, and Ryan Boyles, Associate State Climatologist, for their comments and suggestions.

 

Appendix