Spatial Interpolation Methods for PM2'5 Species Data

1
Spatial Interpolation Methods for PM2.5 Species
Data

• Neil Deardorff
• 5/8/03
• EnvE 424/524

2
Data

• EPA Region VII IMPROVE monitors data (i.e.
  speciation filters)
• SO42- daily average measurements in mg/m3
• Calculated by taking 3 times the measured filter
  sulfur
• Data measured every third day
• Obtained from VIEWS Network

3
Data

• EPA Region VII IMPROVE monitor sites
• Buffer sites used for some of the interpolations

4
Methods

• Inverse Distance Weighting versus Kriging
  interpolation
• Ten Region VII sites versus 21 with additional
  buffer sites
• Three consecutive measurement days to show
  movement of sulfate event
• Comparison of RMSE of all methods

5
8/24/03

• Inverse Distance Weighting

6
8/24/03

• Ordinary Kriging with Gaussian weighting

7
8/27/03

• Inverse Distance Weighting

8
8/27/03

• Ordinary Kriging with Gaussian weighting

9
8/30/03

• Inverse Distance Weighting

10
8/30/03

• Ordinary Kriging with Gaussian weighting

11
Naval Research Lab Model
12
Results and Conclusions

• RMSE comparison
• Buffer sites data improves RMSE of ten Region VII
  sites
• Kriging, in general, is better than IDW

13
Thank you

• Any questions?