[R-sig-Geo] Spatial interpolation for air pollution data

[Clint Bowman]

Pankaj,

Although geospatial techniques should work in the eastern portions of 
North Carolina, I'd doubt they should be applied in the more mountanous 
western portion.  We've used the algorithm in EPA's BenMAP program to fuse 
available CMAQ forecasts with available monitoring to produce a gridded 
map of background concentrations.

<http://lar.wsu.edu/nw-airquest/lookup.html>

Clint

Clint Bowman			INTERNET:	clint at ecy.wa.gov
Air Quality Modeler		INTERNET:	clint at math.utah.edu
Department of Ecology		VOICE:		(360) 407-6815
PO Box 47600			FAX:		(360) 407-7534
Olympia, WA 98504-7600

         USPS:           PO Box 47600, Olympia, WA 98504-7600
         Parcels:        300 Desmond Drive, Lacey, WA 98503-1274

On Mon, 9 Mar 2015, Pankaj Agarwal wrote:

> Dear All,
>
> I am working on figuring out which R package to use for spatial interpolation of air pollutant data.  I am new to the field of geospatial data analysis with a basic background in statistics, so hopefully my questions are not too na??e.
>
> We have a data set of daily air pollutant levels measured at monitoring stations for the state of North Carolina, USA.  The number of monitoring stations are less than the number of counties in the state.  We have the location of these stations and the location of county centroids in Latitude/Longitude.  We would like to interpolate the air pollutant measurement to the centroid of the counties for which the air pollutant measurement is missing.   Furthermore, we also have the centroid Lat/Long at the zip code level for each county and would also like to interpolate at the zip code level.  Once this is done, we would also like to incorporate meteorological data that is available as daily measurements of wind speeds, humidity, min/max/avg temperature etc. into the interpolation as co-factors.
>
> My understanding is that the kriging method of interpolation should work and I have found three packages - "fields", "geoR" and  "gstat" that provide this function.  I am trying to figure out which one of these would best serve the purpose.  From the documentation I found that only "gstat" offers co-kriging to take into account the meteorological co-factors into account.
>
> 1.  Could you please advise which package would provide the capabilities that I need for the spatial interpolation of the air pollutant data at the county and zip code level.
> 2.  The documentation for "gstat" says that the Lat/Long data needs to be projected and I believe rgdal::spTransform can be used for this purpose.  I have two questions related to this.
> 2a. Do the other packages also require projected data or can they accept Lat/Long.
> 2b. When we project Lat/Long data, do we lose any accuracy in the interpolation given that the original air pollutant measurement is at the Lat/Long coordinates.  This may be important because we are trying to study the health effects of the air pollutant levels and the magnitude of this effect may be small so losing accuracy might bias the analysis disproportionately.
>
> Thanks you for any guidance you can provide.
>
> Sincerely,
>
> - Pankaj
>
> ----------------------------------------------------------------
> Pankaj Agarwal, M.S
> Bioinformatician
> Database Analyst II
> Surgical Sciences Applied Therapeutics Section
> Department of Surgery
> Duke University
> 919-244-6389
> p.agarwal at duke.edu<mailto:p.agarwal at duke.edu>
>
>
> 	[[alternative HTML version deleted]]
>
>