Interpolation of 17 data points - temperature data

Question

I have 17 location points here in Iceland: weatherstation points.

It's a 10-year, month-to-month, composite of temperature values for each station which means 120 values for each station. 2040 values overall. Since I only have 17 station I only have 17 values visible on my map.

What I want to do is interpolate these points to create a temperature map for the area and I want to be able to compare them, f.i. compare june for every year and make a map of that - compare june-sep for each year - compare december for 5 of those 10 years etc. I basically want to be able to compare what I want.

The distance between each point varies but the area is fairly small so only a few kilometers between stations.

When the interpolation is done I hope to be able to do a time animation from the data.

This is a school project. I am attending The University of Iceland.

So, which method should I use and how?

Thanks in advance.

Einar Ingi

Answer 1

I'm not a quantitative export, but it seems like Spline interpolation might be a good fit.

Polynomial functions are regarded as a good method for interpolation of monthly and yearly climate elements but are less suitable on higher temporal resolutions like days and hours

This paper goes over the various interpolation methods and how they relate to temperature data.

http://www.knmi.nl/bibliotheek/knmipubIR/IR2009-04.pdf

Answer 2

Similar answer to artwork21.

I answered a question a while back and the same answer seems applicable here.

How to Create Contours for Zonation

I've not used ArcGis, however I have achieved a similar result with SAGA GIS which is an Opensource GIS application.

http://www.saga-gis.org/en/index.html

Use Gridding to interpolate the point data (B-Spline, Cubic Spline, Multi Level B-Spline...) and then using "Contour Lines from Grid" to convert the grid to contours.

I know its not a direct answer to your question but it giveas you options.

Basically follow the directions above but ignore the contour lines from gird part and you have a nicely interpolated grid data set for each month, and its done using SAGA which is opensource.

Answer 3

You can use GDAL and its powerful gdal_grid command for creating regular grids of scattered data.

For example this command will create an interpolation raster in GTiff from a shapefile named "points.shp" storing the values in a field named "value" using the inverse distance to a power algorithm:

$ gdal_grid points.shp output.tif -zfield value -ot Float64 -l points