How to do a looping to Interpolate a huge Rainfall Time-series in GRASS GIS or R

Question

In GRASS GIS I have access to a table in PostgreSQL Database with this structure:

Table Name: rainfall_time_series

Columns: id | year | month | day | value_mm | x | y | gauge_code

Setting g.region to the Watershed limits, I need create for each day one interpolated raster, representing the rainfall. This times-series is very huge, has 17 years, my question is: is it possible to do this operation with a looping? it is probably easier in R, but I'm newbie in this software.

Thanks in advance

Answer 1

The answer has been posted here:

http://lists.osgeo.org/pipermail/grass-user/2013-February/067195.html

Answer 2

Doing this in R would require a number of building blocks:

• A function that extracts the unique days present in the database.
• A function that extracts the data for one day from the database into a data.frame. You can then transform this to a SpatialPoints object (see sp package documentation.
• A function that interpolates the data, and returns an interpolated grid for each day. This can be done using a host of interpolation methods in e.g. the automap/gstat package, the fields package, etc. For a more exhaustive list I refer to the CRAN Task view for Spatial data.

Once you have these blocks, the following pseudo code links them together:

unique_days = get_unique_days()
dat_list = lapply(unique_days, get_data_from_database) 
interpolated_maps = lapply(dat_list, interpolate_daily_data)