I am planning to distribute temperature data using kriging with external drift in R. I have time series data of temperature and i guess i need to make variogram for each day. Is there any way through which we can automatically generate variogram model and run kriging with external drift. Thanks in advance.
Dear Amrit
There is a library automap and the functon autofitVariogram()
Automatically fitting a variogram to the data on which it is applied. The automatic fitting is done through fit.variogram. In fit.variogram the user had to supply an initial estimate for the sill, range etc. autofitVariogram provides this estimate based on the data and then calls fit.variogram.
It's very simple and it works very well
Dear Amrit,
As Alain says, automap (with the autoKrige function to estimate the semivariogram and predict) is a good alternative for automatic estimation. This function estimates the semivariogram from the iterative Gauss-Newton algorithm for weighted least squares. An easy way to use it is (also using gstat):
library(automap)
library(gstat)
#Semivariogram
vgOK
You can write a script using the functions for variogram fitting and estimation by let's say gstat or autoKrig. But observe that any automatic variogram fitting (without human intervention) may cause problems. If one of the input data happen to be problematic (let's say a malfunction of the measuring device) then the variogram fitting will not be optimal, and it may be entirely wrong since your data are polluted with garbage. So, automatic variogram fitting and estimation may be accompanied by a good automatic outlier control of your data (which is not always easy to do). I assume also that, no matter what scheme you use, you will validate first your results on a large dataset from the past before you proceed to operational distribution of data. You can learn a lot and improve your algorithm considerably from such a process.
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping
- Maps