[R-sig-Geo] Summary of multiple rasters

[Tomislav Hengl]

Dear Johannes,

Looks like you are in the field of error propagation [http://spatial-accuracy.org/workshopSUP]. 

Both mean and median are valid statistical parameters to describe the central tendencies. What is often more interesting is the propagated uncertainty in the final estimates. I assume that your output is a binomial variable? In our 
HESS paper [http://www.hydrol-earth-syst-sci-discuss.net/7/767/2010/hessd-7-767-2010.html], we use the Information entropy index to quantify the uncertainty of the model (we deal with a binary/Bernoulli variable). The code and the dataset can be obtained from: [http://geomorphometry.org/content/uncertainty-stream-networks-derived-elevation-data]

A number of (Gaussian) grid statistics parameters can be derived using SAGA GIS:

> library(RSAGA)
> rsaga.get.usage(lib="geostatistics_grid", module=5)
SAGA CMD 2.0.3
library path:   C:/PROGRA~1/R/R-210~1.1/library/RSAGA/saga_vc/modules
library name:   geostatistics_grid
module name :   Statistics for Grids
Usage: 5 -GRIDS <str> [-MEAN <str>] [-MIN <str>] [-MAX <str>] [-VAR <str>] [-STDDEV <str>] [-STDDEVLO <str>] [-STDDEVHI <str>]
  -GRIDS:<str>          Grids
        Grid list (input)
  -MEAN:<str>           Arithmetic Mean
        Grid (optional output)
  -MIN:<str>            Minimum
        Grid (optional output)
  -MAX:<str>            Maximum
        Grid (optional output)
  -VAR:<str>            Variance
        Grid (optional output)
  -STDDEV:<str>         Standard Deviation
        Grid (optional output)
  -STDDEVLO:<str>       Mean less Standard Deviation
        Grid (optional output)
  -STDDEVHI:<str>       Mean plus Standard Deviation



HTH,

T. Hengl



> -----Original Message-----
> From: r-sig-geo-bounces at stat.math.ethz.ch [mailto:r-sig-geo-
> bounces at stat.math.ethz.ch] On Behalf Of Johannes Signer
> Sent: Sunday, May 02, 2010 5:30 PM
> To: r-sig-geo at stat.math.ethz.ch
> Subject: [R-sig-Geo] Summary of multiple rasters
> 
> Dear List,
> 
> my problem is as follows:
> 
> I have a model that predicts me the distribution  of a species in the
> landscape. I run that model several times and now I am not sure which
> summary statistic to use (mean, median). Is there an index to get an idea of
> how the prediction in each cell are distributed?
> 
> Any are greatly appreciated,
> 
> Thanks
> Johannes
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo