I would add (as a sincere question): wouldn't be the case of listwise deletion or multiple imputation? How does pairwise deletion behave for gstat? See http://www.uvm.edu/~dhowell/StatPages/More_Stuff/Missing_Data/Missing.html
On Tue, Dec 1, 2009 at 8:51 AM, Jon Olav Skoien <[email protected]> wrote: > Hi Friederike, > > For which variable do you have NAs? If soil moisture is sm, you are missing > in the first line: > [!is.na(soildmD79.sp$smoist.20.06.2005),] > > If that is not the problem, it would be helpful if you sent a reproducible > example. > > Cheers, > Jon > > [email protected] wrote: >> >> Dear list, >> >> I am having problems with dealing with missing values in gstat computing a >> semivariogram. >> >> I am using R 2.9.1, gstat 0.9-64 on Windows Vista. >> >> I get the following error message: >> >> g <- gstat(id = "sm", formula = smoist.20.06.2005 ~ 1, data = >> soildmD79.sp) >> g <- gstat(g, id = "bd", formula = BD ~ 1, data = >> soildmD79.sp[!is.na(soildmD79.sp$BD),]) >> g <- gstat(g, id = "S", formula = S ~ 1, data = >> soildmD79.sp[!is.na(soildmD79.sp$S),]) >> g <- gstat(g, id = "U", formula = U ~ 1, data = >> soildmD79.sp[!is.na(soildmD79.sp$U),]) >> g <- gstat(g, id = "T", formula = T ~ 1, data = >> soildmD79.sp[!is.na(soildmD79.sp$T),]) >> g <- gstat(g, id = "GMD", formula = GMD ~ 1, data = >> soildmD79.sp[!is.na(soildmD79.sp$GMD),]) >> g <- gstat(g, id = "vcover", formula = vcover ~ 1, data = >> soildmD79.sp[!is.na(soildmD79.sp$vcover),]) >> g <- gstat(g, id = "mean_C", formula = mean_C~ 1, data = >> soildmD79.sp[!is.na(soildmD79.sp$mean_C),]) >> g <- gstat(g, id = "biomass", formula = biomass ~ 1, data = >> soildmD79.sp[!is.na(soildmD79.sp$biomass),]) >> png(file = "sm_soilparamter_cross_20-06-2005.png", width = 1400, height = >> 1000, bg = "white", res = 100) >> plot(variogram(g), main = "Cross variograms Soil moisture and stable soil >> parameters", sub="20/06/2005") >> >> Error: dimensions do not match: locations 200 and data 96 >> >> This message occurs due to 100 x,y coordinate pairs and 96 values for the >> variable soil moisture. >> Even though the function contains already "data = >> soildmD79.sp[!is.na(soildmD79.sp$BD),]" the missing values still cause an >> error. >> >> Thanks for any suggestions or comments, >> Friederike >> >> _______________________________________________ >> R-sig-Geo mailing list >> [email protected] >> https://stat.ethz.ch/mailman/listinfo/r-sig-geo > > _______________________________________________ > R-sig-Geo mailing list > [email protected] > https://stat.ethz.ch/mailman/listinfo/r-sig-geo > _______________________________________________ R-sig-Geo mailing list [email protected] https://stat.ethz.ch/mailman/listinfo/r-sig-geo
