Let us create a simple simulation worthy of gstat::krige with universal kriging, in the language of linear mixed models, i.e.: Where X is a n x n_vars matrix of n_vars additional variables, betas is a n_vars x 1 so-called fixed-effects vector, Z is a n x q indicator matrix of q locations, each repeating say n / q times, e is normal random noise from N(0, sig2e), and b is a q x 1 so-called random-effects vector, with a standard Gaussian q x q covariance matrix or kernel: Where d_i, d_j are a pair of 2D (lat, lon) locations (there are q of these), and sill and range standard params. In code: library(sp) library(gstat) library(MASS) n

Tata Communications has secured a "multi-crore" order from the Goods and Services Tax Appellate Tribunal (GSTAT) to set up an end-to-end digital infrastructure covering network, IoT, cloud, and security solutions.The company said the project, aimed at streamlining GSTAT’s functioning, will provide seamless access across all benches nationwide, strengthen data security, and deliver a simplified user experience, PTI reported."Tata Communications has been awarded a strategic project to deliver a comprehensive digital infrastructure — spanning network fabric, IoT, cloud, and security, with unified management across domains — for the Goods and Services Tax Appellate Tribunal (GSTAT) benches established by the Ministry of Finance, Government of India.

Offers bathymetric interpolation using Inverse Distance Weighted and Ordinary Kriging via the 'gstat' and 'terra' packages. Other functions focus on quantifying physical aquatic habitats (e.g., littoral, epliminion, metalimnion, hypolimnion) from interpolated digital elevation models (DEMs). Functions were designed to calculate these metrics across water levels for use in reservoirs but can be applied to any DEM and will provide values for fixed conditions. Parameters like Secchi disk depth or estimated photic zone, thermocline depth, and water level fluctuation depth are included in most functions.

I am trying to do a kriging interpolation on a fake data set to familiarize myself with gstat. This code has worked before on a larger df but for some reason when I try to make the variogram, the output is empty, and I get a NULL error message. I've tried getting rid of the log() part and that did not help. x

The gstat package supports stratified ordinary kriging using sp objects as shown at https://github.com/r-spatial/gstat/blob/master/demo/examples.R (the code below). This year (https://r-spatial.github.io/gstat/news/index.html) gstat also got support for sf and stars objects. While sf and stars works well for many geostatistical methods, I am unable to make it work for stratified ordinary kriging. # an example of gstat + sp library(gstat) library(sp) data(meuse) coordinates(meuse) = ~x+y data(meuse.grid) gridded(meuse.grid) = ~x+y meuse$part.a = krige(part.a ~ 1, meuse.grid, meuse, nmax = 1)$var1.pred #> [inverse distance weighted interpolation] x1 = krige( log(zinc) ~ 1, meuse[meuse$part.a == 0, ], meuse.grid[meuse.grid$part.a == 0, ], model = vgm(.548, "Sph", 900, .0654), nmin = 20, nmax = 40, maxdist = 1000 ) #> [using ordinary kriging] x2 = krige( log(zinc) ~ 1, meuse[meuse$part.a == 1, ], meuse.grid[meuse.grid$part.a == 1, ], model = vgm(.716, "Sph", 900), nmin = 20, nmax = 40, maxdist = 1000 ) #> [using ordinary kriging] x = rbind(as.data.frame(x1), as.data.frame(x2)) gridded(x) = ~x+y spplot(x["var1.pred"], main = "stratified kriging predictions") The main problem is how to use only selected values from a stars object. library(stars) #> Loading required package: abind #> Loading required package: sf #> Linking to GEOS 3.7.1, GDAL 2.3.2, PROJ 5.2.0 library(sp) data(meuse.grid) gridded(meuse.grid) = ~x+y meuse.grid_stars = st_as_stars(meuse.grid) meuse.grid_stars[meuse.grid_stars$part.a == 1, ] #> Warning in structure(eval(rlang::expr(x[[i]][!!!args])), levels = attr(x[[i]], : Calling 'structure(NULL, *)' is deprecated, as NULL cannot have attributes. #> Consider 'structure(list(), *)' instead. Is it possible to do stratified ordinary kriging using gstat and stars?

Some time ago I posted my own answer to create a grid inside a country to be used in kriging interpolation: 0........0.kriging error messages, grid file problem ? (automap, gstat) This grid was for South Africa. Now I have data point for Lesotho (the hole inside SA), but cant get around it, to work with both together. I need a joint grid for kriging in gstat. I used makegrid in sp: SA0