
Spatial kriging of SHmax axes after van den Boogaart & Schaeben (2002)
Source:R/interpolation_krige.R
krige.RdOrdinary kriging of the axial orientation tensors, with a distance-only
trace variogram (isotropy of measurements). The kriged tensor is decomposed
with the eigenvalue method of ot_eigen2d(). The per-observation
uncertainty, given as a standard deviation in degrees, is used as
measurement error on the diagonal of the kriging matrix.
Usage
circular_krige(
x,
grid = NULL,
lon_range = NULL,
lat_range = NULL,
gridsize = 2,
R_range = 300,
max_data = 50,
min_data = 3L,
min_dist_threshold = 200,
vgm_model = c("spherical", "exponential", "gaussian"),
nugget = 0,
range_km = 300,
sill = NULL,
use_uncertainty = TRUE
)
PoR_circular_krige(
x,
PoR,
grid = NULL,
PoR_grid = TRUE,
lon_range = NULL,
lat_range = NULL,
gridsize = 2.5,
remove_PoR = FALSE,
...
)Arguments
- x
sfobject containing- azi
angles in degrees (e.g. orientation of \(\sigma_\text{Hmax}\))
- unc
(optional) Uncertainties of
aziin degree- type
(optional) Methods used for the determination of the direction of \(\sigma_\text{Hmax}\)
- grid
(optional) Point object of class
sf.- lon_range, lat_range
(optional) numeric vector specifying the minimum and maximum longitudes and latitudes (ignored if
gridis specified).- gridsize
numeric. Target spacing of the regular grid in decimal degree. Default is
2.5. (is ignored ifgridis specified)- R_range
numeric value or vector of search radii in km. Only observations within the radius are used.
- max_data
integer. The number of nearest observations used for each prediction. Default is
50.- min_data
integer. If the number of observations within distance
R_rangeis less thanmin_data, a missing valueNAwill be generated. Default is3forcircular_idw()and4forcircular_idw_stats().- min_dist_threshold
numeric. Distance threshold for smallest distance of the prediction location to the next observation location. Default is
200km.- vgm_model
Variogram model:
"spherical"(default),"exponential"or"gaussian".- nugget
numeric in \([0, 1)\). Nugget of the trace variogram.
- range_km
numeric. Range of the variogram model in km.
- sill
numeric in \((\text{nugget}, 1]\). Sill of the trace variogram. If
NULL(default), it is set to \(1 - \bar{R}^2\), where \(\bar{R}\) is the mean resultant length of the doubled angles ofx$azi.- use_uncertainty
logical. If
TRUE(default),unc(standard deviation in degrees) is used as measurement error.- PoR
Pole of Rotation.
data.frameor object of class"euler.pole"containing the geographical coordinates of the Euler pole- PoR_grid
logical. Whether the grid should be generated based on the coordinate range in the PoR (
TRUE, the default) CRS or the geographical CRS (FALSE). Is ignored ifgridis specified.- remove_PoR
logical. Whether PoR azimuths and coordinates will be removed from final output or not (the default.)
- ...
Arguments passed to
circular_krige()
Value
sf object containing
- lon,lat
longitude and latitude in degrees
- azi
kriged axis direction in degrees (0 to 180), the principal direction from
ot_eigen2d()- sd
circular standard deviation of the axis in degrees, computed from
confasrad2deg(sqrt(-2 log(conf))) / 2- R
search radius in km
- mdr
mean distance between grid point and data / R
- N
number of observations used
- conf
confidence in \([0, 1]\), \(2 \lambda_1 - 1\), i.e. the axial strength relative to the trace of the kriged tensor
- kvar
kriging variance on the trace-variogram scale
- eig1
largest eigenvalue of the kriged tensor (axial strength), in \([0.5, 1]\). The smaller eigenvalue, \(1 = \lambda_1\), is the axial dispersion
Attribute "vgm" records the variogram parameters used.
Details
The method is the simplified ordinary-kriging form derived by van
den Boogaart & Schaeben (2002) for axes in the embedding \(u u^T\). Their
isotropy-of-measurements assumption gives the trace variogram used here.
The kriged tensor is \(\sum_i \lambda_i u_i u_i^T\), computed by
ot_eigen2d() with the kriging weights as weights. The principal
direction is the eigenvector of the largest eigenvalue, and the eigenvalues
sum to one. Their values give the strength and dispersion of the prediction.
Fit nugget, sill, and range_km to the empirical variogram from
variogram_axial().
References
van den Boogaart, K.G. and Schaeben, H. (2002). Kriging of regionalized directions, axes, and orientations I. Mathematical Geology, 34(5), 479-503.
See also
variogram_axial() for computing the empirical variogram of axial data;
ortensor2d() and ot_eigen2d() for the orientation tensor and its eigenvalue decomposition.
Examples
data("san_andreas")
v <- variogram_axial(san_andreas)
res <- circular_krige(san_andreas, R_range = 300,
vgm_model = "spherical",
nugget = 0.25, range_km = 1000)
head(res)
#> Simple feature collection with 6 features and 11 fields
#> Geometry type: POINT
#> Dimension: XY
#> Bounding box: xmin: -114.82 ymin: 24.08 xmax: -108.82 ymax: 26.08
#> Geodetic CRS: WGS 84
#> lon lat azi sd R mdr N conf kvar eig1
#> 1 -112.82 24.08 135.071737 0 300 0.7787910 17 1 0.4427279 1.073801
#> 2 -110.82 24.08 138.899239 0 300 0.4965600 50 1 0.2427488 1.145847
#> 3 -108.82 24.08 5.491198 0 300 0.1915108 50 1 0.2465356 1.188411
#> 4 -114.82 26.08 47.683645 0 300 0.8111352 22 1 0.1676407 12.948122
#> 5 -112.82 26.08 150.694353 0 300 0.5630126 50 1 0.3011651 1.091931
#> 6 -110.82 26.08 96.097099 0 300 0.2681149 50 1 0.4309789 1.046459
#> eig2 geometry
#> 1 -0.07380128 POINT (-112.82 24.08)
#> 2 -0.14584713 POINT (-110.82 24.08)
#> 3 -0.18841052 POINT (-108.82 24.08)
#> 4 -11.94812223 POINT (-114.82 26.08)
#> 5 -0.09193137 POINT (-112.82 26.08)
#> 6 -0.04645894 POINT (-110.82 26.08)
attributes(res)$vgm
#> $model
#> [1] "spherical"
#>
#> $nugget
#> [1] 0.25
#>
#> $range_km
#> [1] 1000
#>
#> $sill
#> [1] 0.5344795
#>
#> $max_data
#> [1] 50
#>