Predictive Surfaces Flashcards
what are predictive surfaces
measurements at a set of locations to predict values in locations that WERE NOT MEASURED
are predictive surfaces used for discrete or continuous data
continuous data
what can predictive surfaces be used to do
interpolate and extrapolate
interpolate vs extrapolate
interpolate
- predicting values between known points
extrapolate
- predicting values outside of the known points
exact vs approximate interpolation
exact
- creates a surface that passes through ALL known points
approximate
- creates a surface that MAY vary from known values
local vs global methods
local
- uses spatially defined data (uses data around specific points)
global
- uses al data in the study area
what are potential predictive surfaces
Inverse Distance Weighting (IDW)
Natural Neighbour
Spline
Trend
is inverse distance weighting local or global interpolation
local
is inverse distance weighting exact or approximate interpolation
exact
benefits and limitations of is inverse distance
benefits
- known influence of proximity
- uniform distribution of points
- can control the smoothness
limitations
- doesn’t handle sharp changes in data
- creates bullseye pattern around points
- does NOT extrapolate
what does IDW predict
values using a weighted combination of sample points
what does the power control in IDW
the significance of points based on their distance
(increased power = more emphasis on nearest points)
contrast fixed vs variable search radius in IDW
fixed
- searched radius will remain constant unless min number of points not met
variable
- search radius will change to include a minimum number of sample points
what is natural neighbor
finds the nearest input samples to grid cells and weights them based on proportionate areas overlapping grid cell area
is natural neighbors local or global interpolation
local
is natural neighbors exact or assumptions interpolation
exact
benefits and limits of natural neighbor
benefits
- ideal for irregular spaced data
- resistance to cluster bias or overrepresentation
limitations
- does not represent peaks, ridges or valleys
- does NOT extrapolate
what is spline
users can control the number of points used to calculate each interpolated cell value
does spline with more points create a smooth or rough surface
smooth
regularized vs tension spline
regular
- allows users to adjust the weight parameter to SMOOTH the surface
tension
- allows users to adjust the weight parameter to STIFFEN the surface
is tension spline exact or approximate interpolation
becomes approximate
what is spline
minimizes the curvature to create a smooth surface
is spline local or global interpolation
local
is spline exact or approximate interpolation
exact but CAN extrapolate
