Chapter 12 - Spatial Estimation Flashcards
What are three common spatial estimation methods?
- Spatial interpolation; 2. Spatial prediction; 3. Core area
This estimation method estimates values at unmeasured locals using measured values of the same variable
Spatial interpolation
This spatial estimation method predicts values at unmeasured locals using measured values of other variables
Spatial prediction
This spatial estimation method predicts the chance of occurence of an object or event using probability theory
Core area
What are three different methods of spatial interpolation?
- Creating isoclines; 2. Polygons; 3. Raster
To test spatial interpolation accuracy what is done with some data?
Some data is withheld
What are four sampling methods used in spatial interpolation?
- Systematic; 2. Random; 3. Cluster; 4. Adaptive
This sampling method collects points on a uniform grid
Systematic
This sampling method uses randomly distributed points
Random
This sampling method uses grouped points
Cluster
This sampling method uses more points in variable areas
Adaptive
What are four types of spatial interpolation?
- Nearest neighborhood interpolation; 2. Fixed radius (or local averaging); 3. Inverse distance weighting; 4. Spline
This type of spatial interpolation creates polygons around sample points
Nearest neighborhood interpolation
What is another name for the polygons created in nearest neighborhood interpolation?
Thiessen polygons
This is conceptually the simplest type of spatial interpolation
Nearest neighborhood interpolation
What type of data are poorly represented by nearest neighborhood interpolation?
Continuous data
In nearest neighborhood interpolation, points inside polygons have this
Same value
Is nearest neighborhood interpolation an exact interpolator?
Yes
In this type of spatial interpolation, a circle is placed over each raster cell
Fixed radius (or local averaging)
What is done with sample points within a circle in fixed radius/local averaging spatial interpolation?
Points are averaged and output to cell
What are cells without points assigned in fixed radius/local averaging spatial interpolation?
Zero or null
What happens if a circle radius is too small in fixed radius spatial interpolation?
Many nulls created
What happens if a circle radius is too large in fixed radius spatial interpolation?
Data is smoothed too much
Is fixed radius/local averaging spatial interpolation an exact interpolator?
No
This type of a spatial interpolation uses a set of known sample points with different weights given to each known point
Inverse distance weighting
In inverse distance weighting spatial interpolation, weights are inversely proportioned to this
Distance
Do distant or near points have less weight in inverse distance weighting spatial interpolation?
Distant points
Is inverse distance weighting an exact interpolator?
Yes
This spatial interpolation method fits a smooth line or surface through a set of points
Spline
Spline spatial interpolation uses these to describe a surface
Polynomial equations
This is a flexible ruler used for drawing smooth curves
Spline
Is spline spatial interpolation an exact interpolator?
Yes
This type of spatial estimation predicts values of variable at missing sites using statistical models
Spatial prediction
Spatial prediction statistical models use these two things to predict values
Coordinate data and independent variables
Spatial prediction uses these two types of correlation
Spatial autocorrelation and cross-correlation
What is the First Law of Geography?
Everything is related to everything else, but near things are more related than distant things
Who is the First Law of Geography attributed to?
Waldo Tobler
This is correlation of a variable with itself over space
Spatial autocorrelation
These measure the degree of spatial autocorrelation
Different indices
This spatial autocorrelation index ranges from -1 to +1
Moran’s index
A positive spatial autocorrelation has this type of trend when plotted
Upward trend
A random spatial autocorrelation has this type of trend when plotted
No trend
What are two types of spatial prediction?
Trend surface and kriging
This type of spatial prediction fits a 2D surface to a set of sample points using statistical methods
Trend surface
Trend surface uses this type of polynomial equation
Low-order polynomial
Is trend surface spatial prediction an exact predictor?
No
This type of spatial prediction is central to geostatistics
Kriging
What type of spatial interpolation is kriging similar to?
Inverse distance weighting
This type of spatial prediction uses neighboring values weighted by distance
Kriging
Kriging uses this to assign weights
Spatial autocorrelation
Are weights arbitrary in kriging?
No
Is kriging an exact predictor?
No
Data is sorted into these in the kriging method
Lag distance bins
This is calculated and plotted for each kriging lag distance bin
Variance
The kriging method produces this graph
Variogram
What are three components of a variogram?
Nugget, sill, range
This component of a variogram is where the line intercepts the y-axis
Nugget
This component of a variogram is just above the asymptote
Sill
This component of a variogram is the distance between the asymptote and the x-axis
Range
This model statistically fits through semi-variance points in the kriging method
Variogram model
What three types of spatial estimation are exact, with predicted values equally observed?
- Thiessen polygons; 2. Inverse distance weighting; 3. Spline
What three types of spatial estimation are not exact, with predicted values that may not be equally observed?
- Fixed-radius; 2. Trend surface; 3. Kriging
This group of spatial estimation methods identifies areas of high density
Core area
This group of spatial estimation methods is based on a distribution of known observations
Core areas
Core area spatial estimation methods create these from polygons
Regions or territories
Core area spatial estimation methods create these from rasters
Density fields
What are the three types of core area spatial estimation?
- Mean center; 2. Hulls; 3. Kernel mapping
This is the simplest core area method
Mean center
This core area method averages x-y coordinates for observed values
Mean center
This is often added to a mean center
Mean circle
Can a mean circle radius mean different values?
Yes
This is the simplest core area method for an irregular core area
Hulls
This core area method creates a polygon enveloping all points
Hulls
Hulls ignore these
Clusters in data points
Hulls can be treated as this for points
A natural boundary
This core area method makes a continuous density surface
Kernel mapping
Peaks are areas of this in kernel mapping
High density
This is a density function assigned to each data point in kernel mapping
Kernel
What type of kernel is the most common?
Gaussian kernel
This is the width of kernel defined by the user
Bandwidth
In kernel mapping, this is applied to each sample point
Kernel function
In kernel mapping, point kernels are summed for this
Composite density function
In kernel mapping, a narrow bandwidth creates this type of surface
Spiky surface
In kernel mapping, a broad bandwidth creates this type of surface
Broader/flatter surface
These are used to determine optimum bandwidth in kernel mapping
Formulas
Kernel mapping is used to create this
2D density surface