Lecture 8. Flashcards
Spatial Predictions
Methods are used to estimate values at a unsampled location
spatial interpolation
is the prediction of variables at unmeasured locations, and based on a sampling of the same variables at known locations
Spatial interpolation rely on
he nearest points to estimate missing values, and use some measure of distance from known to unknown values.
Spatial Prediction
involves the estimation of variables at unsampled locations, but estimates are based at least in part on other variables
Core Area is chatacterized by
high use, density, intensity, or probability of occurrence for a variable or event
Core areas are defined
from a set of samples, and are used to predict the frequency or likelihood of occurrence of an object or event
what are the 2 controls in the sampling process
distributing locations and number of samples
Sampling patterns (4 types)
Systematic
random
cluster
adaptive
Systematic Sampling is when
Samples are spaced uniformly at fixed X and Y intervals and may not be the same in both directions
What are the Advantages of Systematic Sampling
Easy understand, and ease in planning and description.
What are the disadvantages of Systematic Sampling
All samples receive the same attention, difficult to stay on lines and may be biases
Random Sampling is when
Select points based on random number process and plot on a map
What are the Advantages of Random Sampling
Less biased
What are the disadvantages of Random Sampling
Does nothing to distribute samples in areas of high variation and location of the points may be a problematic
What is Cluster Sampling
Cluster centers are established by random or systematic means and then samples are arranged around each center on a map
What are the Advantages of Cluster Sampling
Reduced travel time
What are the disadvantages of Cluster Sampling
Biased samples and prone to higher sampling error than the samples formed using other sampling methods.
What is Adaptive Sampling
More sampling where there is more variability and is two stage sampling from prior knowledge
What are the advantages of Adaptive sampling
More efficient, homogeneous area have few samples, better representation of variable areas.
What are the disadvantages of Adaptive sampling
Need prior information on variability through space
IN selecting a method
No single method has been shown to be more accurate in every application.
Accuracy is judged by
difference between the measured and interpolated values at a number of withheld sample points
Thiessen Polygon is done by
Draw lines connecting the points to their nearest neighbors and then find the bisectors of each line, connect then and then assign the resulting polygon the value of the center point.
Fixed Radius (local averaging)
Cell values estimated based on the average of nearby samples and Samples used depend on search radius.
Inverse Distance Weighted
Estimates the values at unknown points using the distance and values to nearby known points
IDW equation
Zi is value of known point
Dij is distance to known point
Zj is the unknown point
N is a user selected exponent
size of N affects________and larger N means____
the shape of the surface and the closer points are more influential.
Spatial regression
Fitting a statistical model, a trend surface, through the measured points (typically polynomial)
Trend surface equation
Z is dependent variable to be predicted at the point (X,Y).
ai values are coefficients for the predictive equation, usually estimated through a least squares statistical process
X and Y are the coordinate locations.
Q and W are independent variables measured at the point (X, Y)
Spatial prediction
Based on mathematical models often built by a statistical process and Distinction between interpolation and prediction may be viewed as artificial
Prediction Accuracy
using 10% of the sample and create a interpolated surface and compare to the whole sample to see accuracy prediction
Interpolation uses
only the measured target variable and sample coordinates to estimate the variable at unknown locations
Prediction uses
Methods address the presence of spatial autocorrelation
Spatial prediction is more ______ than spatial interpolation.
General
Estimation Exact methods
Theissen
IDW
Spline
Estimation No exact methods
Fixed-Radius
Trend surface
Kriging
