Geostatistical analysis

Question 1

3 types of spatial data

Answer 1

area data, continuous data and point pattern

Question 2

point pattern types?

Answer 2

uniform, random and clustered

Question 3

how to analyze point patterns?

Answer 3

spatial autocorrelation and semivariance, and model with semivariance and kriging

Question 4

Why test spatial similarity?

Answer 4

You often assume samples are independent, but if samples are taken close to each other they may be similar and not independent (pseudoreplication)

Question 5

how to interpret spatial autocorrelation outcomes?

Answer 5

Locations that are close are more similar than those far away. The value is between -1 and 1, and 0 = not correlated. -1 means close by is dissimilar, like trees in a grid formation. (next to each tree is grass). +1 means similar, so when trees are clustered each tree is near a tree and grass is near grass.

Question 6

what is a semivariogram?

Answer 6

function that relates semivariance (or dissimilarity) of locations to distance that separates them.

Question 7

can you estimate unknown locations with semivariogram?

Answer 7

yes

Question 8

What is the nugget in a semivariogram?

Answer 8

intercept with y-axis, unexplained variation.

Question 9

What is the range in a semivariogram?

Answer 9

the point were the line levels off. variation of the data set

Question 10

What is the Sill in a semivariogram?

Answer 10

distance from start to where it levels off. Corresponds to the average patch size

Question 11

what’s on the y axis in a semivariogram?

Answer 11

semivariance (dissimilarity)

Question 12

In a semivariogram, an increase of semivariance (dissimilarity) means

Answer 12

there is clustering, as closer points are more similar.

Question 13

In a semivariogram, a flat line means

Answer 13

homogenous landscape

Question 14

In a semivariogram, an up and down line means

Answer 14

patterns in landscape that repeat, transect was long enough to encounter multiple patches

Question 15

whats an isotropic pattern?

Answer 15

the same (semivariance) in all directions.

Question 16

what’s an anisotropic pattern?

Answer 16

will give different semivariance patterns if you walk fron N to S than if you walk from E to W.

Question 17

what is kriging?

Answer 17

Technique to predict locations that you didn’t measure. (interpolate the value of a variable z at an unobserved location from observations of its value at nearby locations

Question 18

how to test if things are spatially heterogenously distributed over a landscape?

Answer 18

Make semivariograms and test if it has a slope using regression with semivariance as dependent and lag as independent variable (slope should be > 0)

Question 19

how to correct for effect of something with a semivariogram?

Answer 19

test if there is a significant effect with regression, then use residuals of the regression to make the semivariogram