Spatial Statistics Review

Which of the following can't be done with spatial statistics?

• interpolating unknown values

• calculating correlation between points

• analyzing categorical data

• calculating variation between points

Answer explanation

Spatial statistics only works with continuous data. To analyze categorical data, you'd use landscape metrics.

According to Tobler's First Law of Geography, the red point in this image is most closely related to the ___ point.

• green

• blue

• purple

Answer explanation

Tobler's Law states that everything is related to everything else, but near things are more related than distant things.

Which diagram demonstrates zero spatial correlation?

• a

• b

• c

Answer explanation

• - Diagram A shows positive spatial correlation because similar values are clustered.

• - Diagram B shows negative spatial correlation because similar values are dispersed.

• - Diagram C shows zero spatial correlation because there is no clear pattern.

Falsely assuming that your data is spatially independent increases the risk of:

• getting an error in ArcGIS

• finding statistically significant results where none exist (Type 1 Error)

• failing to find statistically significant results where they do exist (Type 2 Error)

Answer explanation

Analyzing your data without accounting for potential spatial dependence increases the risk of Type 1 Errors, or incorrectly finding statistically significant results.

If you want to analyze a landscape at a scale of 100 meters, what should the resolution of your data be?

• less than or equal to 100 meters

• 150 meters

• 200 meters

• greater than 300 meters

Answer explanation

Resolution, or the distance between data points, should be less than or equal to the scale at which you will analyze your data.

If you want to analyze a landscape at a scale of 100 meters, what should the extent of your data be?

• 50 meters

• 100 meters

• 150 meters

• 200 meters

Answer explanation

Extent should be twice the scale at which you will analyze your data.

When you determine that your data is spatially autocorrelated, you can prove causality between variables.

• True

• False

Answer explanation

Spatial autocorrelation does not imply causation.

Which of the following statements is FALSE regarding the assumption of stationarity?

• Stationarity assumes uniformity in mean and variance across a landscape.

• Stationarity is likely to be violated with elevation or precipitation changes across a landscape.

• Using non-stationary data means that results will not be universal.

• Non-stationary data cannot be used in statistical analysis.

Answer explanation

Non-stationary data can be analyzed. You can account for non-stationarity by using detrending methods in your model.

Which of the following landscapes is most likely to violate the assumption of stationarity?

• a plateau

• a mountain

• a plain

Answer explanation

Stationarity assumes uniformity across a landscape. It is likely to be violated with elevation changes, precipitation changes, soil nutrient variability, etc. across a landscape.

What information is used in point pattern analysis? (There are multiple correct answers.)

• location of observations

• metadata

• attributes

• scale

Answer explanation

Location and attributes of observations are needed to conduct point pattern analysis.