• Ch 01:
  • Overview
  • The datasets
• Ch 02
  • R tricks
  • Problems

• Increase in sensors, increase in available data

• Statistical methods and software

• So-called spatial data

• subtleties, but basically geolocated, x and y coordinates

• CRS (Coordinate Reference System[s])

Statistical properties

1. Spatial data has a lot of data points, so power is large even for tiny effect sizes, thus the Null is always rejected (even if practically meaningless)

2. Spatial data points near each other are almost never independent, violating the common assumption that they are (a/k/a spatial autocorrelation)

Ecological properties

• Low ‘ecological resolution’

• High ‘data resolution’

• Complex relationships

(next slide shows Soil moisture, Veg reflectance, Yield…)

1. Geostatistical data x-y point data with a continuous measure (like soil moisture). Extrapolation between measured points is a goal.

2. Areal data points or polygons representing a uniform unit of measure (like the crop planted within a field boundary)

3. Point pattern data what is the spatial pattern (like whether pest outbreaks are random or spatially explained by some feature)

1. Spatial component (x-y)

2. Attribute component (something measured or classified)

3. Scale and sample size (for measuring earthworms, is 1m or 1000m better to sample?)

4. Vectore data versus Raster data

Ex: dataset 1 yellow billed cuckoo habitat

What spatial featured are associated with presence in this species?

Ex: dataset 1 yellow billed cuckoo habitat

What spatial featured are associated with presence in this species?