Spatial Analysis

1
Spatial Analysis
CS 189 Oct 30 Nov 1, 2007
2
Outline

• Today
• Types of operations (Bolstad, ch. 9)
• Vector and raster
• Demo
• Raster analysis (part of) (Bolstad, ch 10)
• Demo
• Thursday and next week
• Terrain analysis (Bolstad, ch 11)
• Spatial estimation (Bolstad, ch 12)
• Cartographic modeling

3
Spatial Analysis

• Spatial analysis applies one or more spatial
  operations to coordinate data to answer a
  question
• Typically involves using data from one or more
  layers to generate output
• Analysis may consist of a single operation on a
  single layer, or many operations using multiple
  data layers to generate desired output

4
Spatial Analysis
5
Scope

• Scope the extent of the input data that are
  used to determine values at the output locations
• Local uses one input location
• Neighborhood uses adjacent locations
• Global uses all locations in input

6
Spatial Operations

• Spatial operations are also referred to as
  geoprocessing or commands, tools...
• Most operations are at least conceptually
  compatible in both vector and raster models
• Almost always vector operations and raster
  operation are implemented in the software with a
  different set of of commands
• Some are easier to use in one model (vector vs.
  raster) than the other

7
Spatial Operations

• Basic spatial operations generically classified
  as
• Selection
• Classification/reclassification
• Dissolving
• Proximity
• Overlay
• Hundreds of other operations exist to generate
  very specific outputs

8
Selection Operations

• Two types of selection operations spatial and
  attribute
• Spatial states entirely north of Arkansas
• Attribute states that are larger than 84,000 sq.
  km.

9
Selection Operations

• Spatial selections can be done by manually
  selecting features on the screen or by using
  spatial selection operations
• Adjacency and containment are the most common
  operations
• Often one performs a spatial selection using a
  combination of on-screen selection, spatial
  selection operations and attribute selection

10
Classification

• Classification or reclassification categorize
  features based on a set of conditions
• Usually involves modifying or adding to attribute
  data
• Example parcels larger than 1 acre in size are
  assigned value of large parcels between 0.1
  and 1 acres assigned value of medium parcels
  smaller than 0.1 acres assigned value of small

11
Classification

• Also used to group features for symbolization on
  maps
• Many methods for classification (manual,
  automatic, binary, equal interval, natural
  breaks, etc.) for numerical data
• Important to realize that different
  classifications change appearance of map
• Bolstad covers in detail read but you won't be
  expected to know details of each method

12
Dissolve

• Dissolve combines like features in a data layer
• Adjacent polygons may have identical values.
  Dissolve removes, or dissolves away the common
  boundary, making the two polygons with identical
  values a single polygon
• Often dissolve is used after performing a
  classification

13
Dissolve
14
Proximity Operations

• Proximity operations examine the distance between
  features
• Buffer is a common proximity operation
• A buffer is a region that is less than or equal
  to a specified distance from one or more features
• Buffers are often used to identify areas that are
  inside/outside a given distance
• Ex all homes within a mile of a fire station

15
Buffers

• Buffers can be performed on vector or raster
  datasets

16
Buffers

• Output of vector buffer always a polygon

17
Overlay

• Operations that combine spatial and attribute
  data from two or more spatial data layers
• Requires common coordinate system

18
Overlay (Vector)

• A new layer is created usually requiring the
  splitting of lines or areas and the creation of
  new features
• Attributes are usually appended to the table but
  not always depends on the specific type of
  overlay operation
• Most overlays involve at least one polygon layer
  but can use points and lines in overlay
• Not all overlay layer combinations will give
  meaningful results look at the command's Help
  file to determine if it is appropriate

19
Overlay (Vector)

• Types
• Clip one layer is used as a cookie-cutter to
  remove features outside of that layer attributes
  are usually not combined
• Intersection the spatial AND - combines data
  from both layers but only where they coincide
• Union the spatial OR - combines data from
  both layers where either layer has data
• AND / OR concept is very important
• Other types of overlay operations that are
  variants of the above (erase, identity, etc.)

20
Overlay (Vector)
21
Overlay (Raster)

• A cell-by-cell combination of two or more layers
• Very sensitive to cell size, coordinate system,
  and origin
• Often datasets must undergo resampling and/or
  coordinate system transformation
• Important to pay attention to number of possible
  outputs more than 256 will increase the dataset
  size

22
Other Types of Analysis

• Network analysis
• Performed on a set of connected features (roads,
  utilities, rivers, etc.)
• Used to represent and analyze the cost, time,
  delivery or accumulation of resources
• Geocoding / linear referencing
• Spatially referencing points in a network based
  on the known value/address along a network
  segment
• Address matching most common application

23
Raster Analysis

• Simple structure of raster data (cells) and the
  flexibility of using cell based data for analysis
  has led to the use of raster analysis for many
  different applications
• Lots of different software products developed
• Basic concepts can be used for advanced analysis
  in many different ways
• Basis for terrain analysis and modeling

24
Raster Analysis

• Raster operations may be categorized as
• Local
• Neighborhood
• Global

25
Local Functions

• Many, many local functions
• Grouped into four classes
• Mathematical functions (, -, , /)
• Boolean/logical operations (AND, OR, NOT)
• Reclassification (conditional statements)
• Overlay (often performed with mathematical
  operations and is the basis for a lot of
  cartographic modeling)

26
Map Algebra

• Type of local raster analysis that uses
  math-based operations on (, -, , /) rasters
  that store numbers as the cell values

27
Logical Operations

• Also known as Boolean operations, typically
  involves the comparison of cell values to other
  cell values, the result being an output grid with
  true (usually represented as 1) and false
  (represented as 0) values
• Three basic logical operators
• AND both input values must be true
• OR only one of the values must be true
• NOT switches true for false, false for true
  (nulls stay the same)

28
Reclassification

• Assigning output values based on input values
• Table matches input values to a
  reclassification table
• Range of values similar to table above but
  range of values instead of 1 to 1
• Conditional a statement that assigns specified
  true and false values based on a condition

29
Overlay

• Easiest if done with binary (0 or 1) data
• Often not a separate set of operations, but
  instead use map algebra
• Union may be performed with addition
• Intersection may be performed with addition
• Clip may be performed with multiplication

30
Outline (Nov 1)

• Demo
• Vector overlay (Union Intersection)
• Raster (Selection, Buffer, Intersection)
• Neighborhood Operations
• Terrain analysis
• Demo (if time)

31
Neighborhood Operations

• Use a moving window
• Usually a 3x3 window, but may be different
• Uses all the cell values in the window to
  calculate the CENTER cell value
• Repeats this, moving across the entire raster

32
Moving Windows

• Variety of functions use moving windows
• Math functions like min, max, mean, range
• Statistical functions, standard deviation, etc
• Boolean logic
• Terrain analysis
• Slope
• Aspect
• Hillshade
• Watersheds

33
Kernels and Filters

• Kernel uses a set of constants in the window
  which are applied across the entire raster
• Edge detection
• Smooth data (noisy data) mean filter
• High-pass filters emphasize differences in
  noisy data

34
Cost Surfaces

• Cost associated of traveling from a source cell
  to other cells
• May simply use only distance to determine cost,
  may incorporate friction surfaces, additional
  costs associated with obstacles (terrain,
  traffic, etc.) in the path of travel
• Calculate a least cost pathway

35
Terrain Analysis (Ch 11)

• Terrain variables (elevation, slope, aspect, etc)
  influence temperature, vegetation, visibility,
  water flow, soil erosion and more and therefore
  are often part of spatial analysis
• Appropriate moving window functions applied to
  raster Digital Elevation Models (DEMs) produce
  many of these terrain variables

36
Slope

• Slope the change in elevation (rise) with a
  change in distance (run)
• May be expressed as degrees or percent
• degrees slope (0 flat 90 vertical 45
  rise equals the run)
• percent slope (0 flat infinite vertical
  100 rise equals run)
• Slope direction at any point is the steepest
  direction of elevation change

37
Slope

• Different methods for calculating slope produce
  different results
• No one method is better than the others in all
  situations

38
Aspect

• Aspect is the direction of the steepest slope
• Reported as azimuth angle between 0 and 360
  degrees
• Flat areas (0 slope) have no aspect

39
Profile Curvature

• Profile index of surface shape of the steepest
  slope
• Smaller values indicate a concave (bowl shaped)
  surface
• Larger values indicate a convex (peaked) shape

40
Plan Curvature

• Plan index of surface shape at right angles to
  the steepest slope / profile
• Smaller values indicate a concave (bowl shaped)
  surface
• Larger values indicate a convex (peaked) shape

41
Hydrologic Functions

• Flow direction the direction (0 360 degrees)
  water will flow if poured onto the ground (often
  same as aspect)
• Flow accumulation / upslope area measurement of
  the area that drains through a given cell
• Watershed the area that contributes flow to a
  given point on the landscape
• Drainage network set of cells through which
  water flows

42
Hydrologic Functions
43
DEM Errors

• Random errors in DEM elevation values often
  create pits (or sinks)
• Pits are cells that are lower than all
  surrounding cells and therefore do not drain
• Sometimes pits are valid sinkholes
• Often they result from small errors in flat
  regions during DEM creation
• Pits cause problems during hydrologic analysis
• Usually, pits are filled before calculating any
  hydrologic function

44
Viewshed

• Viewshed the collection of areas visible from a
  given point

45
Viewshed
46
Shaded Relief

• Also called a hillshade
• Depiction of the brightness of terrain
  reflections given a terrain surface and sun
  location
• Rarely used for analysis, mostly to visualize
  terrain gives terrain a real-world appearance

47
Shaded Relief
48
Next Week

• Tuesday
• Quiz
• Cartographic modeling
• Thursday
• Exercise