ArcGIS Spatial Analyst provides a rich set of tools to perform cell-based (raster) analysis. Of the three main types of GIS data (raster, vector, and TIN), the raster data structure provides the most comprehensive modeling environment for spatial analysis.
Cell-based systems divide the world into discrete uniform units called cells, based on a grid structure. Each cell represents a certain specified portion of the earth, such as a square kilometer, hectare, or square meter. Cells are given values that correspond to the features or characteristics that are located at or describe the locations they represent, such as an elevation value, soil type, or residential classification. In a cell-based system, geographic location is not defined as an attribute but is inherent in the storage structure, known as the locational perspective.
The locational perspective allows ArcGIS Spatial Analyst to store continuous data—for example, elevation, oil concentration, and sound—more effectively. In continuous data, each location has a quantity, magnitude, or intensity assigned to it, and the values are relative to one another. The locational perspective also allows for greater diversity in spatial analysis for both discrete—for example, land use and vegetation type—and continuous data, which will become apparent in the wide variety of discussions accompanying each toolset.
The following table lists the available toolsets in ArcGIS Spatial Analyst and gives a brief description of each.
||The conditional tools allow for control of the output values based on conditions placed on the input values. The conditions that can be applied are either attribute queries or are based on the position of the conditional statement in a list. A simple attribute query might be, if a cell value is greater than five, multiply it by ten; otherwise, assign a value of one to the location.
||By calculating density, you spread point values over a surface. The magnitude at each sample location (line or point) is distributed throughout a landscape, and a density value is calculated for each cell in the output raster. For example, density analysis will take population counts assigned to town centers and distribute the people throughout the landscape more realistically.
||There are two main ways to perform distance analysis in ArcGIS Spatial Analyst: Euclidean distance and cost distance. The Euclidean distance functions measure straight-line distance from each cell to the closest source (the source identifies the objects of interest, such as wells, roads, or a school). The cost distance functions (or cost weighted distance) modify Euclidean distance by equating distance as a cost factor, which is the cost to travel through any given cell. For example, it may be shorter to climb over the mountain to the destination, but it is faster to walk around it.
||The extraction tools allow you to extract a subset of cells by either the cells' attributes or their spatial location. Extracting cells by attribute is accomplished through a where clause. For example, your analysis may require an extraction of cells higher than 100 meters in elevation from an elevation raster. You can also extract by a specified shape. For example, you can extract all cells that fall inside, or outside, a specified circle, rectangle, or polygon.
||Sometimes a raster dataset contains data that is erroneous or irrelevant to the analysis at hand or is more detailed than you need. For instance, if a raster dataset was derived from the classification of a satellite image, it may contain many small and isolated areas that are misclassified. The generalization functions assist with identifying such areas and automate the assignment of more reliable values to the cells that make up the areas.
||The groundwater tools can be used to perform rudimentary advection–dispersion modeling of constituents in groundwater.
||Hydrology functions simulate the flow of water over an elevation surface and create either a stream network or a watershed.
||Surface interpolation functions create a continuous (or prediction) surface from sampled point values. The continuous surface representation of a raster dataset represents height, concentration, or magnitude—for example, elevation, pollution, or noise. Surface interpolation functions make predictions from sample measurements for all locations in a raster dataset whether or not a measurement has been taken at the location.
||In a local function, the value at each location on the output raster is a function of the input values at the location. When computing a local function, you can combine the input rasters, calculate a statistic, or evaluate a criteria for each cell in an output raster based on the values of each cell from multiple input rasters. For example, you can find the mean precipitation for a ten-year period or find how many years the precipitation exceeded 0.5 meters.
||Map Algebra is the analysis language for ArcGIS Spatial Analyst. It is a simple syntax similar to algebraic syntax. For example, to create a slope map from an elevation surface, use the following command: outslope–slope(elevation). Most ArcGIS Spatial Analyst functions can be accessed through Map Algebra.
||ArcGIS Spatial Analyst provides a full suite of mathematical operators and functions. These operators and functions allow for the arithmetic combining of the values in multiple rasters, the mathematical manipulation of the values in a single input raster, the evaluation of multiple input rasters, or the evaluation and manipulation of values in the binary format.
||Multivariate statistical analysis allows for the exploration of relationships between many different types of attributes. There are two main types of multivariate analysis available in ArcGIS Spatial Analyst: (1) supervised and unsupervised classification, and (2) principal component analysis (PCA). A third multivariate analysis, regression, is available in ArcGrid Workstation. Accompanying these analyses are a series of tools to evaluate each step in the analysis process. These tools can be used, for example, to predict the biomass (the dependent variable) at each location given the quantities of precipitation, soil type, aspect, and temperature (the independent variables).
||Neighborhood functions create output values for each cell location based on the value for the location and the values identified in a specified neighborhood. The neighborhood can be of two types: moving or search radius. Moving neighborhoods can be either overlapping or nonoverlapping. Overlapping neighborhood functions, also referred to as focal functions, generally calculate a specified statistic within the neighborhood. For example, you may want to find the mean or maximum value in a 3x3 neighborhood. The nonoverlapping neighborhood functions, or block functions, allow for statistics to be calculated in a specified nonoverlapping neighborhood. Search radius functions perform various calculations based on what is within a specified distance from point and linear features.
||A common spatial analysis query is to identify the suitability of each cell location relative to specific criteria. The criteria can be relative costs, preferences, or risks. Suitability models answer questions such as, Where is the best location to construct a house? What is the cheapest route to build a road? and Which areas should be conserved for deer habitat? The Weighted Overlay tool allows you to easily reclassify your data, weight the datasets, and combine them to create a suitability map.
||The raster creation functions create new rasters in which the output values are based on a constant or a statistical distribution. The Create Constant Raster tool creates an output raster of constant values within a specified map extent and cell size. The Create Normal Raster tool assigns values to an output raster so the values produce a normal distribution. The Create Random Raster (or Map Algebra Rand) tool randomly assigns values to cells on an output raster.
||Reclassifying your data simply means replacing input cell values with new output cell values. There are many reasons why you might want to reclassify your data. Some of the most common reasons are: (1) to replace values based on new information, (2) to group certain values together, (3) to reclassify values to a common scale (for example, for use in a suitability analysis or for creating a cost raster for use in the Cost Distance tool), and (4) to set specific values to NoData or to set NoData cells to a value. There are several approaches to reclassify your data: by individual values, by ranges, by intervals or area, or through an alternative value.
||Using the solar radiation analysis tools, you can calculate incoming solar insolation (global, direct, and diffuse radiation) across a geographic area or for specific point locations. Using an input surface DEM, you can determine the amount of radiant energy that is received from the sun across a landscape for a given period of time.
||With the surface analysis tools, you can gain information by producing a new dataset that identifies a specific pattern within an original dataset. Patterns that were not readily apparent in the original surface can be derived, such as contours, angle of slope, steepest downslope direction (aspect), shaded relief (hillshade), and viewshed.
||Zonal functions take a value raster as input and calculate a function or statistic using the value for each cell and all cells belonging to the same zone. The zonal functions are grouped by how the zones are specified: by a single input value raster or by a second zone raster. You can use the zonal tools to locate the number of endangered species (the value raster) within each parcel (the zone raster) or to find the area or perimeter of each zone in a raster.