GIS/MEA582: Geospatial Modeling and Analysis

This is an unmaintained course material, please see current material at:

Geospatial Analysis I: global, zonal and focal operations, map algebra

Resources: ESRI virtual campus Create folder Map_algebra in your preferred workspace. Include resulting maps into your report (use your judgement to decide which results to include).

Start ArcMap:

Start->Programs->ArcGIS->ArcMap

Compute zonal statistics for feature data

Add landuse feature class data.

File->Add Data->Add Data...  ./ncrast.mdb/landcl96
right click landcl96->Open Attribute Table

Add a field (column) to the attribute table:

left click Table Options menu (upper right corner of Table window) icon->Add Field...

Set Name to 'Acreage' and Type to 'Double'

Calculate acres (populate the Acreages field):

right click the column heading and select 'Field Calculator'

You may see the pop-up window with the warning: You are about to do a calculate outside of an edit session. This method is faster than calculating in an edit session, but there is no way to undo your results once the calculation begins. Do you wish to continue? Click YES

double left click on Count to populate equation in the bottom menu
Click on Multiply * button
Type in 812.25 (i.e. area of each cell in square meters) 
Click 'OK'
Note the new values in the Acreages field
Close the Table

Compute %area for each category in each zipcode zone. Add zipcode data:

File->Add Data->Add Data... ./ncrast.mdb/zipcodes

Check out a Spatial Analysis Extension license:

left click Customize Menu->Extensions...  check Spatial Analysis

Open the ArcToolbox
Select Spatial Analyst Tools->Zonal->Tabulate Area
Set Input Raster or feature zone data to 'zipcodes'
Set Zone field to 'VALUE'
Set Input Raster or feature class data to 'landcl96'
Set Class field to 'VALUE'
Set Output Table to '.\Map_algebra\zip_sum'
Set Processing cell size to 28.5
Click 'OK'

View the Tabulate results:

Click on the 'List by Source' button at the top of the Table of Contents (TOC)
right click zip_sum->Open

Explain how the data are presented.

Handling NULL data

Create a copy of SRTM DSM map, converting cells with elevation=0 to NULL:

File->Add Data->Add Data... ncrast.mdb/elev_srt
right click elev_srt->Data->Export Data...
Set the Location to '.\Map_algebra'
Set the format to 'GRID' and the Name to 'elev_srt2'
Click 'Save'
Click 'Yes' to add the GRID layer to the TOC

Replace cells with elevation = 0 to nodata / NULL (i.e. -9999 for ArcGIS):

Open the ArcToolbox (if not already open)
Select Spatial Analyst Tools->Conditional->Set Null
Set Input Conditional Raster to 'elev_srt2'
Set the Expression to VALUE = 0
Set Input false raster or constant value to 'elev_srt2'
Set Output raster to .\Map_algebra\elev_srt3
Click 'OK'

Reclassification

Recode a 360-direction aspect map to 8 directions (D8):

File->Add Data->Add Data...  .\ncrast.mdb\aspect
select Spatial Analyst->Reclass->Reclassify
Set Input raster to aspect
Click the 'Classify' Button
Set the Method to 'Equal Interval' and Classes to '8'
Click 'OK'
Set the Output Raster to '.\Map_algebra\aspect_d8'
Click 'OK'

Mosaicing

Patching raster data (often called mosaicing):

File->Add Data->Add Data...
Select ncrast.mdb/elid_782 (hold down control key to select multiple layers)
   select ncrast.mdb/elid_783
   select ncrast.mdb/elid_792
   select ncrast.mdb/elid_793

Open the ArcToolbox (if not already open)
Open Data Management Tools->Raster->Raster Dataset->Mosaic To New Raster
Set Input Rasters to 'elid_782' , 'elid_783' , 'elid_792' , 'elid_793'
Set Output Location to '.\Map_algebra\'
Set Raster dataset name to 'elev_6m'
Set Number of Bands to 1
Leave the remaining fields as their defaults
Click 'OK'

Optional - have a look at the patched DEM in ArcScene.

Creating a subset

Create a subset of the map elevation for a subregion:

File->Add Data->Add Data...
Select ./ncrast.mdb/elev
(Note: hold down control key to select multiple layers)
Select ncrast.mdb/ortho01
Open the ArcToolbox (if not already open)
Select Spatial Analyst Tools->Extraction->Extract by Mask
Set Input Raster to 'elev'
Set Input raster of feature mask data to 'ortho01'
Set the out put raster to '.\Map_algebra\elev_mask'
Click 'OK'

Map Algebra

Compute Normalized Difference Vegetation Index (NDVI)

Explain the difference between FP and INT handling.
(Note: make sure you have spaces around operators (=,-,/) ).

File->Add Data->Add Data...
Select ./ncrast.mdb/lsat7_40
Select ./ncrast.mdb/lsat7_30
Open the ArcToolbox (if not already open)
Select 'Spatial Analyst->Map Algebra->Raster Calculator
Type in the expression:
   ("lsat7_40" - "lsat7_30") / ("lsat7_40" + "lsat7_30")
   Set Output raster to './Map_algebra/ndvi1
Click 'OK'
Select 'Spatial Analyst->Raster Calculator
Type in the expression: 
   Float("lsat7_40" - "lsat7_30") / Float("lsat7_40" + "lsat7_30")
   Set Output raster to './Map_algebra/ndvi2
Click 'OK'
Assign the ndvi2 map a color ramp that has green for high values

Compute the difference between the SRTM DSM and lidar-based NED DEM:

File->Add Data->Add Data...
Select ncrast.mdb/elev_ned
Select ncrast.mdb/elev_srt
Open the ArcToolbox (if not already open)
Select 'Spatial Analyst->Map Algebra->Raster Calculator
Type in the expression: "elev_srt" - "elev_ned"  
Set Output raster to './Map_algebra/elev_diff
Click 'OK'

Create a custom color table to distinguish the negative and positive values:

right click elev_diff->Properties
Click on Symbology tab
Select a 'Red to Green' color ramp
Click 'OK'

Working with the conditional function

Create map of urban areas:

File->Add Data->a Add Data...
Select ncrast.mdb/landcl96
Open the ArcToolbox (if not already open)
Select 'Spatial Analyst->Map Algebra->Raster Calculator
Type in the expression: Con( ("landcl96" == 1) | ("landcl96" == 2),"landcl96",0)  
(Note: in ArcGIS 10 map algebra components are case-sensitive - i.e. con != Con) 
Set Output raster to './Map_algebra/calcmap5a
Click 'OK'

Alternatively use the 'pick' operation:

Open the ArcToolbox (if not already open)
Select Spatial Analyst Tools->Conditional->Pick
Set input position raster to landcl96
Set the input rasters as constant values: 1 and 2. 
Enter 1 and then 2 such that there are two entries (each value is on its own line).
Set output raster to '.\Map_algebra\urban_pick'
Click 'OK'

Handling null values

Check whether the land use map correctly classifies lakes as water.

File->Add Data->Add Data...
Select ncrast.mdb/landcl96 (this layer may already be present in your project)
Select ncrast.mdb/lakes
Open the ArcToolbox (if not already open)
Select 'Spatial Analyst->Map Algebra->Raster Calculator
Type in the expression: Con(IsNull("lakes"), "landcl96", "lakes" / 1000) 
Set output raster to '.\Map_algebra\luselakes'
Click 'OK'

Create a map where all land use categories > 1 are set to null:

Open the ArcToolbox (if not already open)
Select 'Spatial Analyst->Map Algebra->Raster Calculator'
Type in the expression: SetNull( "landcl96" > 1, "landcl96" )     
Set output raster to '.\Map_algebra\developed'
Click 'OK'

Create mask for areas with elevation between 60 and 100m and land use 1 or 2:

File->Add Data->Add Data...
Select ncrast.mdb/landcl96 (this layer may already be present in your project)
Select ncrast.mdb/elev   (this layer may already be present in your project)
Open the ArcToolbox (if not already open)
select 'Spatial Analyst->Map Algebra->Raster Calculator'
Type in the expression: 
SetNull( (("elev" < 100) & ("elev" > 60)) & (("landcl96" == 1) | ("landcl96" == 2)), 1)
Set output raster to 'mask'