Introduction to Geographic Information Systems

ERE 450/550

Lab Exercise 5

Lab due on Wednesday December 6 in class

 

Purpose

The purpose of this lab exercise is to address the city’s request for a relative urban temperature analysis.

 

This lab is an extra credit assignment with grade point opportunities worth up to three regular labs (30pts).  The spatial model of this exercise is worth up to 10 points, the production of a map illustrating the analysis is worth up to 10 points, and the tables detailing the results of this analysis are worth up to 10 points. However, the maximum points awarded from this exercise can not improve your lab grade beyond the 40 maximum points awarded for the lab component of this course (see course information on my webpages).

 

The spatial model for this exercise must be hand written and follow the format for minimum steps as illustrated in the four previous labs.  The specifics for the table and map are embedded within the exercise.

 

Elizabeth Nifkin has requested a “hot spots” relative urban temperature analysis for the city of Syracuse based on some very general assumptions about factors that affect urban temperature including issues related to land cover, elevation, slope, aspect, and population. She would like this analysis done using the NAD 1983 State Plane New York Central FIPS 3102 datum and projected coordinate system. She would like this analysis done with an accuracy of a 30-meter resolution. She would like a map with appropriate symbolization for the resulting temperature categories using a black and white printer. She would like two tables, one with the acreage breakdown for each of the temperature categories developed from this analysis, and she would like the spatially weighted mean temperature, as well as a few other statistics describing the temperature layer.

 

You will need the following layers: Land cover for Syracuse (nlcdsyr), the boundary of the city (Syracuse.shp), the 2000 census blocks for the city (syrblocksgcs83.shp), and the DEM for the city (syrdem).

 

Data Management:

 

  1. Project (NOT PROJECT DEFINE) all 4 layers using the NAD 1983 State Plane New York Central FIPS 3102 (Feet).prj file (datum and projected coordinate system).

·        Most of this will be raster work, so you have to have the Spatial Analyst extension on.

·        The project tool for raster and feature(vector) layers are different, make sure you use the correct one and especially NOT define projection.

·        For raster projections, don’t worry about NADCON, the resampling during raster projection takes care of that. Be sure to use the default resamping technique, NEAREST.

·        Don’t forget that you are going to be changing the distance units from meters to feet, so that you’ll get the converted feet numbers for cell resolution. In other words, the resolution numbers (output cell size) you get after projection will be in the units of feet after conversion (so 30m = xx.xxx feet)

·        Name your output layers appropriately and make sure they are getting saved where you want them.

·        When selecting from the State Plane projected coordinate systems use the file in the Nad 1983 (feet) folder and not the Nad 1983 (Intl Feet) folder.

·        Name the projected and reprojected layers as follows: nlcdsyr to syrnlcdsp83; Syracuse to syrbndsp83, syrblocksgcs83 to syrcenblkssp83, and syrdem to syrdemsp83.

 

  1. Derivative Mapping – Create Slope (percent) and Aspect from DEM
    • Use syrdemsp83
    • Use Spatial Analyst Surface Analysis functions to create the aspect and slope (percent) layers (you can access these either from ArcToolbox or the Spatial Analyst toolbar under Surface).
    • When doing slope, you want the Percent option for output measurement and you want the Z-factor to remain 1.
    • Name the layers as follows: the slope layer name syrslopeper; the aspect layer name syraspect.

 

  1. Resample Slope, Aspect, and DEM to fit the land cover parameters (30 meter resolution)
    • Use syrslopeper, syraspect, and syrdemsp83 and name the outputs resulting from resampling: RSslope, RSaspect, and RSdem respectively.
    • Use the NEAREST option for resampling technique
    • For the environment settings (“Environments…” button)

Under General Settings: Set the Output Extent (using the browse button) to the same as your syrnlcdsp83. After the top, bottom, left and right coordinates appear in the window below Output Extent then change the Output Extent to As specified Below, and set the Snap Raster to the same layer as used for the Output Extent (syrnlcdsp83 so the window said Snap to Dataset syrnlcdsp83). Under Raster Analysis Settings: Set the Cell Size to the same so the window reads-- Same as Dataset “syrnlcdsp83.” Just in case, after I got back to the main Resample window, I also changed the optional cell size using the same layer I used above.

 

Analysis

  1. Reclassify the reprojected land cover data (syrnlcdsp83) twice. One time to create a binary mask of water named watermask that you’ll use later (0’s for water (value 11) and 1’s for everything else), and the second time for the land cover temperature factor layer named lctemp according to the following table.

 

Old Values

New Values

11

-1

21

0

22

2

23

4

32

3

41

-4

42

-4

43

-4

81

1

82

2

85

1

91

-3

93

-2

 

·        You can use either the tool from ArcToolbox or reclassify from the Spatial
Analyst toolbar.

 

  1. Reclassify the resampled slope layer (RSslope) to a binary slope layer with 0’s for slopes 0 – 5% and 1’s for slopes 5% and higher. Name it binaryslope.

·        When reclassing, I usually use a value the exceeds the maximum value of the layer for the second set of values.

 

  1. Reclassify the resampled aspect layer (RSaspect) to an ordinal aspect layer using the following guidelines --- 1 for South or Flat (-1), 2 for North, 0 for East and West. Name it ordaspect.

·        In the reclassify table: 1 to 0 = 1, 0 to 45 = 2, 45 to 135 = 0, 135 to 225 = 1, 225 to 315 = 0, 315 to 361 = 2)

 


  1. Create an ordinal layer of the North slopes over 5% and name it ordnorth.

·        Use either the raster calculator (from the Spatial Analyst toolbar) or the raster math “times” tool from ArcToolbox to multiply the binary slope layer (binaryslope) by the ordinal aspect layer (ordaspect).

·        If you use the Raster Calculator you should make your calculation permanent after you do the calculation (right click on the layer name in the TOC and select that option)

 

  1. Create the second temperature factor layer (name it ntemp) that classifies the north slopes over 5% as -2,

·        Use ordnorth and reclass the 2’s to -2, and the rest (0’s and 1’s) to 0.

 

  1. Create the third temperature factor layer in two steps by first reclassing the ordinal aspect layer (ordaspect) so that the 1’s remain the same (1) and everything else (2’s and 0’s) becomes 0’s (name the output binarysflt). Next, multiply that layer times the water mask (watermask) to create the final layer of temperature factor related to southern and flat aspect (name it sflttemp)

 

  1. Create the elevation temperature factor layer (the fourth one) by reclassing the resampled elevation (RSdem) values above the mean to be -1 and below the mean 0 (name it elevtemp)

·        You’ll have to check to see what the mean elevation is. Open the attribute table for the resampled DEM, and write down the mean and use it for your reclass.

 

  1. Finally, create the population density based temperature factor (the fifth one) so that the blocks with population density above the mean are given a value of 1 and the others a value of 0.

·        You’ll first have to calculate the area for the projected census block layer (syrcenblkssp83), then add a field and calculate for population density.

·        Don’t worry about converting to any specific units at this point.

·        Because the area units are in feet, the population density will be a really small number (pop/feet), make your field large enough and with at least 10 decimal places.

·        Determine the mean population density and then add a new field for the temperature factor calculating the records above the mean to be 1 and then below the mean 0.

·        You’ll use the select or query records function to do this.

·        Finally rasterize this layer (convert feature to raster) on your temperature factor field using the environmental settings and cell size settings to be the same as the reprojected land cover layer (syrnlcdsp83) name the output popdenbinary.

·        As you did with the slope layer, multiply the watermask layer times the population density raster layer (popdenbinary) to create your final layer for population density temperature factor (name it popdentemp).

  1. Lastly, use the raster calculator to add all the temperature factor layers together. There should be 5 of them: the population density factor (popdentemp), the elevation factor(elevtemp), the southern and flat aspect factor (sflttemp), the north slope factor(ntemp), and finally the land cover factor (lctemp). You should make that layer permanent, and name it finaltemp.

 

 

 

Output

Create your map from the last layer you created (finaltemp) using a symbolization appropriate for representing temperature range for a black and white printer.

 

Use the attribute table of finaltemp to determine the acreage for each temperature factor class. (Count is the number of cells that have that value…so if you know the resolution then you can do the math).

 

Use the zonal statistics as table tool from ArcToolbox (Spatial Analyst/Zonal) to determine the statistics of the layer finaltemp. The Syracuse city boundary is your input raster or feature zone data (because there is only one record you can pick anything for the zone field) and your final temperature layer is the input value raster.  Provide a table with the min, max, range, mean, standard deviation, variety, majority, minority, and median.