Ib. Creation of 2001 DEM

In February 2001 the Japan International Cooperation Agency (JICA) in cooperation with various Guatemalan agencies, generated new digital topography of some key hazard areas, including Santiaguito and Pacaya, from aerial photographs (JICA, 2003). We used the 2001 digital contour lines from JICA to interpolate 10 m DEMs using the ANUDEM algorithm developed by Hutchinson(2003).   Digital contour lines were obtained from JICA for each geologic quadrangle. The contour lines were given in a CAD format (.dgn).  The data file contained all of the elements on the quadrangle (roads, streams, contour lines etc.).  To only view the contour line, I first identified a contour line to obtain all of the necessary codes and then changed the drawing layers to only reflect the necessary codes. Once the contour lines are selected the data was exported to a shapefile by choosing Data>Export Data. The .dgn files had many grids per each quadrangle; therefore this process was done multiple times to convert the contour lines to shapefiles.
  
Note: The above process was only completed for the contour lines. Additional processing using the previous steps may be used to extract the roads and rivers for the region.

Next all of the contour lines were merged into one layer for each quadrangle. In ArcGIS 9.0 this was done using the append tool.  In ArcGIS 9.2 there is a much easier tool located in ArcToolbox under Data Management Tools>General Tools>Merge.  This allows the user to put in multiple shapefiles and create a new shapefile with the merged contour lines. Once the contour lines are merged into one shapefile or feature class the contour lines can be converted to a DEM. The location of the Topo to Raster tool is shown below:

                         


The Topo to Raster menu has many input parameters and most of the time you will only use three of these. Here you input the feature classes (line or point) you wish to convert to a DEM.  The feature data is input, along with the output raster name and an output cell size. For this study I created a 10 meter DEM using 20 meter contour intervals. The cell size should be based on the original data quality of the dataset. This information is normally obtained from the metadata or in this case I used the information from the project report.  Be sure you are using an appropriate cell size to represent your data resolution. More information on this may be found in the DEM Error References. For this study I did not change the other parameters and used the defaults. I created 3 DEMs for Santiaguito representing the Colombá, Retalhuleu, and San Lorenzo quadrangles for the Santiaguito region and Amatitlán quadrangle for Pacaya. The quadrangle boundaries were used to encompass the same extent as the 1954 DEMs.  

Projection
The 2001 data was originally in a Guatemalan Transverse Mercator coordinate system with the parameters shown below:

Projection: Transverse Mercator
False Easting: 500000.000000
False Northing: 0.000000
Central Meridian: -90.500000
Scale Factor: 0.999800
Latitude of Origin: 0.0000000
Linear Unit: Meter
Geographic Coordinate System: GCS_WGS_1984

Once the data is converted to an ESRI Grid (DEM) the coordinate system will be listed as unknown. First I defined a projection using the GTM coordinate system as shown below:

 


In the define projection menu first specify the file the coordinate system will be assigned to.  Then open the coordinate system menu to choose the coordinate system.  I have already created a GTM projection file, which will be available on the network, so I chose select and navigated to where the projection file is saved. This process will be done for each DEM, in this case I have 4, 3 for Santiaguito and 1 for Pacaya. For my data I decided to process each DEM separately until the data is converted and the projection is defined. I then mosaicked the DEMs together into one raster for Santiaguito, which will be discussed later.  Now adding the 1954 and 2001 DEMs into ArcMap you see they do not line up and are a large distance apart; they are in different coordinate systems. To remedy this I converted the 2001 DEM to the 1954 DEM coordinate system of UTM Zone 15 NAD 1927 as shown below:





When using select to choose the coordinate system, navigate to Projected Coordinate Systems>UTM>NAD1927 and choose NAD 1927 UTM Zone 15N, which is the zone for Guatemala.  Converting the 2001 DEMs to the 1954 coordinate systems aligns the data rather well and from my investigation is the best possible fit for the data without ground control points. Now the 1954 and 2001 datasets should be aligned. The data for Pacaya is now complete since we only have one DEM for each year.  For Santiaguito we have three DEMs (Colombá, Retalhuleu, and San Lorenzo), for 1954 and 2001, which represent the volcano and areas downstream.  Next I mosaicked these DEMs into one single DEM.  This step does have not have to be completed, but may be helpful for analysis on the entire dataset. The mosaic tool is shown below:



In the Mosaic to New Raster menu add the Rasters which will be mosaicked. Next assign an output location, by choosing the folder to save the data in, and then name the raster dataset.  Assign the coordinate system; in this case it would be UTM Zone 15N NAD 1927.  And lastly assign the cell size to whatever your dataset is, in this case 10 m. I leave the other menu defaults as is.
Now the datasets are in the same coordinate system, are aligned on top of one another, and are mosaicked into one DEM.  

Note: The 2006 DEM data will probably be in the GTM coordinate system.  Using the steps above it would be best to convert them to the UTM Zone 15 NAD 1927 coordinate system to align with the 1954 and 2001 datasets. If only the 2001 and 2006 datasets are analyzed both can be left in the GTM coordinate system.

Next I describe the DEM Difference procedures located here.