GIS applications Week 3: Hurricanes

This week we were to look at Key West, Florida's flooding from Hurricane Wilma's storm surge. I was initially panicked, owing to the fact that I did Project 2 first because it wasn't clear to me whether to do both 1 and 2 but I certainly needed to do 2, and only afterwards did I find the Project 1 work doc (I had thought I had printed out all the docs from both projects and hence had all the instructions I had to go on, but apparently not) which explained pretty much everything I should have done. In any case, I did come up with all the deliverables for project 2, although the table and graph are from Excel since it was quicker and the Project 2 directions didn't specify requirements for creating the table/graph). When I got to the second quiz, the one focusing on Hurricane Katrina, I discovered the Project 1 work doc and worked through that to get to the answers I needed. It made the route a lot clearer. Ah, hindsight...

We had to display a map of elevation and bathymetry of Key West, which showed how low the island is. After this we had to determine what proportion of the island was flooded by an 8-foot storm surge, and make a graph of the flooded land by landcover type. Key West turns out to be heavily developed (densely networked with streets and classified as developed land) across nearly the whole island (the landcover map is the 2nd map on this blog). All I know about the Florida Keys comes from Carl Hiaasen, so this was a surprise to me - but correspondingly, most of the land that was affected was developed land. Only two areas were free of water and both were quite small.

Moreover, all but one of Key West's churches and schools, both of its hospitals and its only airfield are in the low-lying (under 8 feet) area (see the first map). This was a surprise to me too. I don't know how many hurricanes have hit the Keys but there must have been a few and there are certain to be more in the future.

http://students.uwf.edu/db27/Bobwhite.ppt

Final Week: the Bobwhite/Manatee Florida Transmission Line Project

As a start: here is a link to the final Powerpoint about the project: http://students.uwf.edu/db27/Bobwhite.ppt

and just in case it's not retrievable otherwise, here is the link to the pdf:

http://students.uwf.edu/db27/Bobwhite.pdf

Here is a map from the Powerpoint ( I had trouble getting this link to show the revised ppt which includes this particular map, so I put it in here just in case):

Hmm. What to say about this that I haven't already said in the essay and the Powerpoint? For this project we were given some background information about Florida Power and Light's efforts to get approval/certification for an electrical transmission line that would run through part of Manatee and Sarasota counties in Florida. We were also given some data about the project (the actual, somewhat revised corridor for the project was approved in late 2008) and asked to find some more.

We were required to do four things: 1) evaluate the extent of the conservation lands and wetlands within FPL's preferred corridor; 2) count the number of housing units in the corridor and a 400-foot buffer around it, and also count the number of land parcels within the corridor and the buffer; 3) count the number of schools and of daycares within the corridor and buffer; and 4) measure the length of the corridor.

This was an enjoyable project because of the variety of work it entailed. Doing the first part (conservation lands, wetlands) was pretty straightforward - see the map above. Counting housing was more problematic - we did it from aerial photographs - because there were some buildings that looked from the air like what in North Carolina we would call tobacco barns. I didn't count every structure; I tried to count the ones that looked as though there was regular access (driveway, space cleared around the building) as housing. I noted a caveat in the essay because of this and the fact that there was no date for the aerial photos. Parcel counts were easy given the selection/calculation functions in ArcMap, though it's hard to say what is within and what is without the corridor when many more parcels intersected than were entirely within the corridor or buffer.

Schools and daycares were easy to exclude given the useful data from the Florida Geographic Library. Measuring the transmission line was a bit more of a challenge. I found it difficult in ArcMap because the measuring line would disappear when I panned the screen to get to the next section. Instead I found the same area in Google Earth, whose measurement tool is more user-friendly for multiple-screen uses (at least for me), and measured it while looking at the ArcMap aerial photos as well to compare the ground view. I later measured it again in ArcMap, adding short distances together, and got the same result but with more effort.

Thanks for the opportunity to do these labs during the semester. This in particular was a really interesting absorbing project.

Week 11: Labels, Spatial Analyst and 3D Analyst

We had several tasks this week and five maps to produce. The first section I worked on was Spatial Analyst. We had to create a model to conduct spatial analysis of an elevation raster - creating slope (steepness) and hillshade (shows shadows based on elevation and orientation relative to the sun) and aspect (shows compass direction of each raster cell). ArcMap turns out to have three different ways of building such models: using the tool in ArcToolbox, which involves filling out the usual dialog box for input and output; dragging-and-dropping using the ModelBuilder (see the graphic with rectangles and ovals above), and creating the code using the command line. I would describe ModelBuilder as clunky and the command line approach as elegant, in part because the code lets you see what you're doing and ModelBuilder just says "Please do something with this input". But they all seem to work. This short lab section was relatively straightforward.

Second, we had to convert vector line and polygon files to raster files and reclassify those raster files. The largely purple graphic above shows the converted and reclassified vegetation map of a vegetation study area in/near Harlan, Kentucky. I was a bit confused because the white spaces within the study area were originally, in the vector file, "altered/developed" land; we had to reclassify this as NoData, which meant that in the reclassified raster version all that land comes out as white as well as the small portion that was NoData before (hence my note at the bottom of the map).

Third, we had to produce two maps of the Tampa, Florida area, using a variety of labels and annotation to create the map. I have used labels a little in ArcMap but never felt I knew what I was doing, so this was directly useful for me. The Cities and Roads map used different symbols for different kinds of roads; the Paddling Trails & Bird Sanctuaries map also used annotation. These maps reminded me of how tricky it can be to get all the information you need on to a map without crowding it - the Paddling Trails map of course needs roads as well as they are such an integral part of the landscape (and how else do you get to the rivers you want to paddle?) although without the roads it is prettier! I had a lot of difficulty passing the ESRI exam on the labels&annotation section; it's confusing and the only easy part was decoding the Vbscript code. I'm looking forward to the programming class.

The final map, at the top left I think, was from the 3D analyst section. It shows a hiker's-eye view of the cabin ( the red dot) and the two "observers" (the other two dots; this made me think of military/police exercises!). The black snaky lines are actually contour lines - I couldn't resist using the "Contours" icon when I found it, although I had no idea how to use it or how to remove the lines. The white contour line at the top is the last one I created, and since I didn't know how to remove it and it was the last one, it's still white. The 3D function in ArcScene was very interesting - when I wasn't frantically trying to stop the navigation tool from turning my mountains round and round - although I don't have an idea of what I could do with it in real life at this point and I do wonder how useful 3D is given what one can interpolate from 2D maps. The other thing that intrigued me greatly was seeing the overlapping transparent 3D rectangles that make up the 3D display before they had finished drawing the map completely.

Week 9: Vector Analysis, Part II

This week's lab was all about creating and manipulating buffers around lines (roads and rivers) and polygons (lakes). Given a set of shapefiles for roads, rivers & lakes and for conservation areas, we had first to create buffers of varying distances around each of the three features (roads, rivers, lakes). Second, we had to combine these buffers through a "union" function, and finally, we were to create a new shapefile that included all the areas that were a) within the different specified distances of each of those three features but b) not within six conservation areas. a) was represented by performing an "intersect" function on the combined buffers, and b) was achieved by performing an "erase" function on the result of a). Along the way there were a few other manipulations of buffers.

This was a pretty straightforward lab for the most part and it was interesting to see that we could create buffers of different distances for different attributes within the same feature. My greatest difficulty was working with ArcGIS to try to calculate the areas of the final polygons because I couldn't initially get the Area field to be wide enough to include all the numerals. Eventually I tricked it (or so it seems to me), by giving the field a longer name, into giving me the answer I needed.

Answers to the lab questions:

1) Using "Intersect" gave me the same result on the map as using the records within the union of Roads and Water that were within both the roads buffer and the water buffer. The table for this layer contained four records, like the buffers_union_export file (before that file was converted to a singlepart layer, which had 82 records).

2) To exclude conservation areas from the project, I used "Erase", because it cuts out the conservation areas that are within (that intersect with) the buffer area.

3) The possible_sites layer, the final result, has four records, but two of them are the same on the map (in the table they are identical except that one has a buffer distance of 150 and the other of 500, and they have different FIDs); this part makes me wonder whether I have done something wrong as I don't understand exactly why I have (almost) duplicate records. The area of the largest feature is 60,783,617 square meters; the area of the smallest (duplicate) feature is 585,748.

Week 8.5 (Spring Break): Gulf County Land Ownership

For the spring break week we could do this as an optional lab. It was largely about using raw data and joining tables: we had to join a land parcel file that had no associated data to a tax roll table that included ownership data. After that we were to calculate the acreage for each parcel and sort the data table to find the four largest landowners, which are shown on the accompanying map.
This was a fascinating lab for me because 1) Gulf County is the county I was assigned in week 5, 2) as is clear from the map, a very small number of landowners own nearly all of the county, and 3) it's clear from the map, in which I left the boundaries of each parcel clearly marked, that the great majority of the county was divided up in a very systematic way - it looks as though the division is in approximately 640-acre sections. I don't know whether the division was for the purposes of development (my first thought) or farming (my second, after I saw the land parcel size) but it was very striking. It reminded me immediately of Carl Hiaasen's terrific novels about rapacious developers and other lowlifes in Florida - not that I know anything about the specifics of this county. It also reminded me of how helpful a map can be in making sense of a complicated table, or at least in highlighting some of the table's more interesting features.
The lab itself was quite straightforward. I think the map is a bit more crowded than I hoped, but I wanted the county itself to be as large as possible while still fitting in the legend. I included public land (the crosshatched part) to show how that was a large chunk of the remaining large-landowner parcels. The St. Joseph acreage also includes the entries listed as "The St Joe Company", since that had the same two contact people and the same address as "St. Joseph Land Development Company". I didn't combine the entries for TIITF with those for TIITF/GFWFC because it wasn't clear to me that they were the same thing.

Week 7: practice with digitizing features

This week we were to take last week's data files (two raster aerial photos of the UWF campus and a buildings and road file) and modify them by a) creating a new file (athletic fields) and digitizing its contents from the aerial map, b) adding one building via digitizing to the buildings file, and modifying the shape of two others, and c) adding in a road from the aerial photo - the one in the southeast corner that connects the larger road's oxbow ends - again, via digitizing.
This was a useful second part of the lab exercise after the first part, which was to complete the editing module in ESRI's Virtual Campus class following ESRI's step-by-step directions. I had a short period of panic when I added my files one by one only to discover that one of the aerial photos was invisible (because I'd forgotten to do the last step in georeferencing in the previous week's lab, I think) but eventually I got that sorted out. I am also glad to learn a bit more about the actual properties of the files we create and edit.