GIS 5935 Lab 5

 Surface Interpolation Lab

This week we covered the general topic of Interpolation and examined  Thiessen, IDW and Spline techniques in greater depth, Spatial Interpolation is the prediction of variables at unmeasured locations, and based on a sampling of the same variables at known locations (Bolstad, 2019, pp. 522). 

For the first portion of this weeks lab we were asked to created a DEM using IDW and Spline interpolation using provided elevation point data. Once we completed creating the DEM's we utilized the raster calculator in order to compare the values from the two DEMs. Below is the comparison map layout from this portion of the lab.

For the second portion of this weeks lab we were tasked with interpolating the water quality data for Tampa Bay. We used a provided data set and focused on the BOD (Biochemical Oxygen Demand) water quality parameter. We used four interpolation methods: Thiessen, IDW (Inverse Distance Weighted), Spline regularized, and Spline tension. We examined the minimum, maximum, mean, and standard deviation values of each of the interpolation methods.  

                 Thiessen

                      IDW

 Spline Regularized

         Spline Tension

Each method differs in its approach. Thiessen interpolation technique assigns an interpolated value equal to the value found at the nearest sample point. It is With IDW the sampling point has a local influence that decreases with distance, meaning that points closer together are weighted more than those further away Spline uses polynomial functions to minimize the surface curvature, resulting in a surface that passes exactly through the input data points. In this lab we used both regularized and tension splines.

For the parameters in this lab I chose Spline tension as the preferred interpolation method. It minimizes curvature and passes through each data point taking consideration of the sampled min and max producing a better and smooth surface. This particular technique showed the most fluid, smooth interpolation and was visually more realistic than the other methods. Also, the standard deviation and data range for this method was also relatively low, meaning it was more precise than Thiessen or Spline Regularized. The IDW result was not as successful at predicting valleys and peaks. There is no right answer to the question of which method is the best. You must make you selection on a case by case basis. 

References:

 Bolstad P. (2019). Gis fundamentals : a first text on geographic information systems (6th ed.). XanEdu.

GIS5935 Lab 4

 Surfaces - TINs and DEMs

This week we examined TINs (triangulated irregular networks) and DEMs (digital elevation models). There are pros and cons associated with each of them and neither is exclusively superior to the other. There are many factors to consider when choosing which one to use for your project. For this lab we created 3D visualizations of elevation models, created and modified a TIN, and compared the TIN and DEM contour lines for our study area.

We began by adding an elevation source to give the TIN a 3D surface and using Vertical Exaggeration to visually enhance the depth.

Next, we utilized a DEM to crate a suitability map for a Ski Run. We used various tools such as Rast to TIN, Reclassify, Slop and Aspect. Once we had our various new layers I chose to use Weighted Overlay to create my final suitability raster using the following weights: 25% aspect, 40% elevation and 35% slope. Below is the 3D representation of my results:

Our next step was to add the Bear Lake TIN and examine the data. TINs vector data models made up of irregular triangles. The three points (nodes) of the triangle are made up of elevations data. The node area connected by line segments which become the edges of the TIN. We examined the data, changed the slope, aspect and drew the edges using Simple symbology the results are pictured below:

Lastly, we created a TIN, utilized the Spline tool, created Contours and examined the differences in the TIN contour lines and the DEM contour lines.

TIN contour lines                                                    DEM contour lines

  

Both sets of contour lines are pictured together here in a two different perspectives:

This view is an overhead "bird's eye" view. As you can see here (above) the greatest difference in the two sets appears to be in the lower elevation areas. The sharp lines of the TIN clearly deviate from the curving lines of the DEM.

This view is a closer straight on view of a higher elevation area:

As you can see there appears to be less variation between the TIN and DEM contour lines in this view.

DEMs are more readily available for general use. TINs are data heavy and can be expensive to obtain and process. TINs are more accurate in areas with a greater number of elevation points available and can be more accurate than DEMs. As a result, it is not possible to say than one is better than the other. It is a case by case decision that has to be made weighting all the factors for each project.

 

GIS 5935 Lab 3

 Data Quality Assessment Lab

This week we continued working in the area of Data Quality.  Prior to the lab we had readings which introduced us to various components of spatial data quality and methodologies used to measure them. We were also introduced to TIGER road networks which are created by the U.S. Census Bureau. 

The goal of this weeks lab is to determine the completeness of two provided road networks; one TIGER network and one Street Centerline network for Jackson County. When examining road networks, it is important to know which areas are well covered and which are not (Haklay 2010) We did this by comparing the total length of the roads in the two networks.

The goal of the accuracy assessment was to visually and numerically compare a summary of differences between The TIGER Roads network and the Jackson County Street Centerlines network. The information was broken down into polygons (1 km x 1 km), inside a county boundary.

 The analysis methodology for the county boundary was a simple length comparison. The greater the length the greater the complete ness. This was quickly attained by calculating the length using the Calculate Geometry tool. TIGER Roads was found to be more complete having a length greater than Street Centerlines (by 570 kilometers).

The analysis methodology at the grid level was more complex but did not follow a standardized protocol. I used multiple ArcGIS tools: Clip, Summarize Within, Calculate Geometry and Table Join; and Excel for formulas.

The map below is the result of the second analysis in this lab:

References:

Haklay, Muki. (2010). How good is OpenStreetMap information? A comparative study of OpenStreetMap and Ordnance Survey datasets for London and the rest of England. Environment and Planning B: Planning and Design. 37. 682-703. 10.1068/b35097. 

GIS 5935 Lab 2

 This weeks lab was on Data Quality - Standards

We learned about various spatial data quality standards focusing on the National Map Accuracy Standards (NMAS), and National Standards for Spatial Data Accuracy (NSSDA). The NSSDA is one o the most recent standards to be issued by the Federal Geographic Data Committee (Minnesota Planning Land Management Information Center, 2019) Specific metrics that are used when measuring positional accuracy within these standards, such as: Sum; Average; Root Mean Square Error (RMSE); 68th, 90th, and/or 95th percentile. 

We were provided with 2 sets of street map data. We used orthophotos as reference to select and digitize a set of points (20) which we would use to test the accuracy of the two provided street map layers. The points chosen were primarily T-intersections. The points were equally distributed throughout the map. As our study area was relatively square I chose to divide it into four quadrants and select five points in each one.

Below you can see the even distribution of my manually selected reference points in yellow. The two provided street map layers are in pink and green.

X and Y coordinates were determined for all points, test and reference. The data from the attribute tables was then exported to excel and we utilized the Horizontal accuracy statistic worksheet from our provided Positional Accuracy Handbook .pdf. The horizontal accuracy worksheet had to be completed twice. Once for each set of test points obtained from the provided street map layers. Once the formulas were added and the calculations were completed the worksheets could be compared to see which map contained the most accurate data. We were tasked with putting our results into a formal accuracy statement (per NSSDA guidelines). The results match what I expected from the visual inspection of the data, the city of Albuquerque data was the most accurate of the two test sets.

Positional Accuracy: Tested 29.22 feet horizontal accuracy at 95% confidence level for the Albuquerque test points.

Positional Accuracy: Tested 369.68 feet horizontal accuracy at 95% confidence level for the Street Map USA test points.