Geodatabase creation for personal spatial project

Lab 2 - Watershed Delineation and Risk Calculation

11/4/2017

Part 1 - Watershed delineation

Introduction 

The purpose of this lab is to familiarize the user with the process of watershed delineation. This method is multipurpose. Water quality is a large component of watershed analysis. All water in a single shed flows towards one point, if there is a source of pollution higher in the watershed, all downstream locations will be impacted. Similarly, the amount of water in a watershed is a crucial measurement for resource use, partitioning, and planning. In times of high amounts of precipitation, watershed analysis can also serve a predictive function in flood and overflow potential. While this lab skims the surface of the capabilities of watershed delineation, a fundamental understanding of function and development are established.

Methods 

The process of watershed delineation is complex and for the purposes of this lab, broken down into three steps. The three steps used are data collection, processing, and delineation. Multiple online sources (see in citation section below) were used in the creation of this delineation.

Data Collection

The area of interest (AOI) wA as Adirondack National Park, in New York. An Adirondack Park boundary shapefile was downloaded from the New York Sate GIS Clearinghouse, a hydrology shapefile was downloaded from Cornell University's Geospatial Information Repository, and a thirty-arc-second digital elevation model (DEM) of North America was accessed from ESRI data. The two downloaded sources were opened and extracted into a file geodatabase made specifically for the lab, and the DEM was added to the map document, and then saved to the same geodatabase. From this point, all data were ready to be utilized in the next step of watershed delineation.

Data Processing 
Following the addition of the boundary shapefile to the blank map document, a 20 kilometer buffer was established to obtain a larger clipped portion of the DEM, along with creating a smoother watershed output, and avoiding edge issues that would be visible at the end of the delineation process. Although all of the files downloaded and added covered a similar area, they were in different projections. In order to obtain an accurate output, a reprojection of certain layers had to be performed. The boundary shapefile used a UTM zone 18N projection, with NAD 1983 as a datum. However, the hydrology shapefile used a latitude/longitude NAD 1927. Using the project tool in ArcMap, the hydrology layer was assigned the same projection as the boundary layer. Following this projection change, the streams in the hydrology layer were clipped to the park boundary polygon. Due to the polygon being the AOI, any data outside of the border was unnecessary for the purpose of this lab. The aforementioned DEM was then added to the map document from ArcGIS online. Similar to the hydrology layer, the DEM required a projection change and clip. Using raster processing the DEM layer was clipped to the park boundary, and the map document was saved, the original added DEM was the removed from the map document. The clipped DEM was then reprojected to have useful units of measurement which in this case was meters. The park boundary projection was imported and the DEM, now matching both other layers (hydrology and boundary) utilizes NAD_1983_Zone_18N as a projection. The DEM was additionally resampled, using a bilinear technique with a 60m output cell size. All previous layers with differing projections were then removed from the map document.

Watershed Delineation 
The first step in delineation involves calculation of flow direction. This is completed using the Flow Direction tool. The tool determines the direction of steepest decent and then assigns the cell in question a value based off of the calculation. From this calculation, sinks are created in an output raster. A sink is a cell with an unassigned flow direction value, and thus impedes the accuracy of the flow raster. The sinks are removed using the fill tool. This tool fills sinks by assigning values equal to neighboring cells. Now that the DEM is without unusable values, the flow direction tool is again used. The output of the previous step is then used to calculate flow accumulation, to show where water accumulates and creates channels.  A source raster is then required for the function of input in the delineation. The source raster is created using a conditional SQL, where a minimum threshold is established. The threshold denotes the minimum value a cell must have to be considered a stream cell. This value determines the amount of watershed delineated and must be chosen with care. For the purposes of this lab, the threshold value chosen was 50,000. The stream link tool, which assigns unique values to the intersection of linear raster features, was used for the identification of individual streams. The watersheds of the AOI were then able to be delineated using the layers previously calculated. The second flow direction raster acts as the input flow, and the source raster functions as the input raster. The watershed layer is then clipped to the park boundary layer, and a stream shapefile is added to aid in visual comprehension of the map. Follow the completion of all steps, the process was  done again using a 120 meter cell size (compared to the previous 60 meter) as well as differing threshold values, Output from the original set of parameters is shown below.

Fig 1: Different watersheds in the Adirondack Park boundary

Discussion/Conclusion

Wetland delineation serves multiple purposes. This portion of the lab focused on familiarization with the process of delineation. From this point, many applications are possible. One of these such applications can be seen in the second portion of the lab, 

Part 2 - Risk Assessment 

Introduction 

The focus of this portion of the lab is on water accumulation in areas that are unable to process uncharacteristically large amounts of precipitation. The AOI for this portion of the lab is Gentofte, Denmark. Denmark as a country has experienced high amounts of cloudbursts. The massive amounts of rain in short periods of time result in large discharge and backup related issues."Bluespots" are areas within the AOI that are in significant danger of overflow if a cloudburst were to happen. All data used in this portion of the lab was downloaded through the online ArcGIS lesson "Find Areas at Risk of Flooding in a Cloudburst". 

Methods 

Inside of the downloaded data were multiple databases, and tools. These tools contained models that will be used later in this portion of the lab. The Gentofte map was added to the viewer, this map shows the AOI, as well as buildings included within in it. From this map it is easy to see that not only are there many buildings within the AOI, but also bodies of water, and large areas of vegetation. A DEM layer is then added, which shows the elevation values of the study area. Although the bluespots involved in the study will expand out of the city boundary, it is important to keep them within the study to analyze potential sources of flooding. Using model maker and following inspection of individual elements and environments, the Identify Bluespots (shown below) model is validated and ran.

Fig 2: blue spot identification and building adjacency model

 This model runs multiple processes on the map document and data to create a useful output. Two layers, one showing bluespots, and the other showing buildings that are located within or adjacent to the bluespots. Following the addition of the new layers to the map document, and several changes to the symbology of the map document, an output is created. 

Fig 3: Bluespots and adjacent buildings. 

It can be seen from the map above that there are multiple locations in danger of flooding in the event of a cloudburst. 

A more advanced version of the model is ran in the next portion of the lab. This model takes the analysis of bluespots a step further by calculating the amount of rainfall needed to make each bluespot fill up.  Filling up is based on the theory that each bluespot has one and only one watershed contributing to its water level. This is useful in identifying which locations are in more danger of flooding than other locations. While not all water received in a watershed reaches the bluespot, a cloudburt results in a vast majority of the water received arriving at the bluespot as a final destination. This model was run on the same map document as the previous analysis (that is, not ontop of the previous model), and is shown below. 

Fig 4: Bluespot fill model

This method is longer in steps than the first due to the extra analysis completed. Following validation and running of the model, the output layers are added to the map document. The symbology of the filling layer is adjusted so that the relative danger resulting from amount of water needed for flooding is clearly conveyed.

Fig 5: Bluespot fill requirements

In relation to the focus of this lab a watershed layer was added to the map document, so that the relation between stream systems, watersheds, and bluepoints could be observed. 

Fig 6: Sheds and Hazards

Using this information, a simple query was used to identify which buildings were at a particularly high risk, due to nearby bluespots.  

Fig 7: High risk buildings

Following the addition of transportation shapefiles, a similar query was also performed to identify regions of transportation systems that have a relatively low danger rating in the event of a cloudburst. 

Fig: 8 Low risk transportation

Discussion/Conclusion

This lab highlighted the importance of watershed delineation by using specific examples involving human safety. While this is not the only application, it conveys the point that knowledge of delineation is important both on a personal and organizational level. This lab greatly aided in the understanding of steps in creating an accurate delineation, the changes in output that come from different parameters, and the practical uses of the knowledge. 

Citations 

“ArcGIS Pro.” How Flow Direction works—Help | ArcGIS Desktop, pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/how-flow-direction-works.htm.

“Cornell University Geospatial Information Repository.” CUGIR, cugir.mannlib.cornell.edu/index.jsp.

“Find Areas at Risk of Flooding in a Cloudburst.” Find Areas at Risk of Flooding in a Cloudburst | Learn ArcGIS, learn.arcgis.com/en/projects/find-areas-at-risk-of-flooding-in-a-cloudburst/.

“GIS.NY.GOV.” NYS GIS Clearinghouse, gis.ny.gov/.