It has been an honor to represent the Environmental Science & Policy Department as the first GMU alum to serve as a COVES (Commonwealth of Virginia Engineering & Science) Policy Fellow for 2020. This is a program run by the Virginia Academy of Science, Engineering and Medicine, so I thank them for their support along the way and I thank my former Ph.D. committee chairman Dr. Alonso Aguirre for recommending me for this program.
Each fellow serves at a different site. I was selected to serve at the Virginia Department of Conservation & Recreation (DCR), where I had the opportunity to work with the wonderful people in the Division of Dam Safety and Floodplain Management. Because of Covid-19, all of the work was done from my home office (Fig. 1).
Fig. 1. The home office of Nick Walker.
My work at DCR was divided into three sections: Dam Safety GIS Data Production, Floodplain Management and a review of Virginia’s Coastal Zone Management (Section 309). The fellowship also included mentoring and weekly meetings with some of the leading experts in science policy in Virginia. I also presented a PowerPoint to the office at DCR, which you may download here.
Fig. 2. Dam locations in Virginia.
My first assignment was to go through the list of dams – over 3,000 of them – and check the GIS data to verify if the map already included a waterbody for the downstream area (Fig. 2). If it did not, I either copied the waterbody data from the National Hydrography Dataset or created one manually. Once all the waterbodies were complete, we generated watersheds using the ArcHydro toolset. I couldn’t get the toolset to run consistently on my computer, so I had to send the dataset over to DCR and then the results were sent back to me for further analysis. This was one of the challenges with working remotely, issues that could probably have been solved in a few minutes in an in-person meeting often required creative workarounds. Finally, we performed quality assurance and quality control on the new watersheds by comparing them to a topography layer and editing as necessary (Fig. 3).
Fig. 3. Watershed generated by ArcHydro and displayed over a topographic map.
The next step was to work on the floodplain data, by combining datasets on flood hazards, insurance, buildings, HUD opportunity zones, social vulnerability and more. Many of the datasets were not in compatible formats and required adjustments in ArcMap and Excel to make them work together. For example, the insurance data do not have a spatial reference (making it impossible to use tools like spatial join) and sometimes the floodplain layers overlapped (e.g. if a building is located in more than one floodplain) so a few Visual Basic scripts were used to generate new columns of data. With over 4 million buildings in the dataset, it would have been impossible to test each location manually. Additionally, because of mismatches with some of the datasets, it was necessary to use a 10 m buffer around the buildings and other polygons. This was my first time using scripting in an ArcGIS project, so even though the coding was simple it was a good learning experience. One of the examples from that output is shown (Fig. 4), with the buildings color coded to each flood zone.
Fig. 4. Buildings and flood zones in Virginia.
Once we had been able to merge the datasets, we were able to combine the VIMS Social Vulnerability layer with the jurisdictional boundaries, as shown (Fig. 5).
Fig. 5. VIMS Social Vulnerability zones for Virginia, aligned with jurisdictional boundaries.
Data from Conserve Virginia provided an additional challenge, since it included four categories (VIMS Marsh, EcoCoastResil, VWC Wetlands and Floodplain) and we needed to overlay those with the three floodplain layers from Fig. 4, as well as the buildings. My first attempt was simply to tally up the numbers for category and join them to buildings, using the ArcGIS functions vector to raster, cell statistics, reclassify and extract by mask. Converting to raster makes it easy to sum each category, even if the output is a shapefile in this case. But it didn’t work that well because it turns out there’s very little overlap between the different categories and so the resulting layer only had three unique values.
So, we tried again, this time using the reclassify function on each of the rasters to weight the values for each category. Theoretically with four factors and three floodways it could be possible to produce dozens of combinations (and thus a colorful map) but in practice overlap is still limited even with a 10 m buffer around the buildings. Thus, only six possible combinations are used (Fig. 6), but this still provided twice the information of the previous method.
Fig. 6. ConserveVA data combining four factors and floodplain data, overlaid with buildings.
In addition to the GIS work, I was asked to review a draft of the Virginia Section 309 Coastal Needs Assessment & Strategies. I did so and I have a few suggestions regarding this document. First, I saw little evidence that the funded programs are achieving the desired results. I recommend groundtruthing and conducting assessments of existing projects before starting anything new. Second, I want to see a greater emphasis on biodiversity and interdisciplinarity regarding threats to coastal areas. For example, restoring salt marshes can both prevent erosion and soak up carbon dioxide in the atmosphere. On a local scale, helping communities build raingardens with native plants can reduce flooding in backyards and common areas. So, there are ways for everyone to get involved.
Third, I wish the document had mentioned the concept of Traditional Ecological Knowledge (TEK), a concept of looking at nature through both the lens of modern science and through the knowledge accrued by the experiences of people who have lived on this land for generations. For anyone interested in learning more about that, I recommend this video called “Two Eyed Seeing” featuring Mi’kmaq Elder Albert Marshall. Mr. Marshall is in his 80s and has a slow, calm speaking style that is the polar opposite of what you’ll find on Twitter or TikTok.
If anyone has questions about my experience, feel free to contact me at email@example.com. Thank you for reading!