 Speaker is Tom. Tom, did you want to start loading your presentation? Okay. So don't forget to fill in the peer review. We'll repost the link in just a minute or scroll up the chat. Don't forget to read the chat messages when it comes to the questions. You might have to scroll up a little bit. Otherwise, whenever you're set and ready, Tom. Sure. So hi everybody. My name is Tom Rubino. And today I will be presenting my capstone proposal titled the wildlife corridor design in the valley and rich physiographic province Virginia using fast technology. When I first began thinking about potential projects for my capstone, I thought that it would be appropriate to combine a lot of what I've learned throughout this program. Obviously, the components are related to GIS, but I also want to integrate some of the other concepts that I was introduced to, including the use of fast GIS software, Python scripting and conservation GIS. I'll begin by running through my goals and objectives before getting into some of the background information methodology and the implications of the project. With that being said, I've decided to integrate everything by building a methodology for the design of wildlife corridors at a micro level that can be repeatable with an automated script. I'll get into this in more detail shortly, but specifically this project will focus on the local migration of amphibians and reptiles. I also decided to bypass the proprietary software that I'm accustomed to and will solely rely on fast technology. In order for this to be achievable, the following objectives must be met. First, based on the size and migration distance of the focus species, it will be necessary to acquire and or process high resolution data sets. The landscape habitat and the threats that these species face all need to be considered. Second, habitat suitability modeling is the foundation of this project. It will be necessary to develop habitat requirements for each species based on academic and professional research. And third, as you will hear me explain shortly, the habitat ranges that are publicly available are much more coarse than the actual migrations taking place. So in order to model movement that will sometimes only be measured by meters, the range needs to be refined based on habitat suitability, breeding sites and potential conservation threats. And finally, the methodology developed in this project will be developed into a series of Python scripts. It will also include detailed documentation on settings, its function and data requirements from users. As I completed research on this process, I came across a number of research projects and journal articles on wildlife corridors, their implementation, their success and their validity. Of those that got the most attention were regional corridor projects focusing on large mammals like the Panther. The Florida Wildlife Corridor Project is an example of an ongoing project that is attempted to mitigate habitat loss and vehicle collisions for the Florida Panther by acquiring lands in southern Florida. An initiative in Banff National Park has also led to a corridor project that has resulted in 40 wildlife overpasses crossing the Trans-Canada Highway. Fortunately, similar projects are prevalent worldwide. Other research detailed specific methods for designing corridors. A study by Parks discussed the prevalence of least-cost analysis and its perceived over-alliance in many studies. The point of the study was that while the analysis was correct, in many cases an animal is going to travel the shortest distance regardless of unfavorable habitat. If it needs to cross a highway or enter an urban area, it's going to do it. This is something that I will certainly keep in mind as I start analyzing my results. Literature also describes some of the alternative methods like using circuit theory for wildlife corridor analysis. In this case, the corridors act as an electrical circuit, meaning habitat patches are represented as nodes in the circuit and the genes of animals are the currents flowing between the nodes. This results in pathways based on habitat resistance. I also looked into the GIS-based tools that are currently available for corridor modeling. There are definitely options out there, but the one thing that remains fairly consistent between them is their reliance on Esri software, which is a luxury that some organizations don't have. The one exception that I found was Mark San, which is a conservation tool for designing reserves that uses QGIS. While it isn't solely a corridor tool, it uses many of the same concepts to identify biodiversity gaps. The corridor design project is an extremely valuable resource that I've used a lot during this project that offers guidance on corridor development and background information, as well as a suite of corridor tools packaged as an Esri art toolbox. Linkage Mapper is another option for corridor design using RGIS. The tool identifies adjacent core habitat areas and links them using least cost analysis. And finally, CircuitScape is a tool that creates corridors using circuit theory. I plan on using each of these tools as a way to validate and compare my results. So for over 10 years, I have exclusively used Esri products and I have become fairly comfortable doing so. So when I was introduced to QGIS about a year ago, I realized that I could easily replicate most of what I was able to do in RGIS. There were even some things that I preferred. I also saw the benefits of open software and the communities of users that encourage collaboration and help whenever possible. The use of FOS technology gives users other options outside of proprietary software that is, like I said earlier, sometimes difficult for organizations to acquire. I plan on completing the majority of the analysis within QGIS. However, I am going to rely on the many plugins available, including RAS GIS for raster processing. Ultimately, the project will culminate with a methodology automated using Python scripting. So I would now like to detail the study area for the project. The valley and ridge physiographic province that runs through the Appalachian Mountains in the western portion of Virginia. Physiographic provinces are distinct regions that share similar landforms and geologic structure. In the case of the valley and ridge province, it consists of a series of parallel ridges and valleys. I decided to focus on this area because of its inconstant geology, its diverse population of wildlife, and the emergence of conservation threats as populations increase. While the province is home to a diverse population of wildlife, I decided to focus my efforts on smaller species that don't get the same attention as larger mammals like the black bear. In fact, I managed to focus on species that are the complete opposite of the black bear. I chose to analyze the migration patterns of amphibians and reptiles. So in this province, there are currently 12 species listed as vulnerable, imperiled, and critically imperiled, according to the Virginia Department of Conservation and Recreations Natural Heritage Program, which is a program that ranks species based on their significance and rarity in the state. These species face the same threats as other wildlife, yet are often overlooked. Specifically, in this project, I focused on the eastern tiger salamander and the wood turtle, both meant to be case studies to determine how successful this process could be. So I decided to use these two particular species because of their similarities, but also because of their differences. As you can see in this slide, they share many characteristics. They are similar in size, both are in need of protection under the Natural Heritage Program, and both migrate to breeding sites annually. You can also see that there are some differences, especially when it comes to their selection of breeding sites. For instance, the eastern tiger salamander spends most of its life on land before migrating anywhere from 300 meters to 1.5 kilometers to vernal pool breeding sites. On the other hand, the wood turtle spends much of its time in forested uplands before traveling up to 3 kilometers to breeding sites and streams. The threats that these species face is also dependent on their habitat. Each species has different characteristics and behaviors that will need to be taken into account in the analysis, and especially during the automation process. The connectivity of habitat plays a prominent role in the preservation of these species. As habitat continues to be threatened, especially in this area, wildlife corridors have become more prevalent. They are defined as habitat that is usually linear and connects two or more larger habitat patches. They can be a natural riparian buffers along a river or man-made structures crossing a busy highway. They allow wildlife to move short and long distances without being hindered by threats or by features blocking their path. Corridors are created to mitigate habitat loss, but are also used as a planning tool for conservation efforts such as wildlife overpasses and underpasses. The creation of corridors involves planning, acquiring land, and working with government, non-governmental groups, and land owners. So before any analysis can be completed, the habitat range for each species needs to be determined. The Virginia Department of Conservation and Recreation holds locational data for each of these species. However, the data is not publicly available and requires a licensing agreement. I've gone through this process and it's still ongoing pretty much the entire semester. So I wanted to take a look at options for data that would be more accessible. The USGS GAP program and the International Union for Conservation of Nature provide habitat ranges, but they are coarse and cover large areas. As an example, the habitat range for the Eastern Tiger Salamander covers the entire county of Augusta, Virginia. But the foundation of a wildlife corridor analysis is in the data sets that will eventually form a suitability model. This will represent favorable habitat based on a number of factors that are derived from research and field work. Most importantly, the data sets must be detailed and have a high resolution, which is especially true in this case because of the scale. The digital elevation models will represent elevation and be used to derive slope and aspect. So the USGS is currently producing a LiDAR derived one meter data set, but unfortunately it is not yet available throughout the state. So for now the analysis will rely on a 10 meter elevation model that was also produced by the USGS. The land cover will be represented with a one meter resolution data set created by the state of Virginia. It classifies 11 different types of land cover, many of which are representative of the habitat requirements. The state also produces a street centerline data set that will be relied on heavily. This high quality data set is used for emergency response and should be sufficient and recommend accurately representing threats from roads. And finally, the national hydrography data set and the national wetlands inventory offer additional ways of representing wetlands and water sources, which are both crucial habitat for many amphibians and reptiles. So during my research on wildlife corridors, I found that there were generally two types of models used. The first was an expert opinion model that relied on the opinions of research and academia and different professionals. The other models were evidence based and the result of field observations, telemetry data and GPS recordings. In most cases, the empirical models are preferred. There are, however, limitations to these types of models, which mostly involve the actual existence of the data or, as I mentioned earlier, access to the data. So for this project, I've decided to rely on an expert opinion model due to the restricted access. In order to find suitable habitat, I went through a number of sources and based off the habitat requirements that I researched, I developed habitat stability scores ranging from 0 to 100, with 100 being the most suitable. The scoring system was based off of a wildlife corridor resource that I mentioned earlier called corridor design, which you can see here on this slide. It's pretty much determined based on the species' ability to travel and to breed. An example of the land cover scoring that I developed is shown here on the right. I also listed the other factors that I see being important in creating a suitability model. For some species, all factors will not be necessary. But for this proposal, I completed scoring for most of the factors listed, but subjectivity based on various opinions is going to be inevitable. As my project progresses, I intend to reach out to experts in the field to ensure that the habitat requirements that I've developed are valid. One of the biggest threats to connectivity is to construction and increase use of roads in the area. To account for increased traffic, I've divided the roads into separate classes, with each representing a speed limit range. While this isn't always an indication of how busy a road is, the well-isolated highways and roads where a slow-moving reptile would have a difficult time processing. And it would place less emphasis on a less-used, low-speed backriff. I've also created a separate factor for roads that are divided that would impede wildlife movement with a median or a barrier. But the goal of this project is to create wildlife corridors. And to do so, starting and ending points for our destination and the source must be established. Since I will be modeling breeding migrations, key habitat and breeding sites will serve as a source and the destination. Key habitat will be defined as areas that meet a suitability score threshold of 80 or above, which is the minimum score for successful breeding as determined by the corridor design project. It is also based on a threshold for habitat patch size, which should not be less than a minimum migration distance. These values will be interchangeable based on the species stuff. The breeding sites themselves can be derived from land cover, hydrography, or in some cases, if available, from outside sources like the Virginia Department of Conservation's database of vernal pools. Specifically, we are most interested in breeding sites that fall outside of key habitat areas, indicating that they are fragmented. Earlier I spoke about the accessible yet very coarse habitat ranges that are available through the USGS and IUCN. Unfortunately, these areas are insufficient when attempting to create corridors on species that may only travel just a few hundred meters. So they will need to be refined considerably. In order to do so, a grid or hexagon overlay will cover the habitat range with a value consistent with the maximum migration range for most species. And I currently have this set at 6,000 acres, but it's possible that that could change in the future. This will allow areas that should be focused on to be isolated based on the presence of key habitat breeding sites and conservation threats like roads. This cost path analysis is the process that will design the corridors in this project. This GIS process determines the most cost effective route between a source and destination over a raster surface. In this case, the surface will be a resistance model. I've already discussed the creation of the suitability model, but the resistance model is essentially the same thing, but with its values inverted. A highly suitable forested wetland with a suitability score of 100 would now have a score of zero in the resistance model, indicating that there is no cost to traveling through this favorable habitat. Lease cost analysis will be conducted in each hexagon study area so that each key habitat area is modeled to the closest breeding site in the habitats range. So at this point, I've given an overview for how the process will work. An additional component of my project will be to tie everything together in the form of a script or most likely a series of scripts that will automate as much of the process as possible. This will allow others to use it without having to manually go through each step. So while there is a Python console available within QGIS, the script will be created as a standalone script using the PyQGIS libraries. A separate module also exists for raster processing using RAS GIS, which I'm going to have to utilize as well. Because multiple species, each with distinct characteristics and behaviors, will be analyzed, I am very aware that there will be a lot of moving parts, and there will need to be a lot of user-defined values. My goal is to make the process as simple as possible and allow the users to set their own requirements based on suitable habitat, threats, migration distances, and species ranges, depending on what species is being analyzed. I'm also aware that for this project to be successful, the detailed document will be needed that not only goes over the methodology, but includes the background and the different settings that will be required. So the results of this project are intended to provide organizations with more information about the species that they are studying and protecting. It's focused on species that are considered vulnerable in Virginia, and often overlooked will offer an additional data point for these organizations. The process will result in modeled wildlife corridors that can identify areas of concern and help with mitigation strategies, including the installation of signage and placement of road culverts. It could also be used as a tool for knowing where key habitat areas exist as economic expansion continues to spread and cause habitat loss. It is the intention of this project to provide a simple way of identifying these areas with data and software that is readily available. So graphic gives a general timeline for the project. I've identified the Society for Conservation GIS conference as a good fit for the presentation. There's also a few smaller conferences that I'm also considering, but most don't take place until spring and summer of 2018. So I've dedicated the majority of my time to writing and testing scripts, but I also need to manually run through each process, test the results, and I want to actually reach out to professionals in the field for feedback on the validity of my methods and my results. Here is my reference page and pending any questions. This concludes my presentation. Thank you very much for listening. Yeah, Justine asks, the scripts you'll be putting together, will you be packaging these up so that they are standalone or as add-ins? Yeah, right now I think they will be a series of scripts that I'll package together. I also need to look into adding plugins and QGIS, which is something that I've read a little bit about, but I don't have any experience doing it, so that is an option. But my goal is to hang them somewhere, whether it's a web page or a GitHub page. Joe says, can you comment on the process for receiving the protected data from the state? Yeah, I've exchanged a lot of emails with them. They've actually been extremely helpful, so I had to give them a list of the species that I wanted. And I also requested the locations of urinal pools, but it's the government and it's taking a very long time, so it's pretty much just hung up after we've signed some papers. Justine says, I think I missed this, but what species will this be applicable to? Specific or arranged? Yeah, I wanted this just to focus on smaller amphibians and reptiles, so the goal of the project is for it to be usable with any amphibian or reptile. For this project, I just used the Eastern Tiger Salamander and Wood Turtle as case studies. They both have some different characteristics that I hope I could figure out how it works with both species. Anyone have a final question? I think that may be it. Thanks, Tom.