1 00:00:08,830 --> 00:00:09,663 - All right. 2 00:00:09,663 --> 00:00:13,270 So our next presenter is Caitlin Drasher 3 00:00:13,270 --> 00:00:14,220 with University of Vermont. 4 00:00:14,220 --> 00:00:16,770 She'll be presenting on using electrical circuit analysis 5 00:00:16,770 --> 00:00:21,160 to map landscape connectivity for wildlife in Vermont, 6 00:00:21,160 --> 00:00:22,882 with implications for transportation, 7 00:00:22,882 --> 00:00:25,220 planning and mitigation. 8 00:00:25,220 --> 00:00:26,590 - Yeah, so my name is Caitlin Drasher. 9 00:00:26,590 --> 00:00:28,760 I'm a graduate student at the University of Vermont, 10 00:00:28,760 --> 00:00:30,573 working with Dr. James Murdoch. 11 00:00:31,880 --> 00:00:33,870 And I'll be discussing how we're using 12 00:00:33,870 --> 00:00:36,040 this electrical circuit theory analysis 13 00:00:36,040 --> 00:00:39,170 to look at landscape connectivity for wildlife in Vermont. 14 00:00:39,170 --> 00:00:42,340 And this analysis is part of a larger effort 15 00:00:42,340 --> 00:00:45,130 that's been ongoing to assess wildlife connectivity 16 00:00:45,130 --> 00:00:48,400 across our transportation networks, here in Vermont. 17 00:00:48,400 --> 00:00:49,770 And this is in partnership 18 00:00:49,770 --> 00:00:52,260 with the Vermont Agency of Transportation, 19 00:00:52,260 --> 00:00:53,880 the Vermont Fish and Wildlife Department 20 00:00:53,880 --> 00:00:55,943 and the Nature Conservancy in Vermont. 21 00:00:57,150 --> 00:00:59,480 Today, I'll be discussing one tool 22 00:00:59,480 --> 00:01:01,500 that we're using to visualize wildlife movement 23 00:01:01,500 --> 00:01:04,320 throughout the state, using electrical circuit theory. 24 00:01:04,320 --> 00:01:05,700 So I'm going to start out 25 00:01:05,700 --> 00:01:08,260 with discussing our use of circuit theory 26 00:01:08,260 --> 00:01:11,200 to model wildlife movement for our road analysis. 27 00:01:11,200 --> 00:01:13,040 And I'll end with some examples 28 00:01:13,040 --> 00:01:14,330 of how this type of analysis 29 00:01:14,330 --> 00:01:16,373 can be used for other purposes. 30 00:01:19,130 --> 00:01:20,990 So beginning with some background 31 00:01:20,990 --> 00:01:23,650 on the impacts of roadways. 32 00:01:23,650 --> 00:01:26,810 Roads can have significant impacts on wildlife populations. 33 00:01:26,810 --> 00:01:30,110 They obviously contribute to direct mortality 34 00:01:30,110 --> 00:01:32,010 through wildlife vehicle collisions, 35 00:01:32,010 --> 00:01:34,300 but they can also lead to some indirect mortality 36 00:01:34,300 --> 00:01:37,280 due to their effects relating to habitat. 37 00:01:37,280 --> 00:01:39,030 Roads fragment habitat, 38 00:01:39,030 --> 00:01:40,490 they divide up the landscape 39 00:01:40,490 --> 00:01:41,700 and prevent some species 40 00:01:41,700 --> 00:01:44,470 from accessing important resources. 41 00:01:44,470 --> 00:01:46,520 They also impede dispersal 42 00:01:46,520 --> 00:01:50,800 and decrease genetic exchange between subpopulations. 43 00:01:50,800 --> 00:01:53,410 And over time, this can lead to less viable populations 44 00:01:53,410 --> 00:01:56,140 and less genetic resilience. 45 00:01:56,140 --> 00:01:58,320 And roads may also get in the way of populations 46 00:01:58,320 --> 00:02:00,110 that are undergoing rain shifts 47 00:02:00,110 --> 00:02:01,790 in response to climate change. 48 00:02:01,790 --> 00:02:04,460 So terrestrial species will need to physically cross 49 00:02:04,460 --> 00:02:06,660 roadways as they gradually shift their ranges 50 00:02:06,660 --> 00:02:08,960 along with those habitats that they depend on. 51 00:02:09,980 --> 00:02:12,480 Vermont is 78% forested, 52 00:02:12,480 --> 00:02:14,470 though we're not exempt from the effects 53 00:02:14,470 --> 00:02:16,570 of habitat fragmentation by roads. 54 00:02:16,570 --> 00:02:20,030 We have over 25,000 kilometers of roadway in Vermont 55 00:02:20,030 --> 00:02:22,950 and this really divides up our good wildlife habitat 56 00:02:22,950 --> 00:02:26,130 into smaller, more disconnected patches. 57 00:02:26,130 --> 00:02:27,950 Our road network does have 58 00:02:27,950 --> 00:02:31,180 over 88,000 state managed transportation structure. 59 00:02:31,180 --> 00:02:34,363 So these are things like bridges, culverts and underpasses. 60 00:02:35,240 --> 00:02:37,990 Around 6,000 of these structures are large enough 61 00:02:37,990 --> 00:02:40,810 to potentially facilitate wildlife passage. 62 00:02:40,810 --> 00:02:43,700 So these are of special interest to the state 63 00:02:43,700 --> 00:02:45,530 when thinking about how we can mitigate 64 00:02:45,530 --> 00:02:47,393 the effects of roads on wildlife. 65 00:02:50,470 --> 00:02:53,280 So some ways that we can mitigate the barrier effects 66 00:02:53,280 --> 00:02:54,790 is to make them more permeable 67 00:02:54,790 --> 00:02:56,710 with these types of structures. 68 00:02:56,710 --> 00:02:59,140 And thankfully transportation networks 69 00:02:59,140 --> 00:03:00,610 have existing infrastructure 70 00:03:00,610 --> 00:03:02,240 that's already serving this purpose, 71 00:03:02,240 --> 00:03:04,230 or has the potential to serve this purpose 72 00:03:04,230 --> 00:03:06,200 with some modifications. 73 00:03:06,200 --> 00:03:08,440 That being said, it's very expensive 74 00:03:08,440 --> 00:03:09,850 to create new infrastructure, 75 00:03:09,850 --> 00:03:12,360 or even just to update our existing infrastructure 76 00:03:12,360 --> 00:03:14,420 to make roads more wildlife-friendly. 77 00:03:14,420 --> 00:03:17,520 So VTrans wants a way of figuring out 78 00:03:17,520 --> 00:03:20,070 which of our existing transportation structures 79 00:03:20,070 --> 00:03:22,050 are most important for wildlife movement, 80 00:03:22,050 --> 00:03:24,740 based on their position in the landscape. 81 00:03:24,740 --> 00:03:26,830 And that way limited resources 82 00:03:26,830 --> 00:03:28,290 can be spent more effectively 83 00:03:28,290 --> 00:03:29,740 by improving the ones found 84 00:03:29,740 --> 00:03:32,460 within important movement areas for wildlife. 85 00:03:32,460 --> 00:03:34,180 And the goal of those improvements would be 86 00:03:34,180 --> 00:03:36,730 to make the structures more usable and appealing, 87 00:03:36,730 --> 00:03:38,750 so that wildlife might choose to cross 88 00:03:38,750 --> 00:03:40,633 beneath the road rather than over it. 89 00:03:41,900 --> 00:03:44,030 And the goal of our research here 90 00:03:44,030 --> 00:03:46,360 is to develop a method of ranking 91 00:03:46,360 --> 00:03:49,120 all of these different state manage structures, 92 00:03:49,120 --> 00:03:50,670 according to their connectivity value 93 00:03:50,670 --> 00:03:52,870 for terrestrial mammals in Vermont. 94 00:03:52,870 --> 00:03:55,880 And we're focusing on eight species listed here. 95 00:03:55,880 --> 00:03:58,700 These are generally larger bodied mammals 96 00:03:58,700 --> 00:04:00,480 that are more wide ranging 97 00:04:00,480 --> 00:04:03,100 and often cross roadways in Vermont. 98 00:04:03,100 --> 00:04:05,080 So these are the ones that we're starting out with. 99 00:04:05,080 --> 00:04:07,230 And the example I go through today 100 00:04:07,230 --> 00:04:10,770 is using data for a ninth species, American Marten. 101 00:04:10,770 --> 00:04:12,210 This is an endangered species 102 00:04:12,210 --> 00:04:14,510 in the Weasel family in Vermont. 103 00:04:14,510 --> 00:04:16,820 And we were able to use existing marten data 104 00:04:16,820 --> 00:04:18,430 that was sort of just ready to go 105 00:04:18,430 --> 00:04:20,290 for a test run of our process. 106 00:04:20,290 --> 00:04:22,480 So all of the maps you see in this presentation 107 00:04:22,480 --> 00:04:25,763 are from a preliminary analysis of American marten. 108 00:04:28,627 --> 00:04:31,670 As I mentioned, we're using this circuit theory approach 109 00:04:31,670 --> 00:04:33,670 to look at wildlife movement in Vermont. 110 00:04:33,670 --> 00:04:35,100 And this method just treats 111 00:04:35,100 --> 00:04:36,050 the movement of animals 112 00:04:36,050 --> 00:04:39,100 like the flow of electricity through a circuit. 113 00:04:39,100 --> 00:04:41,640 The landscape serves as the circuit, 114 00:04:41,640 --> 00:04:43,770 and it's made up of all these different resistances, 115 00:04:43,770 --> 00:04:45,660 which channels the flow of electricity 116 00:04:45,660 --> 00:04:49,060 through less resistant paths in the landscape. 117 00:04:49,060 --> 00:04:51,720 In Vermont, our circuit is made up of forest, 118 00:04:51,720 --> 00:04:53,830 agriculture, residential areas, 119 00:04:53,830 --> 00:04:56,410 roadways, and many other features. 120 00:04:56,410 --> 00:04:58,460 And each one of these land cover types 121 00:04:58,460 --> 00:05:00,190 has its own resistance to movement 122 00:05:00,190 --> 00:05:02,690 for different wildlife species. 123 00:05:02,690 --> 00:05:05,640 So for example, a white tail deer would move pretty easily 124 00:05:05,640 --> 00:05:08,600 through forested areas or open fields. 125 00:05:08,600 --> 00:05:10,180 And so these types of pluses 126 00:05:10,180 --> 00:05:12,770 would have a low resistance to movement. 127 00:05:12,770 --> 00:05:13,603 But on the other hand 128 00:05:13,603 --> 00:05:16,040 areas of really concentrated human development, 129 00:05:16,040 --> 00:05:17,480 would have a high resistance 130 00:05:17,480 --> 00:05:19,623 to movement for most species. 131 00:05:20,540 --> 00:05:23,533 And we're using a tool called Omniscape for this analysis. 132 00:05:23,533 --> 00:05:26,180 Some folks may have heard of Circuitscape. 133 00:05:26,180 --> 00:05:28,920 This is another tool used to model wildlife connectivity 134 00:05:28,920 --> 00:05:30,053 with circuit theory. 135 00:05:30,960 --> 00:05:33,620 And Omniscape is just a newer version of this tool. 136 00:05:33,620 --> 00:05:35,820 It's a little bit more comprehensive 137 00:05:35,820 --> 00:05:38,130 because you can add electricity into the landscape, 138 00:05:38,130 --> 00:05:40,690 wherever a wildlife species occurs 139 00:05:40,690 --> 00:05:43,030 rather than just at predetermined locations, 140 00:05:43,030 --> 00:05:44,300 like protected areas. 141 00:05:44,300 --> 00:05:47,060 So with Circuitscape, you have to define specific areas 142 00:05:47,060 --> 00:05:48,930 of the landscape that you want your electricity 143 00:05:48,930 --> 00:05:51,257 to come from, the source and the ground. 144 00:05:51,257 --> 00:05:53,420 And these could be large forest blocks 145 00:05:53,420 --> 00:05:55,870 or other areas that you wanna connect to one another. 146 00:05:55,870 --> 00:05:58,320 But Omniscape, lets us use additional data 147 00:05:58,320 --> 00:06:00,290 and puts electricity into the landscape. 148 00:06:00,290 --> 00:06:03,700 Wherever wildlife are known to actually occur. 149 00:06:03,700 --> 00:06:05,970 And I'll walk through each step of our analysis 150 00:06:05,970 --> 00:06:07,973 just to illustrate this whole process. 151 00:06:09,680 --> 00:06:13,150 So our transportation structure ranking process 152 00:06:13,150 --> 00:06:16,290 consists of a multi-scale connectivity analysis. 153 00:06:16,290 --> 00:06:17,670 The first part takes place 154 00:06:17,670 --> 00:06:19,555 at the scale of the entire Vermont landscape. 155 00:06:19,555 --> 00:06:22,440 And this is where we model broad species movements 156 00:06:22,440 --> 00:06:25,520 based on occurrence data and landscape features. 157 00:06:25,520 --> 00:06:27,780 The second step involves modeling wildlife movement 158 00:06:27,780 --> 00:06:30,460 in a fixed radius of each individual structure, 159 00:06:30,460 --> 00:06:32,210 using occurrence data again, 160 00:06:32,210 --> 00:06:34,950 and some more detailed land cover data. 161 00:06:34,950 --> 00:06:37,290 And the third step is where we combine results 162 00:06:37,290 --> 00:06:39,720 from both our landscape scale and our structure scale 163 00:06:39,720 --> 00:06:43,000 to get a single ranking of transportation structures 164 00:06:43,000 --> 00:06:45,850 based on their connectivity value at both of these steps. 165 00:06:48,140 --> 00:06:50,070 So starting with our step one, 166 00:06:50,070 --> 00:06:51,952 this takes place at the landscape scale, 167 00:06:51,952 --> 00:06:54,780 we have two inputs, and our output 168 00:06:54,780 --> 00:06:56,170 will tell us where the wildlife 169 00:06:56,170 --> 00:06:57,940 are most likely to move. 170 00:06:57,940 --> 00:07:00,853 Again, this is, this examples for American marten. 171 00:07:02,060 --> 00:07:05,550 So our first input tells us how much marten electricity 172 00:07:05,550 --> 00:07:08,270 is coming out of the landscape, and where it's coming from. 173 00:07:08,270 --> 00:07:10,890 And we will be using species occurrence data 174 00:07:10,890 --> 00:07:14,060 or occupancy data, which in this example, 175 00:07:14,060 --> 00:07:15,990 is the probability that marten occur 176 00:07:15,990 --> 00:07:17,823 in each pixel of the landscape. 177 00:07:19,700 --> 00:07:21,930 Our second input tells us how resistant 178 00:07:21,930 --> 00:07:23,770 different parts of the landscape are, 179 00:07:23,770 --> 00:07:25,500 which determines how that electricity 180 00:07:25,500 --> 00:07:27,670 then moves throughout the landscape. 181 00:07:27,670 --> 00:07:29,130 So for marten, at this scale, 182 00:07:29,130 --> 00:07:31,291 we were able to use genetic data for resistance, 183 00:07:31,291 --> 00:07:32,163 but for our other species 184 00:07:32,163 --> 00:07:34,850 we will be using expert opinion. 185 00:07:34,850 --> 00:07:37,410 But this resistance matches tells us how easy 186 00:07:37,410 --> 00:07:39,030 or how difficult it is for marten 187 00:07:39,030 --> 00:07:41,740 to move through different land cover types. 188 00:07:41,740 --> 00:07:43,350 So Omniscape is essentially 189 00:07:43,350 --> 00:07:45,010 just stacking that resistance map 190 00:07:45,010 --> 00:07:47,180 on top of the occurrence map. 191 00:07:47,180 --> 00:07:48,040 And the occurrence map 192 00:07:48,040 --> 00:07:49,650 is injecting marten electricity 193 00:07:49,650 --> 00:07:51,080 up into the resistance map. 194 00:07:51,080 --> 00:07:53,320 And the marten then traveled through the landscape 195 00:07:53,320 --> 00:07:55,523 wherever there's areas of lower resistance. 196 00:07:57,100 --> 00:07:59,070 And our output here shows us 197 00:07:59,070 --> 00:08:01,530 a map of electrical current density. 198 00:08:01,530 --> 00:08:03,300 So areas of high current density 199 00:08:03,300 --> 00:08:04,740 are places that we would expect 200 00:08:04,740 --> 00:08:07,550 more dispersing wildlife or marten in this case. 201 00:08:07,550 --> 00:08:09,420 And areas with low current density 202 00:08:09,420 --> 00:08:12,320 are places that we would expect fewer dispersing wildlife. 203 00:08:15,120 --> 00:08:18,000 With this output map, we can get 204 00:08:18,000 --> 00:08:19,940 a preliminary ranking of our structures 205 00:08:19,940 --> 00:08:21,550 at the landscape scale. 206 00:08:21,550 --> 00:08:23,400 So for each transportation structure, 207 00:08:23,400 --> 00:08:25,370 we took the mean current density 208 00:08:25,370 --> 00:08:27,910 within a fixed radius of each structure location. 209 00:08:27,910 --> 00:08:29,080 And we then ranked them 210 00:08:29,080 --> 00:08:31,800 from the highest mean current density to lowest. 211 00:08:31,800 --> 00:08:35,210 And this map is just showing the top 100 structures 212 00:08:35,210 --> 00:08:38,203 important for marten connectivity at the landscape scale. 213 00:08:39,810 --> 00:08:41,730 And another thing we can do with this scale 214 00:08:41,730 --> 00:08:44,040 is classify the different types of current density 215 00:08:44,040 --> 00:08:45,820 from the purple and yellow map, 216 00:08:45,820 --> 00:08:48,150 into more descriptive categories 217 00:08:48,150 --> 00:08:49,770 in a normalized current map. 218 00:08:49,770 --> 00:08:52,870 And this normalized map is just comparing the model outputs 219 00:08:52,870 --> 00:08:56,720 to a null model or a perfect landscape of no resistance. 220 00:08:56,720 --> 00:08:59,500 And this just allows us to better interpret the areas 221 00:08:59,500 --> 00:09:01,940 that are actually impeding movement. 222 00:09:01,940 --> 00:09:04,610 The areas that are allowing unrestricted movement, 223 00:09:04,610 --> 00:09:06,750 and which areas are acting as pinch points. 224 00:09:06,750 --> 00:09:08,930 So this map gets used toward the end 225 00:09:08,930 --> 00:09:10,963 when we combine the two scales together. 226 00:09:13,600 --> 00:09:15,580 So for step two, the structure scale, 227 00:09:15,580 --> 00:09:16,940 we're doing the same analysis 228 00:09:16,940 --> 00:09:19,489 but in a smaller radius around each structure. 229 00:09:19,489 --> 00:09:22,980 We still have our source or our occupancy input 230 00:09:22,980 --> 00:09:24,870 which tells us how much marten electricity 231 00:09:24,870 --> 00:09:26,433 is coming out of each pixel. 232 00:09:27,330 --> 00:09:29,050 And we have our resistance input, 233 00:09:29,050 --> 00:09:30,220 but at this scale we're using 234 00:09:30,220 --> 00:09:33,160 very detailed half meter LiDAR data 235 00:09:33,160 --> 00:09:35,642 produced by the spatial analysis lab, EVM. 236 00:09:35,642 --> 00:09:37,780 And we're using expert opinion to assign 237 00:09:37,780 --> 00:09:39,980 the resistance values to each land cover type. 238 00:09:39,980 --> 00:09:42,130 So back in step one we had used genetic data 239 00:09:42,130 --> 00:09:43,890 for marten, for our resistance, 240 00:09:43,890 --> 00:09:46,870 but now we're using a different land cover dataset 241 00:09:46,870 --> 00:09:49,310 and we need to manually assign some resistances 242 00:09:49,310 --> 00:09:50,763 to the fine scale data. 243 00:09:53,290 --> 00:09:55,150 And we run on the scape again, 244 00:09:55,150 --> 00:09:57,853 and we rank structures at this fine scale, 245 00:09:58,860 --> 00:10:01,120 just again by taking the mean current density 246 00:10:01,120 --> 00:10:03,700 in a radius, fixed radius around each structure. 247 00:10:03,700 --> 00:10:05,680 Just this time we're doing it in a smaller radius 248 00:10:05,680 --> 00:10:09,063 and we're making use of really high resolution data. 249 00:10:11,740 --> 00:10:13,820 So now we enter step three, 250 00:10:13,820 --> 00:10:16,260 and this is where we have to combine the information 251 00:10:16,260 --> 00:10:17,490 from steps one and two, 252 00:10:17,490 --> 00:10:19,960 to get a single ranking of transportation structures, 253 00:10:19,960 --> 00:10:22,020 cause right now we have two different rankings. 254 00:10:22,020 --> 00:10:24,690 And it's really important to look at both of these scales, 255 00:10:24,690 --> 00:10:26,530 since the landscape scale ranking 256 00:10:26,530 --> 00:10:28,180 might not match up with what we see 257 00:10:28,180 --> 00:10:30,233 at the structure scale and vice versa. 258 00:10:30,233 --> 00:10:32,030 So I just wanted to include some examples 259 00:10:32,030 --> 00:10:32,950 to illustrate this. 260 00:10:32,950 --> 00:10:35,050 In this example, here, we have a structure 261 00:10:35,050 --> 00:10:36,490 that would rank pretty high, 262 00:10:36,490 --> 00:10:38,750 if we only looked at that step two ranking, 263 00:10:38,750 --> 00:10:41,150 and this is because there's a lot of forested land cover 264 00:10:41,150 --> 00:10:42,150 around it. 265 00:10:42,150 --> 00:10:44,740 Marten are a very forced dependent species. 266 00:10:44,740 --> 00:10:47,630 So this would rank pretty high in that step two. 267 00:10:47,630 --> 00:10:50,360 However, if we zoom out and we look at the step one ranking 268 00:10:50,360 --> 00:10:51,630 we would see that the structure 269 00:10:51,630 --> 00:10:54,230 is not very important for marten connectivity 270 00:10:54,230 --> 00:10:55,761 based on the landscape level 271 00:10:55,761 --> 00:10:58,143 potential movements of marten in Vermont. 272 00:11:01,020 --> 00:11:04,000 Same thing here with this example, but this structure 273 00:11:04,000 --> 00:11:06,330 would rank pretty high at the landscape scale, 274 00:11:06,330 --> 00:11:08,160 and pretty low at the structure scale 275 00:11:08,160 --> 00:11:10,250 due to all of the developed land cover. 276 00:11:10,250 --> 00:11:12,103 So we really need a way of combining this information 277 00:11:12,103 --> 00:11:14,803 and for both of the scales to get one ranking. 278 00:11:17,800 --> 00:11:19,670 And we're exploring a few different methods 279 00:11:19,670 --> 00:11:21,271 of combining our scales together. 280 00:11:21,271 --> 00:11:24,230 One is to just multiply or combine 281 00:11:24,230 --> 00:11:26,650 the mean current density from the statewide scale, 282 00:11:26,650 --> 00:11:29,460 with the mean current density from the structure scale. 283 00:11:29,460 --> 00:11:30,700 Another option is similar, 284 00:11:30,700 --> 00:11:33,010 but involves using a weighting system 285 00:11:33,010 --> 00:11:35,220 with those four current density classifications 286 00:11:35,220 --> 00:11:36,360 I mentioned earlier. 287 00:11:36,360 --> 00:11:37,660 So we would assign a weight 288 00:11:37,660 --> 00:11:39,040 to each one of those classes 289 00:11:39,040 --> 00:11:41,870 based on how well each one facilitates movement, 290 00:11:41,870 --> 00:11:42,950 and then multiply that 291 00:11:42,950 --> 00:11:45,710 by the structure scale current density average. 292 00:11:45,710 --> 00:11:48,830 And there are a few other options we're looking at, 293 00:11:48,830 --> 00:11:51,130 but we will test each possible ranking method 294 00:11:51,130 --> 00:11:52,280 against camera data. 295 00:11:52,280 --> 00:11:53,910 We have a bunch of structures 296 00:11:53,910 --> 00:11:55,750 that have game cameras placed at them 297 00:11:55,750 --> 00:11:57,470 to monitor wildlife use, 298 00:11:57,470 --> 00:12:00,140 and against roadkill data to see how well 299 00:12:00,140 --> 00:12:02,440 our current density predictions and our rankings 300 00:12:02,440 --> 00:12:05,543 match up with areas of higher roadkill and wildlife use. 301 00:12:07,880 --> 00:12:09,610 Some implications. 302 00:12:09,610 --> 00:12:10,670 As I mentioned in the beginning, 303 00:12:10,670 --> 00:12:13,140 it's very expensive to work on these structures. 304 00:12:13,140 --> 00:12:15,800 So these lists of rankings can be referenced 305 00:12:15,800 --> 00:12:17,810 whenever there's a construction project planned 306 00:12:17,810 --> 00:12:19,940 for a structure, and road managers 307 00:12:19,940 --> 00:12:21,290 can decide whether or not adding 308 00:12:21,290 --> 00:12:22,880 some wildlife friendly infrastructure, 309 00:12:22,880 --> 00:12:25,780 would be worthwhile in that location, 310 00:12:25,780 --> 00:12:29,080 or if those funds would be better spent in another location. 311 00:12:29,080 --> 00:12:31,240 And additionally, the species specific models 312 00:12:31,240 --> 00:12:34,110 may highlight that some structures are primarily used 313 00:12:34,110 --> 00:12:36,800 by one species or a specific group of species 314 00:12:36,800 --> 00:12:38,959 based on those landscape level movements. 315 00:12:38,959 --> 00:12:41,540 And the appropriate improvements can be implemented 316 00:12:41,540 --> 00:12:43,100 based on that information. 317 00:12:43,100 --> 00:12:45,160 So some examples would be adding shelving 318 00:12:45,160 --> 00:12:48,040 for bobcats to walk on, or filling in areas 319 00:12:48,040 --> 00:12:51,380 that have large boulders to promote deer use. 320 00:12:51,380 --> 00:12:53,690 And this is also pretty cost-effective analysis. 321 00:12:53,690 --> 00:12:56,850 We're using a combination of existing wildlife data 322 00:12:56,850 --> 00:12:59,150 and expert opinion, and we can run this analysis 323 00:12:59,150 --> 00:13:01,870 in a pretty short amount of time for many species. 324 00:13:01,870 --> 00:13:05,200 So Omniscape can let us see wildlife movement 325 00:13:05,200 --> 00:13:07,890 over large areas without needing to buy 326 00:13:07,890 --> 00:13:09,610 expensive GPS collars, or invest 327 00:13:09,610 --> 00:13:11,570 a lot of time and effort into capturing animals 328 00:13:11,570 --> 00:13:13,263 and monitoring them in that way. 329 00:13:15,550 --> 00:13:17,810 And I just wanted to end with highlighting 330 00:13:17,810 --> 00:13:20,640 some other applications of circuit theory based modeling 331 00:13:20,640 --> 00:13:22,720 since it can be used for many other things 332 00:13:22,720 --> 00:13:25,270 other than just looking at wildlife movement. 333 00:13:25,270 --> 00:13:28,000 Circuit theory has been used to analyze 334 00:13:28,000 --> 00:13:30,170 the potential paths that fire can travel. 335 00:13:30,170 --> 00:13:32,250 So fire connectivity. 336 00:13:32,250 --> 00:13:35,283 The spread of genotypes throughout the landscape. 337 00:13:36,210 --> 00:13:38,169 Dispersal of early human populations 338 00:13:38,169 --> 00:13:41,332 and present day dispersal of many other species. 339 00:13:41,332 --> 00:13:44,460 And it's been used to look at how broader groups of species 340 00:13:44,460 --> 00:13:47,113 might respond to future climate change scenarios. 341 00:13:49,240 --> 00:13:52,070 And this last paper on the bottom right, 342 00:13:52,070 --> 00:13:54,570 is a review of 459 studies 343 00:13:54,570 --> 00:13:55,867 that have used Circuitscape 344 00:13:55,867 --> 00:13:58,820 or circuit theory based modeling. 345 00:13:58,820 --> 00:14:00,630 So the Omniscape program itself 346 00:14:00,630 --> 00:14:02,850 is a very new update to all of this. 347 00:14:02,850 --> 00:14:04,330 There haven't been many Omniscape papers 348 00:14:04,330 --> 00:14:06,100 published using it yet. 349 00:14:06,100 --> 00:14:08,440 Many of them have been using Circuitscape, 350 00:14:08,440 --> 00:14:10,300 but I do expect we'll see some more 351 00:14:10,300 --> 00:14:12,720 in the near future since it's a more thorough way 352 00:14:12,720 --> 00:14:13,670 to look at connectivity 353 00:14:13,670 --> 00:14:15,770 than some of the other programs available. 354 00:14:17,430 --> 00:14:20,100 And I would like to thank all of our partners 355 00:14:20,100 --> 00:14:22,320 for making this research possible. 356 00:14:22,320 --> 00:14:24,850 Glenn Gingras and Chris Slesar with VTrans, 357 00:14:24,850 --> 00:14:27,260 Paul Marangelo, Dr. Kim Hall, and Aaron Jones 358 00:14:27,260 --> 00:14:29,090 with the Nature Conservancy. 359 00:14:29,090 --> 00:14:31,680 Jens Hilke with the Vermont Fish and Wildlife Department. 360 00:14:31,680 --> 00:14:34,520 My advisor, Dr. James Murdoch, and Vincent Landau 361 00:14:34,520 --> 00:14:36,473 with Conservation Science Partners. 362 00:14:38,160 --> 00:14:40,763 And I'm happy to take any questions. 363 00:14:43,585 --> 00:14:44,570 - Hi, Caitlin. 364 00:14:44,570 --> 00:14:45,870 I really appreciate the talk. 365 00:14:45,870 --> 00:14:48,000 My question is, has this been used 366 00:14:48,000 --> 00:14:52,390 for aquatic species at all to, to test connectivity, 367 00:14:52,390 --> 00:14:53,340 the Omniscape tool? 368 00:14:54,500 --> 00:14:55,500 - Yeah. 369 00:14:55,500 --> 00:14:57,210 Not that Omniscape tool specifically, 370 00:14:57,210 --> 00:14:58,310 again that one's pretty new, 371 00:14:58,310 --> 00:15:01,180 but Circuitscape has been used. 372 00:15:01,180 --> 00:15:02,960 I almost included a study on that slide 373 00:15:02,960 --> 00:15:06,360 about aquatic, like shrimp populations. 374 00:15:06,360 --> 00:15:08,380 And I think their resistances 375 00:15:08,380 --> 00:15:11,210 were based on different currents, 376 00:15:11,210 --> 00:15:14,100 and how the currents were affecting their dispersal. 377 00:15:14,100 --> 00:15:17,080 Not a hundred percent sure, but yeah, 378 00:15:17,080 --> 00:15:18,520 I know it's been used for aquatic species 379 00:15:18,520 --> 00:15:20,830 and amphibians, I've seen a couple 380 00:15:20,830 --> 00:15:23,060 like frog papers using Circuitscapes. 381 00:15:23,060 --> 00:15:26,970 So it definitely has broad applications 382 00:15:26,970 --> 00:15:29,260 for a lot of different species, plants as well. 383 00:15:29,260 --> 00:15:30,410 From Sandy. 384 00:15:30,410 --> 00:15:33,720 How long before AOT will be able to use the data? 385 00:15:33,720 --> 00:15:35,180 Also, do you have plans to work with towns 386 00:15:35,180 --> 00:15:37,120 that are also trying to modify structures? 387 00:15:37,120 --> 00:15:38,514 Great question. 388 00:15:38,514 --> 00:15:39,347 Yeah. 389 00:15:39,347 --> 00:15:40,880 So this project has been ongoing 390 00:15:40,880 --> 00:15:42,470 and this is one component 391 00:15:42,470 --> 00:15:45,380 of how we'll be assessing all of these structures. 392 00:15:45,380 --> 00:15:47,980 Again, we're, we're looking at this camera data too. 393 00:15:49,050 --> 00:15:53,900 And I expect my, this tool portion of the analysis 394 00:15:53,900 --> 00:15:55,400 will be finishing up this spring. 395 00:15:55,400 --> 00:15:58,423 So hopefully there'll be able to implement it pretty soon. 396 00:15:59,450 --> 00:16:01,160 And as far as town structures go, 397 00:16:01,160 --> 00:16:04,670 yeah it's, some town structures that are located 398 00:16:04,670 --> 00:16:06,520 in towns but managed by the state 399 00:16:06,520 --> 00:16:08,700 are included in this analysis, 400 00:16:08,700 --> 00:16:11,400 but there are definitely some town datasets 401 00:16:11,400 --> 00:16:13,890 that we, they're all over the place. 402 00:16:13,890 --> 00:16:16,020 So it was hard to combine them into this analysis, 403 00:16:16,020 --> 00:16:18,450 but that would be a great use in the future 404 00:16:18,450 --> 00:16:20,720 if we can combine all of those data sets into one, 405 00:16:20,720 --> 00:16:22,520 and look at town structures as well. 406 00:16:26,420 --> 00:16:27,880 One from Sasha. 407 00:16:27,880 --> 00:16:30,380 What is the basis for your broad scale occupancy map 408 00:16:30,380 --> 00:16:31,500 and how will you generate this 409 00:16:31,500 --> 00:16:33,240 for all of your target species? 410 00:16:33,240 --> 00:16:34,073 Great question. 411 00:16:34,073 --> 00:16:34,906 Yeah. 412 00:16:34,906 --> 00:16:35,739 So for this presentation, 413 00:16:35,739 --> 00:16:38,452 we were using existing data, from Cody Aylward. 414 00:16:38,452 --> 00:16:40,440 Cody's given presentations at this conference 415 00:16:40,440 --> 00:16:43,240 in the past on the marten data. 416 00:16:43,240 --> 00:16:44,950 So his (indistinct) model 417 00:16:44,950 --> 00:16:47,113 was based off of expert opinion as well, 418 00:16:48,230 --> 00:16:49,440 but for our future species 419 00:16:49,440 --> 00:16:51,990 we'll be using data that's already been collected, 420 00:16:51,990 --> 00:16:54,530 through a really extensive expert opinion survey, 421 00:16:54,530 --> 00:16:57,309 conducted by Skye Pearman Gillman 422 00:16:57,309 --> 00:16:59,805 Pearman Gillman et al. (2020) 423 00:16:59,805 --> 00:17:03,680 And that looked at expert opinion of occupancy 424 00:17:03,680 --> 00:17:07,090 from many different experts all across the Northeast. 425 00:17:07,090 --> 00:17:08,420 And so those models are available 426 00:17:08,420 --> 00:17:10,400 for all of the species, thankfully. 427 00:17:10,400 --> 00:17:12,900 So all of that data should be ready to go as well.