1 00:00:01,400 --> 00:00:03,050 [Amanda] Hi, welcome to Week 6 2 00:00:03,050 --> 00:00:05,020 of Remote Sensing Foundations. 3 00:00:05,020 --> 00:00:06,340 This week, we will start to learn 4 00:00:06,340 --> 00:00:08,550 about some applications of remote sensing, 5 00:00:08,550 --> 00:00:10,200 particularly that of the landscape 6 00:00:10,200 --> 00:00:11,713 are terrestrial applications. 7 00:00:12,850 --> 00:00:14,330 Upon completion of this module, 8 00:00:14,330 --> 00:00:15,690 I hope you'll be able to understand 9 00:00:15,690 --> 00:00:17,010 the different types of data sets 10 00:00:17,010 --> 00:00:19,740 used to monitor wildfire and recovery, 11 00:00:19,740 --> 00:00:20,680 list three ways 12 00:00:20,680 --> 00:00:23,860 in which remote sensing is used in wildlife conservation, 13 00:00:23,860 --> 00:00:24,860 explain the difference 14 00:00:24,860 --> 00:00:27,670 between forest degradation and deforestation, 15 00:00:27,670 --> 00:00:30,190 and also to conceptualize how to measure 16 00:00:30,190 --> 00:00:34,140 and monitor change in forests using remote sensing. 17 00:00:34,140 --> 00:00:35,830 We have a few associated readings, 18 00:00:35,830 --> 00:00:38,960 including an article from Jeremy Hance from 2014, 19 00:00:38,960 --> 00:00:42,070 about remote sensing in the realm of conservation. 20 00:00:42,070 --> 00:00:42,940 It's a little bit dated, 21 00:00:42,940 --> 00:00:44,560 but provides a good perspective 22 00:00:44,560 --> 00:00:48,143 on how remote sensing has been and is used in conservation. 23 00:00:50,160 --> 00:00:53,270 All right, so given that many forest environmental variables 24 00:00:53,270 --> 00:00:54,990 and other terrestrial variables 25 00:00:54,990 --> 00:00:58,030 can be measured directly in the field, 26 00:00:58,030 --> 00:01:01,020 why has remote sensing become an important data source? 27 00:01:01,020 --> 00:01:03,280 So this might seem kind of like an obvious question, 28 00:01:03,280 --> 00:01:08,280 but I think it's an important context to why and how 29 00:01:08,410 --> 00:01:12,260 remote sensing has been developed and used in conservation, 30 00:01:12,260 --> 00:01:14,260 and sort of greater understanding 31 00:01:14,260 --> 00:01:16,400 of Earth's processes in general. 32 00:01:16,400 --> 00:01:18,290 So there's six fundamental reasons for this. 33 00:01:18,290 --> 00:01:19,670 And firstly, 34 00:01:19,670 --> 00:01:24,490 remote sensing imagery provides a more synoptic view. 35 00:01:24,490 --> 00:01:27,770 So the vantage provided by earth-observing sensors 36 00:01:27,770 --> 00:01:29,550 ensures that imagery captures 37 00:01:29,550 --> 00:01:33,223 a complete picture of the environment and its field of view. 38 00:01:34,380 --> 00:01:35,640 Because of this complete survey, 39 00:01:35,640 --> 00:01:37,430 remote sensing allows wall-to-wall mapping 40 00:01:37,430 --> 00:01:40,320 and monitoring of important ecological variables, 41 00:01:40,320 --> 00:01:42,470 such as land cover change. 42 00:01:42,470 --> 00:01:45,450 Secondly, remote sensing data 43 00:01:45,450 --> 00:01:47,360 are available over the entire planet 44 00:01:47,360 --> 00:01:50,460 and often at a range of spatial and temporal scales. 45 00:01:50,460 --> 00:01:52,990 So we've talked about this a lot before, 46 00:01:52,990 --> 00:01:56,100 but I think it's just important to sort of remind ourselves, 47 00:01:56,100 --> 00:01:58,370 so key environmental remote sensing systems, 48 00:01:58,370 --> 00:02:00,600 such as those carried out by like Landsat, 49 00:02:00,600 --> 00:02:04,380 and some of the other systems have been able to provide 50 00:02:04,380 --> 00:02:09,380 pretty constantly updateable stream of imagery 51 00:02:09,650 --> 00:02:13,203 for the entire planet since the 1970s or early 80s. 52 00:02:14,970 --> 00:02:17,247 Thirdly, remote sensing imagery 53 00:02:17,247 --> 00:02:20,090 has a high degree of ho homogeneity. 54 00:02:20,090 --> 00:02:21,930 So, critically, 55 00:02:21,930 --> 00:02:24,530 data from key environmental remote sensing systems 56 00:02:24,530 --> 00:02:27,540 are acquired under relatively fixed conditions. 57 00:02:27,540 --> 00:02:30,040 And the data captured relate to the way 58 00:02:30,040 --> 00:02:32,640 in which the radiation interacts with the environment, 59 00:02:32,640 --> 00:02:34,950 which is constant in space and time. 60 00:02:34,950 --> 00:02:36,950 So there are no human-induced complications 61 00:02:36,950 --> 00:02:39,410 such as differences in measurement practices 62 00:02:39,410 --> 00:02:40,450 from one country or another, 63 00:02:40,450 --> 00:02:43,300 or for instance, user error, 64 00:02:43,300 --> 00:02:46,440 if you're out in the field and you're measuring tree height, 65 00:02:46,440 --> 00:02:48,230 that can be a really difficult thing to do, 66 00:02:48,230 --> 00:02:53,230 and also, difficult thing to do in a consistent way 67 00:02:53,460 --> 00:02:56,280 across systems and across different types of ways 68 00:02:56,280 --> 00:02:58,470 that we measure, for instance, tree height. 69 00:02:58,470 --> 00:02:59,913 So with remote sensing, 70 00:03:01,710 --> 00:03:03,210 while we've talked about some of the things 71 00:03:03,210 --> 00:03:04,770 that we have to do to data 72 00:03:04,770 --> 00:03:07,160 to get them to be readily usable 73 00:03:07,160 --> 00:03:11,510 and applicable to what we're measuring, in general, 74 00:03:11,510 --> 00:03:15,640 those algorithms that we develop for pre-processing, 75 00:03:15,640 --> 00:03:19,000 like subtracting out atmosphere conditions 76 00:03:19,000 --> 00:03:21,460 or accounting for things like sensor angles 77 00:03:21,460 --> 00:03:22,600 and things like that, 78 00:03:22,600 --> 00:03:25,000 all of those things are applied to the data set 79 00:03:25,000 --> 00:03:26,990 in a very homogenous way. 80 00:03:26,990 --> 00:03:30,510 So that's also an important sort of key factor 81 00:03:30,510 --> 00:03:32,700 in why remote sensing 82 00:03:32,700 --> 00:03:35,010 has become such an important data source. 83 00:03:35,010 --> 00:03:38,670 So fourth, the image contains or can easily be converted 84 00:03:38,670 --> 00:03:40,753 to digital numbers or digital images, 85 00:03:41,690 --> 00:03:45,010 and because of this, it can be really easily integrated 86 00:03:45,010 --> 00:03:48,530 with other spatial data sets in GIS. 87 00:03:48,530 --> 00:03:50,460 So we've talked about this as well before, 88 00:03:50,460 --> 00:03:52,750 but again, I think it's just a good reminder 89 00:03:52,750 --> 00:03:57,750 in review of how widely roster data sets can be used 90 00:03:57,780 --> 00:04:00,430 and applied and sort of joined 91 00:04:00,430 --> 00:04:02,840 with other different types of data sets. 92 00:04:02,840 --> 00:04:06,890 So fifth, per unit area or per gigabyte of data, 93 00:04:06,890 --> 00:04:09,160 remote sensing is a really inexpensive way 94 00:04:09,160 --> 00:04:12,640 to acquire large, large amounts of data. 95 00:04:12,640 --> 00:04:15,750 So, of course, we can think about the financial costs 96 00:04:15,750 --> 00:04:19,560 associated with something like developing 97 00:04:19,560 --> 00:04:20,860 and building and launching 98 00:04:20,860 --> 00:04:25,160 and operating satellite remote sensing platform. 99 00:04:25,160 --> 00:04:27,220 That is definitely for sure expensive, 100 00:04:27,220 --> 00:04:31,060 but when we sort of divide that out 101 00:04:31,060 --> 00:04:34,360 by the amount of coverage both temporally and spatially, 102 00:04:34,360 --> 00:04:36,870 it actually becomes really, really inexpensive. 103 00:04:36,870 --> 00:04:40,260 And it's also sort of important to point out 104 00:04:40,260 --> 00:04:41,970 that there's been this increasing trend 105 00:04:41,970 --> 00:04:43,080 to make these data sets 106 00:04:43,080 --> 00:04:46,180 for all of the different environmental science research 107 00:04:46,180 --> 00:04:51,110 that we do with remote sensing freely open and available. 108 00:04:51,110 --> 00:04:54,730 And in addition to that, things like, 109 00:04:54,730 --> 00:04:56,250 resources like Google Earth Engine 110 00:04:56,250 --> 00:04:58,620 are able to provide easy access 111 00:04:58,620 --> 00:05:01,770 to all these freely available data sets. 112 00:05:01,770 --> 00:05:03,470 So sixth, and finally, 113 00:05:03,470 --> 00:05:06,260 not only are data more readily available, 114 00:05:06,260 --> 00:05:09,760 but there also has been an increasing trend 115 00:05:09,760 --> 00:05:12,490 towards the sharing of data products 116 00:05:12,490 --> 00:05:14,300 as well as the image data itself. 117 00:05:14,300 --> 00:05:18,610 So sort of related to the fifth one, of it becoming cheaper. 118 00:05:18,610 --> 00:05:20,210 It's also important to note 119 00:05:20,210 --> 00:05:22,790 that there's a whole bunch of data sharing 120 00:05:22,790 --> 00:05:25,910 that has developed in the world of remote sensing, 121 00:05:25,910 --> 00:05:28,060 and that is a benefit to all of us. 122 00:05:28,060 --> 00:05:30,500 So, it greatly reduces the need 123 00:05:30,500 --> 00:05:33,730 for both expert knowledge of remote sensing 124 00:05:33,730 --> 00:05:36,500 and also of image analysis. 125 00:05:36,500 --> 00:05:39,430 And it's sort of given rise 126 00:05:39,430 --> 00:05:41,720 to a greater amount of communication 127 00:05:41,720 --> 00:05:44,180 between experts and environmental scientists, 128 00:05:44,180 --> 00:05:46,683 and sometimes, also with land managers as well, 129 00:05:47,720 --> 00:05:51,900 has been the case, historically speaking. 130 00:05:51,900 --> 00:05:53,290 So, environmental scientists 131 00:05:53,290 --> 00:05:56,630 can now easily access science data quality products 132 00:05:58,295 --> 00:05:59,540 obtained from other people 133 00:05:59,540 --> 00:06:01,233 and obtained from remote sensing. 134 00:06:02,150 --> 00:06:04,840 And platforms like Google Earth Engine 135 00:06:04,840 --> 00:06:05,850 have given rise to this. 136 00:06:05,850 --> 00:06:10,420 So, if you look through the different data 137 00:06:10,420 --> 00:06:12,260 available on the catalog, 138 00:06:12,260 --> 00:06:14,880 you can see a lot of what's called mature data products, 139 00:06:14,880 --> 00:06:16,310 where people have already done 140 00:06:16,310 --> 00:06:17,550 all of the different processing, 141 00:06:17,550 --> 00:06:19,180 pre-processing of the imagery, 142 00:06:19,180 --> 00:06:21,400 and then they've used that to calculate 143 00:06:21,400 --> 00:06:23,100 these really great global products 144 00:06:24,050 --> 00:06:25,990 that you can now use yourself. 145 00:06:25,990 --> 00:06:28,060 So those were the six things 146 00:06:28,060 --> 00:06:31,310 that I just wanted to kind of remind us and review us about 147 00:06:31,310 --> 00:06:35,430 in terms of why we're here and what's important, 148 00:06:35,430 --> 00:06:37,810 and why remote sensing's become 149 00:06:37,810 --> 00:06:39,800 such an important data source. 150 00:06:39,800 --> 00:06:41,390 So today, we're gonna talk about 151 00:06:41,390 --> 00:06:44,380 three terrestrial applications for remote sensing, 152 00:06:44,380 --> 00:06:46,420 wildfire, wildlife conservation, 153 00:06:46,420 --> 00:06:49,793 and forest degradation and deforestation. 154 00:06:51,250 --> 00:06:55,000 Okay, so starting with remote sensing of wildfires, 155 00:06:55,000 --> 00:06:56,640 we know that the number frequency 156 00:06:56,640 --> 00:06:59,380 and severity of wildfires worldwide 157 00:06:59,380 --> 00:07:01,720 have rapidly increased in recent years, 158 00:07:01,720 --> 00:07:03,170 being driven by warmer temperatures 159 00:07:03,170 --> 00:07:04,890 caused by climate change. 160 00:07:04,890 --> 00:07:07,640 Identifying where and when these fires might ignite 161 00:07:07,640 --> 00:07:10,600 can prevent them from ravaging the surrounding environment, 162 00:07:10,600 --> 00:07:13,760 protects infrastructure, and also can save lives. 163 00:07:13,760 --> 00:07:16,000 So remote sensing is now used 164 00:07:16,000 --> 00:07:21,000 for a couple of different areas around surrounding wildfire. 165 00:07:22,510 --> 00:07:25,030 For instance, there's active fire monitoring, 166 00:07:25,030 --> 00:07:28,560 which, different land management agencies 167 00:07:28,560 --> 00:07:30,710 deploy a whole suite of different technologies, 168 00:07:30,710 --> 00:07:32,360 including thermal imagery, drones, 169 00:07:32,360 --> 00:07:35,530 Lidar monitoring systems, airborne platforms, 170 00:07:35,530 --> 00:07:37,630 all to monitor the spread 171 00:07:37,630 --> 00:07:40,650 and severity of fires that are occurring. 172 00:07:40,650 --> 00:07:41,970 There's also fire recovery. 173 00:07:41,970 --> 00:07:45,980 So there's a huge steady area of science 174 00:07:45,980 --> 00:07:48,150 that is basically geared 175 00:07:48,150 --> 00:07:52,290 towards understanding how vegetation regrows, 176 00:07:52,290 --> 00:07:55,363 using things like classification, time-series analysis, 177 00:07:56,220 --> 00:07:59,730 kind of combining Lidar and optical technologies as well. 178 00:07:59,730 --> 00:08:00,780 And then finally, 179 00:08:00,780 --> 00:08:03,600 there's a huge area of science and research 180 00:08:03,600 --> 00:08:06,050 that's committed towards predicting fire 181 00:08:06,050 --> 00:08:07,990 and also mitigating it. 182 00:08:07,990 --> 00:08:11,453 So, concerning things like fuel-load tracking, 183 00:08:12,400 --> 00:08:15,570 climatic conditions, identifying climatic conditions 184 00:08:15,570 --> 00:08:18,920 that might increase the likelihood of fire occurring. 185 00:08:18,920 --> 00:08:21,420 Also, identifying vegetation stress, 186 00:08:21,420 --> 00:08:23,150 which can be done pretty readily 187 00:08:23,150 --> 00:08:25,780 with remote sensing technologies. 188 00:08:25,780 --> 00:08:30,190 So couple of examples, drones equipped with Lidar sensors 189 00:08:30,190 --> 00:08:32,060 are being used by Spanish researchers 190 00:08:32,060 --> 00:08:34,920 to help manage the nation's forests. 191 00:08:34,920 --> 00:08:37,210 And instead of detecting fires themselves, 192 00:08:37,210 --> 00:08:39,740 the drones have advanced mapping capabilities 193 00:08:39,740 --> 00:08:41,620 to gather data that can be used 194 00:08:41,620 --> 00:08:44,470 to help lower the risk of fires occurring, 195 00:08:44,470 --> 00:08:45,440 and when they're occurring, 196 00:08:45,440 --> 00:08:49,990 having them be absolutely devastating. 197 00:08:49,990 --> 00:08:51,560 And they do this by identifying 198 00:08:51,560 --> 00:08:54,960 where the most intense wildfires and rapidly spreading zones 199 00:08:54,960 --> 00:08:56,850 are most likely to occur, 200 00:08:56,850 --> 00:09:01,173 based on all these conditions that I just listed here. 201 00:09:03,510 --> 00:09:05,283 So like fuel-load trafficking, 202 00:09:06,763 --> 00:09:11,204 vegetation stresses, drought conditions, things like that. 203 00:09:11,204 --> 00:09:13,000 On the other hand, in California, for instance, 204 00:09:13,000 --> 00:09:14,940 which is a known hotspot for wildfires, 205 00:09:14,940 --> 00:09:17,370 the US Department of Homeland Security 206 00:09:19,944 --> 00:09:21,490 Science and Technology of Directorate 207 00:09:21,490 --> 00:09:24,090 recently tested for prototype technologies 208 00:09:24,090 --> 00:09:26,510 for early wildfire detection. 209 00:09:26,510 --> 00:09:28,710 So among these was a sensor 210 00:09:28,710 --> 00:09:33,450 that monitors particulates and air quality, 211 00:09:33,450 --> 00:09:35,960 so, meaning smoke and air pollution, 212 00:09:35,960 --> 00:09:38,400 alongside with optical remote sensors, 213 00:09:38,400 --> 00:09:40,400 thermal imagery, and photoelectric sampling. 214 00:09:40,400 --> 00:09:42,930 So, all of these kind of combined together 215 00:09:42,930 --> 00:09:45,950 to try to predict wildfire 216 00:09:45,950 --> 00:09:50,950 and detect the spread of wildfires when they do occur. 217 00:09:57,550 --> 00:09:59,360 So by way of another example, 218 00:09:59,360 --> 00:10:04,360 we can also talk about Australia's 2019-2020 fire season 219 00:10:04,940 --> 00:10:07,820 as an example of governments 220 00:10:07,820 --> 00:10:10,700 and other land management agencies as well as scientists 221 00:10:10,700 --> 00:10:11,940 kind of quickly coming together 222 00:10:11,940 --> 00:10:16,940 to sort of develop and use technologies using remote sensing 223 00:10:17,530 --> 00:10:21,680 surrounding wildfire monitoring and mitigation. 224 00:10:21,680 --> 00:10:26,680 So it was known as the Black Summer, and massive blazes 225 00:10:27,010 --> 00:10:30,770 charred more than 186,000 square kilometers of land. 226 00:10:30,770 --> 00:10:32,890 There were also large amounts of smoke 227 00:10:32,890 --> 00:10:35,900 that billowed Southeast over the Tasmanian Sea. 228 00:10:35,900 --> 00:10:40,240 And smoke was disrupting both air traffic 229 00:10:40,240 --> 00:10:43,990 and also air quality in as far away places 230 00:10:43,990 --> 00:10:47,970 as the Western coast of South America. 231 00:10:47,970 --> 00:10:50,080 So it was obviously a really, really big blaze 232 00:10:50,080 --> 00:10:52,340 or a bunch of different blazes, 233 00:10:52,340 --> 00:10:53,600 sort of an interesting aside 234 00:10:53,600 --> 00:10:56,740 that kind of involved the remote sensing community 235 00:10:56,740 --> 00:10:58,500 and a little bit of a scandal. 236 00:10:58,500 --> 00:11:00,580 There was a whole bunch of different graphics designers 237 00:11:00,580 --> 00:11:02,160 that were sort of enhancing 238 00:11:02,160 --> 00:11:06,170 and kind of sensationalizing remote sensing imagery 239 00:11:07,010 --> 00:11:10,120 for the use of sort of sensational news stories 240 00:11:10,120 --> 00:11:11,670 about just how bad the fires were. 241 00:11:11,670 --> 00:11:15,130 So I put an example up here, at the lower left, 242 00:11:15,130 --> 00:11:16,680 of a BBC news article, 243 00:11:16,680 --> 00:11:19,110 and that's not actually a real remote sensing image, 244 00:11:19,110 --> 00:11:20,780 if you look at it closely. 245 00:11:20,780 --> 00:11:23,520 It might have been applied on a remote sensing image, 246 00:11:23,520 --> 00:11:24,970 but that image has been doctored, 247 00:11:24,970 --> 00:11:29,540 and those fires are not actually what was seen from space. 248 00:11:29,540 --> 00:11:31,950 What's usually seen from space, optically speaking, 249 00:11:31,950 --> 00:11:36,950 is a lot less bright when it comes to fires, in general. 250 00:11:37,240 --> 00:11:39,420 So just sort of an interesting aside there, 251 00:11:39,420 --> 00:11:44,420 but, in general, so I put up a couple of different examples, 252 00:11:45,820 --> 00:11:47,940 but this was an example of 253 00:11:47,940 --> 00:11:50,210 sort of operational and experimental satellite, 254 00:11:50,210 --> 00:11:54,500 remote sensing systems, kind of providing a capability 255 00:11:54,500 --> 00:11:58,350 to provide both regional and global monitoring of fires. 256 00:11:58,350 --> 00:12:00,910 So the systems that were developed, 257 00:12:00,910 --> 00:12:02,780 they provide different types of fire information 258 00:12:02,780 --> 00:12:04,710 for estimation of fire danger, 259 00:12:04,710 --> 00:12:06,890 also detecting of active fires 260 00:12:06,890 --> 00:12:09,310 and estimating burn area afterwards. 261 00:12:09,310 --> 00:12:11,540 They're also used to quantify emissions 262 00:12:11,540 --> 00:12:16,540 from fires themselves, and used to estimate fire damage, 263 00:12:17,770 --> 00:12:20,500 and they can also be used 264 00:12:20,500 --> 00:12:23,400 to monitor post-fire ecosystem recovery. 265 00:12:23,400 --> 00:12:26,210 So I put a couple of different examples here, 266 00:12:26,210 --> 00:12:29,990 the Sentinel Bushfire Monitoring System and the Firewatch, 267 00:12:29,990 --> 00:12:34,990 which are both developed by the Australian government 268 00:12:35,800 --> 00:12:37,410 and scientists. 269 00:12:37,410 --> 00:12:41,620 And these are examples of these capabilities, 270 00:12:41,620 --> 00:12:45,530 that are, satellite sensing based systems 271 00:12:45,530 --> 00:12:47,550 for accomplishing all of these different things 272 00:12:47,550 --> 00:12:49,423 with respect to wildfire. 273 00:12:50,970 --> 00:12:55,860 So in general too, as I've sort of said, 274 00:12:55,860 --> 00:12:58,930 fire monitoring is definitely a good example 275 00:12:58,930 --> 00:13:01,920 of land management and disasters communities 276 00:13:01,920 --> 00:13:03,890 kind of coming together with the scientific community 277 00:13:03,890 --> 00:13:08,230 to produce real-time or near real-time monitoring systems. 278 00:13:08,230 --> 00:13:09,310 Wildfire is definitely one 279 00:13:09,310 --> 00:13:12,260 of those examples of disturbance 280 00:13:12,260 --> 00:13:17,260 that are highly critical for knowledge on the ground 281 00:13:18,040 --> 00:13:22,880 in near real-time to be able to mitigate 282 00:13:22,880 --> 00:13:27,880 and also warn existing human habitation and infrastructure. 283 00:13:29,170 --> 00:13:30,480 And in the case of Australia, 284 00:13:30,480 --> 00:13:33,400 there was a lot of wildlife that was lost as well. 285 00:13:33,400 --> 00:13:38,070 So this is an example of kind of these communities 286 00:13:38,070 --> 00:13:41,090 coming together to work together to produce these systems. 287 00:13:41,090 --> 00:13:43,240 And they are mostly remote sensing based, 288 00:13:43,240 --> 00:13:46,420 but with a huge element of GIS as well. 289 00:13:46,420 --> 00:13:49,530 And that's kind of unique and new 290 00:13:49,530 --> 00:13:53,593 in terms of the context of history, 291 00:13:53,593 --> 00:13:55,660 and historically, governments and scientists, 292 00:13:55,660 --> 00:14:00,030 and these land management and conservation communities 293 00:14:00,030 --> 00:14:02,080 may not have come together as quickly 294 00:14:02,940 --> 00:14:05,493 to be able to do this type of work together. 295 00:14:10,010 --> 00:14:13,470 All right, so what types of satellite imagery do we use 296 00:14:13,470 --> 00:14:15,983 to monitor active fires? 297 00:14:17,320 --> 00:14:18,930 Different sources of satellite imagery 298 00:14:18,930 --> 00:14:22,720 can be used to visualize fire conditions and progressions, 299 00:14:22,720 --> 00:14:24,870 also calculate band ratios 300 00:14:24,870 --> 00:14:26,870 reflecting disturbance and severity, 301 00:14:26,870 --> 00:14:28,670 and map burned areas 302 00:14:28,670 --> 00:14:31,540 with training informed classification algorithms. 303 00:14:31,540 --> 00:14:34,050 There are a lot of different fire data sets 304 00:14:34,050 --> 00:14:36,630 that have already been made and shared 305 00:14:36,630 --> 00:14:38,540 in platforms like Google Earth Engine. 306 00:14:38,540 --> 00:14:39,870 And they're readily available 307 00:14:39,870 --> 00:14:43,130 for monitoring fire locations extends and progressions 308 00:14:43,130 --> 00:14:44,710 on a global scale. 309 00:14:44,710 --> 00:14:46,210 So this week in the tutorial, 310 00:14:46,210 --> 00:14:48,970 your goals are gonna be to access 311 00:14:50,140 --> 00:14:52,950 and visualize fire monitoring data sets 312 00:14:52,950 --> 00:14:54,480 in Google Earth Engine. 313 00:14:54,480 --> 00:14:58,070 So you're gonna look at four different types of data sets, 314 00:14:58,070 --> 00:15:00,660 also adjust previously drafted code 315 00:15:00,660 --> 00:15:04,100 to calculate fire characteristics for a fire of your choice, 316 00:15:04,100 --> 00:15:05,800 somewhere around the globe. 317 00:15:05,800 --> 00:15:09,970 And then finally, you're gonna also explore fire metrics 318 00:15:09,970 --> 00:15:13,050 and sort of visualization possibilities 319 00:15:13,050 --> 00:15:14,200 in Google Earth Engine. 320 00:15:16,420 --> 00:15:20,160 All right, next for our terrestrial applications this week, 321 00:15:20,160 --> 00:15:23,380 we're gonna talk a little bit about wildlife applications. 322 00:15:23,380 --> 00:15:26,590 And while there isn't a chapter per se 323 00:15:26,590 --> 00:15:29,460 in the textbook that you all are reading, 324 00:15:29,460 --> 00:15:33,090 concerning wildlife applications with remote sensing, 325 00:15:33,090 --> 00:15:37,430 there's a huge body of research and information out there 326 00:15:37,430 --> 00:15:39,960 to kind of understand how remote sensing 327 00:15:39,960 --> 00:15:43,350 and GIS technologies together, especially, 328 00:15:43,350 --> 00:15:46,710 are typically used in the areas of research 329 00:15:46,710 --> 00:15:49,900 and conservation for wildlife management. 330 00:15:49,900 --> 00:15:54,900 So three main areas of research concerning wildlife 331 00:15:56,190 --> 00:15:58,600 are inventory and monitoring, 332 00:15:58,600 --> 00:16:00,320 predictive modeling, 333 00:16:00,320 --> 00:16:04,637 which are things like range expansion or shifting of ranges, 334 00:16:05,890 --> 00:16:08,670 and also, wildlife management itself. 335 00:16:08,670 --> 00:16:10,910 So there's a whole bunch of different applications 336 00:16:10,910 --> 00:16:13,610 for these three areas 337 00:16:13,610 --> 00:16:18,610 in both terrestrial and aquatic wildlife studies. 338 00:16:22,430 --> 00:16:25,850 So to delve a little deeper and provide some examples, 339 00:16:25,850 --> 00:16:28,260 and within those three main areas, 340 00:16:28,260 --> 00:16:31,140 basically, remote sensing and other geospatial tools 341 00:16:31,140 --> 00:16:34,290 can provide things like an analysis of wildlife distribution 342 00:16:34,290 --> 00:16:35,970 and population health, 343 00:16:35,970 --> 00:16:38,480 assessment of physical and environmental conditions 344 00:16:38,480 --> 00:16:40,940 to determine habitat suitability to species 345 00:16:40,940 --> 00:16:42,640 and changes to habitat, 346 00:16:42,640 --> 00:16:45,800 measurement of factors that determine habitat suitability, 347 00:16:45,800 --> 00:16:50,370 like climate, vegetation, land use, other surface features, 348 00:16:50,370 --> 00:16:53,310 analysis of habitat fragmentation, 349 00:16:53,310 --> 00:16:55,210 the connectivity of wildlife corridors 350 00:16:55,210 --> 00:16:57,310 in urban and suburban areas, 351 00:16:57,310 --> 00:17:00,930 and also things like migration patterns and range shifts. 352 00:17:00,930 --> 00:17:05,030 So as an example of that latter one, especially two, 353 00:17:05,030 --> 00:17:08,160 at the bottom here, that first figure on the left 354 00:17:08,160 --> 00:17:10,540 shows the location of test sites 355 00:17:10,540 --> 00:17:14,310 and what their land cover composition looks like. 356 00:17:14,310 --> 00:17:16,140 And these test sites were for a study 357 00:17:16,140 --> 00:17:18,330 that was looking at the migration patterns 358 00:17:18,330 --> 00:17:22,490 of white stork across Southwestern Europe. 359 00:17:22,490 --> 00:17:26,950 So the graph at the right shows the relative distribution 360 00:17:26,950 --> 00:17:30,120 of movement data points per class 361 00:17:30,120 --> 00:17:35,120 based on a land cover data set that's from way back in 2012. 362 00:17:36,640 --> 00:17:41,160 So the figure shows winter in blue and summer in orange, 363 00:17:41,160 --> 00:17:43,660 migration for this white stork individuals 364 00:17:43,660 --> 00:17:45,933 that were collared and measured in the study. 365 00:17:47,110 --> 00:17:50,130 So the points were registered over agriculture, 366 00:17:50,130 --> 00:17:53,490 and suggested that the animals spent a lot of time 367 00:17:54,730 --> 00:17:57,700 that was recorded within that land cover type, 368 00:17:57,700 --> 00:18:00,370 and so, they were sort of looking at 369 00:18:00,370 --> 00:18:03,040 also where nesting sites were as well, 370 00:18:03,040 --> 00:18:06,490 and kind of developing a greater understanding 371 00:18:06,490 --> 00:18:11,490 of what the habits were of this white stork population 372 00:18:12,020 --> 00:18:15,970 as it migrated and moved in different areas 373 00:18:15,970 --> 00:18:18,380 throughout Southwestern Europe. 374 00:18:18,380 --> 00:18:19,760 So this is an example 375 00:18:19,760 --> 00:18:24,600 that covers sort of a few of these different topics, 376 00:18:24,600 --> 00:18:28,360 including things like, maybe perhaps looking at rain shift 377 00:18:28,360 --> 00:18:30,450 or also looking at the distribution 378 00:18:30,450 --> 00:18:33,180 and sort of what their habitat, 379 00:18:33,180 --> 00:18:35,730 what their conditions were looked like 380 00:18:35,730 --> 00:18:37,630 and what their suitable habitats were. 381 00:18:40,810 --> 00:18:41,760 So we can also talk 382 00:18:41,760 --> 00:18:46,170 about each of these different applications that I mentioned 383 00:18:46,170 --> 00:18:49,180 in a little bit more detail. 384 00:18:49,180 --> 00:18:53,070 First, if we're talking about inventorying and monitoring, 385 00:18:53,070 --> 00:18:56,380 scientists and conservationists and land managers as well 386 00:18:56,380 --> 00:19:00,360 often accomplish these remote sensing-based inventories 387 00:19:00,360 --> 00:19:03,970 in conjunction with GIS location records. 388 00:19:03,970 --> 00:19:05,340 And sometimes, these location records 389 00:19:05,340 --> 00:19:07,453 can also be crowdsourced as well. 390 00:19:08,900 --> 00:19:13,900 They are usually species presence and absence measurements, 391 00:19:14,670 --> 00:19:17,700 and they're usually overlaid with land cover class, 392 00:19:17,700 --> 00:19:21,420 which can help then kind of inform habitat ranges. 393 00:19:21,420 --> 00:19:24,080 They can also use surveys and inventories, 394 00:19:24,080 --> 00:19:28,560 and scale them to the pixel level to create range maps, 395 00:19:28,560 --> 00:19:32,980 and also calculate species richness and diversity 396 00:19:32,980 --> 00:19:34,893 within these areas that they're found. 397 00:19:35,930 --> 00:19:38,230 So the figure at the bottom shows a simple workflow, 398 00:19:38,230 --> 00:19:40,430 where researchers overlaid animal movement data 399 00:19:40,430 --> 00:19:42,920 with time-series, remote sense and imagery. 400 00:19:42,920 --> 00:19:45,610 After joining them together, they extracted the pixel value 401 00:19:45,610 --> 00:19:48,963 for each animal position through time for analysis. 402 00:19:50,820 --> 00:19:55,820 And this was all done using R and also Google Earth Engine. 403 00:20:00,470 --> 00:20:04,720 Some applications for wildlife management 404 00:20:04,720 --> 00:20:05,553 using remote sensing, 405 00:20:05,553 --> 00:20:09,310 including to protect and enhance wildlife populations, 406 00:20:09,310 --> 00:20:12,880 research wildlife population ranges in movement, 407 00:20:12,880 --> 00:20:16,700 also understand habitats and change to habitats, 408 00:20:16,700 --> 00:20:18,420 and also to promote opportunities 409 00:20:18,420 --> 00:20:23,420 for kind of the coordination of recreational research 410 00:20:23,660 --> 00:20:26,690 and education kind of surrounding wildlife. 411 00:20:26,690 --> 00:20:28,620 So can we protect areas 412 00:20:28,620 --> 00:20:33,620 that also include protecting wildlife 413 00:20:34,070 --> 00:20:39,060 as well as providing areas of recreation for humans 414 00:20:40,540 --> 00:20:41,890 in addition? 415 00:20:41,890 --> 00:20:46,890 So this is actually an example in the right of a workflow 416 00:20:46,990 --> 00:20:51,990 that was basically creating a habitat suitability map, 417 00:20:52,140 --> 00:20:57,140 and also then using that to monitor habitat change 418 00:20:58,430 --> 00:21:03,430 for a particular species of bird that was of interest. 419 00:21:09,110 --> 00:21:09,943 So also, 420 00:21:09,943 --> 00:21:12,800 a little bit more about habitat suitability and change. 421 00:21:12,800 --> 00:21:16,340 So habitat suitability and home ranges are, again, 422 00:21:16,340 --> 00:21:20,360 also determined usually in conjunction with GIS 423 00:21:20,360 --> 00:21:22,010 using remote sensing. 424 00:21:22,010 --> 00:21:26,140 So they typically involve combining results by year 425 00:21:26,140 --> 00:21:27,700 to add a temporal dimension, 426 00:21:27,700 --> 00:21:31,050 to kind of understand range shifts as well. 427 00:21:31,050 --> 00:21:34,890 And when we track and monitor them through time, 428 00:21:34,890 --> 00:21:36,110 it can shed light on 429 00:21:37,030 --> 00:21:41,460 how or why migration patterns may shift. 430 00:21:41,460 --> 00:21:45,670 So here is a study by Crego et al in 2021, 431 00:21:45,670 --> 00:21:50,570 which extracted NDVI values across a year of data 432 00:21:50,570 --> 00:21:53,210 for range areas, 433 00:21:53,210 --> 00:21:57,690 sort of in conjunction with a couple of different species, 434 00:21:57,690 --> 00:22:01,650 including African Buffalo elephant and wilderbeest 435 00:22:01,650 --> 00:22:05,750 that were being tracked using GPS telemetry units. 436 00:22:05,750 --> 00:22:07,750 So the results sort of showed 437 00:22:07,750 --> 00:22:11,260 a bunch of different discrepancies between annual mean NDVI, 438 00:22:11,260 --> 00:22:13,530 which is in red of these graphs, 439 00:22:13,530 --> 00:22:16,170 and the time match NDVI, which is in blue, 440 00:22:16,170 --> 00:22:20,610 at the animal locations across the species and the regions. 441 00:22:20,610 --> 00:22:21,860 So they were sort of using this 442 00:22:21,860 --> 00:22:26,490 to kind of understand where these animals were ranging 443 00:22:26,490 --> 00:22:30,070 and what the health of those ecosystems looked like 444 00:22:30,070 --> 00:22:31,970 for the time that they were measuring. 445 00:22:34,640 --> 00:22:37,050 So the third area of terrestrial application 446 00:22:37,050 --> 00:22:38,360 that we're gonna talk about today 447 00:22:38,360 --> 00:22:41,350 is forest degradation and deforestation. 448 00:22:41,350 --> 00:22:44,450 Forests are fundamentally important ecosystems 449 00:22:44,450 --> 00:22:45,283 on our planet. 450 00:22:45,283 --> 00:22:47,820 They occupy 1/3rd of the planet's land mass, 451 00:22:47,820 --> 00:22:50,670 covering nearly 4 billion hectares. 452 00:22:50,670 --> 00:22:52,010 The largest forested areas 453 00:22:52,010 --> 00:22:55,050 are located in the boreal and equatorial zones. 454 00:22:55,050 --> 00:22:57,770 And the current distribution is not fixed, 455 00:22:57,770 --> 00:23:02,770 but has a continuously changing sort of area over time 456 00:23:03,820 --> 00:23:05,890 due to the influence of environmental change 457 00:23:05,890 --> 00:23:08,420 and also human impacts. 458 00:23:08,420 --> 00:23:11,060 So 1/3rd of a current forest areas 459 00:23:11,060 --> 00:23:14,630 are made up of forest that is considered primary are intact. 460 00:23:14,630 --> 00:23:15,870 And the remaining 2/3rds 461 00:23:15,870 --> 00:23:18,030 are subjected to a lot of anthropogenic activity 462 00:23:18,030 --> 00:23:21,790 and have, as a result, an uncertain future. 463 00:23:21,790 --> 00:23:24,453 The global forest condition, as visualized here, 464 00:23:26,400 --> 00:23:28,690 was developed using satellite-derived 465 00:23:28,690 --> 00:23:32,490 and also modeled current and future 466 00:23:32,490 --> 00:23:34,560 or potential, sorry, forest cover. 467 00:23:34,560 --> 00:23:39,140 It's a map that was developed by Mitchell et al. in 2017. 468 00:23:39,140 --> 00:23:41,900 And it was for the purpose of 469 00:23:41,900 --> 00:23:44,220 sort of understanding carbon balance 470 00:23:44,220 --> 00:23:46,033 across these different ecosystems. 471 00:23:48,430 --> 00:23:49,610 All right, so in order to be able 472 00:23:49,610 --> 00:23:52,280 to measure forest change from space, 473 00:23:52,280 --> 00:23:53,750 it's really important to distinguish 474 00:23:53,750 --> 00:23:55,680 between deforestation and degradation 475 00:23:55,680 --> 00:23:57,510 and what we mean by those terms, 476 00:23:57,510 --> 00:23:59,760 and especially, how we would measure the two of them 477 00:23:59,760 --> 00:24:01,550 from remote sensing. 478 00:24:01,550 --> 00:24:04,580 So each year, several million hectares of forest disappear, 479 00:24:04,580 --> 00:24:07,970 and vast areas are subjected to degradation as well. 480 00:24:07,970 --> 00:24:11,493 Deforestation is the complete removal of forest cover. 481 00:24:12,330 --> 00:24:15,820 Conversion into agricultural and pasture lands 482 00:24:15,820 --> 00:24:18,610 are the principle causes of deforestation. 483 00:24:18,610 --> 00:24:20,880 But once certain areas are abandoned, 484 00:24:20,880 --> 00:24:23,050 sometimes, forest cover can once again develop 485 00:24:23,050 --> 00:24:26,100 leading to what is known as secondary forests. 486 00:24:26,100 --> 00:24:27,370 Unlike deforestation, 487 00:24:27,370 --> 00:24:30,530 degradation is partial destruction of forest cover. 488 00:24:30,530 --> 00:24:33,100 It's defined as a loss of the forest capacity 489 00:24:33,100 --> 00:24:34,740 to provide ecosystem services 490 00:24:34,740 --> 00:24:36,740 like carbon storage or forest products, 491 00:24:36,740 --> 00:24:40,063 following the anthropogenic activity, 492 00:24:40,920 --> 00:24:45,920 which are things like unsustainable logging, agriculture, 493 00:24:46,730 --> 00:24:51,730 sort of the encroachment of invasive species, fire, 494 00:24:52,560 --> 00:24:55,043 fuel wood gathering, and livestock grazing. 495 00:24:56,820 --> 00:24:59,620 So, degraded forests 496 00:24:59,620 --> 00:25:02,300 include a large diversity of forest types 497 00:25:02,300 --> 00:25:04,060 depending on the nature, intensity, 498 00:25:04,060 --> 00:25:05,790 and frequency of the degradation. 499 00:25:05,790 --> 00:25:08,950 And they can be also, as I'm sure you can imagine, 500 00:25:08,950 --> 00:25:13,513 a bit more difficult to measure and characterize from space. 501 00:25:14,560 --> 00:25:19,130 So an example of this is the map and graph at the bottom. 502 00:25:19,130 --> 00:25:20,400 So on the left, 503 00:25:20,400 --> 00:25:25,400 the map is showing degradation between the years 1987-2019 504 00:25:26,210 --> 00:25:28,270 for the country of Georgia. 505 00:25:28,270 --> 00:25:32,180 And those two insights show regions within the country 506 00:25:32,180 --> 00:25:36,913 that are showing a large expansion of forest degradation. 507 00:25:37,980 --> 00:25:42,230 The graph at the right is pixel counting for the entire map 508 00:25:42,230 --> 00:25:45,530 of both deforestation and degradation 509 00:25:45,530 --> 00:25:47,600 between those years. 510 00:25:47,600 --> 00:25:49,130 And as you can see, 511 00:25:49,130 --> 00:25:52,070 degradation accounts for many, many more pixels 512 00:25:52,070 --> 00:25:55,300 than deforestation accounts for. 513 00:25:55,300 --> 00:25:57,790 So, many of these areas, 514 00:25:57,790 --> 00:26:01,220 degradation is much more of an issue or a problem 515 00:26:01,220 --> 00:26:02,753 than deforestation. 516 00:26:04,340 --> 00:26:06,850 So how do we characterize and measure degradation, 517 00:26:06,850 --> 00:26:08,620 particularly from space? 518 00:26:08,620 --> 00:26:10,720 Here is an example of the type of information 519 00:26:10,720 --> 00:26:12,270 that we might try to collect. 520 00:26:12,270 --> 00:26:15,407 So in the inset 'A,' 521 00:26:17,088 --> 00:26:21,150 that's maybe known as moderate forest degradation, 522 00:26:21,150 --> 00:26:22,900 there's cutting sites in there, 523 00:26:22,900 --> 00:26:25,090 and in the process of regeneration. 524 00:26:25,090 --> 00:26:29,690 And so, using a Landsat image with a color-composite, 525 00:26:29,690 --> 00:26:31,870 with mid infrared, near infrared 526 00:26:31,870 --> 00:26:35,083 in red and the blue channel respectively, 527 00:26:35,945 --> 00:26:40,080 we could perhaps be able to gain a better understanding 528 00:26:40,080 --> 00:26:43,163 of how degraded that inset is. 529 00:26:44,240 --> 00:26:46,250 'B' shows more intense degradation, 530 00:26:46,250 --> 00:26:48,560 with large portions of the ground visible, 531 00:26:48,560 --> 00:26:51,120 following forest exploitation. 532 00:26:51,120 --> 00:26:54,350 And 'C' shows the average degradation 533 00:26:54,350 --> 00:26:57,540 with areas of exploitation 534 00:26:57,540 --> 00:27:01,010 and also old roads which are growing in. 535 00:27:01,010 --> 00:27:05,610 So all of these areas pictured are in Brazil. 536 00:27:05,610 --> 00:27:07,420 And this was from a study 537 00:27:07,420 --> 00:27:10,640 where they were talking about how to kind of differentiate 538 00:27:10,640 --> 00:27:13,240 between these different patterns of degradation 539 00:27:13,240 --> 00:27:16,140 and these different levels of recovery 540 00:27:16,140 --> 00:27:18,570 from degraded areas as well. 541 00:27:18,570 --> 00:27:22,070 So the most common vegetation indices that we use 542 00:27:22,070 --> 00:27:25,360 are ones that differentiate between bare ground 543 00:27:25,360 --> 00:27:26,193 and vegetation. 544 00:27:26,193 --> 00:27:28,510 Those are sort of the main areas. 545 00:27:28,510 --> 00:27:29,977 So I guess the next question is, 546 00:27:29,977 --> 00:27:34,647 "Which images should be used for measuring deforestation?" 547 00:27:37,167 --> 00:27:38,727 "Which images should be used 548 00:27:38,727 --> 00:27:43,250 "for measuring deforestation and degradation?" 549 00:27:43,250 --> 00:27:46,960 As with many of the other areas of remote sensing 550 00:27:46,960 --> 00:27:48,390 that we've talked about, 551 00:27:48,390 --> 00:27:51,700 there are trade-offs and benefits of using each type of data 552 00:27:51,700 --> 00:27:55,003 to characterize both deforestation and degradation. 553 00:27:56,270 --> 00:27:57,690 Optical data, for instance, 554 00:27:57,690 --> 00:28:01,420 is acquired in the visible and near infrared at wavelengths, 555 00:28:01,420 --> 00:28:04,990 and it's sensitive to operating variables of vegetation 556 00:28:04,990 --> 00:28:06,900 like chlorophyll activity, 557 00:28:06,900 --> 00:28:11,360 also structure and internal water content of leaves. 558 00:28:11,360 --> 00:28:12,693 Radar, on the other hand, 559 00:28:13,620 --> 00:28:15,910 which can be acquired at multiple frequencies 560 00:28:15,910 --> 00:28:16,960 with longer wavelengths, 561 00:28:16,960 --> 00:28:19,700 are sensitive to the structure of the vegetation, 562 00:28:19,700 --> 00:28:22,400 and also can be sensitive to the structure of the ground. 563 00:28:22,400 --> 00:28:23,720 And structural properties 564 00:28:23,720 --> 00:28:26,170 include things like basal area stand structure 565 00:28:26,170 --> 00:28:27,870 and the presence of water as well. 566 00:28:29,200 --> 00:28:32,670 So this figure shows common remote-sensing platforms 567 00:28:32,670 --> 00:28:37,670 and sensor combinations of remote sensing data 568 00:28:38,070 --> 00:28:43,070 that we can use to measure deforestation and degradation. 569 00:28:43,310 --> 00:28:45,030 On the left, we see these platforms 570 00:28:45,030 --> 00:28:47,250 and the most commonly utilized sensors 571 00:28:47,250 --> 00:28:49,090 for specific platforms, 572 00:28:49,090 --> 00:28:53,060 and also the relative altitude at which they're deployed. 573 00:28:53,060 --> 00:28:55,700 On the right, we show, at the top, 574 00:28:55,700 --> 00:28:57,740 a true-color digital aerial photography 575 00:28:57,740 --> 00:29:00,103 and false-color with near infrared sensing, 576 00:29:02,820 --> 00:29:07,120 which is an example of an optical data set 577 00:29:07,120 --> 00:29:08,850 in the middle of a Lidar point cloud 578 00:29:08,850 --> 00:29:10,500 of vegetation near river. 579 00:29:10,500 --> 00:29:14,670 And finally, SAR data, or Synthetic Aperture Radar Data, 580 00:29:14,670 --> 00:29:17,990 for two polarizations from Sentinel one at the bottom. 581 00:29:17,990 --> 00:29:21,400 So a key advantage of Lidar and SAR sensors overall, 582 00:29:21,400 --> 00:29:22,800 and especially in this context, 583 00:29:22,800 --> 00:29:25,500 is their ability to penetrate clouds and smoke 584 00:29:25,500 --> 00:29:27,310 and operate at night. 585 00:29:27,310 --> 00:29:30,140 SAR sensors can differentiate land cover features 586 00:29:30,140 --> 00:29:32,210 according to the surface roughness 587 00:29:32,210 --> 00:29:37,200 and 3D structure of the forest features, 588 00:29:37,200 --> 00:29:38,863 as well as water content too. 589 00:29:39,770 --> 00:29:41,930 So, depending on the wavelength of a sensor, 590 00:29:41,930 --> 00:29:44,960 whether it's X band, L band, or C band, 591 00:29:44,960 --> 00:29:47,520 so those are three different types of radar, 592 00:29:47,520 --> 00:29:51,763 the signal can penetrate vegetation, canopy, and also soils. 593 00:29:52,830 --> 00:29:55,930 Conversely, Lidar systems, as we know, 594 00:29:55,930 --> 00:29:57,710 admit pulses from lasers, 595 00:29:57,710 --> 00:30:00,143 and because of that, 596 00:30:01,060 --> 00:30:02,340 their ability to measure the distance 597 00:30:02,340 --> 00:30:04,100 and the reflective light, 598 00:30:04,100 --> 00:30:05,330 it allows for the capability 599 00:30:05,330 --> 00:30:09,423 for measuring or making 3D representations of the system. 600 00:30:11,430 --> 00:30:12,530 And degraded systems 601 00:30:12,530 --> 00:30:15,200 are able to be accurately captured by Lidar, 602 00:30:15,200 --> 00:30:18,840 provided that they're flown at the proper resolution 603 00:30:18,840 --> 00:30:23,830 for being able to capture changes and structural features. 604 00:30:23,830 --> 00:30:24,930 So in recent years, 605 00:30:24,930 --> 00:30:27,910 as a result of developments of sensor technology, 606 00:30:27,910 --> 00:30:30,100 passive and active sensors now have versions 607 00:30:30,100 --> 00:30:32,290 that can be mounted on all platforms. 608 00:30:32,290 --> 00:30:37,030 So, including things like ground-based drone or UAV, 609 00:30:37,030 --> 00:30:39,900 which is Unmanned Aerial Vehicles, 610 00:30:39,900 --> 00:30:42,913 airborne or spaceborne platforms as well. 611 00:30:44,020 --> 00:30:44,853 Although, 612 00:30:47,350 --> 00:30:50,530 larger platforms such as satellites and larger planes, 613 00:30:50,530 --> 00:30:52,380 they can herein carry heavier play loads 614 00:30:52,380 --> 00:30:54,730 so they can allow for larger sensor systems 615 00:30:54,730 --> 00:30:58,093 that have a higher quality and accuracy of data capture. 616 00:30:59,530 --> 00:31:03,713 However, because of this sort of sensor miniaturization, 617 00:31:04,950 --> 00:31:07,640 we can also find them now mounted 618 00:31:07,640 --> 00:31:11,040 kind of in newly found combinations. 619 00:31:11,040 --> 00:31:15,280 So such as mounting a synthetic aperture radar on a UAV, 620 00:31:15,280 --> 00:31:17,290 which wasn't something that was available 621 00:31:17,290 --> 00:31:18,650 maybe 10 years ago. 622 00:31:18,650 --> 00:31:23,420 So we're able to start kind of providing more, 623 00:31:23,420 --> 00:31:26,550 more options for data set capture. 624 00:31:26,550 --> 00:31:29,070 And in addition, it's also becoming possible 625 00:31:29,070 --> 00:31:32,276 to mount two sensor types on the same platform 626 00:31:32,276 --> 00:31:35,380 and develop sort of analysis systems. 627 00:31:35,380 --> 00:31:38,560 The user can process multiple different data sets 628 00:31:38,560 --> 00:31:40,810 to produce higher quality maps 629 00:31:40,810 --> 00:31:45,430 and do things like develop methods to unmix spectral data, 630 00:31:48,350 --> 00:31:49,390 for instance. 631 00:31:49,390 --> 00:31:51,280 So this approach considers pixels 632 00:31:51,280 --> 00:31:56,280 made up of more than one type of soil, covered to be mixed. 633 00:31:56,630 --> 00:31:58,820 This is sort of an example. 634 00:31:58,820 --> 00:32:03,300 So for example, a pixel location of on the edge of a forest 635 00:32:03,300 --> 00:32:05,380 that contains a portion of bare soil 636 00:32:05,380 --> 00:32:06,850 and a portion of vegetation 637 00:32:08,870 --> 00:32:12,210 might be contrasted to a pixel 638 00:32:12,210 --> 00:32:16,380 that is considered to be more pure, is how it's termed, 639 00:32:16,380 --> 00:32:20,130 where it might just be the entire pixel is made up of soil 640 00:32:20,130 --> 00:32:22,700 or the entire pixel is made up of vegetation. 641 00:32:22,700 --> 00:32:25,730 So spectrum on mixing methods are able to determine 642 00:32:25,730 --> 00:32:28,790 the internal composition of mixed pixels 643 00:32:28,790 --> 00:32:31,130 on the basis of spectral characteristics 644 00:32:31,130 --> 00:32:35,010 of these pure pixels, which are also known as end members, 645 00:32:35,010 --> 00:32:39,630 and the pure pixel end members are able to show 646 00:32:39,630 --> 00:32:44,630 all the different classes of cover and change in one image. 647 00:32:52,520 --> 00:32:55,330 So, kind of similar to this concept 648 00:32:55,330 --> 00:32:58,060 of being able to fly more than one sensor, 649 00:32:58,060 --> 00:33:01,160 to be able to accomplish collection 650 00:33:01,160 --> 00:33:06,160 of sort of more spectorally-diverse data sets, 651 00:33:06,540 --> 00:33:09,030 and sort of putting them together 652 00:33:09,030 --> 00:33:12,210 and performing operations like spectral and mixing, 653 00:33:12,210 --> 00:33:13,740 which I sort of went into a little bit 654 00:33:13,740 --> 00:33:14,710 in the last slide there, 655 00:33:14,710 --> 00:33:19,710 where you might have the ability to characterize pixels 656 00:33:20,970 --> 00:33:24,090 that have a have mixed vegetation and soil 657 00:33:24,090 --> 00:33:27,963 or vegetation and other type in them, 658 00:33:29,570 --> 00:33:31,870 we've also started to think about 659 00:33:31,870 --> 00:33:35,010 how we can apply hyperspectral data sets 660 00:33:35,010 --> 00:33:40,010 or data sets that have a much larger spectral capture 661 00:33:40,030 --> 00:33:42,890 to study things like forest degradation, 662 00:33:42,890 --> 00:33:44,750 because as you can imagine, 663 00:33:44,750 --> 00:33:47,530 looking at all the broccoli from space, 664 00:33:47,530 --> 00:33:51,140 so sometimes, I have to refer to forest 665 00:33:51,140 --> 00:33:55,800 or healthy intact forest as broccoli, 666 00:33:55,800 --> 00:33:57,900 you're looking at the broccoli from space, 667 00:33:58,820 --> 00:34:00,370 there's a whole bunch of different difficulty 668 00:34:00,370 --> 00:34:01,410 that kind of arises 669 00:34:01,410 --> 00:34:06,410 when you have all these different types of disturbances 670 00:34:06,970 --> 00:34:09,880 or human interaction that's going on below the broccoli, 671 00:34:09,880 --> 00:34:13,220 so, creation of roads and other things like that. 672 00:34:13,220 --> 00:34:16,280 And some of it can be seen from space and some of it can't. 673 00:34:16,280 --> 00:34:19,730 So how can we measure from the top down 674 00:34:19,730 --> 00:34:20,690 all of these different things 675 00:34:20,690 --> 00:34:23,100 that are sometimes occurring from the bottom up? 676 00:34:23,100 --> 00:34:25,130 So hyperspectral data sets are one way 677 00:34:25,130 --> 00:34:28,050 that we can also kind of approach 678 00:34:28,050 --> 00:34:32,360 measuring forest degradation and deforestation, 679 00:34:32,360 --> 00:34:34,520 and especially degradation. 680 00:34:34,520 --> 00:34:37,030 So the figure at the top is kind of a reminder 681 00:34:37,030 --> 00:34:38,770 of the different bands or channels 682 00:34:38,770 --> 00:34:41,210 that are measured by common sensors. 683 00:34:41,210 --> 00:34:44,360 So the two at the top show Landsat and Sentinel, 684 00:34:44,360 --> 00:34:45,490 and at the bottom, 685 00:34:45,490 --> 00:34:47,520 it shows the near continuous measurement 686 00:34:47,520 --> 00:34:49,690 of hyperspectral data sets. 687 00:34:49,690 --> 00:34:52,930 So the use of optical data like Sentinel, Landsat 688 00:34:52,930 --> 00:34:55,240 to identify deforestation or forest degradation 689 00:34:55,240 --> 00:34:57,700 takes into account both spatial 690 00:34:57,700 --> 00:35:01,190 and temporal variations in reflectance, 691 00:35:01,190 --> 00:35:02,860 but there's also an increasing trend 692 00:35:02,860 --> 00:35:04,230 in the development of remote sensing 693 00:35:04,230 --> 00:35:06,540 in many tropical countries to increase 694 00:35:06,540 --> 00:35:10,400 from these optical or panchromatic wavelengths 695 00:35:12,520 --> 00:35:14,853 to multispectral satellite systems, 696 00:35:15,960 --> 00:35:20,960 and also to especially hyperspectral sensing technologies 697 00:35:21,120 --> 00:35:22,650 on airborne platforms. 698 00:35:22,650 --> 00:35:24,490 So like other remote sensing technologies, 699 00:35:24,490 --> 00:35:26,910 hyperspectral offers certain advantages 700 00:35:26,910 --> 00:35:29,830 over conventional multispectral data. 701 00:35:29,830 --> 00:35:33,500 So hyperspectral data or imagery can be used 702 00:35:33,500 --> 00:35:36,500 or can be considered as kind of a single-image data set 703 00:35:36,500 --> 00:35:39,820 with a continuous spectrum of radiance or reflectance 704 00:35:39,820 --> 00:35:43,470 as values associated with each pixel. 705 00:35:43,470 --> 00:35:45,750 Research and development of hyperspectral imagery 706 00:35:45,750 --> 00:35:49,690 has been an active area of research during the past decade 707 00:35:49,690 --> 00:35:51,150 and even further back as well. 708 00:35:51,150 --> 00:35:54,270 But sort of this application with forests 709 00:35:54,270 --> 00:35:56,250 and disturbance in forest 710 00:35:57,140 --> 00:35:59,550 has been on the rise in the past decade. 711 00:35:59,550 --> 00:36:01,033 So, hyperspectral imagery, 712 00:36:02,256 --> 00:36:03,940 it may be better suited 713 00:36:03,940 --> 00:36:06,860 to resolve things like drought effects on tropical forest, 714 00:36:06,860 --> 00:36:09,140 because they are highly sensitive 715 00:36:09,140 --> 00:36:12,943 to canopy leaf water content and light use efficiency. 716 00:36:13,970 --> 00:36:15,730 So the use of hyperspectral data 717 00:36:15,730 --> 00:36:19,500 also allows the retrieval of specific biochemical compounds 718 00:36:19,500 --> 00:36:22,090 associated with plant function and structure, 719 00:36:22,090 --> 00:36:25,080 including things like foliar nitrogen, 720 00:36:25,080 --> 00:36:28,310 carbon and nitrogen ratios, chlorophyll concentration, 721 00:36:28,310 --> 00:36:31,460 and structural compounds like lignin or cellulose. 722 00:36:31,460 --> 00:36:34,310 And in terms of monitoring degradation, 723 00:36:34,310 --> 00:36:36,330 the derivation of functional types 724 00:36:36,330 --> 00:36:39,190 has commonly centered on the use of fine spectral resolution 725 00:36:39,190 --> 00:36:40,940 or hyperspectral data 726 00:36:40,940 --> 00:36:44,500 in order to discriminate among different plant attributes. 727 00:36:44,500 --> 00:36:46,670 So within a highly-heterogeneous landscape 728 00:36:46,670 --> 00:36:50,910 like that of a top view of a degraded forest, 729 00:36:50,910 --> 00:36:55,390 having more spectra means less possibility of overlap 730 00:36:55,390 --> 00:36:58,250 when characterizing these vegetation types, 731 00:36:58,250 --> 00:36:59,450 and also their health, 732 00:36:59,450 --> 00:37:03,010 and other features like soils and non-tree vegetation. 733 00:37:03,010 --> 00:37:05,030 Basically, with hyperspectral data, 734 00:37:05,030 --> 00:37:07,390 spectral fingerprints are features 735 00:37:07,390 --> 00:37:11,000 are able to be more finally detailed 736 00:37:11,000 --> 00:37:13,713 to the remote-sensing specialist. 737 00:37:15,670 --> 00:37:18,530 So, remote sensing can be used, as we know, 738 00:37:18,530 --> 00:37:20,700 for mapping and measuring virtually 739 00:37:20,700 --> 00:37:23,910 all types of landscape variables, 740 00:37:23,910 --> 00:37:26,750 but especially key forest variables 741 00:37:26,750 --> 00:37:29,280 like density and basal area. 742 00:37:29,280 --> 00:37:31,770 So I wanted to include this table here, 743 00:37:31,770 --> 00:37:33,350 because it highlights 744 00:37:33,350 --> 00:37:35,510 really particularly common applications 745 00:37:35,510 --> 00:37:38,360 for specific remote-sensing systems. 746 00:37:38,360 --> 00:37:42,530 So the double check marks refer to more common applications 747 00:37:42,530 --> 00:37:43,610 and the single check marks 748 00:37:43,610 --> 00:37:45,650 refer to less common applications. 749 00:37:45,650 --> 00:37:46,680 So it's important to note 750 00:37:46,680 --> 00:37:48,790 that there are examples in the literature 751 00:37:48,790 --> 00:37:51,803 for nearly all sensor and platform combinations. 752 00:37:52,980 --> 00:37:56,030 So as you can see from here, some of the columns, 753 00:37:56,030 --> 00:37:59,066 which are multispectral, fine spatial resolution, 754 00:37:59,066 --> 00:38:02,498 multispectral medium or coarse spatial resolution, 755 00:38:02,498 --> 00:38:06,090 hyperspectral, synthetic-aperture radar, and Lidar, 756 00:38:06,090 --> 00:38:08,270 those are the different divisions that we're thinking about. 757 00:38:08,270 --> 00:38:09,960 And then we're also looking at things 758 00:38:09,960 --> 00:38:12,180 like land use and land cover change, 759 00:38:12,180 --> 00:38:15,230 how they're applied, cover estimates, 760 00:38:15,230 --> 00:38:17,830 and there's a whole bunch of subcategories for that. 761 00:38:19,470 --> 00:38:21,650 Also, vegetation structure, 762 00:38:21,650 --> 00:38:23,943 vegetation chemistry and moisture, 763 00:38:25,810 --> 00:38:29,060 biodiversity, disturbance, and soil. 764 00:38:29,060 --> 00:38:30,230 So I wanted to kind of 765 00:38:32,130 --> 00:38:36,670 include all of these different attributes in here, 766 00:38:36,670 --> 00:38:38,520 because I thought it was a good way 767 00:38:38,520 --> 00:38:40,470 to kind of pictorially represent 768 00:38:40,470 --> 00:38:42,710 what these different types of remote sensing platforms 769 00:38:42,710 --> 00:38:45,940 that we've talked about a bit this semester, 770 00:38:45,940 --> 00:38:48,880 how they're kind of used with respect to forests 771 00:38:48,880 --> 00:38:52,510 and measurement of forest degradation and deforestation. 772 00:38:52,510 --> 00:38:55,420 I'm sure similar tables might be out there 773 00:38:55,420 --> 00:38:57,980 for thinking about how they might be applied 774 00:38:57,980 --> 00:39:00,170 to things like wildlife management, 775 00:39:00,170 --> 00:39:02,560 as well as buyers as well. 776 00:39:02,560 --> 00:39:04,110 But, if you have a chance, I guess, 777 00:39:04,110 --> 00:39:07,040 spend a few minutes reading these different slides 778 00:39:07,040 --> 00:39:12,040 to kind of look at how they're put together, 779 00:39:12,230 --> 00:39:16,510 and kind of what different examples of their uses might be, 780 00:39:16,510 --> 00:39:19,693 which is in the column at the far right. 781 00:39:22,660 --> 00:39:27,140 So finally, as kind of a case study example, 782 00:39:27,140 --> 00:39:29,330 I wanted to bring up a study area 783 00:39:30,210 --> 00:39:34,260 that kind of been on the rise in the past couple decades, 784 00:39:34,260 --> 00:39:37,490 and that's the Democratic Republic of the Congo. 785 00:39:37,490 --> 00:39:42,490 So, in many forest loss hotspots across our planet, 786 00:39:43,360 --> 00:39:45,160 and the DRC is one of those, 787 00:39:45,160 --> 00:39:47,460 accurate classification of forest cover and type 788 00:39:47,460 --> 00:39:50,530 are needed before we can even measure loss. 789 00:39:50,530 --> 00:39:52,930 There's many different places across the tropics 790 00:39:52,930 --> 00:39:55,690 that's home to where all of the rainforest in the world are. 791 00:39:55,690 --> 00:39:58,730 And many of the countries that contain the rainforest 792 00:39:59,720 --> 00:40:01,540 don't necessarily have access 793 00:40:01,540 --> 00:40:06,540 to sophisticated remote-sensing image processing systems, 794 00:40:09,816 --> 00:40:12,100 and so they rely heavily, and increasingly, 795 00:40:12,100 --> 00:40:15,470 they rely heavily on platforms like Google Earth Engine 796 00:40:15,470 --> 00:40:19,410 to be able to interpret what the free imagery out there, 797 00:40:19,410 --> 00:40:21,050 like Landsat, is measuring. 798 00:40:21,050 --> 00:40:26,050 So this is a study from 2020, and it was basically a map 799 00:40:28,060 --> 00:40:31,370 where the study region was at the forest edge 800 00:40:32,270 --> 00:40:37,253 of the North Batéké Chiefdom. 801 00:40:38,096 --> 00:40:39,520 And I kind of wanted to include this here 802 00:40:39,520 --> 00:40:41,700 just to kind of give you an idea 803 00:40:41,700 --> 00:40:45,870 of how complicated the workflows are 804 00:40:45,870 --> 00:40:50,870 to get to be able to accurately classify forests 805 00:40:52,420 --> 00:40:55,010 and also classify forest health as well. 806 00:40:55,010 --> 00:41:00,010 So the resulting map is at the right side, 807 00:41:00,670 --> 00:41:04,770 and the workflow to be able to get from, 808 00:41:04,770 --> 00:41:06,993 using Sentinel 2 data, 809 00:41:08,090 --> 00:41:13,030 and being able to understand what is included 810 00:41:13,030 --> 00:41:16,210 in forest versus non forest areas, 811 00:41:16,210 --> 00:41:18,950 and what the NDVI 812 00:41:18,950 --> 00:41:21,940 of these different land cover classes might be, 813 00:41:21,940 --> 00:41:25,270 as well as where communities and where roads 814 00:41:25,270 --> 00:41:26,910 and things exist. 815 00:41:26,910 --> 00:41:29,650 This is sort of modern map-making. 816 00:41:29,650 --> 00:41:32,700 And it's important to kind of understand 817 00:41:32,700 --> 00:41:34,460 that there's a lot of work 818 00:41:34,460 --> 00:41:35,980 that goes into just making these maps, 819 00:41:35,980 --> 00:41:39,250 and here, they were also, the study was also then 820 00:41:39,250 --> 00:41:44,250 looking at areas where agriculture was on the rise 821 00:41:44,750 --> 00:41:49,750 and deforestation by the agriculture was occurring. 822 00:41:56,840 --> 00:41:59,470 Okay, so we've kind of gone through 823 00:41:59,470 --> 00:42:01,860 these three different areas, 824 00:42:01,860 --> 00:42:03,810 and just to kind of go back 825 00:42:03,810 --> 00:42:07,350 to what we're doing with the readings, 826 00:42:07,350 --> 00:42:10,070 there is a chapter on forest degradation and deforestation 827 00:42:10,070 --> 00:42:11,400 in Google Earth Engine. 828 00:42:11,400 --> 00:42:13,923 And the goals of that chapter are to, 829 00:42:14,900 --> 00:42:16,850 at the end of the tutorials, 830 00:42:16,850 --> 00:42:20,930 be able to detect deforestation and forest degradation 831 00:42:20,930 --> 00:42:23,113 using spectral unmixing, 832 00:42:24,070 --> 00:42:28,727 and interpreting the fraction of images and NDFI 833 00:42:30,340 --> 00:42:33,540 using a temporal color composite, 834 00:42:33,540 --> 00:42:38,240 implementing an NDFI-based change map 835 00:42:38,240 --> 00:42:41,110 to monitor deforestation and forest degradation, 836 00:42:41,110 --> 00:42:43,570 and finally, running a time-series change detection 837 00:42:43,570 --> 00:42:47,670 to detect deforestation and forest degradation, 838 00:42:47,670 --> 00:42:49,763 and also revegetation through time. 839 00:42:53,680 --> 00:42:55,750 Concurrently to what you're gonna be doing 840 00:42:55,750 --> 00:42:58,770 in the tutorial this week, 841 00:42:58,770 --> 00:43:03,440 I also wanted to kind of highlight a website 842 00:43:03,440 --> 00:43:06,179 that's out there, that's called globalforestwatch, 843 00:43:06,179 --> 00:43:08,040 and it uses a whole bunch 844 00:43:08,040 --> 00:43:10,370 of these different types of approaches 845 00:43:10,370 --> 00:43:15,370 to detect forest cover, forest gain, and forest loss, 846 00:43:16,330 --> 00:43:20,410 as a bunch of other different variables as well, 847 00:43:20,410 --> 00:43:23,260 pertaining to forest worldwide. 848 00:43:23,260 --> 00:43:25,100 And it has a really kind of fun 849 00:43:25,100 --> 00:43:28,690 and interesting graphical user interface. 850 00:43:28,690 --> 00:43:32,010 So for the Yellowdig assignment, 851 00:43:32,010 --> 00:43:35,090 you're gonna be going to Global Forest Watch, 852 00:43:35,090 --> 00:43:38,900 and you're going to be working with the map a little bit 853 00:43:38,900 --> 00:43:43,900 to kind of figure out how to measure forest loss 854 00:43:46,000 --> 00:43:48,290 in an area that you choose. 855 00:43:48,290 --> 00:43:51,980 So just to kind of show you a couple of things 856 00:43:53,840 --> 00:43:54,940 on the webpage, 857 00:43:54,940 --> 00:43:58,220 so this is what occurs when you open up the map, 858 00:43:58,220 --> 00:43:59,480 this is what you'll see. 859 00:43:59,480 --> 00:44:02,180 You can also switch, I don't know if you can see my mouse, 860 00:44:02,180 --> 00:44:06,670 but you can switch down to the planet satellite imagery 861 00:44:06,670 --> 00:44:11,200 down on the bottom of the pop out that's on that map, 862 00:44:11,200 --> 00:44:12,500 if you wanna look at it, 863 00:44:12,500 --> 00:44:15,800 and it's sort of a satellite background, 864 00:44:15,800 --> 00:44:19,050 which is what I'll be showing in the next slide. 865 00:44:19,050 --> 00:44:22,180 So here, I've zoomed into an area of Madagascar 866 00:44:22,180 --> 00:44:24,123 that is near and dear to my heart. 867 00:44:25,646 --> 00:44:29,520 And so, I wanted to look at tree cover loss through time, 868 00:44:29,520 --> 00:44:30,660 and what you can do, 869 00:44:30,660 --> 00:44:34,590 if you see, these are the sort of baseline variables 870 00:44:34,590 --> 00:44:35,980 that are included. 871 00:44:35,980 --> 00:44:39,510 You can exclude them by pressing the X, 872 00:44:39,510 --> 00:44:41,050 or you can include others 873 00:44:41,050 --> 00:44:44,020 by clicking on forest change or land cover, 874 00:44:44,020 --> 00:44:46,670 clicking on some of those different icons over there 875 00:44:46,670 --> 00:44:48,350 at the far left. 876 00:44:48,350 --> 00:44:50,730 There's also climate and biodiversity as well. 877 00:44:50,730 --> 00:44:53,480 So you can include different variables 878 00:44:53,480 --> 00:44:55,030 and you can exclude other ones. 879 00:44:55,030 --> 00:44:57,140 And if you look at the cover loss, 880 00:44:57,140 --> 00:44:58,440 there's a little start arrow, 881 00:44:58,440 --> 00:45:01,070 and you can actually play a little movie for yourself 882 00:45:01,070 --> 00:45:05,150 as to what cover loss has looked like 883 00:45:05,150 --> 00:45:09,053 in the region that you choose between 2001 and 2021. 884 00:45:10,050 --> 00:45:13,220 So here, I also wanted to kind of show you, 885 00:45:13,220 --> 00:45:15,320 down at the left side, you can see, 886 00:45:15,320 --> 00:45:17,910 there's a popup that I clicked on, 887 00:45:17,910 --> 00:45:20,680 so you can look at, it says image type, 888 00:45:20,680 --> 00:45:22,340 and if you click on it, 889 00:45:22,340 --> 00:45:24,520 right now, it's natural color, 890 00:45:24,520 --> 00:45:26,030 but you can also click on it 891 00:45:26,030 --> 00:45:28,640 to show a false color, near infrared 892 00:45:29,480 --> 00:45:31,850 of the satellite background. 893 00:45:31,850 --> 00:45:34,560 So that's what that looks like if you do that. 894 00:45:34,560 --> 00:45:38,253 And then, okay, so going back to just the natural color. 895 00:45:39,300 --> 00:45:42,230 If you pop out the forest change, 896 00:45:42,230 --> 00:45:45,260 you can see that a bunch of the other variables 897 00:45:45,260 --> 00:45:47,820 will come up, things that you can click on. 898 00:45:47,820 --> 00:45:49,330 And when you toggle them, 899 00:45:49,330 --> 00:45:51,270 when you click on the grayed-out areas 900 00:45:51,270 --> 00:45:52,360 and turn them on and off, 901 00:45:52,360 --> 00:45:54,420 that's when they'll get included in your map 902 00:45:54,420 --> 00:45:55,880 and sort of overlay it. 903 00:45:55,880 --> 00:45:57,880 You can also play with the opacity 904 00:45:57,880 --> 00:46:00,240 of the variable that you choose 905 00:46:00,240 --> 00:46:02,860 if you wanna sort of dial it up or dial it back. 906 00:46:02,860 --> 00:46:05,313 Right now, it's sort of all set at the defaults. 907 00:46:07,200 --> 00:46:12,200 So here, if you see, I removed forest gain, 908 00:46:12,670 --> 00:46:15,973 'cause I just wanted to focus on cover and loss. 909 00:46:16,860 --> 00:46:20,960 And another interesting thing that you can do, 910 00:46:20,960 --> 00:46:23,000 so if you turn on another data set, 911 00:46:23,000 --> 00:46:28,000 so here I turned on tree cover loss by Dominant Driver, 912 00:46:28,540 --> 00:46:31,700 which has a color-coordinated key, 913 00:46:31,700 --> 00:46:33,080 and you'll notice that my map 914 00:46:33,080 --> 00:46:35,370 suddenly became very pixelated, 915 00:46:35,370 --> 00:46:38,520 and that's because not all of the data sets 916 00:46:38,520 --> 00:46:42,280 are calculated using the 30-meter Landsat 917 00:46:42,280 --> 00:46:43,630 or Sentinel data sets, 918 00:46:43,630 --> 00:46:45,950 some of them are calculated using other data sets. 919 00:46:45,950 --> 00:46:49,010 So this is a tree cover loss by Dominant Driver dataset. 920 00:46:49,010 --> 00:46:50,940 And if you look at the metadata 921 00:46:50,940 --> 00:46:53,970 down where the variables listed, 922 00:46:53,970 --> 00:46:56,560 you can see that it's got a 10-meter resolution. 923 00:46:56,560 --> 00:46:59,900 So that's why, it suddenly looks very pixelated on my map. 924 00:46:59,900 --> 00:47:01,250 There's also some other ones in there 925 00:47:01,250 --> 00:47:03,750 that are 500-meter Motus, 926 00:47:03,750 --> 00:47:06,850 375-meter Veers, things like that. 927 00:47:06,850 --> 00:47:09,440 So just be aware as you're clicking around and playing 928 00:47:09,440 --> 00:47:11,990 that that's why things might get pixelated for you. 929 00:47:14,360 --> 00:47:16,700 So finally, if you go back, 930 00:47:16,700 --> 00:47:20,470 I took us back to just looking at what I was interested in, 931 00:47:20,470 --> 00:47:23,970 in looking at, which was tree cover loss in the area. 932 00:47:23,970 --> 00:47:28,680 If you click on the analysis tab at the top, next to legend, 933 00:47:28,680 --> 00:47:30,960 what you'll be able to do is, 934 00:47:30,960 --> 00:47:32,923 it'll ask you if you wanted to upload, 935 00:47:34,690 --> 00:47:36,100 it'll ask you if you wanna upload a shape file 936 00:47:36,100 --> 00:47:38,230 or you can actually draw one interactively. 937 00:47:38,230 --> 00:47:39,320 So I drew a box, 938 00:47:39,320 --> 00:47:41,970 you can sort of see the black box around the region. 939 00:47:41,970 --> 00:47:43,660 And then you can click 'analyze,' 940 00:47:43,660 --> 00:47:47,480 and it will calculate and produce a map for you 941 00:47:47,480 --> 00:47:52,480 of what the forest loss looks like between 2001 and 2021. 942 00:47:57,430 --> 00:48:00,640 So that's sort of along the lines 943 00:48:00,640 --> 00:48:04,730 of what you will be doing for your Yellowdig assignment. 944 00:48:04,730 --> 00:48:06,660 There's some questions that go along with that, 945 00:48:06,660 --> 00:48:08,060 so just pay attention 946 00:48:08,060 --> 00:48:11,510 and look forward to reading and interacting 947 00:48:11,510 --> 00:48:14,920 with your discussions of sort of why you chose things 948 00:48:14,920 --> 00:48:16,080 and what you're seeing there, 949 00:48:16,080 --> 00:48:18,033 and how accurate you think it might be. 950 00:48:19,060 --> 00:48:21,830 All right, well, thank you for your attention, 951 00:48:21,830 --> 00:48:24,500 and let me know if you have any questions, 952 00:48:24,500 --> 00:48:26,563 and we'll see you next week.