class: center, middle, inverse, title-slide .title[ # Mapping global mining activities using
open Earth observation data .erc-logo-left[] .fineprint-logo-left[] .wu-logo-right[] ] .subtitle[ ##
] .author[ ### Dr Victor Maus
.font80[Institute for Ecological Economics
Vienna University of Economics and Business]
.font80[
victor.maus@wu.ac.at
]
.font80[
vwmaus.github.io
] ] .date[ ### .right[.font80[UNODC and UNEP, Apr 19, 2023, Vienna].footnote-right[Except where otherwise noted, content on these slides is licensed under
CC BY-SA 4.0 license
]] ] --- layout: false class: clear, center background-image: url(./../img/kayapo_gold_mining-Ibama-via-Wikimedia-CC-BY-SA-2-0.jpg) .cc-bottom-right.font-light[Illegal gold mine in the Kayapo indigenous territory in Brazil by IBAMA via Wikimedia Commons <a href="https://creativecommons.org/licenses/by-sa/2.0/deed.en">CC BY-SA 2.0</a>] .font-dark.center.font250.font-light[**Currently, there is not enough geospatial information available to conduct comprehensive sustainability assessments for mining**] ??? - A link to the slides will be provided at the end of the talk, so you don't need to bother taking notes - Companies and governments do not disclose sufficient geospatial information - This hinders comprehensive sustainability and risk assessments of the sectors activities, particularly at global scale --- layout: false class: clear, center background-image: url(https://www.esa.int/var/esa/storage/images/applications/observing_the_earth/copernicus/sentinel-2/19729272-10-eng-GB/Sentinel-2_pillars.jpg) background-size: cover .cc-bottom-left.font-light[Image: © <a href="https://www.esa.int/Applications/Observing_the_Earth/Copernicus/Sentinel-2">European Space Agency (ESA)</a>] .center.font150.font-light[**Earth observation is the only viable data source for monitoring mining areas**] ??? - Earth observation satellites are the only viable data source to close this information gap - And the open data policy by several space agencies makes it easy to access and use satellite images for a range of applications, including mapping mines --- layout: false class: clear, middle, center .center.font180[**With little training, humans can identify mining in satellite images**] .pull-left.center[ .bg-washed-blue.b--dark-blue.ba.bw0.br3.shadow-5.mh1.mt0.font100[ **Open cut of the Carajás Iron Mine, Brazil** ] <div class="copyright-container"> <img src="./../img/carajas-mine-satellite.png" height="450"/> <div class="copyright-bottom-left2 font-light">Figure: <a href="https://www.fineprint.global/viewer">FINEPRINT Viewer</a></div> </div> ] .pull-right.center[ .bg-washed-blue.b--dark-blue.ba.bw0.br3.shadow-5.mh1.mt0.font100[ **Gold mining in Kayapo indigenous territory, Brazil** ] <div class="copyright-container"> <img src="./../img/mining-Kayapo-indigenous-territory-brazil.png" height="450"/> <div class="copyright-bottom-left2 font-light">Figure: <a href="https://www.fineprint.global/viewer">FINEPRINT Viewer</a></div> </div> ] ??? - My research team has been exploring this rich open satellite data archives to map mines - As currently there are no automated algorithms that can map mines globally - We have mapped small-scale and industrial mines via visual interpretation of satellite images - Automated global scale mapping mines is an open scientific challenge --- layout: false class: clear, middle, center background-image: url(./../img/mine-app-gui.png) background-size: contain .footnote-right[.font-light[[User interface of the open-source mine mapping tool](https://doi.org/10.5281/zenodo.3580739)]] ??? - We then develop a tool to support systematic visual interpretation of mining regions - The tool brings several layers together to support experts interpretation of images --- layout: false class: clear, middle, center background-image: url(./../img/mine-app-data-sources.png) background-size: contain .footnote-right[.font-light[[User interface of the open-source mine mapping tool](https://doi.org/10.5281/zenodo.3580739)]] ??? - We then develop a tool to support systematic visual interpretation of mining regions - The tool brings several layers together to support experts interpretation of images --- layout: false class: clear, middle, center background-image: url(./../img/global-mining-map.png) .footnote-right[.font-light[Figure: [FINEPRINT Viewer www.fineprint.global/viewer](https://www.fineprint.global/viewer)]]] .footnote-left[[Maus et al. (2020, 2022)](http://doi.org/10.1038/s41597-022-01547-4)] .left-column.pl1.mt0.font120.left.bottom.font-light[ <br><br><br><br><br><br><br><br><br> **45,000 Polygons** **100,000 km²** ] ??? - As a result we constructed a database including about 45,000 polygons covering about 100,000 km² of mines globally - In this map you can see the spatial distribution of the polygons - This database untapped a range of global land use and ecological assessments --- layout: false class: clear, middle, center .center.font120[**The database includes industrial mining as well as small-scale artisanal mining globally**] <div class="copyright-container"> <img src="https://media.springernature.com/full/springer-static/image/art%3A10.1038%2Fs41597-022-01547-4/MediaObjects/41597_2022_1547_Fig1_HTML.png?as=webp" height="500"/> <div class="copyright-bottom-left font-dark"><a href=http://doi.org/10.1038/s41597-022-01547-4>Maus et al., 2022</a></div> </div> ??? - In this slide we see the examples from South America, in Chile and Brazil - And it also may include illegal mining --- layout: false class: clear, middle, center .center.font180[**Mining land use distribution across countries**] <div class="copyright-container"> <img src="https://media.springernature.com/full/springer-static/image/art%3A10.1038%2Fs41597-022-01547-4/MediaObjects/41597_2022_1547_Fig3_HTML.png?as=webp" height="500"/> <div class="copyright-bottom-left font-dark"><a href=http://doi.org/10.1038/s41597-022-01547-4>Maus et al., 2022</a></div> </div> ??? - The polygons are distributed across 145 countries - And the area is unevenly distributed, 6 countries account for more than 50% of the mapped mining area - This includes, Russia, China, Australia, United States, Indonesia, and Brazil - It is important to note that, the database covers LSM and SAM globally --- layout: false class: clear, middle, center .center.font180[**Mining in indigenous land in Brazil**] .pull-left.center[ <div class="copyright-container"> <img src="./../img/illegal-mining-munduruku.png" height="450"/> <div class="copyright-bottom-left font-dark">Map by Matthias Humer</div> </div> .bg-washed-blue.b--dark-blue.ba.bw0.br3.shadow-5.mh1.mt0.font100[ **Munduruku indigenous territory, Brazil** ] ] .pull-right.center[ <div class="copyright-container"> <img src="./../img/illegal-mining-kayapo.png" height="450"/> <div class="copyright-bottom-left font-dark">Map by Matthias Humer</div> </div> .bg-washed-blue.b--dark-blue.ba.bw0.br3.shadow-5.mh1.mt0.font100[ **Kayapó indigenous territory, Brazil** ] ] ??? - As we can see the mining areas in red in this slide are within protected areas, outside the consecrations in orange - Although, we have not check this areas with Brazilian authorities, they are likely illegal activities - My team is currently conducting a more comprehensive study to check how much mining is outside concession - But this is still challenging for several countries as we do not always have open information about concessions - Our database has also served as bassis for other studies, for example --- layout: false class: clear, middle, center .center.font180[**Mining-related forest loss by country accumulated from 2001 to 2019**] <div class="copyright-container"> <img src="./../img/top-10-contry-mrd.png" height="500"/> <div class="copyright-bottom-left font-dark"><a href=https://www.wwf.de/fileadmin/fm-wwf/Publikationen-PDF/Wald/WWF-Studie-Extracted-Forests.pdf>Kramer et al., 2023</a></div> </div> ??? - Based on this database, we have published last week a new study together with WWF Germany - Where we have quantified the mining related deforestation globally - Some main highlights of this study are shown in this map - Showing the countries contribution to mining related deforestation from 2001 to 2019 - In this study we have also linked deforestation to individual commodities --- layout: false class: clear, middle, center .center.font180[**Mining-related forest loss by commodity, 2001-2019**] <div class="copyright-container"> <img src="./../img/mineral-commodity-deforestation-wwf.png" height="500"/> <div class="copyright-bottom-left font-dark"><a href=https://www.wwf.de/fileadmin/fm-wwf/Publikationen-PDF/Wald/WWF-Studie-Extracted-Forests.pdf>Kramer et al., 2023</a></div> </div> ??? <!-- --- --> <!-- layout: false --> <!-- class: clear --> <!-- .font-light.font200.center.bd[Monitoring mining requires] --> <!-- .middle.center[ --> <!-- <img src="./../img/example-cnn-coal-mine-classification.png" height="300"/> --> <!-- <br>.font180[**Deep neural networks show promising results in delineating coal mining open pit with overall accuracy between 85% and 90%**] --> <!-- ] --> <!-- .footnote-left[Maus et al. (Ongoing)] --> <!-- ??? --> <!-- - Automatizing mining monitoring shows promising results using a class of algorithms called, Convolution Neural Networks --> --- layout: false class: clear, center, middle background-image: url(./../img/towards-automated-mining-monitoring.png) background-size: contain .font.dark.font120[**Towards monitoring legal/illegal mining**<br><br><br><br>] ??? - Sentinel-2 images provide sufficient spatial and temporal resolution for near real time monitoring, but supporting information with higher spatial resolution is critical - Current automated approaches work well for homogeneous regions and mining types but not for large scales applications - Cloud cover in tropical regions can substantially delay the detection - Local expert knowledge is critical - Open Earth observation images (Sentinel-2) provide sufficient spatial and temporal resolution for near real time monitoring - However, higher spatial resolution is very useful, e.g. openly available for the tropics from Norway's International Climate and Forests Initiative Satellite Data Program - How can we transition from snapshots of mining land use to monitoring mining activities over time? - Our global mining database was very useful to calculate some ecological footprints - However, can only provide snapshots of land use, as it is very time-consuming - Monitoring mining activities requires automation <!-- <div class="copyright-container"> --> <!-- <video height="320" autoplay muted loop> --> <!-- <source src="https://assets.planet.com/web/videos/nicfi/nicfi_tropics_basemap_map_large_video_d2-1.mp4" type="video/mp4"> --> <!-- </video> --> <!-- </div> --> <!-- <div class="copyright-bottom-left2 font-light">Video: <a href="https://www.planet.com/nicfi">NICFI https://www.planet.com/nicfi</a></div> --> --- layout: false class: clear, hide-count, middle, left .erc-logo-left[] .fineprint-logo-left[] .wu-logo-right[] <br><br><br><br><br> .pull-left.left[ Dr Victor Maus<br>.font100[Institute for Ecological Economics<br>Vienna University of Economics and Business]<br>.font100[victor.maus@wu.ac.at]<br>.font100[<a href='https://vwmaus.github.io'>vwmaus.github.io</a>]<br> <img src="./../img/qrcode_www.victor-maus.com.png" height="190"/> ] .right[ .font80[UNODC and UNEP, Apr 19, 2023, Vienna]<br> .font150[Thank you!] ]