background-color: #006DAE class: middle center hide-slide-number <div class="shade_black" style="width:60%;right:0;bottom:0;padding:10px;border: dashed 4px white;margin: auto;"> <i class="fas fa-exclamation-circle"></i> These slides are viewed best by Chrome and occasionally need to be refreshed if elements did not load properly. </div> <br> .white[Press the **right arrow** to progress to the next slide!] --- background-image: url(images/Werombi_Bushfire.jpg) background-size: cover class: hide-slide-number split-70 title-slide count: false .column.shade_black[.content[ <br> # .monash-blue.outline-text[Using Remote Sensing Data to Understand Fire Ignition during the 2019-2020 Australia Bushfire Season] <h2 class="monash-blue2 outline-text" style="font-size: 30pt!important;"></h2> <br> <h2 style="font-weight:900!important;"></h2> .bottom_abs.width100[ *Weihao Li, Chang L., **Di Cook**, Emily Dodwell*
<i class="fas fa-link faa-float animated "></i>
[https://bit.ly/VicBushFireIgnition](https://bit.ly/VicBushFireIgnition) ACEMS Retreat, Nov 6 2020 <br> ] ]] <div class="column transition monash-m-new delay-1s" style="clip-path:url(#swipe__clip-path);"> <div class="background-image" style="background-image:url('images/large.png');background-position: center;background-size:cover;margin-left:3px;"> <svg class="clip-svg absolute"> <defs> <clipPath id="swipe__clip-path" clipPathUnits="objectBoundingBox"> <polygon points="0.5745 0, 0.5 0.33, 0.42 0, 0 0, 0 1, 0.27 1, 0.27 0.59, 0.37 1, 0.634 1, 0.736 0.59, 0.736 1, 1 1, 1 0, 0.5745 0" /> </clipPath> </defs> </svg> </div> </div> --- # People and motivation <table> <tr> <td> <img src="images/weihao.jpg" style="width: 200px; border-radius: 30%"> </td> <td width="100px"> </td> <td> <img src="images/emily.jpg" style="width: 180px; border-radius: 50%"> </td> </tr> <tr> <td> Weihao Li </td> <td> </td> <td> Emily Dodwell </td> </tr> <tr> <td> Monash EBS Honours </td> <td> </td> <td> AT&T research </td> </tr> </table> <br> *Motivation*: Spatio-temporal visualisation and analysis of emergency call data. This is private so the bushfire data was collected because it has some similar form and structure. --- # πΊ Background <h2>β‘ Lightning or π₯Arson?</h2> <img src="images/1602783588.png" width = "50%" style = "float: left"/> <img src="images/bushfire-inforgraphic-not-normal-768x768.jpg" width = "50%" style = "float: right"/> <!-- https://twitter.com/MRobertsQLD/status/1220588928706568193 --> --- # π‘ Remote sensing data Japan Aerospace Exploration Agency provides a hotspot product (reflected energy from the earth) taken from the **Himawari-8** satellite. .footnote[[Example code to access data provided in a G. Williamson gist post](https://gist.github.com/ozjimbob/80254988922140fec4c06e3a43d069a6)] <img src="images/hotspots_before.png" style="width: 80%; float:center"/> --- # π Data Sources .monash-red2[**π₯ Historical fire origins**]: 2000-2019 .font_my_2[[Department of Environment, Land, Water and Planning](https://discover.data.vic.gov.au/dataset/fire-origins-current-and-historical)] .monash-red2[**π‘ Remote sensing data**]: .font_my_2[[Japan Aerospace Exploration Agency](https://www.eorc.jaxa.jp/ptree/userguide.html)] .font_my[ **Wind speed data**: 1-day, 7-day, ..., 2-year averages from .font_my_2[[Commonwealth Scientific and Industrial Research Organisation and Automated Surface Observing System](https://doi.org/10.25919/5c5106acbcb02)] **Temperature, Rainfall and Solar exposure**: 1-day, 7-day , 14-day, 28-day, ..., 720-day averages computed from .font_my_2[[Bureau of Meteorology](https://CRAN.R-project.org/package=bomrang)] **Fuel layer**: Forest type, forest height class, forest crown cover from .font_my_2[[Australian Bureau of Agricultural and Resource Economics](https://www.agriculture.gov.au/abares/forestsaustralia/forest-data-maps-and-tools/spatial-data/forest-cover)] **Road map**: Proximity to the nearest road using .font_my_2[[OpenStreetMap](%20https://www.openstreetmap.org%20)] **Fire stations**: Proximity to the nearest CFA station .font_my_2[[Department of Environment, Land, Water and Planning](https://discover.data.vic.gov.au/dataset/cfa-fire-station-vmfeat-geomark_point)] **Recreation sites**: Proximity to the nearest camping site .font_my_2[[Department of Environment, Land, Water and Planning](https://discover.data.vic.gov.au/dataset/recreation-sites)] ] --- # π» Data fusion <img src="images/data_fusion.png" style="width: 100%;"/> --- # π» Detect ignitions by clustering hotspot data <img src="images/clustering.png" style="width: 100%"/> --- # π» Estimated ignitions 76,000 hotspots reduced to 1,000 ignition sites. <img src="images/hotspots_after.png" style="width: 100%; float:left"/> <!-- <img src="images/hotspots_before_summary.png" style="width: 50%; float:right"/> --> --- # π Exploratory analysis of historical fire origins .font_my_2[ Text processing of 26 causes, reduced to four major causes. Lightning and accident were the two main sources of historical bushfire ignitions, which took up 41% and 34% respectively. There were 17% bushfires caused by arson. ] <img src="images/ignition_summary.png" style="width: 50%; float:left"/> <img src="images/ignition_year.png" style="width: 50%; float:right"/> --- # π Spatial distribution of historical fire origins .font_my_2[ Roughly different spatial locations of ignition causes. Lightning bushfires were concentrated in the east of Victoria. Bushfires caused by arson were near Bendigo! ] <img src="images/density.png" style="width: 100%; float:left"/> --- # π Proximity of historical fire origins .font_my_2[ Lightning-caused bushfires were further away from the CFA stations and roads. In contrast, bushfires caused by arson were closer to CFA stations and roads. ] <img src="images/density_cfa.png" style="width: 100%; float:left"/> --- # π Modelling .monash-blue[A **random forest** model outperformed other model choices to classify different causes of bushfire ignition. ] .monash-blue[.font_my_2[80% of the data used as training set, 7497 observations, and the remaining 1872 observations was used as test set.]] <br> .font_my_2[ Model performance was compared using multi-class AUC (Hand and Till, 2001).] <table class=" lightable-classic table" style="font-family: Cambria; width: auto !important; margin-left: auto; margin-right: auto; font-size: 20px; margin-left: auto; margin-right: auto;"> <thead> <tr> <th style="text-align:left;"> Model </th> <th style="text-align:right;"> Accuracy </th> <th style="text-align:right;"> Muti-class AUC </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> Multinomial logistic regression </td> <td style="text-align:right;"> 0.53 </td> <td style="text-align:right;"> 0.74 </td> </tr> <tr> <td style="text-align:left;"> GAM multinomial logistic regression </td> <td style="text-align:right;"> 0.68 </td> <td style="text-align:right;"> 0.82 </td> </tr> <tr> <td style="text-align:left;"> Random forest </td> <td style="text-align:right;"> 0.75 </td> <td style="text-align:right;"> 0.88 </td> </tr> <tr> <td style="text-align:left;"> XGBoost </td> <td style="text-align:right;"> 0.74 </td> <td style="text-align:right;"> 0.88 </td> </tr> </tbody> </table> --- # π Model performance .font_my[ .monash-blue[The overall accuracy of our model was 74.95%.] - High accuracy with lightning and accident ignitions. - Less accurate predictions for arson and burning off. ] <br> <table class=" lightable-classic table" style="font-family: Cambria; width: auto !important; margin-left: auto; margin-right: auto; font-size: 20px; margin-left: auto; margin-right: auto;"> <thead> <tr> <th style="text-align:left;"> </th> <th style="text-align:left;"> Lightning </th> <th style="text-align:left;"> Accident </th> <th style="text-align:left;"> Arson </th> <th style="text-align:left;"> Burning_off </th> <th style="text-align:right;"> Total </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> Prediction:Lightning </td> <td style="text-align:left;"> 703 (0.9) </td> <td style="text-align:left;"> 77 (0.12) </td> <td style="text-align:left;"> 50 (0.15) </td> <td style="text-align:left;"> 44 (0.32) </td> <td style="text-align:right;"> 874 </td> </tr> <tr> <td style="text-align:left;"> Prediction:Accident </td> <td style="text-align:left;"> 51 (0.07) </td> <td style="text-align:left;"> 494 (0.78) </td> <td style="text-align:left;"> 89 (0.27) </td> <td style="text-align:left;"> 38 (0.28) </td> <td style="text-align:right;"> 672 </td> </tr> <tr> <td style="text-align:left;"> Prediction:Arson </td> <td style="text-align:left;"> 18 (0.02) </td> <td style="text-align:left;"> 55 (0.09) </td> <td style="text-align:left;"> 175 (0.54) </td> <td style="text-align:left;"> 22 (0.16) </td> <td style="text-align:right;"> 270 </td> </tr> <tr> <td style="text-align:left;"> Prediction:Burning_off </td> <td style="text-align:left;"> 5 (0.01) </td> <td style="text-align:left;"> 8 (0.01) </td> <td style="text-align:left;"> 11 (0.03) </td> <td style="text-align:left;"> 32 (0.24) </td> <td style="text-align:right;"> 56 </td> </tr> <tr> <td style="text-align:left;"> Total </td> <td style="text-align:left;"> 777 </td> <td style="text-align:left;"> 634 </td> <td style="text-align:left;"> 325 </td> <td style="text-align:left;"> 136 </td> <td style="text-align:right;"> 1872 </td> </tr> </tbody> </table> --- # π Model interpretation .font_my_2[ Variable importance assessed using [Local Interpretable Model-agnostic Explanations (lime)](https://lime.data-imaginist.com). Proximity to the nearest road, proximity to the nearest road and average wind speed had largest influence on the prediction. ] <img src="images/varimp.png" style="width: 100%; float:left"/> --- # π **Prediction for 2019-2020 Australia bushfires**
--- # π Summary of findings .font_my_2[ .monash-blue[- Majority of the bushfires in 2019-2020 season were caused by **lightning**.] - 138 bushfires caused by accidents which took up 14% of the total fires. Most of them were ignited in March. - 37 bushfires were caused by arsonists, and over half of them were in March. - Very few planned burns were predicted after October 2019 which suggests the correctness of our model. ] <br> <table class=" lightable-classic table" style="font-family: Cambria; width: auto !important; margin-left: auto; margin-right: auto; font-size: 20px; margin-left: auto; margin-right: auto;"> <thead> <tr> <th style="text-align:left;"> Cause </th> <th style="text-align:right;"> Oct </th> <th style="text-align:right;"> Nov </th> <th style="text-align:right;"> Dec </th> <th style="text-align:right;"> Jan </th> <th style="text-align:right;"> Feb </th> <th style="text-align:right;"> Mar </th> <th style="text-align:left;"> Total </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> Lightning </td> <td style="text-align:right;"> 19 </td> <td style="text-align:right;"> 57 </td> <td style="text-align:right;"> 315 </td> <td style="text-align:right;"> 266 </td> <td style="text-align:right;"> 32 </td> <td style="text-align:right;"> 149 </td> <td style="text-align:left;"> 838 (0.82) </td> </tr> <tr> <td style="text-align:left;"> Accident </td> <td style="text-align:right;"> 3 </td> <td style="text-align:right;"> 8 </td> <td style="text-align:right;"> 34 </td> <td style="text-align:right;"> 13 </td> <td style="text-align:right;"> 0 </td> <td style="text-align:right;"> 80 </td> <td style="text-align:left;"> 138 (0.14) </td> </tr> <tr> <td style="text-align:left;"> Arson </td> <td style="text-align:right;"> 2 </td> <td style="text-align:right;"> 2 </td> <td style="text-align:right;"> 10 </td> <td style="text-align:right;"> 2 </td> <td style="text-align:right;"> 0 </td> <td style="text-align:right;"> 21 </td> <td style="text-align:left;"> 37 (0.04) </td> </tr> <tr> <td style="text-align:left;"> Burning_off </td> <td style="text-align:right;"> 7 </td> <td style="text-align:right;"> 0 </td> <td style="text-align:right;"> 2 </td> <td style="text-align:right;"> 0 </td> <td style="text-align:right;"> 0 </td> <td style="text-align:right;"> 0 </td> <td style="text-align:left;"> 9 (0.01) </td> </tr> </tbody> </table> --- # π Policy implications .font_my[ <!-- - Our prediction may help CFA and FFMV to improve fuel management in remote area. --> - CFA and FFMV may need to consider fire prevention solutions in remote areas of Victoria to reduce the risk of lightning-ignited bushfire. Perhaps new strategically located CFA stations might improve accessibility and speed of response. - CFA should be asked about the increasing number of accident-caused bushfires. Could it be a policy/coding change. Are there substantially more careless campers, or logging operations? - Arson is a small source of fires, and primarily a problem closer to populated areas. Possibly sensor network might be recommended. ] --- # Shiny app: https://ebsmonash.shinyapps.io/VICfire/ <iframe src="https://ebsmonash.shinyapps.io/VICfire/?showcase=0" width="110%" height="550px"></iframe> --- # π Summary 1. Algorithm to detect bushfire ignition from hotspot data 2. Model to predict the cause of bushfire ignition 3. Prediction of the causes of the 2019-2020 Australia bushfires 4. A complete and adaptable workflow for monitoring and understanding new ignitions from hotspot data 5. Shiny app for exploration of historical fire origins, predicted causes of 2019-2020 fires and future fire risk maps 6. All work conducted with open data and open source software. --- background-image: url(images/Werombi_Bushfire.jpg) background-size: cover class: hide-slide-number split-70 count: false .column.shade_black[.content[ <br><br> ## Acknowledgements Slides produced using [Rmarkdown](https://github.com/rstudio/rmarkdown) with [xaringan](https://github.com/yihui/xaringan) styling. Monash style by the kunoichi, Dr Emi Tanaka. Data and code for analysis is available on [Weihao's GitHub repo](https://github.com/TengMCing/bushfire-paper). Shiny app code is available at [Chang's GitHub repo](https://github.com/timtam3/Bushfire/tree/master/VICfire). # Thanks for listening! <br> <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/"><img alt="Creative Commons License" style="border-width:0" src="https://i.creativecommons.org/l/by-sa/4.0/88x31.png" /></a><br />This work is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/">Creative Commons Attribution-ShareAlike 4.0 International License</a>. ]] <div class="column transition monash-m-new delay-1s" style="clip-path:url(#swipe__clip-path);"> <div class="background-image" style="background-image:url('images/large.png');background-position: center;background-size:cover;margin-left:3px;"> <svg class="clip-svg absolute"> <defs> <clipPath id="swipe__clip-path" clipPathUnits="objectBoundingBox"> <polygon points="0.5745 0, 0.5 0.33, 0.42 0, 0 0, 0 1, 0.27 1, 0.27 0.59, 0.37 1, 0.634 1, 0.736 0.59, 0.736 1, 1 1, 1 0, 0.5745 0" /> </clipPath> </defs> </svg> </div> </div>