Downloaded 16 times
Assessing Reef Health Using a Low Altitude Sensing Platform
This document discusses using low-altitude aerial surveys with unmanned aerial vehicles to assess coral reef health in Kaneohe Bay, Hawaii. The goals are to develop a cheap, fast, and spatially accurate technique to survey entire reef areas at a high resolution. Test flights with a DJI Phantom 2 drone captured imagery over 1500 square meters that was processed into orthomosaics and used texture/color models to classify and map coral species, bleaching, and health. This could create a reef atlas for Kaneohe Bay, inform management, and increase understanding of reef dynamics by providing long-term monitoring at various spatial scales.
More Related Content
Similar to Assessing Reef Health Using a Low Altitude Sensing Platform
More from Hawaii Geographic Information Coordinating Council
Assessing Reef Health Using a Low Altitude Sensing Platform
- 1. Assessing reef healthusing a low-altitude remote sensing platform Joshua Levy UH Biology 10/15/15
- 2. Coral Reefs • Biodiversityhotspots • Economically important: Fisheries, tourism, coastal protection, ecosystem services – Global • $29.8 billion/year(Cesar et al .2003) – Hawaii • $360 million/year (Cesar et al. 2002)
- 3. The Problem • Coralreefs are in global decline – 27% already substantially degraded – Loss of additional 31% by 2030 (Wilkinson, 2000) • 3rd global coral bleaching event – “Bleachapalooza”
- 4.
- 5. Causes of CoralBleaching – High temperature • Thermal anomalies – El Niño – PDO – Anthropogenic climate change…? – Coastal runoff • Sedimentation • Freshwater • Nutrients Clark Little
- 6. Effects of coralbleaching • Reduced reef resilience • Ecosystem resilience “The capacity of an ecosystem to absorb disturbances or shocks, and adapt to change while retaining function and structure”. McClanahan et al. 2012
- 7.
- 8. Hawaiian Reefs • 85%of US coral reefs • Huge economic importance • Decrease in reef resilience after bleaching could impact Hawaiian economy
- 9. Hawaii coral bleachingevents • 2014 – First recorded mass bleaching event on Hawaiian Islands – Extensive bleaching in MHI’s, NWHI’s • Moderate-high recovery • Lisianksi mortality – 1.5 miles • 2015 – Recorded throughout MHI’s, NWHI’s • Kahe (first recorded bleaching) • Kaneohe Bay* • 22 other Oahu sites
- 10. How can coralsrecover? • “Normal” SST • Good water quality • High herbivore biomass
- 11. How can wehelp corals recover? • Good water quality – Reduce pollutants/ coastal runoff • High herbivore biomass – Reduce fishing Protect affected areas – Where are they?
- 12. Reef Survey Methodologies •Small scale – In situ (cm-mm) • Snorkel/scuba • Large scale (km-m) – High Altitude Remote Sensing • Satellite /Airplane • Incomplete spatial assessments
- 13. Goals • Develop cheap,fast, spatially accurate, survey technique – High resolution data – Entire large reef area (16,000m2) – Coral species – Coral bleaching – Coral health at various spatial scales
- 14. UAV’s • Unmanned AerialVehicles – Cheap (~$2000) – Portable (4kg) – Low altitude flight (1-400m) • Large resolution range (m-mm’s) – 15-20min flights – Programmable autonomy
- 15. Study Site: KaneoheBay • 50+ patch reefs • History of pollutants/coastal runoff – Cultivation and agriculture – Stream modification – Dredging – Sewage outfalls
- 16. Bleaching in KBay •2014 – Extensive bleaching – Reports of high recovery • 2015 – Bleaching extent? – Recovery? – Management?
- 17. Low Altitude AerialSurveys • Ground station • Datalink package • iPad mini • External battery • DJI P2 • H3-3D Gimbal • GoPro 3
- 18.
- 19. Post Processing • Orthomosaics –Agisoft Photoscan ~1500m2
- 20. Structure from Motion •Estimate 3-D structures from multiple overlapping 2-D image sequences – Identify POI’s using SIFT • Edges, gradients • Insensitive to varying orientation/ illumination • # depends on image quality, resolution, heterogeneity of scene
- 21. Post Processing • ImageClassification – Texture + color models – Distinguish • Coral species • Coral vs. sand vs. algae • Quantify color change • Spatial analysis using ArcGIS • Geo-reference • Spatial measurements Caras & Karnieli. Ground-Level Classification of a Coral Reef Using a Hyperspectral Camera. Remote Sens. 7, 7521–7544 (2015).
- 22.
- 23. Predicted Outcomes • K-Baypatch reef atlas – Highlight severe bleaching – Monitor recovery • Inform Kaneohe Bay reef management groups
- 24. Management tool • Longterm monitoring of reef condition – Predict mortality events • Locate and better manage reefs impacted by local anthropogenic stress • Increase reef resilience to coral stressors
- 25. Future Projects • Expandto other shallow water coastal habitats • Automated workflow – Cloud processing – Machine-learning image classification models • Neural network
- 26. Citizen science • Affordable,user- friendly, low altitude remote sensing package for coastal communities – Hawaii • Coastal communities depend on healthy reefs • Near real-time reef assessment • Local (self) enforcement of dynamic, situational management policies
- 27. Scientific advancement • Citizenscience data collection – Generates large quantity of accurate, unbiased data from various locations – Global reef database – Various temporal and spatial scales Increase overall understanding of coastal reef system dynamics
- 28. Acknowledgments C. Hunter, UHMBiology E. Hochberg, BIOS E. Franklin, UHM/HIMB R. Bidigare, UHM SOEST R. O’Connor, NOAA/UHM WRRC DAR UH MOP NOAA #NA10NM4520163
Editor's Notes
- #3 As some of you may know, coral reefs are known as the rainforests of the ocean. They are an extremely biodiverse, highly productive marine ecosystem that have a large economic impact on coastal communities.
- #5 Although many of you may have heard the term floating around recently due to recent events, let me explain exactly what we mean by coral bleaching.
- #7 Ecosystem reselience is defined as..
- #8 Phase shift usually results in a permanent loss of coral cover These photos are both from Kbay patch reefs. Showing an example of a phase shift from a healthy reef system to one overgrown by invasive algae
- #9 So now we know what coral bleaching is and how it effects coral reef health, but lets take a look at the reefs literally in some of our back yards.
- #11 Some of these things are out of our control (cant pour enough ice cubes into the ocean to cool it off), but others are not..
- #12 Streamline resource allocation to specific areas where time and effort put in will have the greatest effect
- #13 As of now, we only have the ability to look at reef systems at extremely small or large scale. There is no moderate scale that can help us connect information collected by the two extreme methods.
- #15 As many of you are aware, unmanned aerial vehicles/ UAV’s, UAS’s etc, are becoming a huge asset in the field of low altitude mapping, and terrestrial conservation. The benefits of these platforms is the ability to acquire high resolution images that can help answer the small-scale questions, while still looking at the bigger picture.
- #16 K-Bay has a long history of abuse from anthropogenic manipulations ranging from mass cultivation, agriculture, and stream modifications in the 1940s, to direct sewage outfalls that pumped sewage into the bay for over 30 years. All of which have caused significant impacts on water quality and coral health throughout the years.
- #17 However, after the relative improvement in water quality, recent surveys show there is a heterogeneous distribution of coral between, and even within patch reefs. There are various lines of thought as to why this is the case, including proximity to shore and therefore vulnerability to acute stressor events, and water flow throughout Kaneohe Bay dictating areas of good and bad water quality. But as of now, no one has been able to conduct patch reef assessments at the spatial scale needed to answer this question.
- #18 P2 with 3-D gimbal, GoPro, extra accessory: Data link connects GS app on ipad to the copter Note simplicity and portability of the set up. Everything fits inside of the Pelican Case (seen in the picture above)
- #19 These are images collected on Reefs 42, 44 in Kaneohe Bay. Two top are raw, no editing whatsoever. Note detail, ability to discern coral spp, color differences, orange float is a ground control point used to collect GPS coordinates for geo referencing the complete orthomosaic Bottom images were taken recently on reef 44. these images have been corrected for fish eye distortion, and minor contrast adjustments. Bleaching is very apparent
- #20 Solves camera orientation
- #21 60% overlap Scale invariant feature transformations
- #22 After creating the orthomosaics, we need to differentiate between coral species, and between coral and other benthic substrate (sand, algae). I also need to quantify color change of each colony throughout the course of a year. Although I have not yet attempted it, I plan to use packages in ENVI for this analysis.
- #24 This work should result in an increased understanding of the dynamic between environmental stressors and patch reef condition in Kaneohe Bay, and maybe prove that the relationship between coral condition and local environment occurs at a smaller scale than previously thought. Hopefully these results and new survey methodology will provide Kaneohe Bay management groups with novel tools and information to better tackle the issue of coral reef management
- #26 As technology advances, I would like to….
- #29 For their continued advice and support throughout this project WATER RESOURCES RESEARCH CENTER