networking friday remote sensing applications in ocean ...€¦ · the use of hyperspectral data:...
TRANSCRIPT
Networking Friday
Remote Sensing Applications in Ocean Monitoring, Corals and Mangrove Studies
MILTON KAMPEL, [email protected]
Networking Friday, São José dos Campos, May 15th, 2020
What brings me here today?• Much of the research effort undertaken in recent years has focused on the use of Earth
Observation data associated with the development and validation of (ocean) remote sensing products for various applications.
• The basic idea here is to comment on some EO/RS activities of INPE´s MOceanS Lab., tocomment on some relevant aspects in this field, evaluate the limitations that still exist, tofurther expand international partnerships, collaborations and personnel training at aninternational level helping AIR Centre network stakeholders to stay active, connected,informed and inspired.
2 AIR Centre Networking Friday, May 15th, 2020
Outline• Who we are? What we do?
• Remote Sensing of Coral Reefs/Benthic Mapping.• Metocean Monitoring.
• Land-Ocean Interactions.
• Remote Sensing of Mangroves.
MissionTo develop, operate and use space systems for the advancement of science, technology and applications in the areas of outer space and Earth environment, and offer innovative products and services for the benefit of Brazil.
Institutional Presentation
3 AIR Centre Networking Friday, May 15th, 2020
http://www.inpe.br/institucional/sobre_inpe/instalacoes.php#sede35
Space & Atmospheric Science Weather & Climate
Space Technology Earth System Sciences
Satellite Tracking & Control
Integration & Testing Lab.Associated Laboratories
Earth Observation
http://www.inpe.br/institucional/pesquisa_desenvolvimento/
http://www.inpe.br/institucional/sobre_inpe/missao.php
Staff 1,000 people + ~600 students + [post-docs + other fellowships]
4 AIR Centre Networking Friday, May 15th, 2020
5 AIR Centre Networking Friday, May 15th, 2020
MUX (20 m)WFI (64 m)
PAN (5 m)
CBERS-4
MUX (20 m)
Cameras MUX WPM WFI
Band 1 0.45 - 0.52 µm 0.45 - 0.52 µm 0.45 - 0.52 µm
Band 2 0.52 - 0.59 µm 0.52 - 0.59 µm 0.52 - 0.59 µm
Band 3 0.63 - 0.69 µm 0.63 - 0.69 µm 0.63 - 0.69 µm
Band 4 0.77 - 0.89 µm 0.77 - 0.89 µm 0.77 - 0.89 µm
Band 5 (PAN) 0.45 - 0.90 µm
Resolution 16 m 2 m, 8 m 55 m
CBERS-04A
6 AIR Centre Networking Friday, May 15th, 2020
Sun synchronous orbit
Altitude = 628 km
Inclination = 97.89º
Revisiting rate = 31 days
Descending node at 10h30 local time
Launch: 20th Dec. 2019
Cameras MUX WPM WFI
Manufacturer Brazil China Brazil
Type Push broom Push broom TDI Push broom
Revisiting rate 31 days 31 days 5 days
Quantization 8 bits 10 bits 10 bits
Swath 95 km 92 km 684 km
7 AIR Centre Networking Friday, May 15th, 2020
CBERS 04A
WFI (55 m)MUX (16 m)
WPM (2 m) WPM (2 m)
8 AIR Centre Networking Friday, May 15th, 2020
01-MAR-2018
Amazonia 1First MMP missionForest monitoringLaunch: 2020 (tbc)
Amazonia 1BMMP missionSABIA-Mar (Oceans)MapSARLaunch: tbd
CubeSats
9 AIR Centre Networking Friday, May 15th, 2020
Center for Weather Forecast and Climate Studies
10 AIR Centre Networking Friday, May 15th, 2020
Earth Observation
11
It involves scientific and technological knowledge in the fields of remote sensing and geoinformatics, survey of natural resources and monitoring of the environment. It conducts research, development and applications in the fields of Remote Sensing and Digital Image Processing (83 employees: 80% PhD and 15% MSc).
GEOINFORMATICS
SATELLITE MISSIONS&
DATA CENTER
REMOTE SENSING
PROGRAMS
UN SDGs
EDUCATION & DIFFUSION
AIR Centre Networking Friday, May 15th, 2020
NE Regional Center (CRN)
S Regional Center (CRS)
Amazonia Regional Center (CRA)
Education at INPE
12 AIR Centre Networking Friday, May 15th, 2020
INPE’s Environmental Monitoring Systems: Land Cover, Deforestation Alerts and Fire
Courtesy: Maurano, LE
13 AIR Centre Networking Friday, May 15th, 2020
2019 2020
14 AIR Centre Networking Friday, May 15th, 2020
BLUE AMAZON
EEZ
Continental
Shelf
EEZ +
Continental
Shelf
67% Brazilian
TerritoryAnother
Amazone
2,100,000 km2
5,700,000 km2
5,700,000 km2
Would such an approach like this possible or reasonable?
15 AIR Centre Networking Friday, May 15th, 2020
Satellite Remote Sensing• Collecting and interpreting
information about the environment and the surface of the Earth from a distance, primarily by sensing radiation that is naturally emitted or reflected by the Earth’s surface or from the atmosphere, or by sensing signals transmitted from a device and reflected back to it.
• Examples of remote-sensing methods include aerial photography, radar, and satellite imaging.
16 AIR Centre Networking Friday, May 15th, 2020
Corals/Reefs Mangroves Aquaculture Algae blooms Climate Health
Marine SpatialMapping
Bays & Estuaries
Wind PowerAir-Sea Ocean ColourFisheries Inland waters Sat Missions
Oil & Gas
Land-Ocean
Electromagnetic Spectrum
18 AIR Centre Networking Friday, May 15th, 2020
Robinson 2010
Methods and Types of Sensors
Coral Reefs Studies• Coral reefs are the most biodiverse
and productive ecosystems in marine environments.
• Human activity and natural climate trends constitute a major threat to coral reefs worldwide.
• Models predict a significant reduction in reef spatial extension together with a decline in biodiversity in the relatively near future.
• In this context, monitoring programs to detect changes in reef ecosystems are essential.
19 AIR Centre Networking Friday, May 15th, 2020
https://www.nature.org/en-us/what-we-do/our-priorities/protect-water-and-land/land-and-water-stories/8-easy-ways-you-can-help-coral-reefs/
Coral Reefs Studies by RS
20 AIR Centre Networking Friday, May 15th, 2020
https://www.geospatialworld.net/news/satellite-mount-lidar-sensor-helps-researchers-develop-new-understanding-planktons/
https://eos.com/sentinel-2/Photos: MOceanS Lab.
• Remote sensing approaches to acquiring data in coral reef ecosystems are the most cost-effective and allow for synoptic monitoring of large areas, including places with difficult access.
• In recent years, coral reef mapping using remote sensing data has benefited from instruments with better resolution and computational advances in storage and processing capabilities.
RS Data Sources
21 AIR Centre Networking Friday, May 15th, 2020
Passive Sensors
Active Sensors
Purkins & Roelfsema, 2015
RS in the Visible
22 AIR Centre Networking Friday, May 15th, 2020
Good water penetration capacity
HyperspectralSensors
MultiespectralSensors
Bottom types
Bathymetry
Chlorophyll
Suspended Matter
Water Quality
Several studies
Ground truth
Visible Spectrum
Some Potentialities and Limitations• Bottom reflectance (ρb) is the
central parameter in the remote sensing of coral reefs and, depends on the physical structure and chemical substrate composition.
• Although RS has a great potential in studies of the sea bottom, extracting the reflectance spectrum from the data of orbital optical sensors is complex.
• Several processes affect the satellite signals.
23 AIR Centre Networking Friday, May 15th, 2020
Phinn et al. 2012
Physical Processes in Water Column
a. Absorption coefficients spectra (m-1) measured in a productive oceanic environment (1 mg m-3 chl-a): Total absorption (black), TSM absorption (red), water molecules (blue), phytoplankton (green), and CDOM (yellow).
b. Absorption spectra of phytoplankton (green) and TSM (red) are plotted, for a better visualization.
24 AIR Centre Networking Friday, May 15th, 2020
Kirk 2011
Light Penetration in the Water Column
Light decay modeled along water column expressed as percentage of incident light as function of depth (m).
a. Curves represent different (nm) in an environment considered as Case-1 water, where [chl-a] = 0.01 mg m-3.
b. Light at 400 nm but in different environments: Case-1 waters (chl-a=0.1 mg m-3); French Polynesia Case-1 waters (Kd=0.14 m-1); Case-2 waters in Abrolhos Coral Reef Bank, Brazil (Kd=0.18 m-1); Case-1 waters (chl-a=1 mg m-3); Case-2 waters (chl-a=0.5 mg m-3, Kd=0.3m-1, minerals concentration=0.5 g m-3).
25 AIR Centre Networking Friday, May 15th, 2020
Zoffoli et al. 2014
RS of Submerged Substrates
26 AIR Centre Networking Friday, May 15th, 2020
a. Reflectance of different substrates at 0 m depth, as function of (nm): coral sand (blue), brown algae (red) and green algae (green).
(b-d). Reflectance above water vs. (nm) simulated for the same substrates in clear Case-1 waters (chl-a = 0.05 mg m-3) at different depths: 1, 5, 20 m.
(e-g). Reflectance above water vs. (nm) simulated for the same substrates in Case-2 waters (chl-a: 1 mg m-3; aCDOM (440) = 0.3 m-1; minerals: 1 g m-3) at depths: 1, 5, 20 m.
Zoffoli et al. 2014
Multispectral & Hyperspectral Sensors(a) Reflectance spectra using a very high-
resolution spectral spectrometer –Hyperspectral.
(b) The same thing with a Multispectral spectrometer.
27 AIR Centre Networking Friday, May 15th, 2020
Phinn et al. 2012
Examples of spectral signatures of the same stretch of live coral
1. In situ reflectance from field spectrometry; 2. Simulated surface reflectance with 1.0 m of water column; 3. Reflectance obtained with the CASI-2 hyperspectral sensor; 4. Reflectance (x 10,000) obtained from the QuickBird-2 multispectral orbital sensor.
28 AIR Centre Networking Friday, May 15th, 2020
Phinn et al. 2012
Multispectral sensorsWith different spatial and temporal resolutions
29 AIR Centre Networking Friday, May 15th, 2020
Hedley et al. (2016)
Spatial vs. Temporal Resolution
Consider the type of application of interest.
30 AIR Centre Networking Friday, May 15th, 2020
Hedley et al 2016
Multispectral Application
(a, b) Geomorphological zone; (c, d) mapsof benthic classification. (b, d) show theenlarged area.
The authors applied currentprocessing algorithms toSentinel-2 data from variouslocations on the Great BarrierReef and performed directcomparisons with Landsat -8.
(a, b) RGB composition from originalimages using bands 2, 3 and 4; (c, d)bathymetric map; (e, f) RGB composition inbands 1, 2 and 3.
32 AIR Centre Networking Friday, May 15th, 2020
High resolution (~m) benthic habitat andbathymetric maps derived from WorldView-2satellite images and calibrated with in situobservations.
Study area distributed in an equatorialtransect covering 65,000 km2 of habitatsdominated by coral reefs in 11 countries.
Maps of the reefsof Gizo Island,Solomon Islands.(a) Location ofGizo Island inNew Georgia.b. True-colorWorldView-2 ofthe reefs.c. Bathymetricmap derivedfrom theWorldView-2image calibratedwith in situ data.d. Map of benthichabitat.
Multispectral Application
33 AIR Centre Networking Friday, May 15th, 2020
Multispectral Application
34 AIR Centre Networking Friday, May 15th, 2020
Coral Reefs (2020) submitted
Abrolhos Coral Reef Bank (Brazil) mapping using high
spatial resolution WorldView-2 satellite imagery
María Laura Zoffoli1,2*, Milton Kampel1, Robert Frouin3, Thais Andrade Galvão de
Medeiros1
The authors assessed the benefits of using dataacquired by the WV-2 satellite for producing thefirst bottom type map in the ACRB MarineNational Park and for characterizing landscapeecology of reefs distribution.
What is the problem?Resolving the Complexity of Coastal Waters
35 XIX SBSR, Santos, April 2019
• Extensive studies using in-itumeasurements and remote sensing imaging have shown that visible hyperspectral imaging is one of the best available tools to resolve the complexity of the coastal ocean from space.
• There is a need for high spatial (and temporal) resolution in the near costal ocean.
Hyperspectral Sensors
36 XIX SBSR, Santos, April 2019
The use of hyperspectral data:• Allows to derive a greater number of classes with higher accuracy using classification
techniques.• Allows to simultaneously separate information on depth, bottom type and water quality
parameters.
Example of resampled coral andmacroalgae reflectance data for:
(a) Blue, green and red LandsatTM bands.
(b) Typical CASI bandconfiguration, with 10 bands.
The dashed lines show high-resolution reflectance spectra of acoral (Porites sp.) and algae(Caulerpa sp.).
Gray bars showing locations andwidths of the bands and solid linesare resampled spectra according toinstrument's bands.
Hedley, 2012
Comparison between (a) the Heron Reefpseudo RGB (b) the HOPE-LUT classificationand (c) the BRUCE classification map. RegionsA and B in (a) highlight the inner lagoon andthe reef, respectively.
Benthic map with higher classification accuracy using hyperspectral remote sensing inversion models in hyperspectral aerial images of Heron Reef, Great Barrier
Accuracy of the five depth classifiers for (a) seagrass and algae; (b) coral classes.
Hyperspectral Application
37 XIX SBSR, Santos, April 2019
Hyperspectral Application
38 XIX SBSR, Santos, April 2019
Evolution of the coral reef habitat classesbetween 2009 (1) and 2015 (2) for threezones (A, B and C) of the Saint-Gilles coralreef unit. Left, habitat map for the entireSaint-Gilles flat reef unit in 2015.
Approach based on the analysis ofhigh resolution multispectral andhyperspectral images.
Detect and quantify changes inbenthic cover on a highlyheterogeneous reef in thesouthwest Indian Ocean.
Thermal Infrared
• Passive detection of radiationemitted by the source object.
• Allows to derive the surface seatemperature (SST) from radiometricobservations in two bands located at~4 µm (night) and 10–13 µm (dayand night).
• Measures the skin temperature ofthe sea surface (~10 µm) - SSTskin.
• Daily global SST maps + temporalcomposites.
39 AIR Centre Networking Friday, May 15th, 2020
Schematic profile of thetemperature of the oceanclose to the surface,showing the skin layerand the depths andmeasurements relevantfor satellite SSTmonitoring.
(https://oceanservice.noaa.gov/facts/sea-surface-temperature.html)
Heron et al. 2012
TIR Applications
40 AIR Centre Networking Friday, May 15th, 2020
Ecological processes associated with a possibleinteraction between warming oceantemperatures and water quality
Dissolved inorganic nitrogen (DIN) stimulatesthe increase in coral symbiont populations byincreasing metabolic demands during periods ofhigh irradiance and temperature.
High levels of DIN make coral-algae symbiosisless stable leading to bleaching
(a) Historical thermal index based on the typical summermaximum sea surface temperature (TSM). (b) Water qualityindex based on terrestrial runoff rich in (DIN). (c) Thermalstress index based on the maximum SST occurring in anythree-day period. (d) Expected areas of low resistance to coralbleaching (in red).
Scott A. Wooldridge, Terence J. Done
TIR Applications
41 AIR Centre Networking Friday, May 15th, 2020
NOAA Coral Reef Observation program (CRW)has developed a global 5 km daily product suitebased on satellite observations to monitorthermal stress on coral reefs.
SST
SST Anomaly
Bleaching Hot Spots
Degree HeatingWeek (DHW)
Daily Global 5km Satellite Coral
Bleaching Heat Stress Alert Area
(7-day max)
MW - RADAR
42 AIR Centre Networking Friday, May 15th, 2020
• Detection of sea surface roughness → measurement of the backscatter of the signalactively transmitted by the remote sensor.
• Operation during day and night; not much sensitive to cloud cover.
• Backscatter parameters → indicate ocean state → transmission time and return ofradar pulse frequency change, phase difference and polarity change.
• Monitoring of the environmental conditions around coral reefs.
High Frequency (HF) and Very High Frequency (VHF)
Ground-wave Radar
Heron et al. 2012
Synthetic Aperture Radar (SAR)
HF RADAR• HF radars → remotely measure
surface currents, exploring Bragg'sresonant backscatter phenomenon.
• Networks of HF radar systems →map surface currents every hour inranges of ~ 200 km (possibly more),with horizontal resolution of ~ km.
a) Detection and monitoring of oilspills on the sea surface;
b) Evaluation of the connectivity oflarvae populations in near-realtime.
43 AIR Centre Networking Friday, May 15th, 2020
Doppler spectrum of an ocean surface current radar system (OSCR) operating at 25.4 MHz.
HF-RADAR Application
44 AIR Centre Networking Friday, May 15th, 2020
✓ Assessment of connectivity betweenMarine Protected Areas (MAPs) alongthe central California coast.
✓ Network of HF Radar stations forocean currents mapping.
✓ The authors computed 40-dayreversed trajectories during 1 yearbetween potential regions of larvalorigin and destinations withinestablished MAPs.
Maps of the hourly back-projections of waterparticles, out through 40 days (960 hours) in thepast, for each MPA region; repeated daily for each 1-km grid-point in the given MPA with forcing fromsurface currents measured by HF-radar during 2008These water particle track-points are color coded perthe legend in panel #10 to show the travel time (upthrough 40 days) the waters took to reach the MPAs.
Synthetic Aperture Radar (SAR)
45 AIR Centre Networking Friday, May 15th, 2020
✓Emit pulses of electromagnetic radiationat wavelengths of ~cm and detectbackscattered radiation from the Earth'ssurface.
✓Register the phase and amplitude ofreturned signal.
✓Oceanic phenomena that affect seasurface roughness on a small scale canproduce a detectable signal in a SARimage.
✓As e.g., wind, waves, oil slicks andseeps, other surfactants, upwelling,shallow-water bathymetry
✓With this set of parameters and relatively highspatial resolution, the use of SAR data can beuseful in monitoring environmental conditionsaround coral reefs.
Chaturvedi, 2019
Heron et al. 2012
SAR Application
46 AIR Centre Networking Friday, May 15th, 2020
Analysis of the potential of SAR images(RADARSAT-1) to detect coral reefsexposed in the Environmental ProtectionArea of Costa dos Corais, Brazil.
Differences in processes such as radarwaves interact with the surface of the seaand with emerging reefs.
1. Smooth sea surfaces are specularreflectors - little energy is returned tothe sensor - dark tones in the images.
2. Coral reefs have a more diffusescattering - greater amounts of energyreturn to the sensor - intermediateshades of gray in the images.
A) Raw RADARSAT-1 image; B) Filtered SAR image (3x3 Kuan filter)
SAR Application
47 AIR Centre Networking Friday, May 15th, 2020
Corals spawn at night, usuallyin calm environmentalconditions. These conditionscan potentially be predictedwith relative reliability.
(a) SAR features on 25/OCT/1998 and outlines ofthe features on 16/APR/1998. (b) Multibeambathymetry of the area showing the submarinechannels underlying the features in the SAR image.
Features observed in SAR image of 16/APR/1998on: (a) Vulcan and Goeree Shoals; (b) BarracoutaShoal; (c) Demarcation of the feature on VulcanShoal and isobathimetric contours obtainedfrom a nautical chart.
SAR features observed oncarbonate reefs in the TimorSea were interpreted ascaused by a coral spawningevent.
LIDAR (Light Detection and Ranging)
48 AIR Centre Networking Friday, May 15th, 2020
✓Emits laser beams at different wavelengths. Common inthe near infrared band. Sensors used for bathymetryoperate at two frequencies.
✓Installed on board manned and unmanned platforms.
✓Allows the generation of 3D-Digital Elevation Models.
✓Provides useful information to improve our knowledgeof the functional relationships betweengeomorphological structure and ecological processes inthe marine environment.
✓Provides support for coastal research and mappingefforts.
✓More specifically, it contributes to the understanding ofthe three-dimensional geomorphology of coral reefs NASA Experimental Advanced Airborne Research Lidar (EAARL).
(Brock et al. 2004)
Purkis and Brock, 2012
LIDAR Application
49 AIR Centre Networking Friday, May 15th, 2020
High-resolution Scanning Hydrographic Operational AirborneLidar Survey (SHOALS) → Used to define the morphology ofspur-and-groove structures on the fringing reef off the southcoast of Molokai, Hawaii.
Example of bathymetric data derived fromSHOALS. (A) Visualization on a shaded reliefmap of the SHOALS bathymetry. (B) Exampleof a bathymetric profile parallel to the coastalong the 10 m isobath.
Morphology of the reef and the island platform south of Molokai fromSHOALS bathymetric data and from the National Ocean Serviceoverlaid with the location of 36 transects used in the analysis.
LIDAR Application
50 AIR Centre Networking Friday, May 15th, 2020
NASA Experimental Advanced Airborne Research Lidar(EAARL)→ Used to measure the topography of shallow reefsubstrates, on a sub-metric scale.
Map depicting color-codedspot NAVD88 referencedelevations over Alina Reef(highlighted spots populatethe LIDAR raster transectselected to demonstrate theanalysis).
Wire mesh plot depicting a digitalelevation model created for AlinaReef (site e584207_n2807950_17)based upon the NASA EAARLsurvey that was conducted on 5August 2002. The viewing angle isoblique and from the southwest
LIDAR Application
51 AIR Centre Networking Friday, May 15th, 2020
Digital (a) surface, (b) intensity (532 nm wavelength),and (c) coral reef state classification derived frombathymetric LiDAR soundings (40,364 × 34,588 pixelsat 0.5 m pixel size).
A VHR airborne bathymetric LIDAR system was used tocompute the coral reef ecosystem´s surface andreflectance in combination with a multispectral cameramounted on a drone that produced a BGRorthorectified.A coral reef ecological map was generated for the firsttime at submeter scale in the lagoon of Moorea Island,French Polynesia.Five ecological states were classified through an ANNcalibrated with 275 samples to determine the class ofcoral state with satisfactory accuracy.
Spatio-temporal characterization of satellite-derived surface fields in the Abrolhos Coral Reef Bank
52 AIR Centre Networking Friday, May 15th, 2020
Sea Surface Temperature
53 AIR Centre Networking Friday, May 15th, 2020
SST
54 AIR Centre Networking Friday, May 15th, 2020
CoroaVermelha
Parcel dasParedes
Parcel deAbrolhos
Raw data AnomaliesDecadal average
Timbebas
Chlorophyll-a Concentration
55 AIR Centre Networking Friday, May 15th, 2020
CHL
56 AIR Centre Networking Friday, May 15th, 2020
CoroaVermelha
Parcel dasParedes
Parcel deAbrolhos
Raw data AnomaliesDecadal average
Timbebas
Surface Ocean Wind Vectors
57 AIR Centre Networking Friday, May 15th, 2020
OWV
58 AIR Centre Networking Friday, May 15th, 2020
Mixed Layer Depth
59 AIR Centre Networking Friday, May 15th, 2020
Euphotic Depth
60 AIR Centre Networking Friday, May 15th, 2020
MLD vs. ZEU Dynamics
61 AIR Centre Networking Friday, May 15th, 2020
Net Primary Production
62 AIR Centre Networking Friday, May 15th, 2020
Environmental Characterizations
63 AIR Centre Networking Friday, May 15th, 2020
SST - Campos CHL - AMBES ZEU - Santos MLD - SEAL
OWV - SEAL PPO - Santos
64 AIR Centre Networking Friday, May 15th, 2020
Chlorophyll-a
Ocean Wind Vectors
Precipitation
Operational MetOcean MonitoringOriginal Algorthm 2.5C Adjusted Algorthm 0.9C
Algorithms Cal/Val
Frontal Systems
Wind PPT
Sea Surface Temperature
Environmental monitoring –Bays and Estuaries
• Eutrophication index• Environmental studies• Surface circulation
NPP/VIIRS – 20/APR/2019 AVHRR/MetOp – 5/JUN/2019
ABI/GOES16– 2/FEB/2019 ASCAT/MetOpA– 2/FEB/2019 • Water mass indicator• Surface circulation• Water renewal
•Accumulated precipitation in watersheds
• Synoptic monitoring• Cold fronts passages• Atm. instabilities• Navigation
AIR Centre Networking Friday, May 15th, 202065
66
VIIRS– 12/AUG/2018SENTINEL 1 – 12/AUG/2018
Synergy –MultiSensors
Monitoring of coastal plumes and efluents
Sentinel2 – Landsat8 – CBERS4
Land Cover – Land Use, Urban Expansion, Vegetation Coverage
AIR Centre Networking Friday, May 15th, 202067
LANDSATTime series with ~300 images of Guanabara Bay
Processing showing the expansion of the urban area (in red) between 1983 (left) and 2018 (right)
RS of Mangroves
Mangroves are coastal ecosystems typical of tropical and subtropical regions subject to the tidal regime.
Considered important sources of support for the conservation of terrestrial and marine biodiversity, providing habitat, spawning grounds, nurseries and food for various animals.
68 AIR Centre Networking Friday, May 15th, 2020
Photos: Moceans Lab.
Distribution and Diversity of Mangrove Species
69 AIR Centre Networking Friday, May 15th, 2020
(World Atlas of Mangroves, Spalding et al. 2010)
Total area ~ 150,000 km² extension of mangroves
1st Indonesia 31,894 km² (20.9%)2nd Brazil with 13,000 km² (8.5%)3rd Australia with 9,910 km² (6.5%)
Mangrove Species
70 AIR Centre Networking Friday, May 15th, 2020
Laguncularia racemosa (mangue branco)
Rhizophora mangle(mangue vermelho)
Avicennia schaueriana(mangue preto)
Photos: Moceans Lab.Acrostichum aureumHibiscus pernambucensis Spartina alterniflora
Global Map of Mangroves Above Ground Biomass
71 AIR Centre Networking Friday, May 15th, 2020
The structural development of mangrove forests depends on the intensity and periodicity of numerous environmental factors.Regions with a topography subject to a wide range of tides, wide freshwater input by effluents, abundant precipitation, insolation, nutrients and sediments provide maximum structural development of mangroves.
(Hutchison et al. 2014)
AGB estimates are important for estimating mangrove productivity, determining the stock and cycling of elements in this ecosystem, forest carbon, estimating the degree of maturity, structural development and stress level.
Below Ground Biomass
72 AIR Centre Networking Friday, May 15th, 2020
Mangroves fix and store hugeamounts of carbon in the soil(49% to 98%).
Donato et al. 2011Photos: Moceans Lab.
Threatened Ecosystem
73 AIR Centre Networking Friday, May 15th, 2020
Population growth in coastal areas has led tomajor disturbances in mangroves, especiallywith deforestation and shrimp farming,which pose serious threats.One third of the Earth's mangrove forestshave already been destroyed, mainly byincreasing human population density.
Photos: Moceans Lab.
RS of Mangroves
74 AIR Centre Networking Friday, May 15th, 2020
Each type of data has advantages and disadvantages.Making it possible to obtain specific information according to the objective of each study: biomass estimation, change detection, mapping, species discrimination.
Active Systems
LIDAR SAROPTICAL IMAGERY
Passive System
Spectral Signature – VIS/NIR/SWIR
75 AIR Centre Networking Friday, May 15th, 2020
Relative Spectral Radiative Response
76 AIR Centre Networking Friday, May 15th, 2020
WorldView-2 Relative Radiative Response (nm)
Each sensor captures the energy reflected from thetarget within a wavelength range (bands).
Normalized Difference Vegetation Index
77 AIR Centre Networking Friday, May 15th, 2020
✓ Observation of the difference in reflectance of plants between the red and near infrared bands.
✓ Plant leaf pigments like chlorophyll absorb visible light (0.4-0.7μm) during photosynthesis.
✓ The cellular structures of the leaf reflect in the infrared (0.7-1.1 μm).
✓ This index can be associated with biophysical parameters of vegetation, such as biomass.
SAR System
78 AIR Centre Networking Friday, May 15th, 2020
Rosenqvist et al. (2007)
L-band Polarimetric Synthetic Aperture Radar System (PALSAR)
✓ Active system: generates its own electromagnetic energy that is transmitted from the sensor to the surface, interacting with the surface and producing a backscatter of the energy that returns to the sensor.
✓ Operates in the microwave region.✓ Data acquisition regardless of weather
conditions.
SAR System
79 AIR Centre Networking Friday, May 15th, 2020
The backscatter coefficient (σ˚) determines the amount of electromagnetic energy reflected by a resolution cell for the radar.
Types of surface-scattering associations: (a) flat-specular, (b) flat-corner reflection, (c) rough-diffuse
The intensity of energy returned tothe radar can vary depending onthe sensor and target parameters:
WavelengthPolarization (HH, VV, VH, HV)Incidence angle
Dielectric constantRoughnessGeometry of acquisition
Incident Wave Penetration Capability
80 AIR Centre Networking Friday, May 15th, 2020
The penetration of the electromagnetic wave varies dependingon the wavelength and dielectric properties of the target
Banda X: λ= 3 cm
Banda C: λ= 5 cm
Banda L: λ= 23 cm
Banda P: λ= 75 cm
Small branches and leaves
Branches, trunks and soil
SAR & Vegetation
81 AIR Centre Networking Friday, May 15th, 2020
Mechanism and behavior of forest backscatter:
A. Canopy surface backscatteringB. Interior canopy scattering, volumetricC. Canopy-soil interactionD. Double-bounce, trunk-soil interactionE. Trunk scatteringF. Soil scattering
-1st Return
-Last Return(X, Y, Z)
-2nd Return
LIDAR
82 AIR Centre Networking Friday, May 15th, 2020
✓ Light Detection and Ranging (LIDAR) is an activesensor that emits optical beams towards theground and receives the backscattered laserportion.
✓ With the synergistic use of GPS and inertialmeasurement (IMU) systems, the exact locations(x, y) and aircraft movements (roll, pitch and yaw)are documented at the time the pulse is sent andreceived.
✓ The system measures the time of the pulse pathemitted and received by the sensor.
✓ Calculates the distance covered by the returns.
✓ Making it possible to estimate the height of thevegetation.
LIDAR
83 AIR Centre Networking Friday, May 15th, 2020
It makes it possible to obtain the height of the vegetationwith high precision and thereby estimate the AGB.
LIDAR Point CloudDescriptive Metrics Digital Terrain Model
Canopy Height Model
84 AIR Centre Networking Friday, May 15th, 2020
LIDAR AGB Estimation
85 AIR Centre Networking Friday, May 15th, 2020
Auto-PLS model
AGB = 561,374.0 tonsArea of 5,432.0 hectaresAverage AGB of 103.4 t / ha.
Quantitative and qualitative information on the current biomass of the mangrove forest in the Guapimirim EPA.Carbon estimates.Contributing to decision making and formulation of public policies within the scope of coastal management and REDD+
Pereira et al. 2018
86 AIR Centre Networking Friday, May 15th, 2020
Classification of mangrove cover
PALSAR L-band image, HH HV VH VV polarization and incoherent attributes
Frequency-based contextual classification
Accuracy > 80%
87 AIR Centre Networking Friday, May 15th, 2020
Distinguish physiographic types of mangrove forests anddifferent degrees of structural development.
Physiographic mangrove types: BasinProgradation MatureGap
88 AIR Centre Networking Friday, May 15th, 2020
89 AIR Centre Networking Friday, May 15th, 2020
• Researchers: Milton Kampel, Natalia Rudorff Oliveira, Rosio Camayo
• Post-Docs: Aline M. Valerio, Francisca R. Pereira, Fabio Dall Cortivo, Amalia Detoni, Francisca Pereira, Thais Menezes
• PhD students: Philipe R. Leal, Andrea L. Oliveira, Gabriel Moiano
• MSc students: Vitor Galazzo, Raissa Teixeira, Gabriel Lucas
• BSc students: Elias R. Krausz• Former students: Maria Laura Zoffoli, Joao Felipe
C. Santos, Lucas B. Freitas, Larissa Valerio, Jean F. Silva, Gustavo Molleri, Leandro R. Freitas, FredericoRudorff
• Former Collaborators: Marcelo Franco, CaioFonteles, Catarina Cecilio, Soyla Moraes, Elaine Oliveira
• CPTEC/INPE: Luiz Augusto Machado, RogerioBatista, Mario Figueiredo, Ricardo Braga, Guilherme Andrade Monitoring Oceans from Space
90 AIR Centre Networking Friday, May 15th, 2020