Remote Sensing of Vegetation Properties

Size: px

Start display at page:

Download "Remote Sensing of Vegetation Properties"

Augustine Byrd
5 years ago
Views:

1 Remote Sensing of Vegetation Properties K. Tansey, H. Balzter, S. Johnson, Paul Arellano and many others Department of Geography University of Leicester

2 Research at the University of Leicester Strong track record in Space and EO Science Home to the National Centre for Earth Observations Office for EMBRACE G-STEP (early catapult model) New Research Institute (LISEO) & National Space Centre Enjoy the services and support of a NERC KEF Centre for Landscape and Climate Research Funding from EC, ESA, NERC, InnovateUK Interested in exploring Agri-Tech opportunities

3 Earth Observation to See Vegetation Earth observation methods provide a high spectral resolution information about the canopy and vegetation biochemical and biophysical characteristics.

Hydrocarbons - Natural Geological Process Bacteria in soil consume (oxidize) the hydrocarbons Increase CO2 in soil Decrease Oxygen concentration Low ph values

4 Hydrocarbons - Natural Geological Process Bacteria in soil consume (oxidize) the hydrocarbons Increase CO2 in soil Decrease Oxygen concentration Low ph values Mineral are dissolved and mobilized then increased their concentrations A generalised map of hydrocarbon microseepage and its surface expressions (from Yang, 1999: p. 8)

5 Soil-gas sampling (Geochemistry) Vegetation sampling: chlorophyll and other pigments content, structure, phenology. Reflectance measurements in field Airborne and satellite images Vegetation stress detection

6 Deepest, Darkest Ecuador Oil spills in the Amazon forest

7 Fieldwork: Sampling Parameters at leaf level: Chlorophyll a and b Water content Dry matter content Leaf internal structure Reflectance (Visible and near Infrared) Transmittance (Visible and near infrared) Spectrophoto meter Chlorophyll meter

8 Fieldwork campaign in the Amazon Vegetation sampling Vegetation sampling in the vertical profile of the forest Biophysical and biochemical parameters extraction

9 Biophysical and biochemical alterations of the forest Results show lower levels of chlorophyll content across the vertical profile of the fores increase foliar water content

Canopy Models approach R i = f (A i, B i, C i, D i, E i ) Ground {D i } Sensor {E i } Canopy {C i } Source {A i } Atmosphere {B i } A i : Spectral intensity, Wavelength (λ), location angles (θ, ψ) B

C i : Optical parameters (reflectance and transmittance) Pigments contents (chlorophyll a and b, etc.

10 Canopy Models approach R i = f (A i, B i, C i, D i, E i ) Ground {D i } Sensor {E i } Canopy {C i } Source {A i } Atmosphere {B i } A i : Spectral intensity, Wavelength (λ), location angles (θ, ψ) B i : Wavelength (λ) absorption and scattering properties of aerosol particles, water vapour and ozone). C i : Optical parameters (reflectance and transmittance) Pigments contents (chlorophyll a and b, etc.) Structural parameters (geometrical shapes and positions) of vegetation components (leaves, stems, etc.), LAI. D i : Reflectance and absorption, roughness, texture, density, moisture E i : Spatial and spectral resolution, calibration parameters, location angle. In order to detect stress of vegetation affected by pollution the Inverse canopy model technique can be applied: C i = f (A i, B i, R i, D i, E i )

11 Method Scaling-up model 3D canopy model (Flight or DART) Leaf Model (PROSPECT) Chlorophyll Water content Organic matter Internal structure Tree high DBH LAI Crow shape and size Leaf size Canopy level Leaf level

12 EO-1 Hyperion hyperspectral image preprocessing RAW RADIANCE CORRECTED REFLECTANCE

Image processing: Vegetation indices Analysis of variance and pairwise comparison - Holm adjustment method Discriminant function analysis INDEX Polluted forest vs.

13 Image processing: Vegetation indices Analysis of variance and pairwise comparison - Holm adjustment method Discriminant function analysis INDEX Polluted forest vs. Secondary Pristine forest forest Secondary forest vs. Pristine forest BROAD-BAND VEGETATION INDICES 1 Simple Ratio SR *** *** *** 2 Normalized Difference Vegetation NDVI *** *** ** 3 Green Normalized Difference Vegetation Index GNDVI *** *** ** 4 Atmospherically Resistant Vegetation Index ARVI ns *** *** 5 Enhanced Vegetation Index EVI ** *** * NARROW-BAND VEGETATION INDICES: Greenness / Chlrorophyll / REP 6 Sum Green SG *** *** ns 7 Pigment Specific Simple Ratio-Chla PSSRa *** *** *** 8 Red Edge Normalized Difference Index NDVI 705 *** *** *** 9 Modified Red Edge Simple Ratio msr 705 ns *** *** 10 Modified Red Edge Normalized Difference Index mndvi 705 ns *** *** 11 Carter Index 2 CRT2 *** *** *** 12 Lichtenthaler Index 1/Pigment Specific Normalized Difference LIC1/PSNDa *** *** ** 13 Optimized Soil-Adjusted Vegetation Index OSAVI *** *** * 14 Modified Chlorophyll Absorption Ratio Index MCARI ns ns ns 15 Ratio of derivatives at 725 and 702 nm Der ns *** *** 16 Red Edge Position REP ** ns ** 17 Vogelmann Red Edge Index VOG1 *** *** *** 18 Chlorophyll Index CI 590 * *** *** 19 MERIS Terrestrial Chlorophyll Index MTCI *** *** *** NARROW-BAND VEGETATION INDICES: Other Pigments 20 Structure Insensitive Pigment Index SIPI * *** *** 21 Red Green Ratio RG ns ns ** 22 Anthocyanin Reflectance Index 1 ARI1 * ** *** 23 Anthocyanin Reflectance Index 2 ARI2 ns *** *** NARROW-BAND VEGETATION INDICES: Water Indices 24 Water Band Index WBI ns *** *** 25 Normalized Difference Water Index NDWI. *** * 26 Moisture Stress Index MSI * *** ns 27 Normalized Difference Infrared Index NDII ns *** *** 28 Normalized Heading Index NHI ns ns ns *** Strongly significant (0.1%) ** Higly significant (1%) * Significant(5%). Lowest significant (10%) ns No significant VI LD1 relative contribution SG 53.0% NDVI 21.6% NDVI % CTR2 4.6% GNDVI 4.6% LIC1 3.5% VOG1 2.4% OSAVI 0.9% MTCI 0.3% PSSRa 0.3% SR 0.2% 75%

14 Chlorophyll content at canopy level MTCI and chlorophyll content (μg cm -2 ) at canopy level estimated in the study sites

15 Using Radars to Estimate Biomass

16 Monitoring and Managing Global Forest Resources

17 Invasive Species Mapping

18 Crops monitoring using high temporal resolution satellites Sentinel-1 and 2 & Landsat + others VENUS satellite image (France and Israel) 2 day revisit time 10 m resolution (3 m resolution) Super-spectral camera

19 Moving it Forward Very interested in Gathering and Understanding User Requirements Scale, Region, Application Track record of working with SMEs through InnovateUK or other business engagement models Amazing level of support available: EMBRACE -> Catapult NERC Knowledge Exchange Fellow (Sarah) Business development managers in Energy and Environment (Maggy) and Space (Sarah) EC H2020 Bid managers

March 2017 DATA-DRIVEN INSIGHTS FOR VINEYARDS

March 2017 DATA-DRIVEN INSIGHTS FOR VINEYARDS What do great wine, water on mars and drones have in common? Today: Drone Technologies in Viticulture AGENDA Technology Context: big data, precision ag, drones