The ESA  BIOMASS Mission

Carsten Pathe

About this presentation

This presentation has been compiled within the project EEBiomass, which is funded by the Federal Ministry of Economics and Technology (BMWi) with the project number 50EE1904.
This project is a collaboration between the Max-Planck Institute for Biogeochemistry, the Microwaves and Radar Institute (DLR-HR) of the German Aerospace Centre (DLR), the Friedrich Schiller University Jena and the Helmholtz-Centre for Environmental Research – UFZ Leipzig.

 

Links

Project website: https://eebiomass.org

Twitter: https://twitter.com/eebiomass                

Recommended prior knowledge

Recommended prior knowledge

What you should know before viewing this presentation:

How to use this presentation

  • This presentation is using the reveal.js presentation framework.

  • You can move through the presentation using the cursor keys in two directions:






     

  • To see an overview of all slides, press ESC

  • In the lower right corner of the screen, you can see the possible directions:

  • To jump to the table of content, click on the symbol in the lower left 

 

Chapter 1

Chapter 2

Chapter 3

Chapter x

...

Ch. 1, Slide 1

Ch. 1, Slide 2

Ch. 2, Slide 1

Ch. 2, Slide 2

Ch. 2, Slide 3

Ch. 3, Slide 1

Ch. x, Slide 1

Ch. x, Slide 2

Ch. x, Slide y

...

Content

Introduction

Climate change

Introduction

The carbon cycle is fundamental to the functioning of Earth, involving many intermeshed processes through which carbon is exchanged between the atmosphere, land, and ocean. Quantifying this global cycle is essential to understanding many of the dramatic changes taking place in the Earth system. In particular, the disturbance to the carbon cycle from the burning of fossil fuel and land-use change is the most significant driver of global change 

(IPCC, 2013)

"The unprecedented intrusion of human beings into nature has given rise to worldwide devastations. Animals and plants have been suffering on a massive scale. Newer problems keep coming into the picture on a daily basis, damaging the planet massively. Climate change has been the front runner among all the problems. It is caused by a rise in the amount of greenhouse gases, deforestation, waste generation, and population explosion" (Kumar et al., 2021, p. 1).

Therefore, we need "to improve our understanding of global change and how it will affect the Earth system and the feedbacks within the system. This is important so that societies can predict, mitigate and adapt to any likely impacts" (ESA, 2015, p. 6).

Introduction

  • Biomass is a fundamental parameter to describe the spatial distribution of carbon in the biosphere.  Therefore, biomass is a basic accounting unit for carbon (ESA, 2015).
  • Recognizing these fundamental facts, ESA’s Earth Observation Programme Board selected the proposed Biomass mission to become the 7th Earth Explorer mission on 7th of May 2013.
  • This presentation will provide you with a comprehensive overview of the upcoming ESA Biomass Mission. It starts with the basic intentions to design and build the system, will then list the mission objectives and requirements and provide some insights into the important technical details of the platform and the instrument. This will be followed by an overview of the mission products, algorithms to generate them, dissemination possibilities and some issues like RFI or ionospheric distortions.

1

1

Objectives of the BIOMASS Mission

I - Primary Objectives

II - Secondary Objectives

  1. Reduction in Large Uncertainties in Land-use Change Carbon Flux
  2. Providing Scientific Support for International Treaties and Agreements
  3. Landscape Carbon Dynamics and Prediction
  4. Initializing and Testing the Land Element of Earth System Models
  5. Forest Resources and Ecosystem Services
  6. Biodiversity and Conservation
  1. Subsurface Geology
  2. Terrain Topography under Dense Vegetation
  3. Glacier and Ice Sheet Velocities

The Biomass Mission Requirements Document defines:

(European Space Agency (2015). Biomass Mission Requirements Document. In, EOP-SM/1645. Noordwijk, The Netherlands, p. 27-29)

What is Biomass?

Given its importance for the Earth system, Biomass has been recognized as an Essential Climate Variable (ECV) by GCOS and is defined as:

A temperate forest

  • ECVs are defined by GCOS:
    https://gcos.wmo.int/en/essential-climate-variables
  • An ECV is a physical, chemical or biological variable or a group of linked variables that contributes significantly to the characterization of the Earth's climate.
  • Currently, there are 54 ECVs

"Biomass is defined as mass per unit area of live or dead plant material. Unit of measure is g/m2 or multiples" (Bombelli et al., 2009).

What is Biomass?

IPCC Good Practice Guidance for LULUCF 2003 (Penman et al., 2003) divides biomass in terrestrial ecosystems into:

All living biomass above the soil including stem, stump, branches, bark, seeds, and foliage.

All living biomass of live roots. Includes fine roots (< 2 mm diameter), small roots (2 – 10 mm diameter), and large roots (> 10 mm diameter).

Biomass in terrestrial ecosystems

Dead mass

Litter

Includes all non-living woody biomass not contained in the litter, either standing, lying on the ground, or in the soil. Dead wood includes wood lying on the surface, dead roots, and stumps larger than or equal to 10 cm in diameter and greater than 1 m in length.

Above-ground biomass

Below-ground biomass

Definitions by Bombelli et al., 2009:

Includes all non-living biomass with a diameter less than a minimum diameter chosen by a given country (for example 10 cm), lying dead, in various states of decomposition above the mineral
or organic soil.

Biomass is important because:

it is a
raw material

  • food
  • fiber
  • fuelwood

is related to vegetation structure

Biodiversity

determines rate and magnitude of autotrophic respiration

 = respiration of roots, which is another source of CO2

determines amount of carbon emitted to the atmosphere when ecosystems are disturbed

ecosystems may be disturbed due to:

  • logging
  • land conversion
  • fire
  • storms
  • pests

Biomass as dry weight is about 50% carbon.

(after Houghton et al. 2009)

Forests and biomass

Why do most people think of forests when hearing the word biomass?

  • Forest hold 70-90% of the terrestrial above-ground and below-ground biomass.
  • Above-ground biomass accounts for 70-80 % of total forest biomass (most of it in trees).

  • Above-ground biomass is vulnerable to fire, logging, land conversion (land use change), storms, pests, etc., and thus its carbon is easily released to the atmosphere.

(Houghton et al. 2009)

How can we measure biomass?

Only above-ground biomass can be measured with some accuracy at the broad scale. There are mainly four ways to measure biomass and combinations thereof (Bombelli et al., 2009):

  • In-situ destructive direct biomass measurement
  • In-situ non-destructive biomass estimations (using allometric equations or conversion factors)
  • Derive from remote sensing data
  • Vegetation Models

In-situ measurements

 

Tree age

  • Time elapsed since germination of the seed
  • Important parameter for assessing tree growth and yield
  • Can be measured by counting the tree or growth rings
    • ​Tree rings may also be called annual ring as they correspond to
    • Every year of growth, which is not completely true, as the
      formation of tree rings only occurs during the vegetation period,
      which may last four to five months depending on the latitude and

      local climatic conditions.

  • Can be counted at dead felled tree or with an increment borer  for
    living trees with which
    an increment  core is extracted from the
    tree at breast height, (Van Laar and Akça 2007).

Tree rings

Increment borer

1

1

2

2

In-situ measurements

Tree diameter

  • In forestry, the diameter of a tree stem is used as a direct measure that, together with other variables, can be used to compute cross-sectional areas, surface area and tree volume via allometric formulas.
  • Tree diameter is typically measured at breast height, abbreviated DBH (= diameter at breast height)
    • In the USA, this is a height of 4.5 feet above ground.
    • In countries using the metric system, the breast height
      is measured at 1.3 m above ground level
    • Instruments commonly used in forestry to
      measure BHD are calipers or simple measuring
      tapes

 

Beam caliper

Husch et al. 2002

In-situ measurements

Tree height

  • Most important vertical distance measure
  • Is an ambiguous term if not clearly defined
    • Total tree height h is defined as the vertical distance along the axis of the tree between the tree top and the ground
    • Bole height hb is defined as the vertical distance along the axis of the tree between the crown point and the ground
      • Crown point is the point, where the first crown‐forming branch is located.
    • Merchantable height hm is defined as the vertical distance along the axis of the tree between the terminal position of the last usable position of the stem and the ground

 

(De Young 2016)

Different measures for tree height

Remote Sensing of Biomass

optical data

hyperspectral data

LiDAR

active microwave data

LiDAR

Further reading

  • Kumar, L., Sinha, P., Taylor, S., & Alqurashi, A.F. (2015). Review of the use of remote sensing for biomass estimation to support renewable energy generation. Journal of Applied Remote Sensing, 9, 097696
  • Shi, L., & Liu, S. (2017). Methods of Estimating Forest Biomass: A Review. In J.S. Tumuluru (Ed.), Biomass volume estimation and valorization for energy (p. 516). London: IntechOpen (https://www.intechopen.com/books/5393)
  • Sinha, S., Jeganathan, C., Sharma, L.K., & Nathawat, M.S. (2015). A review of radar remote sensing for biomass estimation. International Journal of Environmental Science and Technology, 12, 1779-1792
  • Timothy, D., Onisimo, M., Cletah, S., Adelabu, S., & Tsitsi, B. (2016). Remote sensing of aboveground forest biomass: A review Tropical Ecology, 57, 125-132

Carbon Cycle & The Need for Biomass data

Atmosphere

Land

Ocean

Exchange of C as result of diffusion of CO2 across ocean surface

  • The carbon cycle is a global biogeochemical cycle, where carbon (C) is exchanged between the three major reservoirs: Atmosphere, Land, Ocean.

Carbon sequestration through photosynthesis of plants

Release of carbon through respiration of plants and soils

Carbon Cycle without anthropogenic activities

Carbon Cycle & The Need for Biomass data

Atmosphere

Exchange of C as result of diffusion of CO2 across ocean surface

Carbon sequestration through photosynthesis of plants

Release of C through respiration of plants and soils

Carbon Cycle with anthropogenic activities

Release of C through burning of fossil fuels, land use change, agricultural practices, cement production

Since the beginning of the industrial revolution around 1750, humans released large amounts of greenhouse gases (CO2, CH4, N2O) to the atmosphere.

Ocean

Land

Facts and Figures

(IPCC, 2013)

  • Since the start of the industrial revolution, the concentration of atmospheric CO2 increased by 40% from 278 ppm in 1750 to 390.5 ppm in 2011.
     
  • The concentration of other greenhouse gases increased too – during the same time interval, CH4 increased by 150% from 722 ppb to 1803 ppb, and N2O by 20% from 271 ppb to 324.2 ppb in 2011.
     
  • Current concentrations of atmospheric CO2, CH4 and N2O exceed any level measured for at least the past 800,000 years (period covered by ice cores)
     
  • Furthermore, the average rate of increase of these three gases observed over the past century exceeds any observed rate of change over the previous 20,000 years.

Facts and Figures

  • Global carbon budget averaged over the last half-century revealed that 82 % of the total emissions were caused by fossil CO2 emissions and 18 % by land use change (Friedlingstein et al. 2019)
     
  • One-third of the CO2 released from burning fossil fuels, is absorbed by forests every year (Harris et al. 2021)
     
  • Research found that the world’s forests sequestered about twice as much CO2 as they emitted between 2001 and 2019; forests provide a “carbon sink” that absorbs a net 7.6 billion metric tonnes of CO2 per year (Harris et al. 2021)

"With a very high level of confidence, the increase in CO2 emissions from fossil fuel burning and those arising from land use change are the dominant cause of the observed increase in atmospheric CO2 concentration" (IPCC, 2013, p. 467).

Facts and Figures

The net terrestrial land-atmosphere flux is not measured directly, but is inferred by subtracting the fossil fuel and net ocean-atmosphere fluxes from the observed atmospheric increase.

 

This inevitably causes the net land‑atmosphere flux to have the largest uncertainty amongst all the net fluxes. The land surface is found to be a net carbon sink, with uncertainty of the same order as its magnitude.

Further Information Carbon Cycle

IPCC, Climate Change 2013: The Physical Science Basis, Chapter 6: Carbon and Other Biogeochemical Cycles
https://www.ipcc.ch/report/ar5/wg1/carbon-and-other-biogeochemical-cycles/

IPCC, Climate Change 2001: The Scientific Basis, Chapter 3: The Carbon Cycle and Atmospheric Carbon Dioxide
https://www.ipcc.ch/report/ar3/wg1/the-carbon-cycle-and-atmospheric-carbon-dioxide/

Source: NOAA

Illustration of the Carbon Cycle

Why do we need the Biomass Mission?

The ESA P-band Biomass mission is designed "to reduce the major uncertainties in carbon stocks and fluxes associated with the terrestrial biosphere, including carbon fluxes associated with Land Use Change, forest degradation and forest regrowth." (Quegan et al. 2021)

 

 

"The main goal of the Biomass mission is contributing to reduce the uncertainty in the worldwide spatial distribution and dynamics of forest biomass, with the final aim to improve current assessments and future projections of the global carbon cycle." (Carbone et al. 2021)

Uncertainties?

Uncertainty

Accuracy

Precision

Accuracy refers to the closeness of the measurements to a specific value and describes systematic errors, a measure of statistical bias; low accuracy causes a difference between a result and a "true" value.

 

 

Precision refers to the degree to which individual measurements vary around a central value and describe  random errors, a measure of statistical variability.
Measurements with high
precision are highly reproducible as repeated measurements will reliably lead to similar results; however.

Measurements may be precise but accurate, not precise but accurate, precise and accurate or not precise and not accurate.

Uncertainties?

  • National forest inventory data are the basis for annual global reports on forest resources and carbon stocks.
  • In most developed countries there are operational methodologies for biomass inventories by national authorities (e.g. forest services) using different techniques (e.g. field based surveys, LiDAR, remote sensing) -each method comes with its own uncertainties.
  • In most developing countries, such biomass information are uncertain or do not exist at all. This is especially unfortunate because these are the countries experiencing great changes and fast rates of deforestation.
  • Furthermore, different definitions, standards, and qualities in national inventories lead to biases and uncertainties. If supra-national bodies (e.g. FAO) collect country reports (national reports/inventories) and, together with remote sensing based biomass assessments at selected sampling sites, creates global reports like the Global Forest Resources Assessment, uncertainties sum up.

 

(Bombelli et al., 2009)

In-situ measurements

 

  • AGB estimated from in-situ measurements of variables using an allometric model as a function of ρ as stem wood density of the focal tree, D as trunk diameter at breast height, and H as its total height:



     
  • To get AGB values at plot* level, tree-level AGB estimates are summed across trees and over the plot area
  • Errors due to the estimation of height, wood density, and choice of the allometric model usually result in a plot-based AGB uncertainty that is non-negligible.
    • AGB is estimated with ~50% absolute error for a single tree 
    • When propagated at stand level, error is ~10% at the 1-ha scale 

(Chave et al., 2019, Chave et al., 2014, Réjou-Méchain et al. 2014)

AGB=f(\rho,D,H)

Errors / Uncertainties

* A plot is a parcel of woodland or forest

Mission Requirements

In order to comply with the objectives, a number of mission requirements have been formulated in the Biomass Mission Requirements Document (BMRD).

Objectives

Geophysical Product Requirments

Level-1 Data Requirements 

ESA (2015): Biomass Mission Requirements Document, EOP-SM/1645

1

1

2

Level-1 Data: SAR data are available at different levels of processing and usability for the user, please see also Mission Products

2

Geophysical Product Requirements

MR-010: "[The] Biomass [mission] shall deliver maps of

  1. Above ground forest biomass, defined as dry weight of woody matter, expressed in t ha-1
  2. Forest height, defined as upper canopy height according to the H100 standard used in forestry expressed in meters
  3. Deforestation/logging area, defined as an area where an intact patch of forest has been cleared, expressed in a binary classification of intact/deforested or logged areas."

MR-020: "All products shall be processed and archived up to Level-2."

1. Product Definition

(ESA 2015, p. 30)

Geophysical Product Requirements

MR-030: "To address primary and secondary mission objectives Biomass shall provide full global coverage throughout the mission as a goal. As a threshold the mission shall cover forested land areas below 75° N and above 56° S as defined in  Figure 4.1 [refers to BMRD] by areas in grey color."

2. Coverage

Fig. 4.1 of the Biomass Mission Requirements Document showing spatial coverage requirement of the mission 

Source: ESA Biomass Mission Requirements Document 

(ESA 2015, p. 30)

Geophysical Product Requirements

MR-040: "The BIOMASS system shall provide geophysical products (Level-2 products) with a resolution of 4 ha (biomass and forest height) and 0.25 ha (disturbance)."

3. Spatial Resolution

4. Temporal Sampling

MR-050: "The Biomass system shall provide the data products defined in MR-010 for the area defined in MR-030 once every 6 months as goal / once every year as threshold during the mission lifetime.  ."

(ESA 2015, p. 36)

(ESA 2015, p. 36)

Geophysical Product Requirements

MR-060: "The BIOMASS system shall provide products (Level-2 products) with an accuracy of

  • 10 t ha   for biomass < 50 t ha    and better than 20% for biomass values above this for the forest biomass product;
  • 20% (G) / 30% (T) in height for the forest height product;  
  • 90% detection probability for deforested areas."

5. Uncertainty

-1

-1

(ESA 2015, p. 37)

Geophysical Product Requirements

6. Latency

MR-070: "The Biomass data latency shall be better than 1 month."

What is latency?

Generally speaking, latency can be seen as time delay between the cause and the effect of some physical change in a system under observation.

In the context of the Biomass Mission, data latency means the time required from the data acquisition by the sensor and the availability of the products to the users. When the mission was designed, it was decided that there is no need for near-real-time processing.

(ESA 2015, p. 37)

Geophysical Product Requirements

MR-080: "The Biomass mission duration shall be a minimum of 5 years."

7. Mission Duration

(ESA 2015, p. 37)

Level-1 Data Requirements

MR-100: "The Biomass observing technique is a P-band SAR in side-looking mode. The centre frequency is to be selected based on the ITU allocation: 435 MHz with 6 MHz  bandwidth."

1. Observing technique and selection of frequency

2. Polarisation

MR-110: "The mission shall operate in full polarimetry."

(ESA 2015, p. 46)

(ESA 2015, p. 48)

Level-1 Data Requirements

MR-120: "The incidence angle shall be greater than 23˚(T) / 25˚(G)."

"The threshold requirement for the incidence angle of the Biomass mission is therefore set to be greater than 23°, and the goal to be greater than 25°."

3. Incidence Angle

4. Geolocation Requirements

MR-130: "The Biomass Level-1 product shall be geolocated to an accuracy of 25 m at
1 sigma."

(ESA 2015, p. 48)

(ESA 2015, p. 49)

Level-1 Data Requirements

MR-140: "The dynamic range shall cover -30dB to +5dB. "

5. Dynamic Range

6. Radiometry

MR-200: "The spatial resolution (ground range × azimuth) shall be 50×50 m² (G) / 60×50 m² (T) with ≥ 6 ENL at 25 degrees incidence angle. "

(ESA 2015, p. 49)

(ESA 2015, p. 50)

6.1. Resolution and number of looks 

Level-1 Data Requirements

MR-210: "The absolute radiometric bias shall be ≤0.3 dB at 1 sigma."

(ESA 2015, p. 52)

6.2. Radiometric bias and stability

MR-220: "The radiometric stability shall be ≤ 0.5 dB (T) / 0.17 dB (G) at 1 sigma which shall be achieved through system and vicarious calibration."

(ESA 2015, p. 52)

6.3. Noise Equivalent Sigma Nought

MR-230: "The Noise Equivalent Sigma Nought shall be ≤ -27 dB (T) / -30 dB (G)."

(ESA 2015, p. 52)

Level-1 Data Requirements

MR-240: "The total ambiguity ratio TAR shall be ≤ -18 dB."

(ESA 2015, p. 53)

6.4. Range and azimuth ambiguities

MR-250: "The cross talk of the instrument  shall be ≤ -30 dB."

(ESA 2015, p. 54)

6.5. Cross Talk and Channel Imbalance

MR-252: "The channel imbalance shall be better than -34 dB in transmission and in reception."

(ESA 2015, p. 54)

Level-1 Data Requirements

MR-254: "The standard deviation of the residual phase error over the pulse travel time shall be less than or equal to 10°. "

7. Phase Errors

(ESA 2015, p. 54)

MR-258: "The standard deviation of the residual phase error over the data take time shall be less than or equal to 10°. "

Level-1 Data Requirements

MR-260: "The ratio between the interferometric Doppler common bandwidth and the SAR Doppler bandwidth shall be equal or higher than 0.98. "

8. Interferometric Requirements

(ESA 2015, p. 54)

Level-1 Data Requirements

9. Orbit/Mission Phases

MR-300: "The Biomass mission shall support an interferometric phase and a tomographic phase. The tomographic phase shall be at the beginning of the mission and shall provide one global coverage."

MR-305: "During the interferometric phase the Biomass system shall provide L1 data for the area defined in MR-030 every 6 month. The requirement can be relaxed to 7.5 month to achieve a 3 to 4 days repeat cycle."

(ESA 2015, p. 55)

MR-308: "The Biomass system shall observe during both the ascending and descending passes, thus providing two coverage maps within the required periods."

Level-1 Data Requirements

9. Orbit/Mission Phases

(ESA 2015, p. 55)

MR-310: "The repeat cycle of the interferometric phase shall be ≤  4 days."

9.1. Interferometric phase repeat cycle and baseline

MR-320: "In the interferometric phase the baseline B between the i-th orbit of cycle n and the i-th orbit of cycle n+1shall be 30% of the critical baseline Bc over the Equator. "

MR-330: "The interferometric phase shall acquire three consecutive acquisitions complying with requirements MR-310 and MR-320 to support dual baseline interferometric processing. "

MR-335: "During the interferometric phase, the orbit ground track shall be controlled to within ±10 % of the critical baseline Bc. "

(ESA 2015, p. 56)

Level-1 Data Requirements

MR-340: "The repeat cycle of the tomographic phase shall be ≤ 4 days."

9.2. Tomographic phase repeat cycle and baseline

(ESA 2015, p. 57)

MR-350: "The baseline of the tomographic phase between orbit i of cycle n and that of cycle n+1 shall be  15% of the critical baseline Bc at the equator."

MR-360: "The tomographic phase shall acquire seven consecutive acquisitions following MR-340 and MR-350 to support tomographic processing."

MR-365: "During the tomographic phase, the orbit ground track shall be controlled to within ±7.5%  (G) / ±10 % (T) of the critical baseline Bc."

Level-1 Data Requirements

9.3. Local time of the orbital node at the equator

(ESA 2015, p. 58)

MR-370: "The Biomass orbit shall be sun synchronous with a dawn dusk or near dawn dusk (06:00 ± 15 min) equator crossing time. "

Why P-band?

Why P-band has been selected as information carrier for the ESA Biomass Mission? Before this question is answered, we will take a closer look at radar frequency bands.

 

They are small sections of the electromagnetic spectrum, which integrates electromagnetic radiation at all wavelengths or frequencies. Active remote sensing techniques, like the P-band SAR Biomass Mission, are using the microwave region of the electromagnetic spectrum. This covers a wide range of wavelengths λ or frequencies f ranging from wavelengths of 1 mm (300 GHz) to 1 m (0.3 GHz).
 

The microwave range is divided into a number of so-called radar bands, designated with a capital Latin letter.

 

  • The ESA Biomass Satellite will be the first operational space-borne P-band Synthetic Aperture Radar (SAR)
     
  • P-band
    • Radar bands defined by the International Telecommunication Union
    • Defined as electromagnetic frequency band with
      • Wavelength 30 - 100 cm or
      • Frequency 1 - 0,3 GHz
    • Sometimes P-band is also called UHF standing for Ultra High Frequency

Why P-band?

Why P-band?

P-band SAR as an active remote sensing technique using the microwave region of the electromagnetic spectrum is (almost) independent of:

  • Sunlight and thus time of the day
  • Weather conditions as microwaves can penetrate clouds (with some limitations)

 

Why P-band?

ESA SAR Earth Observation Satellite Sentinel-1 acquiring C-band SAR data through clouds

  • The ability to penetrate a vegetation canopy mainly depends on the used radar band = frequency of the radiation (or its wavelength)
     
  • As a rule of thumb, it can be said that the larger the wavelength, the deeper the penetration depth.

Why P-band?

Sensitivity of SAR measurements to forest structure and penetration into the canopy
at different wavelengths

  • The ability to penetrate a vegetation canopy mainly depends on the used radar band = frequency of the radiation (or its wavelength)
     
  • As a rule of thumb, it can be
    said that ​at large wavelengths,

    only large parts of a tree
    contribute to radar
    backscatter signals.

Why P-band?

Main scatterers at different frequencies

leaves, twigs

leaves, small branches

branches and trunk

trunk

Main scatterers:

Why P-band?

Examples of SAR imagery at C-, L-, and P-band (AIRSAR) over tropical forests along the Ja River in Papua New Guinea showing differences of penetration and impacts of forest structure as SAR false color composite (HH, HV, VV) imagery.

Why P-band?

An example of a P-band SAR image recorded by the DLR E-SAR airborne instrument over Reminstorp in Sweden. Acquired in September 2010 during the BioSAR campaign.

 

RGB Colour Composite: R = HH, G = HV, B = VV The forest stand and site characteristics are reflected in the different tones and colours of the image.

Why P-band?

P-band radar backscatter:

  • Shows highest sensitivity to biomass compared to all other frequencies that can be exploited from space
  • Can penetrate the canopy in all forest biomes and interacts preferentially with the large woody vegetation elements in which most of the biomass resides
  • Is more sensitive to biomass than at higher frequencies (X-, C-, S- and L-bands); lower frequencies (e.g. VHF) display even greater sensitivity but present formidable challenges for space borne SAR because of ionospheric effects
  • Displays high temporal coherence between passes separated by several weeks, even in dense forest - allows use of Polarimetric interferometric SAR (PolInSAR) to retrieve forest height and use of Tomographic SAR (TomSAR) to retrieve vertical structure
  • P-band is highly sensitive to disturbances of the biomass

(Quegan et al. 2019, Carbone et al. 2021)

Mission Planning

  • With the Living Planet Programme (LPP) ESA developed a new strategy for Earth Observation, where Earth observation missions are divided into Earth Explorer and Earth Watch missions (Simon et al., 2006)

Earth Explorer Missions

Earth Watch Missions

Research and demonstration

User driven operational missions

  • Copernicus Programme (The Sentinels)
  • Cooperation of ESA with EUMETSAT

Links:

GOCE, SMOS, CryoSat, Swarm, ADM-Aeolus, EarthCare, Biomass

LPP

Mission Planning

Typical phases on an ESA mission

The Biomass launch is expected around the end of 2023 and is planned to be a five-year mission.

Mission Planning & Preparation Timeline

March 2005

May 2006

Submission of 24 proposals -> Scientific and Technical Assessment

ESA Call for Proposals for Earth Explorer Core Missions

Selection of six Candidate Missions with Biomass being one of them

Biomass

Phase 0
Preliminary Feasibility Studies

Mission Planning & Preparation Timeline

January 2009

2010/2011

Presentation of the 6 Candidate Missions at a User Consultation Meeting

Subsequent down-selection of 3 Candidate Missions for a Phase-A Feasibility Study

Conduction of study; formally completed by Preliminary Requirements
Review (PRR)

Mission Planning & Preparation Timeline

January 2013

Extension of Phase-A studies by 6 months to address critical issues found during studies and to further consolidate the scientific dossier of the proposed candidates

Results of extension published in "Addendum to the Report for Mission Selection"

Reports for Mission Selection for the 3 Candidate Missions issued

June 2012

Earth Science Advisory Committee (ESAC) recommended implementing Biomass as the 7th Earth Explorer Mission

ESA Earth Observation Envelope Programme

March 2013

Mission Planning & Preparation Timeline

May 2013

April 2016

ESA and Airbus Defence and Space UK signed a €229 million contract to build the next Earth Explorer: the Biomass satellite, due to begin its mission in 2021

Approval of Biomass Mission as the 7th Earth Explorer Mission in the frame of Earth Observation Envelope Programme

152nd meeting of
ESA Earth Observation Programme Board

(Sources:
ESA 2015,
ESA, May 7 2013,"N° 13–2013: ESA’s next Earth Explorer satellite", URL:
https://www.esa.int/Newsroom/Press_Releases/ESA_s_next_Earth_Explorer_satellite,
Scipal, 2017,

ESA, May 3, 2016: ”Ready to build the BIOMASS forest mission”, URL: http://www.esa.int/Our_Activities/Observing_the_Earth/Ready_to_build_the_Biomass_forest_mission,

ESA, 28 Oct 2019: "Contract seals deal for Biomass satellite’s ride into space", URL: https://www.esa.int/Applications/Observing_the_Earth/Biomass/Contract_seals_deal_for_Biomass_satellite_s_ride_into_space)

ESA and Arianspace sign Launch Contract and secure launch of ESA Earth Explorer Biomass Satellite with lift-off from French Guiana scheduled for 2022

October 2019

Mission Planning & Preparation Timeline

April 2021

Biomass satellite successfully completed a suite of mechanical tests including sine vibration, acoustic, shock and clamp-band release tests.

(Source:
ESA, 28th April 2021: "Forest measuring satellite passes tests with flying colours
", URL:
https://www.esa.int/Applications/Observing_the_Earth/Biomass/Forest_measuring_satellite_passes_tests_with_flying_colours,
ESA, 29th October 2021: "ESA and NASA launch revolutionary open-source platform", URL: https://www.esa.int/Applications/Observing_the_Earth/ESA_and_NASA_launch_revolutionary_open-source_platform,
ESA - biomass - ESA's forest mission, URL: https://www.esa.int/Applications/Observing_the_Earth/FutureEO/Biomass)

The Biomass launch is expected around 2023 and is planned to be a five-year mission.

2023

ESA and NASA launch MAAP open-source platform

Ocotber 2021

Biomass Mission Phases

(Source: Sedehia et al., 2021)

  • BIOMASS mission is designed for lifetime of 5 years
  • Will consist of two phases:
    • Tomographic phase (TOM) of 1 year followed by the
    • Interferometric phase (INT)

launch

Tomographic
Phase
(TOM)

Interferometric Phase (INT)

Commis-
sioning
Phase

launch

+

4 months

End of
planned
lifetime

end of year 1

end of year 1

end of year 2

end of year 3

end of year 4

end of year 5

duration:
12 months

Biomass Mission Phases

Biomass Mission Phases

TOM

  • will last 1 year starting after commissioning phase
  • during the TOM phase, 6 spatial baselines resulting in 7 SAR acquisitions for a given location will be obtained
    • For separating 3 – 4 layers in forests of 50 m height using tomographic techniques, at least 5–6 passes are required

 

INT

  • The INT phase will provide interferometric polarimetric observations to achieve the mission objectives.
  • during INT phase, 3 passes, each separated by 3 days, will be combined in 3-baseline Pol-InSAR.
  • each cycle of the INT will require 228 days (~7 months) to provide global coverage.

(ESA, 2012, Sedehia et al., 2021)

Biomass Mission Ground Facilities

  • Mission control: ESA's European Space Operations Centre (ESOC) in Darmstadt, Germany
  • Communication links: ESA's ground station in Kiruna, Sweden, via X-band downlink (310/520* Mbit/s) for science data; via S-band uplink (64 kbit/s) and downlink (128 kbit/s) for tracking, telemetry and command
  • Data: processing at ESA's Centre for Earth Observation (ESRIN) in Frascati, Italy
  • Project and commissioning: managed at ESA's European Space Research and Technology Centre (ESTEC) in Noordwijk, Netherlands

(Source: Background imagery provided by  GoogleEarth with data from Data SIO, NOAA, U.S. Navy, NGA, GEBCO, Image Landsat/Copernicus, Image IBCAO, Image U.S. Geological Survey)

Location of BIOMASS mission ground facilities

Former Missions and Projects

The Biomass mission was not just thought up and planned out of the blue, but is based on many years of research and technical development. These have helped to understand and model the relationship between radar backscatter signals from vegetated areas and various parameters for characterizing a forest. This allows a number of forestry-relevant products to be derived from active radar data.


In the following, some Earth observation campaigns and scientific research projects that helped to pave the way for the Biomass mission are briefly introduced.
 

GEDI

  • GEDI = Global Ecosystem Dynamics Investigation
  • Selected by NASA in July 2014 under the Earth Venture Instrument-2 program
  • System development and science programme began in November 2014
  • Launched on Dec. 5th, 2018 and subsequently installed on the International Space Station (ISS)

GEDI's Optical Bench (OB) being lowered carefully into the GEDI box structure during integration activities inside the Spacecraft Checkout and Integration Area (SCA) at Goddard Space Flight Center.

(Source: https://www.nasa.gov/feature/goddard/2018/gedi-to-measure-earths-forests, Credits: NASA Goddard/Barbara Lambert)

GEDI

  • Instrument:
    • Full waveform light detection and ranging (Lidar) laser system, consisting of three lasers that produce eight parallel tracks of observations.
    • Allows location of GEDI waveforms on the Earth’s surface with horizontal accuracy within +/- 9 m
    • Each laser fires 242 times per second and illuminates a 25 m spot (a footprint) on the surface over which 3D structure is measured. Each footprint is separated by 60 m along track, with an across-track distance of about 600 m between each of the eight tracks.

GEDI

Along track lidar return energy showing vertical distribution of vegetation

ICESat

  • ICESat = Ice, Cloud, and land Elevation Satellite
  • Earth observation mission by NASA (USA)
  • Launched 13th January 2003
  • Main sensor is GLAS (Geoscience Laster Altimeter System) instrument
  • in operation until GLAS failure in 2009
  • Satellite burnt up in the atmosphere on 30th August 2010
  • Objectives:
    • Primary objectives:
      • Determine the mass balance of the polar ice sheets and their contributions to global sea level change
      • Obtain essential data for prediction of future changes in ice volume and sea-level.
    • Secondary objectives:
      • Measure cloud heights and the vertical structure of clouds
      • Aerosols in the atmosphere,
      • Map the topography of land surfaces, roughness, reflectivity
      • Vegetation heights,
      • Snow-cover and sea-ice surface characteristics.

GLAS on the ICESat spacecraft immediately following its initial mechanical integration on June 18th, 2002

ICESat

GLAS instrument when receiving data

  • GLAS
    • Was the first laser-ranging (lidar) instrument for continuous global observations of Earth.i
    • Includes

      • Laser system to measure distance,

      • Global Positioning System (GPS) receiver,

      • Star-tracker attitude determination system

    • Laser transmits short pulses (4 nanoseconds) of infrared light (1064 nanometres wavelength) and visible green light (532 nanometres)

    • Photons reflected back to the sensor from Earth's surface and from the atmosphere, including the inside of clouds, will be collected in a 1-metre diameter telescope

    • Laser pulses at 40 times per second illuminate spots (footprints) with a diameter of 70 meters and spaced at intervals along Earth's surface of 170-metre

SIBERIA 1 & 2

  • The SIBERIA project (SAR Imaging for Boreal Ecology and Radar Interferometry Applications), consisted of two phases - SIBERIA 1 & 2.
     
  • Forest maps covering Russia were only available at a scale of
    1:2.5 Million before the project; latest issues printed in 1990.
     
  • For about 100 million hectares of this map, only data from rough surveys have been carried out during the period 1948-1956. Due to numerous large-scale disturbances, both natural and human induced (fires, insect, illegal logging), the state of vast territories was unknown.
     
  • Major goal of SIBERIA: produce an extensive forest map of a large geographical region in Russia, for which only limited information was at the time available  

 

A tartar-mongol magic drum top has been selected as the logo of the SIBERIA project. Such drums were used by various Tartar and Mongol tribes in religious ceremonies.

The logo is a symbolical map of the Universe – maybe the earliest image of such type in human history. The central, cross-like figure represents the “Spirit-Master of the drum”. The space of the picture divided into two important zones: above you see the sky (Upper World) with stars. Below the horizontal line, there is a human world (Middle World). In the left part, the shaman, holding the drum. Above him – Mountain rams. In the right part: The horse beneath the tree, this animal ready for sacrifice. Above, the same animal after being sacrificed. Its skin is attached to a special ritual construction called Tayilga. (Andrei M. Sagalev)

(Santoro et al., 2003)

Links:

SIBERIA 1 & 2

  • The SIBERIA project was a huge undertaking and combined forces of the German Aerospace Centre (DLR), the European Space Agency (ESA), and the Japanese Space Agency (NASDA)
     
  • 900,000 square kilometres were mapped
     
  • In 1997,  DR shipped two containers with equipment for a mobile receiving  EO data station to Ulaanbaatar (Mongolia) for one of the biggest remote-sensing campaigns ever undertaken in Europe: Imaging the vast forests of central Siberia using three SAR remote sensing satellites – ERS-1 & ERS-2 (C-band) and JERS-1 (L-band). They were used to collecting SAR data simultaneously throughout autumn 1997 and then again in summer 1998.
     
  • The SIBERIA-I forest map was derived primarily from satellite data and remote sensing techniques. These include multi-temporal and interferometric ​data from C- and L-band space borne radar instruments
     
  • Based on extensive database analyses, the classes “Water”, “Smooth open areas” (including bogs, agriculture and grassland) and four total growing stock classes, “≤20 m3/ha”, “20-50 m3/ha”, “50-80 m3/ha” and “>80 m3/ha” were defined as target classes of the map.

(Santoro et al., 2003, Wagner et al., 2003)

SIBERIA 1 & 2

(Santoro et al., 2003, Wagner et al., 2003)

Analysis of scatterplots of ERS coherence and JERS intensity lead to typical patterns and resulted in this figure here.

It shows a clear separation between water and the other classes and to a lesser extent between smooth fields and the various forest volume classes

SIBERIA 1 & 2

  • Two exponential models were used to estimate signatures of coherence and backscatter for the four forest classes.
     
  • Each image was classified using a maximum likelihood algorithm trained on the model input, followed by a new contextual classification algorithm, the Iterated Contextual Probability (ICP) algorithm;
     
  • Total area was divided into 98 map sheets with a scale 1: 200,000
     
  • Validation:
    • Confusion matrices of all test sites were generated by tabulating the correspondence of forest inventory polygons with the radar-derived classes to assess map accuracy
    • Weighted Kappa coefficient was used to estimate the agreement between classified map and ground data.
    • Coefficient varied between 0.73 and 0.97 (pooled κw = 0.94).
    • User accuracies for each individual class were all > 80%
    • Russian experts concluded that the radar-derived forest cover map had a satisfactory quality for practical applications, e.g. for monitoring of reforestation and updating old forest inventory data. 

 

(Santoro et al., 2003, Wagner et al., 2003, Balzter et al., 2000)

SIBERIA 1 & 2

SIBERIA-I:
Histogram matched RGB Mosaic of ERS Coherence, JERS Intensity ans ERS Intensity

SIBERIA 1 & 2

SIBERIA-I:
Classification Algorithm Mosaic of Forest Stem Volume and Land use

Siberia 1 & 2

  • The overall objective of SIBERIA-II is to develop a combined monitoring system to yield estimates of carbon sources, sinks and pools at multiple spatial and temporal scales, from regional scales to those relevant to land-use policy and resource management.

BIOMASAR I & II

(Santoro et al., 2010, Santoro et al., 2011)

  • The BIOMASAR I & II projects were motivated by the fact that datasets of Growing Stock Value (GSV) available at the time contained gaps and errors, leading to the implication that carbon stocks assessment suffers from major uncertainties.
  • Here, active radar remote sensing may help.
  • BIOMASAR I & II made use of the fact that for the boreal zone, ESA ENVISAT ASAR ScanSAR data acquired in C-band are available at an almost daily frequency. Using a novel hyper-temproal retrieval approach, it was possible to retrieve GSV even from a large number of C-band data with a substantially reduce retrieval error when compared to an GSV estimate based on a radar backscatter acquisition.
  • The retrieval model is based on a Water Cloud Model (WCM) with gaps
    • The original WCM is a ​simple model describing the relationship between forest backscatter and forest
      parameters introduced by Attema & Ulaby, 1978; vegetation is represented as a homogeneous medium that is filled with small water droplets positioned over a horizontal plane surface representing the ground
    • But the WCM does not include gaps between trees to represent the horizontal inhomogeneity and the
      varying structure of the canopy along the vertical profile of a forest leading to a vertical inhomogeneity
    • Therefore, the WCM with gaps was introduced by Askne et al., 1997 and is used in the BIOMASAR GSV retrieval approach

BIOMASAR I & II

BIOMASAR algorithm use of a Water Cloud model with gaps:

Forest backscatter = sum of:

  • Direct scattering from the ground through the canopy gaps, weighted by the percentage of ground seen through the gaps in the canopy expressed as (1 -    ), where     is the area-fill factor, representing the percentage of ground covered by vegetation 
  • Ground scattering attenuated by the tree canopy attenuated by the vegetation layer
  • Direct scattering from the vegetation weighted by the area-fill factor and the two-way tree transmissivity         
     

forest backscatter

direct scattering from ground

ground scattering attenuated by trees

direct scattering from vegetation

\eta
\eta
\eta T_{tree}
T_{tree}

(Santoro et al., 2010, Santoro et al., 2011)

\sigma_{for}^0=(1-\eta )\sigma_{gr}^0+\eta\sigma_{gr}^0T_{tree}+\eta\sigma_{veg}^0(1-T_{tree})

BIOMASAR I & II

For retrieval purposes it is more convenient to use the expression reported by Pulliainen et al. (1994), in which the two-way forest transmissivity has been expressed as          , where V is the GSV and    is an empirically defined coefficient expressed in              :

ha/m^3
\beta
e^{-\beta V}

(Santoro et al., 2010, Santoro et al., 2011)

\sigma ^0 _ {gr}
\sigma ^0 _ {veg}

= radar backscatter from ground

= radar backscatter from vegetation

  • Estimation of both parameters using MODIS VCF percent tree cover
  • Estimation of the parameter        :
    • Set equal to the mean value of the backscatter for pixels labelled as unvegetated within an estimation window centred at the pixel of interest (Santoro et al., 2011).
    • Pixels = unvegetated if percent tree cover in MODIS VCF product < spatially adaptive threshold (15% up to 30%) within an estimation window 
    • VCF threshold have to satisfy requirement: at least 2% of pixels with a SAR backscatter measurement within the estimation window belong to class of unvegetated pixels!

Parameters:

\sigma ^0 _ {gr}

1

1

(1)

\sigma_{for}^0=\sigma_{gr}^0e^{-\beta V}+\sigma_{veg}^0(1-e^{-\beta V})

BIOMASAR I & II

(Santoro et al., 2010, Santoro et al., 2011, Santoro et al., 2015)

  • Estimation of the parameter          :
    • Obtained from an estimate of the average SAR backscatter for dense forest         (Santoro et al., 2013).
    • Dense forest = pixels within a 200 × 200 pixels of estimation window with a MODIS VCF percent tree cover above 75% of the maximum value
    • Estimation of        requires: at least 1% of the pixels in the estimation window with a SAR backscatter measurement belong to dense forest class
    •         was then obtained by compensating         for residual  ground backscatter contributions by inversion of Eq. 1 (prev. slide) to express          as a function of the estimates of        ,        ,         (value representative of GSV in dense forest) and the coefficient β (Santoro et al., 2013):
\sigma ^0 _ {veg}
\sigma ^0 _ {df}
\sigma ^0 _ {df}
\sigma ^0 _ {veg}
\sigma ^0 _ {df}
\sigma ^0 _ {veg}
\sigma ^0 _ {df}
\sigma ^0 _ {gr}
V_ {df}
  • To finally obtain GSV estimate from a SAR backscatter measurement Eq. 1 (prev. slide) was inverted to express GSV as a function of a measured backscatter            :
\sigma ^0 _ {meas}

(2)

(3)

\sigma_{veg}^0=\frac{\sigma_{df}^0-\sigma_{gr}^0e^{-\beta V_{df}}}{1-e^{-\beta V_{df}}}
\widehat{V}=-\frac{1}{\beta}ln\frac{\sigma_{veg}^0-\sigma_{for}^0}{\sigma_{veg}^0-\sigma_{gr}^0}

BIOMASAR I & II

(Source: Santoro et al., 2011)

Flowchart BIOMASAR algorithm

Flowchart of the BIOMASAR algorithm on the left hand side:

  1. Generation of stacks of radiometrically calibrated, geocoded and co-registered C-band SAR backscatter images
  2. Inversion of individual backscatter measurements using Water–Cloud type of model to derive GSV
  3. Then multi-temporal combination of individual estimates of GSV to get final GSV

(Santoro et al., 2011)

BIOMASAR

Central Siberia Forest Biomass Map:

  • 1 km resolution
  • 2,400,000 km2
  • ENVISAT ASAR – Global Monitoring mode (Jan. 2005 – Feb. 2006)
  • GLC 2000 land cover used as background

BIOMASAR

Map showing growing stock volume (GSV) – the amount of wood expressed in cubic metres per hectare over the Northern Hemisphere, derived from about
70,000 ESA ENVISAT radar acquisitions in C-band  in ScanSAR mode from October 2009 to February 2011.
Dark blue and purple represent areas of high GSV, while orange areas have low GSV. White areas have none.

(Santoro et al., 2015)

ESA DUE GlobBiomass

  • Thee main  purpose  of  the  ESA DUE  GlobBiomass  project  was to  better characterize  and  to  reduce  uncertainties  of  AGB  estimates.
  • Achieved by  developing an innovative synergistic mapping approach for five regional sites (Sweden, Poland, Mexico, South Africa, Kalimantan) for the epochs 2005, 2010 and 2015 and for  one  global  map  for  the  year  2010.
  • For regional mapping, different approaches
    and data sources were used
  • For global mapping, hyper-temporal model
    developed in BIOMASAR further improved
  • Global mapping results available at
  • Pangaea:
    https://doi.pangaea.de/
    10.1594/PANGAEA.894711

ESA DUE GlobBiomass regional test sites

(Source: Departm. Earth Observation, University of Jena)

ESA DUE GlobBiomass

ESA DUE GlobBiomass

(Source: https://globbiomass.org/products/regional-mapping/regional-biomass-mapping-sweden/)

ESA DUE GlobBiomass

ESA DUE GlobBiomass

(Source: Santoro et al. ,2021, https://essd.copernicus.org/
articles/13/3927/2021/
,
 CC-BY-4.0)

ESA DUE GlobBiomass Global map product

ESA CCI Biomass

  • The primary science objective of ESA’s Climate Change Initiative (CCI) Biomass project is to provide global maps of above-ground biomass (Mg ha-1) for four epochs (mid 1990s, 2010, 2017 and 2018), with these being capable of supporting quantification of biomass change.
  • CCI Biomass products are derived from
    • C-band (ESA: Sentinel 1A & B) and L-band SAR data (JAXA: ALOS-2 PALSAR-2) and
    • Additional information from space borne LIDAR (e.g. NASA's Global Ecosystem Dynamics Investigation Lidar GEDI).

 

"Climate Change Initiative (CCI) Biomass project picks up where GlobBiomass left off, but with a much sharper focus on the role of biomass in climate."

"Although the GlobBiomass algorithms form the backbone of CCI-Biomass, there will be a major effort to identify the weaknesses in this approach in different biomes, and develop methods to correct these."

Preparatory Campaigns

A number of preparatory campaigns have been carried out in preparation of the Biomass mission. This comprises airborne campaigns using the E-SAR system operated by the German DLR and the ONERA system operated by the French ONERA. Within these airborne campaigns, fully-polarimetric L- and P-band data were acquired. Additionally, there are also a number of ground-based data collection campaigns.

In the following, a chronological overview of these airborne and ground-based campaigns is presented.

11/2004

03/2007

05/2007

INDREX-2

BioSAR-1

10/2008

BioSAR-2

08/2009

TropiSAR

BioSAR-3

09/2010

TropiScat

08/2012

12/2012

AfriSAR

07/2015

02/2016

07/2016

AfriScat

12/2016

BorealScat

ongoing

INDREX-2

Campaign Indrex-2
Objectives Height retrieval in tropical forest; measurement of repeat-pass temporal decorrelation
Test sites Sungai-Wai & Mawas, Borneo, Indonesia
Time Nov 2004
Forest conditions Tropical rain forest. Sungai-Wai: lowland, AGB ≤600 t/ha; Mawas: peat swamp, AGB ≤200 t/ha
Platform, Sensor, Operator Do228 aircraft, E-SAR, DLR (Germany)
Acquired SAR data Pol-InSAR data set in P & L-band
Outcomes INDREX-II was a successful campaign in terms of experiment planning, radar data and ground measurement acquisition and radar data quality. All suggested experiments could be performed as described in the Experimental Plan submitted to ESA. The planned time schedule could be kept and no major technical problem occur during the INDREX-II campaign.
Website https://earth.esa.int/eogateway/campaigns/indrex-2
Final Report https://earth.esa.int/eogateway/documents/20142/37627/INDREX-II-final-report.pdf

BioSAR

Campaign BioSAR-1
Objectives Biomass estimation and measurement of multimonth temporal decorrelation
Test sites Remningstorp, southern Sweden
Time Mar - May 2007
Forest conditions Hemi-boreal forest, low topography, AGB ≤300 t/ha
Platform, Sensor, Operator Do228 aircraft, E-SAR, DLR (Germany)
Acquired SAR data Pol-InSAR data set in P & L-band
PolSAR data set in P-band
Outcomes  
Website https://earth.esa.int/eogateway/campaigns/biosar
Final Report https://earth.esa.int/eogateway/documents/20142/37627/biosar-2007-final-report-no-synthesis.pdf

BioSAR

Campaign BioSAR-2
Objectives Topographic influence on biomass estimation
Test sites Krycklan, northern Sweden
Time Oct. 2008
Forest conditions Boreal forest, hilly, AGB ≤300 t/ha
Platform, Sensor, Operator Do228 aircraft, E-SAR, DLR (Germany)
Acquired SAR data Pol-InSAR data set in P & L-band
Outcomes The BioSAR 2008 campaign collected in-situ and airborne SAR data in support of decisions taken on satellite instrument configurations for the BIOMASS satellite mission. It provided an important database for the study of longer term mission concepts. While the BioSAR 2007 campaign collected data with the objective to investigate the effect of temporal decorrelation at P-band with 100MHz and 6MHz bandwidth in southern Sweden, BioSAR 2008 recorded data at boreal forests with strong topographic effects in northern Sweden to investigate the effect on forest height estimation and radar backscatter signal variation.
Website https://earth.esa.int/eogateway/campaigns/biosar-2
Final Report https://earth.esa.int/eogateway/documents/20142/37627/BIOSAR2_final_report.pdf

TropiSAR

Campaign TropiSAR
Objectives Biomass estimation in tropical forest; temporal stability of coherence
Test sites Paracou & Nouragues, French Guiana
Time Aug 2009
Forest conditions Tropical rain forest, AGB 300–500 t/ha, lowland and hilly terrain
Platform, Sensor, Operator Falcon 20 aircraft, SETHI, ONERA (France)
Acquired SAR data Pol-InSAR data set in P & L-band
Outcomes Seven SAR flights were conducted with the SETHI system from ONERA during the 23 days of the SAR campaign, which lasted from 10 August to 1 September 2009. The selected waveform is characterized by P and L simultaneous acquisitions with a range resolution of around 1.5 m. During these flights, a temporal dataset characterised with a zero spatial baseline was acquired to allow the quantification of the temporal decorrelation, one key parameter for the performance evaluation of the PolInSAR technique in a single satellite configuration. A PolInSAR and tomographic database was also acquired with vertical baselines of 50, 100, 150, 200 and 250 feet.
Website https://earth.esa.int/eogateway/campaigns/tropisar
Final Report https://earth.esa.int/eogateway/documents/20142/37627/TROPISAR-FINAL-REPORT-2011-V2.1.pdf

BioSAR

Campaign BioSAR-3
Objectives Forest change and multi-year coherence relative to BioSAR-1
Test sites Remningstorp, southern Sweden
Time Sep 2010
Forest conditions Hemi-boreal forest, low topography, AGB ≤400 t/ha (more high biomass stands than in BIOSAR-1)
Platform, Sensor, Operator Falcon 20 aircraft, SETHI, ONERA (France)
Acquired SAR data Pol-InSAR data set in P & L-band
Outcomes The BioSAR 2010 campaign was planned and successfully implemented during the summer and fall of 2010 over the Remningstorp forest test site in southern Sweden. The planning and execution of the campaign was done with the aim to address all of the agreed primary objectives.
Website https://earth.esa.int/eogateway/campaigns/biosar-3
Final Report https://earth.esa.int/eogateway/documents/20142/37627/BIOSAR_2010_FinalReport_v1.0.pdf

TropiScat

Campaign Tropiscat: Ground-based high temporal resolution measurement
Objectives Measurement of long-term temporal coherence and temporal variation of backscatter in tropical forest
Test sites Paracou, French Guiana
Time Aug. 2011 - Dec. 2012
Forest conditions Tropical rain forest, AGB ca. 400 t/ha
Outcomes The overall results of this study prove that the 6 MHz bandwidth limitation is not critical concerning P-band SAR observations of tropical forests. However, in order to have a full assessment of the end-to-end performances, different types of noise sources still need to be considered. Those are for example the ionospheric disturbances, the seasonal effects (the TropiSAR data were acquired during the dry season), the temporal decorrelation. Reversely, some of the effects prevailing in airborne data will be less significant in space borne data. This is the case for example of the large variation of the incidence angle across the swath, the perturbations in the flight tracts of the aircraft.
Website https://earth.esa.int/eogateway/campaigns/tropiscat
Final Report https://earth.esa.int/eogateway/documents/20142/37627/TropiScat-Final-Report.pdf

AfriSAR

Campaign AfriSAR
Objectives Biomass estimation in tropical forest; temporal stability of coherence
Test sites Sites at Lopé, Mondah, Mabounie and Rabi, Gabon
Time July 2015 (SETHI) Feb. 2016 (F-SAR)
Forest conditions Tropical forest and savannah, AGB from 50 to 500 t/ha
Platform, Sensor, Operator Falcon 20 aircraft, SETHI, ONERA (France),
Do228 aircraft, E-SAR, DLR (Germany)
Acquired SAR data Pol-InSAR and TomoSAR data set in P & L-band
Outcomes During the AfriSAR campaign, shared between ONERA (dry season, July 2015) and DLR (wet season 2016), Pol-InSAR and TomoSAR airborne data set were collected over four test sites in Gabon (Africa), therefore covering different forest structures, biomass levels and disturbances. Although the interferometric / tomographic baselines were optimized for P-band acquisitions, L-band data were collected simultaneously as well.
Website https://earth.esa.int/eogateway/campaigns/afrisar-2015
Final Report https://earth.esa.int/eogateway/documents/20142/37627/AfriSAR-Final-Report.pdf

AfriScat

Campaign Afriscat: Ground-based high temporal resolution measurements
Objectives Measurement of long-term temporal coherence and temporal variation of backscatter in tropical forest
Test sites Ankasa, Ghana
Time July 2015–July 2016
Forest conditions Tropical forest, low topography, AGB from 100 to 300 t/ha
Outcomes  
Website https://earth.esa.int/eogateway/campaigns/afriscat
Final Report https://earth.esa.int/eogateway/documents/20142/37627/AfriScat-Campaign-Final-Report.pdf

AfriScat

Campaign Borealscat: Ground-based high temporal resolution measurements
Objectives Time series of backscatter, tomography, coherence and environmental parameters in boreal forest.
Test sites Remningstorp, southern Sweden
Time Dec. 2016, ongoing
Forest conditions Hemi-boreal forest, spruce-dominated stand, low topography, AGB = 250 t/ha
Outcomes  
Website http://www.borealscat.se
Final Report  

The Satellite

The Biomass mission will bring the first P-band synthetic aperture radar to space.

 

BIOMASS satellite with the Large Deployable Reflector Antenna

The Satellite

  • Launch planned for 2023
  • Satellite will be launched with the VEGA launcher from Europe's Spaceport, the Guiana Space Centre near Kourou in French Guiana.
  • VEGA:

    • Eexpendable launch system in use by Arianespace

    • Jointly developed by the Italian Space Agency (ASI) and the European Space Agency (ESA)
    • Designed to launch payloads from 300 to 2500 kg
    • Total height of the rocket is 30 m with a max. diameter of 3 m,
    • Launch mass of 137 tons
    • Launch thrust of 2700 kilonewtons

Vega launch pad in 2017

Vega launcher

Launch

The Satellite

The BIOMASS satellite will circle the Earth:

  • In a sun-synchronous low-Earth orbit,
  • At an altitude of 666 km and an
  • Inclination of 98°
  • With near dawn-dusk: 06:00/18:00 local equatorial
    crossing time at the equator (± 15 min)
  • And Local Time of the Ascending Node (LTAN)
    on the dawn-side
  • As a left-looking system

 

(Carbone et al., 2021, Geudtner et al., 2021, Sedehi et al., 2021, Quegan et al., 2019)

Orbit

Illustration of orbit inlcination

(Source: https://earthobservatory.nasa.gov/features/OrbitsCatalog, NASA’s Earth Observatory for the Public Domain)

The Satellite

  • Sensor platform will carry out roll manoeuvres to successively generate three sub-swaths of 54.32 km, 54.41 km and 46.06 km width.
  • This will result in a range of incidence angles across the combined swaths from 23° to 33.9°.
  • The Biomass mission will operate  in a near 3 day repeat cycle orbit  in both TOM and INT mission phases in order to minimize temporal de-correlation
  • The global coverage is accomplished by a three swaths imaging approach which requires spacecraft attitude roll and orbit repositioning manoeuvres resulting in a total swath of about 150 km.

 

(Carbone et al., 2021, Geudtner et al., 2021, Sedehi et al., 2021, Quegan et al. 2019)

Orbit

The Satellite

ESA’s Biomass satellite undergoing a thermal elastic distortion test

©

Designing and building the satellite platform is a joint effort by many players:

  • Airbus Defence and Space, Ltd., UK as the prime contractor
  • OHB, Italy
  • APCO Technologies, Switzerland
  • original plan was that OHB also builds the satellite
  • due to Covid restrictions consortium of engineers were impaired in their work

The satellite structure was ready by the end of 2020 and then shipped to a testing facility in Toulouse (France) in early 2021.

The Satellite

The Satellite

(Source: Carbone et a. 2021)

 

The Instrument

The Biomass mission P-band will be the first satellite to carry a fully polarimetric P-band synthetic aperture radar for interferometric imaging.

 

In the electromagnetic spectrum, P-band covers the  frequencies from 0.3 to 1 GHz, which corresponds to a wavelength range from 30 to 100 cm.
 

1

Radar band designations and frequency/wavelength ranges are defined in an IEEE Standard: IEEE Std 521-2002 (Revision of IEEE Std 521-1984), doi: 10.1109/IEEESTD.2003.94224

https://ieeexplore.ieee.org/document/1160089

Overview of different radar bands

The Instrument

The Biomass P-band SAR sensor will be operated in a Quad-Pol mode:

  • V- polarization and H-polarisation pulses are transmitted alternatively while
  • backscattered signals are received simultaneously

in both polarizations.

Electromagnetic waves are transverse waves, where the electric and magnetic fields change (oscillate) in a plane that is perpendicular to the direction of propagation of the wave.
Polarization refers to the
geometrical orientation of the oscillations:

  • Vertical polarization: electric field oscillates vertically in relation to the ground Earth's surface
  • Horizontal polarization: electric field oscillates horizontally in relation to the Earth's surface

The Instrument

Parameter Requirement
Instrument P-band full polarimetric interferometric SAR
Instrument mass 215 kg
Power consumptiom 250 W
Data rate
Center frequency 435 MHz or 0.435 GHz (P-band, 70 cm wavelength)
Bandwidth  6 MHz (ITU allocation)
Near incidence angle >23º (threshold); 25º (goal)
Spatial resolution (≥6 looks) ≤ 60 m (across-track) x 50 m (along-track)
Radiometric stability ≤ 0.5 dB (1σ)
Radiometric bias ≤ 0.3 dB (1σ)

Sensor parameters

(Quegan et al., 2019, Ramongassié et al., 2014)

The Instrument

Instrument Architecture

  • INES = Instrument Electronics System
  • DCU = Digital Control Unit
  • CDN = Calibration and Distribution Network
  • PAS = Power Amplifier Subsystem
  • RAS = Receive Amplifier Subsystem
     
  • FA = Feed Array
     
  • LDR = Large Deployable Reflector

(Source: after Carbone et al. 2021, p. 291)

Biomass Satellite Instrument Architecture

The Instrument

INES - Instrument Electronics System

Digital Control Unit - DCU

  • Command and control functions and to ensure communication to the spacecraft (and ultimately to ground)
  • Generates, after digital-to-analog conversion, the radio frequency (RF) transmit waveform in P-band, and receives and conditions the RF received signals before analog-to-digital conversion.
  • Generation of the digital chirps, the radar timing control signals and digital signal processing in receive

Calibration and Distribution Network - CDN

  • Major element of the internal calibration concept of the Instrument
  • Serves as RF signal switch-board, routing RF signals from/to the DCU to/from PAS and RAS, according to selectable nominal or calibration paths

The Instrument

INES - Instrument Electronics System

Power Amplifier Subsystem - PAS

  • Amplify the RF transmit chirp to a total nominal peak power level (mean power over the pulse duration) of 52.5 dBm, with precise phase settings

Receive  Amplifier Subsystem - RAS

  • Amplify, with low added noise, the received radar echoes and to route them to the CDN
    and from there to the DCU.

The Biomass receive amplifier subsystem

The Biomass Power amplifier subsystem

The Instrument

Instrument Architecture

Feed Array

  • Developed by Thales Alenia Space Italia SpA
  •  2x1 (elevation x azimuth) array
  • Main function: transduce the radio frequency (RF) signals from/to the INES to/from radiating patterns towards the LDR and ultimately the target on-ground

Feed Array

Biomass mock up being tested

The Instrument

Instrument Architecture

Large Deployable Reflector (LDR)

  • Satellite will carry a highly accurate and stable 12-metre aperture wire-mesh reflector, the LDR
  • During launch, the LDR is folded up under the payload fairing
  • Once the satellite reaches its orbit, the LDR will be unfolded
  • Reflector will transmit P-band data

LDR - 12-metre reflector in the cleanroom at L3Harris Technologies in Florida, USA, during a test of the deployment procedure

The Instrument

Antenna and Reflector

P-band radar piercing through forest canopy

The Instrument

SAR Acquisition Techniques of the sensor

Biomass P-band sensor offers major advances compared to all previous SAR missions because it will use the three techniques SAR Polarimetry (PolSAR), Polarimetric SAR Interferometry (Pol-InSAR) and SAR Tomography (TomoSAR) complementary to provide information on forest characteristics.

SAR Techniques offered by the Biomass Satellite

Mission Products

Level-1

Sub-surface geology
Terrain topography under vegetation
Glacier and ice sheets velocities

Forest biomass (AGB)

Forest height

Forest disturbance

Primary Mission Products

Secondary Mission Products

Raw Data

Level 1a   Single-look Complex Slant (SCS)

Level-2

Level-0

Level 1b     Detected Ground Multi-look (DGM)

(ESA, 2015, p. 60)

Mission Products

Level-0

"Time-ordered signals series for each of the four data products (i.e. four polarisations, each with in-phase and quadrature-phase channels) with appended ancillary data as necessary (e.g. datation, location, temperatures, pointing). (ESA, 2015, p. 60)"

Level-1

Level-1a

"Single-look Complex Slant (SCS) data, consisting of SAR focused data, internally radiometric calibrated, in zero-Doppler slant range-azimuth geometric projection, at natural geometric spacing with associated ancillary data. The ionospheric corrections shall also be implemented at this level. (ESA, 2015, p. 60)"

Mission Products

Level-1

Level-1b

"Level 1b products refer to the Detected Ground Multi-look (DGM) data, consisting of SAR focused data radiometrically calibrated, amplitude detected, projected in zero-Doppler ground range-azimuth onto a DEM, re-sampled at a regular spacing on ground with associated ancillary data. The ionospheric corrections shall also be implemented at this level. (ESA, 2015, p. 60)"

Products

(Source: Quegan et al., 2019, Table1, p. 47)

Level-2 Primary mission products

Level-2 product

Definition

Information Requirements

Forest Above Ground Biomass (AGB)

Defined as the dry weight of live organic matter above the soil, including stem, stump, branches, bark, seeds and foliage woody matter per unit area, expressed in t/ha (FAO, 2009). It does not include dead mass, litter and below-ground bio-mass. Biomass maps will be produced with a grid-size of 200 m × 200 m (4 ha).

  • 200 m resolution
  • RMSE of 20% or 10 t ha−1 for biomass
  • 1 biomass map every observation cycle
  • Global coverage of forested areas

Forest height

Defined as upper canopy height according to the H100 standard used in forestry expressed in m, mapped using the same 4 ha grid as for biomass. H100 is defined as the average height of the 100 tallest trees/ha (Philip, 1994).

  • 200 m resolution
  • Accuracy required is biome-dependent, but RMSE should be better than 30% for trees higher than 10 m
  • 1 height map every observation cycle
  • Global coverage of forested areas

Severe disturbance

An area where an intact patch of forest has been cleared, expressed as a binary classification of intact vs deforested or logged areas, with detection of forest loss being fixed at a given level of statistical significance.

  • 50 m resolution
  • Detection at a specified level of significance
  • 1 map every observation cycle
  • Global coverage of forested areas

Products

(Source: Quegan et al., 2019)

Following the definition of the secondary mission objectives, there will be map products treating topics like:

  • Sub-surface geology
  • Terrain topography under dense vegetation
  • Glacier and ice sheet velocities
  • Properties of the ionosphere

Algorithms

  • In the following, algorithms for generating the Level-2 Products of the Biomass mission will be introduced.
  • These are:
    • Above-Ground Biomass
    • Forest Height
    • Forest Disturbance

Level-2 product

Definition

Information Requirements

(Source: Quegan et al., 2019, Table1, p. 47)

Forest Above Ground Biomass (AGB)

Defined as the dry weight of live organic matter above the soil, including stem, stump, branches, bark, seeds and foliage woody matter per unit area, expressed in t/ha (FAO, 2009). It does not include dead mass, litter and below-ground bio-mass. Biomass maps will be produced with a grid-size of 200 m × 200 m (4 ha).

  • 200 m resolution
  • RMSE of 20% or 10 t ha−1 for biomass
  • 1 biomass map every observation cycle
  • Global coverage of forested areas

Forest Biomass (AGB)

Forest Biomass (AGB)

CASINO

Apply ground cancellation technique to isolate backscatter from forest canopy and minimize ground contributions

1

Describe forest canopy backscatter  dependence on AGB using a power law model (PLM)

2

Use PLM to estimate AGB for each polarization channel and combine them into weighted average

3

Soja et al. 2021

The Forest Biomass product (AGB) is based on the CASINO (CAnopy backscatter estimation, Subsampling, and Inhibited Nonlinear Optimization) algorithm, developed by  Soja et al. (2021). 

Forest Biomass (AGB)

CASINO

(Ulander et al., 2021)

CASINO based on the assumption:

  • There is a relation between radar backscatter and AGB
  • In polarimetric data, HV is the channel with the highest sensitivity to AGB
  • Relation between HV and AGB can be modelled using a power function

Graphs of some power functions

y=x^a

A power function is a function                        with:

 

where a is a constant number.

f:x\mapsto y

Forest Biomass (AGB)

Apply ground cancellation technique to isolate backscatter from forest canopy and minimize ground contributions

1

  • Forest canopy backscatter is estimated using a ground cancellation technique as introduced by d’Alessandro et al. (2020)
  • This technique decreases ground contributions to the backscattered signals arising from
    • Direct ground backscatter
    • Double-bounce interactions between ground and vegetation
    • Contributions are strongly affected by:
      • Ground topography
      • Soil moisture
      • Surface roughness
    • These contributions are independent of AGB and difficult to model
  • Ground cancellation technique:
    • Suppresses radar backscatter signals coming from the ground
    • Enhance the radar backscatter signals coming from a vegetation layer

(d'Alessandro et al., 2020, Soja et al., 2021)

Forest Biomass (AGB)

  • Ground cancellation technique:
    • Uses two co-registered interferometric SAR (InSAR) data sets:
      • Ground-steered SLC master image         (ground-steered = removed interferometric phase due to topography)
      • SLC slave image 
    • To form ground-cancelled SLC image          :
       

 

(d’Alessandro et al. 2020, Soja et al., 2021)

Apply ground cancellation technique to isolate backscatter from forest canopy and minimize ground contributions

1

(Source: d'Alessandro et al., 2020, p. 6414)

 

Relation between AGB and radar HV polarization power

Ground cancelled power [dB]

SLC power [dB]

 

S_S
S_M
S_{GC}=S_S - S_M
S_{GC}
\sigma^0_{cnp}=c\langle |s_{GCE}|^2 \rangle cos\psi

(1)

  •  
    • Normalization of ground-cancelled and equalized backscatter intensity 

 

|S_{GCE}|^2