Regen Network

Ecological State Protocols

B. Deriemaeker, G. Booman, T. Kazantsev, N. Assareh, M. Rada

Version 0.2
May 13, 2018

1
Contents
List of Figures 4

1 Data Sources 5
1.1 Remote Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1.1 Optical and Near Infrared (NIR) sensors . . . . . . . . . . 5
1.1.2 Radar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1.3 LiDAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2 Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.1 Unmanned Aerial Vehicles (Drones) . . . . . . . . . . . . 6
1.3 Satellites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.1 Landsat . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.2 MODIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.3 Sentinel-2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.4 Sentinel-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.5 Commercial satellites . . . . . . . . . . . . . . . . . . . . 7
1.3.6 Synthetic Aperture Radar (SAR) . . . . . . . . . . . . . . 8
1.4 GIS Datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 IoT Ecological Monitoring Sensors . . . . . . . . . . . . . . . . . 9
1.6 User Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2 Pending Protocols 10
2.1 Endangered Species Habitat Protocol . . . . . . . . . . . . . . . . 10
2.2 Pollinator Density Protocol . . . . . . . . . . . . . . . . . . . . . 11
2.3 Water Quality Protocol . . . . . . . . . . . . . . . . . . . . . . . 11
2.4 Habitat Quality Protocol . . . . . . . . . . . . . . . . . . . . . . 11
2.5 Erosion and Sediment Delivery Protocol . . . . . . . . . . . . . . 12
2.6 Urban Tree Protocol . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.7 Aquifer Carbon Sinks Protocol . . . . . . . . . . . . . . . . . . . 12
2.8 Air Quality Protocol . . . . . . . . . . . . . . . . . . . . . . . . . 13

3 Carbon Sequestration Indicators 13

3.1 Soil Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2 Biological Indicators . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.3 Above Ground Biomass (AGB) . . . . . . . . . . . . . . . . . . . 15
3.4 Surface Water Quality . . . . . . . . . . . . . . . . . . . . . . . . 15
3.5 Water and Nutrient Runoff . . . . . . . . . . . . . . . . . . . . . 16
3.6 Land Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.7 Biodiversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.8 Sediment Delivery and Soil Loss Due to Erosion . . . . . . . . . 18
3.9 Net Primary Productivity (NPP) . . . . . . . . . . . . . . . . . . 19
3.10 Covariables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.11 Climate: Rainfall . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.12 Soil Moisture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2
4 Main Indicators for Carbon Sequestration 21

References 24

3
List of Figures
1 Estimated Carbon Sequestration Potential [NM08] . . . . . . . . 17

4
1 Data Sources
1.1 Remote Sensing
There are three main types of sensors used in the protocols of Regen Network:

• Optical and Near Infrared

• Microwave (Radars)
• Laser scanners (LiDARs)

1.1.1 Optical and Near Infrared (NIR) sensors

Optical and NIR sensors are widely used in the assessment and analysis of a
wide range of characteristics of different ecosystems. This is due to a high
variety of reflective properties of plants under different growth stages and envi-
ronmental conditions in the spectral regions across different species. Optical (or
visual) reflectance (400-750 nm) of plants is determined mainly by pigments—
predominantly chlorophyll and carotenoids and to a lesser degree by minor pig-
ments (i.e. xanthophylls). NIR reflectance (750-1000 nm) is determined by
cellulose from plant cell walls. Going further to short-wave infrared (SWIR)
spectrum (1000-2500 nm), it becomes possible to detect water effects by water-
absorption bands at 1400 and 1900 nm. Optical and near infrared (NIR) sensors
can capture light in just a few or in many spectral bands. Dependent on the
quantity of spectral bands, the sensors are classified into multispectral (usu-
ally 2-5 wide bands) and hyperspectral (usually 100 or more narrow bands).
Specially-designed multispectral cameras are widely used. In contrast to regular
RGB cameras where the signal is split inside the sensor by filtering, professional
multispectral cameras usually have separate lenses and sensors for each spectral
channel. In hyperspectral sensors, the division of the signal into many narrow
bands is achieved by light diffraction by prisms or grates. Light intensity is
also divided between the channels, which causes a need for longer exposure or
a bigger sensor matrix. Due to such complexity, hyperspectral sensors are used
less often, although they are superior in many ways. RGB, multispectral, and
hyperspectral imagery all enable the acquisition of ecological information by
applying algorithms to the different spectral bands of the imagery.

1.1.2 Radar
Microwave sensors are also known as radar. They measure the strength of the
back-scattered signal from a surface. Synthetic-aperture radars (SARs) use the
motion of an antenna to create high-resolution 2-D and 3-D maps. SARs have
a lower signal-to-noise ratio compared to optical sensors but can be operated
under almost all weather and light conditions. SARs are used in cases when
optical data is not available, for example at night or above clouds.

5
1.1.3 LiDAR
LiDAR stands for Light Imaging, Detection, and Ranging. It emits and detects
a reflected laser beam (usually within 600-1000 nm) to reconstruct an object‘s
shape by the creation of point clouds. LiDAR sensors are not limited by the
same signal saturation for the estimation of Above Ground Biomass (AGB)
as optical and radar sensors. This is because LiDAR retrieves canopy height
from the distance measurements between the sensor and the target in contrast
to optical and radar, which correlate AGB with spectral reflectance or radar
backscatter signals. A high LiDAR point density allows for more ground re-
turns to be obtained through gaps in the canopy. In particular, airborne and
ground-based imaging LiDARs provide direct and very accurate measurements
of canopy height. LiDAR is used primarily on UAVs. There is no LiDAR
satellite in orbit at the present time, but some are in the development stage.

1.2 Platforms
1.2.1 Unmanned Aerial Vehicles (Drones)
Using drones for remote sensing has become extremely popular because of
progress in robotics. Relatively affordable amateur drones such as the DJI
Mavic Pro or 3DR Solo can be effective remote sensing instruments providing
hundred-hectare maps within hours. In addition to the RGB camera that the
drone is usually equipped with, manufacturers provide NIR-modified cameras
and multispectral cameras designed specially for drones. One of the simplest yet
most effective cameras is MapIR in which the blue channel is replaced with NIR
by means of filters. Different versions have either a Red-NIR or a Red-Green-
NIR channel combination to allow the mapping of various vegetation indices
utilizing these bands. Tetracam is one of the most experienced manufacturers
of airborne multispectral cameras. They started by producing cameras for pi-
loted planes and heavy drones. Recent models such as Tetracam ADC-Micro
can be mounted on a small UAV. Tetracam served as the prototype for several
novel developments such as Micasense and Sentera. These cameras also have
multi-lens construction and capture signals in separate channels. For instance,
Micasense Parrot Sequoia has Green, Red, Red-Edge, and NIR bands. In addi-
tion, it is equipped with a regular RGB sensor. Despite hyperspectral cameras
being much more complicated and expensive, there are some devices developed
specially for UAV—for instance, Gamaya with about 50 narrow spectral bands
in visual and NIR regions. Another approach for obtaining hyperspectral data
onboard a UAV is by using a single-beam spectrophotometer instead of a camera.
For example, Ocean Optics STS developers kit has been used for this purpose.
UAVs can be equipped with LiDAR—for example RIEGL VUX-1UAV or Yel-
lowScan Mapper. No commercial implementation of Synthetic Aperture Radar
(SAR) onboard a light drone has been reported. There are some prototypes
that hopefully will be mass-produced in future.

6
1.3 Satellites
1.3.1 Landsat
Landsat has circled the earth every 99 minutes and scanned the whole surface
of the earth every 16 days since 1972. Satellite images are freely available and
come in 30 m spatial resolution and 11 spectral bands.

1.3.2 MODIS
MODIS stands for Moderate Resolution Imaging Spectroradiometer and is a
sensor often used for ecosystem monitoring. The sensor is carried onboard the
TERRA satellite launched in 1999. MODIS has 36 spectral bands in visual and
infrared regions with spatial resolution between 250-1000 m. 250 m resolution
refers to the Red and NIR bands which are the most useful for plant study.
There are also 5 other bands in the Blue, Green and SWIR regions with 500
m resolution. With the 99-minute orbit of TERRA, MODIS maps the whole
Earth every 1 - 2 days. MODIS data is freely available.

1.3.3 Sentinel-2
In 2015, Sentinel-2a satellite was launched and was followed by the Sentinel-2b
in 2017. This constellation of satellites has been developed by the European
Space Agency for land monitoring with a strong focus on vegetation. Both
satellites are equipped with a Multi-Spectral Instrument (MSI) with 13 spectral
channels in the visible, NIR and SWIR regions. Spatial resolution in Blue,
Green, and NIR bands is equal to 10 m. There are 3 Red Edge bands, 1 narrow
NIR band, and 2 SWIR bands with 20 m resolution. The other 3 bands refer to
atmospheric study and have a 60 m resolution. By combining both satellites,
an update of freely available data occurs every 5 days.

1.3.4 Sentinel-1
In 2014 and 2016, Sentinel-1a and Sentinel 1b were launched. The satellites
carry a single C band SAR capable to deliver information about land cover
with 5 m spatial resolution every 16 days.

1.3.5 Commercial satellites

Besides those listed above, there are a number of commercial satellites with their
image data available for purchase. Spatial resolution of commercial satellites is
much higher (1 to 2 meters). For example, Pleiades HR 1A and Pleiades HR
1B constellation developed by AIRBUS Defence and Space provides data in
Blue-Green-Red-NIR bands with 2 m resolution and 26-day update. World-
View-3, owned by DigitalGlobe, updates every day and has a super-spectral,
high resolution camera with 29 spectral bands. IKONOS, QuickBird, GeoEye1,
and Worldview 1-2 and 4 are other commercial satellites operated by Digital
Globe. Planet Labs has over 192 satellites in orbit and refreshes its dataset

7
daily. They operate the PlanetScope, RapidEye, and SkySat constellations and
have a maximum resolution of 1 meter and 4 or 5 spectral bands.

1.3.6 Synthetic Aperture Radar (SAR)

Satellites with Synthetic Aperture Radar (SAR) orbit the Earth in a sun-
synchronous polar Low Earth Orbit (LEO) [UNA18]. Data acquisitions can
be made at any time of day or night and independent of cloud coverage, col-
lecting both amplitude and phase data. The SAR satellites have repeating
paths which, using two phase datasets for the same location at different times,
allows for interferometric SAR (InSAR) showing relative ground displacements
between the two datasets along the direction of the radar beam. The SAR satel-
lites operate at designated frequencies with L-band, C-band, and X-band being
the predominate wavelengths. Below is a chart of past, present, and projected
SAR satellite missions.
Various agencies support the different SAR missions:

• European Space Agency (ESA): ERS-1, ERS-2, Envisat, Sentinel-1

• Japan Aerospace Exploration Agency (JAXA): JERS-1, ALOS-1, ALOS-2
• Canadian Space Agency (CSA): Radarsat-1, Radarsat-2, Radarsat con-
stellation
• Deutsches Zentrum fr Luft- und Raumfahrt e.V. (DLR): TerraSAR-X,
TanDEM-X
• Indian Space Research Organization (ISRO): RISAT-1, NISAR (w/ NASA)
• Comisin Nacional de Actividades Espaciales: SAOCOM
• Italian Space Agency (ASI): COSMO-Skymed
• Instituto Nacional de Tcnica Aeroespacial (INTA): PAZ
• Korea Aerospace Research Institute (KARI): KOMPSat-5

• National Aeronautics and Space Administration (NASA): NISAR (w/

ISRO)

1.4 GIS Datasets

Open GIS datasets enable an integration of global, regional, and local data.
The inclusion of these information sets enables correlation of ecosystem data, en-
hanced classification, and identification of specific areas of interest. The datasets
that will be integrated fall into the following categories:

8
 • Hydrological data • Historical changes in global land
cover
• Conservation data
• Socioeconomic data
• Point-clouds and Digital Eleva-
tion Models ( DEM) • Coastal datasets

• Climate data • Benthic cover

• Air quality data • Different ecological classifica-

tions
• Exceptional biodiversity
hotspots • Watershed boundaries

• Regional and national soil • Forest cover and evapotranspira-

databases tion datasets

• Threatened species • Agriculture

1.5 IoT Ecological Monitoring Sensors

The sensors used in the different algorithms and protocols enable the acquisition
of real time ecological information in a variety of different categories. Internet
of Things (IoT) environmental monitoring is especially relevant and important
in the domains of air pollution, agriculture, soil, water quality and weather.
Below is a list of relevant sensors for Regen Network protocols and algorithm
development. Sensors allow cross correlation of different data sources and a
realtime data feed based on actual time-stamped ecosystem data.
List of the relevant sensors for Regen Network protocol and algorithm de-
velopment:

• Air pollution and emissions (in- • Agriculture

cluding forest fires):
– Nitrogen
– Oxygen – Soil methane emission
– NO2 – Soil spectrometer
– SO2 – Soil probe
– Rainfall
– Hydrocarbons
– Rainfall drop size and dis-
– CO tribution
– CO2 – Canopy temperature
– PM1, PM2.5, PM10 – Leaf wetness
– Ozone – Soil heat flux
– Hydrogen Sulfide – Stomatal conductance

– Ammonia • Soil
– Acetylene – Soil water

9
 – Soil carbon – Calcium Ion (Ca2+)
– Soil nitrate – Chloride Ion (Cl-)
– Soil potassium – Cupric Ion (Cu2+)
– Soil phosphorus – Fluoride Ion (F- )
– Soil aeration – Iodide Ion (I-)
– Soil respiration – Chlorophyll a
– Soil nitrogen – Blue-Green Algae
– Soil texture – Rhodamine WT
– Soil organic matter – Colorimeter pH
– Photosynthetic photon flux – Full wavelength spectrum
density spectrophotometer
– Gas flux – UV-VIS spectrophotometer
– Carbon exchange • Weather stations
– Color
– Wind speed and direction
– pH
– Rainfall distribution
• Water Quality – Temperature
– Bromide Ion (Br-) – Humidity

1.6 User Input

In addition to data collected from the technological sources above, there is also
a strong role for data generated through human observations. Organizations
such as Savory Institute are leading the way with dynamic and peer reviewed
methodologies for land stewards to measure ecological outcomes and use that
information to adapt management practices to increase both profit and ecosys-
tem health [Eyk17]. User Input also refers to data collected for the Data Quality
Protocol such as ratings of other data or algorithms, data collected through third
party apps, and Eco-Apps built on top of Regen Ledger. These data sources are
especially importable for the Ecological Supply Protocols (ESPs) so as to link
customers back to the source of the products they enjoy.

2 Pending Protocols
2.1 Endangered Species Habitat Protocol
As of 2017, the IUCN reported over 25,000 threatened species; those include
species listed as Critically Endangered (CR), Endangered (EN), or Vulnerable
(VU) [IUC17]. The endangered species habitat protocol would use satellite data,
GIS datasets, and smart IoT monitoring to track and map habitat for endan-
gered species, identify and predict their movements, and help direct conservation

10
efforts towards certain areas. The protocol could allow manufacturers to make
claims about not containing materials derived from either endangered species or
having directly impacted the habitat of endangered species. It can also inform
botanists, wildlife managers and governments on how to optimize the boundaries
of reserves and fund conservation efforts based on updated species, topography,
soil, precipitation, landcover and climatic data. Thus, the endangered species
habitat protocol could optimize to achieve the highest rate of preservation while
balancing social and economic needs with habitat and species conservation.

2.2 Pollinator Density Protocol

Globally, thirty-five percent of food crop production depends on pollinators.
Agricultural intensification jeopardizes wild bee communities and their stabi-
lizing effect on pollination services at the landscape scale [KVC+ 07]. For bees
to persist on a landscape, they need two things: suitable places to nest and
sufficient food (predicated by floral quality indicators) near their nesting sites
[inv18]. By pairing land use to nesting suitability and floral resources across
seasons, this protocol would be able to verify likeliness of pollinator density and
match sensor measurements with satellite predicted pollinator visitation rates.

2.3 Water Quality Protocol

Water pollution has many different causes; these include but are not limited to
sewage, nutrients, plastics, chemical, radioactive compounds, oil, and wastewa-
ter. By developing this water quality protocol, the quality of both surface and
groundwater could be tied to verifiable sensor input and satellite data. Water
quality monitoring systems could measure dissolved oxygen, temperature, tur-
bidity, Chlorophyll a (chl-a), and many more factors correlated to global water
pollution. Because of the high temporal coverage, Landsat imagery has been
widely used for chl-a assessment. For rivers and other cases that need more spec-
tral and spatial resolution, the multispectral ALOS/AVNIR-2, IKONOS, and
hyperspectral CASI and AISA imagery has successfully been used to determine
turbidity and Total Suspended Solids (TSS).

2.4 Habitat Quality Protocol

Habitat is of a high quality when it is relatively intact and has the structure
and function within the range of historic variability. Habitat quality depends
on a habitat?s proximity to human land uses and the intensity of these land
uses. Generally, habitat quality is degraded as the intensity of nearby land-use
increases [NVJS01] [McK02] [For03]. The habitat quality protocol would mea-
sure the distance to the anthropogenic impact, identify the location (upstream
or downstream), and the magnitude of the landscape threat. In order to assess
externalities from the surroundings, the choice of appropriate spatial units is
a key decision that will influence most aspects of the habitat quality protocol.
The watershed is a relevant eco-geographic spatial unit to analyze because it

11
possesses biogenic and hydroclimatic integrity as well as having a socioeconomic
and cultural identity. By using smart IoT monitoring combined with high res-
olution satellite imagery and GIS datasets, the habitat quality protocol could
help protect the world?s most sensitive habitats and allow decision makers to
incorporate this protocol into their legal frameworks.

2.5 Erosion and Sediment Delivery Protocol

We have lost a third of our arable soils to erosion. Human activity and related
land use change continues to be the primary cause of accelerated soil erosion,
which has substantial implications for nutrient and carbon cycling, land produc-
tivity and in turn, worldwide socio-economic conditions [BLMP17]. The erosion
and sediment delivery protocol would rely on the latest erosion and sediment
delivery models, GIS datasets, and real-time erosion sediment delivery monitor-
ing using satellite and UAV. It has never been more important to re-engineer
our agricultural system and this protocol will incentivize and reward ecological
regeneration that combats erosion and rebuilds our soils.

2.6 Urban Tree Protocol

In our modern cities, trees provide cleaner air, lower stress, cooler temperatures,
reduced flood risk, and increased biodiversity. While the importance of trees is
clear, they still get cut down in order to make space for city expansion. The
urban tree protocol could calculate the canopy projected areas of trees based
on high resolution satellite data and identify the cooling effect of trees using
UAV thermal imagery. The protocol could help tree advocacy groups to better
protect our urban forests and help governments prioritize and incentivize certain
areas for planting.

2.7 Aquifer Carbon Sinks Protocol

Massive aquifers beneath the world?s deserts might store more carbon than
all living plants [LWHT15]. When fossil fuels are burned, 30 percent of the
CO2 is trapped in the atmosphere, causing warming, 40 percent ends up in
the oceans, and the rest winds up elsewhere, mostly in plants which absorb it
through photosynthesis. Not all CO2 taken up by plants is used and converted
into sugars and oxygen. Scientists have been trying to figure out where all the
?leftover carbon? ends up in the planet?s system. One of these many places
might be beneath the world?s deserts. Chinese researchers sampled water from
an underground aquifer in the Tarim Basin and found it stores vast quantities
of carbon dioxide as a result of human activities, particularly farming. If the
same holds true for all the desert aquifers around the world, the trapped carbon
would amount to about a quarter more than the amount stored in living plants
on land. Previously, the carbon trapped in aquifers was thought to be negligible.
Clearly, there is significant carbon stored in aquifers that could be washed up in
the atmosphere if disturbed. Moreover, if the process identified in the new study

12
happens in the same way in other regions around the world where agriculture
and over-irrigation are present, the process could account for the storage of
about 1 trillion tons of carbon.
In sandy soil, when plants soak up CO2, some of it leaches into the ground.
Microbes that break up plant nutrients also contribute. Because conditions are
arid, desert farmers have to irrigate more; the extra water dissolves the CO2
and deposits it in the aquifer below. The carbon is then stored in these geo-
logical structures covered by thick layers of sand and may never return to the
atmosphere. Collectively, the world?s underground desert aquifers cover an area
the size of North America and may account for at least a portion of the ?miss-
ing carbon sink?. Knowing the precise location of these underground carbon
sinks will thus prove extremely important to improve carbon-stock models. If
the extent of the carbon trapping is really this large, then farmers could work
together with authorities to manage the carbon that goes underground.
The aquifer C-sinks protocol would calculate the area of arid and semiarid
aquifers that could be potentially farmed based on integrated remote sensing and
GIS. The Gravity Recovery and Climate Experiment (GRACE) mission should
provide accurate data of aquifer size and distribution when data are missing.
The protocol would help governments to prioritize and incentivize certain areas
for farming.

2.8 Air Quality Protocol

Air pollution is currently one of the most important environmental issues in
many regions around the world. Apart from its direct impacts on human health
and climate change, air pollutants can also adversely affect ecosystems, which
can indirectly impact human health and welfare through food and water con-
tamination. Satellite remote sensing is a good complement for ground-based
data and air quality models. While ground-based air quality sites provide the
most accurate measurements of air quality at a specific location, satellite im-
agery provides data with global, consistent coverage. Since variations in air
quality are complex in time and space, the air quality protocol would take into
account many different data sources and identify pollution hotspots, highlight
relationships between the spatial and temporal distribution of pollution, and al-
low ecological claims to be made about an increase in air quality and a decrease
in pollution.

3 Carbon Sequestration Indicators

3.1 Soil Carbon
The different techniques used for determining Soil Organic Carbon (SOC) based
on remote sensing employ the shape of the reflectance spectrum, for example by
using band depth analysis and principal component analysis. Alternatively, mul-
tivariate regression modeling such as Partial Least-Square Regression (PLSR)

13
and multiple linear regression can be used. By means of these methods, different
topsoil parameters are determined from the spectral signature contained in a
single imaging spectrometer image, where the various variables are represented
by different combinations of absorption features across the spectra.
Spaceborne imaging spectrometer data have not often been used for pre-
dicting soil organic carbon, but advanced spectral unmixing methods applied to
Hyperion data have obtained similar SOC fractions as those in field observations
[MDBSM11]. When mapping soil organic carbon on a large scale without ex-
tensive calibration with soil samples, a solution could be to use indices based on
spectral reflectance. The amount of SOC is then detected with reflectance spec-
troscopy based on the constituents of SOC: cellulose, starch, and lignin. Good
relations have been found for indices based on the visible part of the spectrum
(R2=0.80) and for the absorption features related to cellulose (around 2100
nm) (R2=0.81) [BSK+ 08]. Alternative approaches to determining exact soil
carbon include Regression kriging of predictor variables (Tasseled cap bright-
ness, greenness and wetness indices, NDVI, Vegetation Temp. Cond. Index
[VTCI], DEM, slopes, Compound Topography Index [CTI] and Leaf Area In-
dices for grasslands) [MKK+ 17], and the construction of soil indices based on
brightness, darkness, and greenness [QZT+ 17].

3.2 Biological Indicators

The relation between nutrient requirements of plants and nutrient availabil-
ity in soils can be used to derive soil attributes. Accordingly, the concept of
plant functional types (PFT) can be used to derive the specific type or group of
species that grow on typical soils. Functional types can be distinguished largely
on the basis of optical properties detectable by remote sensing. To fully utilize
the potential of remote sensing, data must be combined with ecological mod-
els linking structural, physiological, and phenological traits based on resource
constraints. Hence, PFT regulate or are regulated by ecosystem processes and
have discrete different functions within the ecosystems. Different PFT have a
particular distribution in relation to geography or environment, e.g. species of
ultramafic soils or acidophilus bog species. Therefore, PFT could be explained
by the DEM derived terrain variables which describe the landscape structure
[UG10].
In addition to PFT, Ellenberg indicator values can be used as a numerical
system to classify species? habitat niches and their peak occurrence along gra-
dients. By finding correlations of Ellenberg indicator values with morphological
or eco-physiological properties, it is possible to identify determinants of species
distributions with respect to environmental factors. Schmidtlein showed that
imaging spectroscopy can be used as a tool for mapping Ellenberg indicator
values for soil water content, soil pH, and soil fertility [Sch05]. The Ellen-
berg indicator values scale the flora of a region along gradients reflecting light,
temperature, moisture, soil pH, fertility, and salinity. In this way, the flora
can be used to monitor environmental change and thereby changes in the soil
[WMS+ 14].

14
3.3 Above Ground Biomass (AGB)
Accurate measurement and mapping of biomass is a critical component of the
proposed carbon sequestration protocol. The Intergovernmental Panel on Cli-
mate Change (IPCC) has listed five terrestrial ecosystem carbon pools involving
biomass: above-ground biomass, below-ground biomass, litter, woody debris,
and soil organic matter. Of these five, above-ground biomass (AGB) is the most
visible, dominant and dynamic pool of the terrestrial ecosystem, constituting
around 30 percent of the total terrestrial ecosystem carbon pool. While detailed
estimates of biomass are necessary for accurate carbon accounting (biomass as
dry weight is 50 percent carbon), there are few reliable estimation methods.
Biomass derived from field data measurements is the most accurate, but it is
not a practical approach for broad-scale assessments. Using remote sensing has
a key advantage; it can provide data over large areas at a fraction of the cost
associated with extensive sampling and enables access to inaccessible places.
Data from remote sensing satellites are available at various scales, from lo-
cal to global, and from a number of different platforms. Optical remote sensing
probably provides the best alternative to biomass estimation through field sam-
pling due to its global coverage, repetitiveness, and cost-effectiveness. Optical
Remote Sensing data is available from a number of platforms, such as IKONOS,
Quickbird, Worldview, SPOT, Sentinel, Landsat, and MODIS. New space-borne
sensors to be launched in the coming years will allow accurate measurements
of AGB in high biomass forests (less than 200 t ha-1) for the first time across
large areas [RVWL+ 17]. Recent developments in high resolution space-borne
and airborne satellite data have provided an opportunity to better estimate and
map AGB across different spatial and temporal scales.
The use of drones and UAVs has opened up avenues for super-fine resolu-
tion biomass estimation for targeted applications. Recent sensors, such as the
Worldview series, now provide meter level spatial resolution while Sentinel and
Landsat 8 provide free data for the whole world, opening up accessibility and
more applications of Remote Sensing data, including for biomass estimation.
Radar Remote Sensing has gained prominence for above-ground biomass esti-
mation in recent years due to its cloud penetration ability as well as detailed
vegetation structural information. Light Detection and Ranging (LiDAR) has
the ability to sample the vertical distribution of canopy and ground surfaces,
providing detailed structural information about vegetation. This leads to more
accurate estimations of basal area, crown size, tree height, and stem volume. A
number of studies have established strong correlations between LiDAR param-
eters and above-ground biomass [ABP+ 18] [KM17].

3.4 Surface Water Quality

By monitoring water quality parameters (i.e. suspended sediments [turbidity],
chlorophyll, and temperature), it is possible to assess wetland degradation or
changes in their carbon sequestration capacity [JFH12]. Optical and thermal
sensors can provide both spatial and temporal information needed to monitor

15
changes in water quality parameters. Integration of remotely-sensed data, GPS,
and GIS technologies provide a valuable tool for monitoring and assessing wa-
terways and wetlands [RZE03]. Remotely-sensed data can be used to create
a permanent geographically located database to provide a baseline for future
comparisons. The spectral characteristics of water and pollutants—which are
functions of the hydrological, biological and chemical characteristics of water—
are essential factors in the monitoring and assessment of water quality. The
different methodologies to interpret images and to evaluate the turbidity are
non-linear multiple regression, principal components analysis (PCA) and neu-
ral networks. Colored Dissolved Organic Matters (CDOM) in water can be
determined using hyperspectral imagery like EO-1/Hyperion, EO-1/ALI, and
ALOS/AVNIR-2. In addition, high-resolution spectroradiometer can be used
for in situ hyperspectral measurements for validation purposes [GMR16].

3.5 Water and Nutrient Runoff

Agriculture intensification and expansion causes soil deterioration, which implies
a lower capacity of soil carbon and nutrient retention, and thus higher export
loadings of nutrients from non-point sources to downstream surface water bodies
during and after storm events. Nutrients are then transported by surface and
subsurface runoff, contributing to non-point source pollutions of surface waters.
Water runoff volumes can be estimated by the Curve Number Method. The
Soil Conservation Service Curve Number (SCS-CN or NRCS-CN) method is
a simple, widely used, and efficient procedure for determining the expected
amount of runoff from rainfall in a particular area. Coupled with thematic
maps like LULC maps, soil types and climate maps, and hydrological tools in
GIS, the NRCS-CN method can be used to calculate the runoff transport and
accumulation through the watershed. CN is an empirical parameter used for
predicting direct runoff or infiltration from rainfall excess. Regardless of some
weaknesses, the CN method presents some advantages such as quantification
of the effect of land use changes on runoff formation [RH00]. The widespread
popularity of the NRCS-CN method attributes to the wide availability of the
required data and its simplicity. As result, the NRCS-CN method, which was
originally intended for the study of agricultural land, became a fundamental part
of hydrological practice and was adopted for application in different climates
and conditions. Moreover the CN method has been integrated into different
hydrological models, including CREAMS, FEST, EPIC, AGNPS, HEC-HMS,
and SWAT. Some of these models have been used used to calculate runoff in
GIS, like the AnnAGNPS or the SWAT (ArcSWAT).
Empirical or locally measured event mean concentrations (EMCs) for dif-
ferent water quality parameters such as BOD, COD, ammonia, nitrate, TKN,
hardness, TDS, TSS, chlorides, sulfates, phosphate, fluorides, and TC are often
available for many agricultural areas and can be used to calculate the amounts of
nutrients in the runoff after a storm event in GIS. Also, models like AnnAGNPS
and SWAT have shown good results when estimating nutrients in runoff.

16
3.6 Land Conversion
Land-cover change and management can alter the amount of organic carbon
stored in the soil and this in turn affects both soil fertility and atmospheric
carbon dioxide concentrations. There is empirical information that could be
used along with LULC maps to globally map and monitor changes in carbon
stocks due to land conversion.

Figure 1: Estimated Carbon Sequestration Potential [NM08]

3.7 Biodiversity
Recognizing the imperative need for biodiversity protection, the Convention
on Biological Diversity (CBD) has recently established new targets towards
2020, the so-called Aichi targets, and updated proposed sets of indicators to
quantitatively monitor the progress towards these targets. There are generally
three ways to measure ecosystem variables:

1. functional processes measured as fluxes, using in situ sensors,

17
 2. precise monitoring of composition, abundance, extent and change, is com-
monly done by in situ monitoring through habitat surveillance combined
with vegetation plots,

3. structural change monitoring using in situ combination with remote sens-

ing from space, aircraft, or drone [JSM+ 17].
Some of the important biodiversity indicators that are required for building
prediction models could be measured in enough detail by using high-resolution
remote sensing data collected with UAVs [SVN+ 18]. Remote sensing has been
increasingly contributing to timely, accurate, and cost-effective assessment of
biodiversity-related characteristics and functions during the last years. Novel
approaches integrating multi-sensor acquisitions can help to improve the un-
derstanding of the various environmental, physical, climatic, and human factors
influencing biodiversity, by monitoring spatial and temporal variations in species
composition.
Several biodiversity-related international projects have recently been imple-
mented, such as the 7th European Framework Programme (FP7) , MS.Monina
(Multi-scale Service for Monitoring Natura 2000 Habitats of European Com-
munity Interest), and BIO SOS (Biodiversity multi-Source monitoring System:
from Space to Species); the latter two focusing on biodiversity monitoring from
space [PMS15].

3.8 Sediment Delivery and Soil Loss Due to Erosion

The loss of soil leads to a decline in organic matter and nutrient content, the
breakdown of soil structure and a reduction of the available soil water stored,
which can lead to an enhanced risk of flooding and landslides in adjacent areas.
Nutrient and carbon cycling can be significantly altered by mobilization and
deposition of soil, considering eroded soil may lose 75 - 80 percent of its carbon
content, with consequent release of carbon to the atmosphere [eur18]. The total
land area subjected to human-induced soil degradation is estimated at about
2 billion hectares. From this, the land area affected by soil degradation due
to erosion is estimated at 1100 Mha by water erosion and 550 Mha by wind
erosion. Using conventional methods to assess soil erosion risk is expensive and
time-consuming.
The soil loss model, Revised Universal Soil Loss Equation (RUSLE), can
be integrated with GIS in order to estimate soil loss [EJBG+ 17] [And10]. The
RUSLE model can predict erosion potential on a cell-by-cell basis, which is ef-
fective when attempting to identify the spatial pattern of the soil loss present
within a large region. GIS can then be used to isolate and query these loca-
tions to identify the role of individual variables contributing to the observed
erosion potential value. The soil erosion assessment depends upon the regional
characteristics of the area, namely climate, soil condition, land use/land cover,
topography, and lithology. A DEM (Digital Elevation Model) is one of the es-
sential inputs required for soil erosion modelling. Other inputs include thematic

18
factor maps that can be generated in GIS from climate and geological data, and
landsat images.

3.9 Net Primary Productivity (NPP)

Net primary productivity (NPP) is an indicator of land productivity. NPP
represents how much carbon dioxide vegetation takes in during photosynthesis
minus how much carbon dioxide the plants release during respiration (metaboliz-
ing sugars and starches for energy). Land productivity can be assessed through
estimates of NPP (tDM/ha/yr), where a change in the absolute numerical value
may be positive or negative. NPP can be quantified using indices derived from
Earth observation data such as the Normalized Difference Vegetation Index
(NDVI) or Enhanced Vegetation Index (EVI). Dry Matter Productivity (DMP)
represents the overall growth rate or dry biomass increase of the vegetation, ex-
pressed in kilograms of dry matter per hectare per day (kgDM/ha/day). DMP
is directly related to NPP, but its units are customized for agro-statistical pur-
poses. According to Atjay et al. (1979), the efficiency of the conversion between
carbon and dry matter is on the average 0.45 gC/gDM. DMP images available at
Copernicus Hub, derive from SPOT-VGT (until December 2013) and PROBA-
V (from January 2014) imagery combined with (modeled) meteorological data
from ECMWF, are made available at 1km resolution and are updated every
10 days. Also new products are being prepared at 300m and 100m resolution.
Similar products could be built for moderate to high resolution images by fol-
lowing methods from the Copernicus Global Land Operations ?Vegetation and
Energy” [WWLF].

19
3.10 Covariables

Covariables Scale/type Source of infor- Indicators af-

mation fected
Climate: hu- Landscape/ Multitemporal se- Changes in area
midity, rainfall remote sensing ries data of rainfall of surface waters.
and pluvio- events, or accu- Changes in runoff
metric maps mulated rainfall volumes between
thematic maps years. Changes
from meteorolog- in hydric erosion
ical databases. gullies (increase in
Intensity of rainfall number or sizes)
events.
Soil moisture Landscape/ re- Topographic AGB, NPP, soil
mote sensing Wetness Index erosion, runoff
(TWI), Temper-
ature?Vegetation
Dryness Index
(TVDI), Vegeta-
tion Temperature
Condition Index
(VTCI)

3.11 Climate: Rainfall

When analyzing changes in ecosystems or species that are directly affected by
the availability of water, it might be necessary to perform very long-term mon-
itoring in order to detect changes due to other causes, like land management or
other anthropic impacts. Otherwise, the weather variable should be somehow
considered or subtracted from the data in order to ?denoise? the data and find
patterns. For example, changes in the area of surface water bodies could vary
between years according to both anthropic and climatic causes. A wetland that
shows erratic changes in size through a 10 year period could be indeed shrinking
or increasing its area due to drainage, irrigation, or channelization practices in
the area. However, some rainy years could mask the real trend so that it could
be detected after more than 15 years, which may be too late for mitigation.

3.12 Soil Moisture

Soil moisture influences many potential indicators for carbon monitoring like
soil erosion, AGB, NPP, and water and nutrient runoff. So, measuring and
monitoring soil moisture variations in time could help to denoise the measures
from these indicators and identify changes due to human actions in moderate
time periods.
Soil moisture changes can be monitored by remote sensing and GIS:

20
 • Soil Water Index (SWI) with a resolution of 1 km?tracks relative soil water
content
• Recently launched microwave SMOS (Soil Moisture and Ocean Salinity)
• Future satellite SMAP (Soil Moisture Active Passive), temporal resolution
of 3-5 days and 1 km resolution
• Surface energy balance models with ASTER and MODIS images as surface
variables
– Soil Energy BALance (SEBAL),
– Two-Source Energy Balance (TSEB)
– Surface Energy Balance System (SEBS)

4 Main Indicators for Carbon Sequestration

Indicator Predictors Data type Data sources

Above Combination of Radar, LiDAR, RapidEye, UAV
Ground different sources and optical re- datasets, World-
Biomass of spectral data, mote sensing view, SPOT,
(AGB) radar (SAR), Land imagery. Some- Sentinel, Landsat
Use/Land Cover times terrestrial and MODIS,
(LULC) maps, samplings are SAR platforms.
terrain sampling, needed.
Regression mod-
els, classifications.
Algorithms (e.g.
BIOMASAR).
Land Con- Changes in LULC Spaceborne sen- Landsat TM,
version maps, vegetation sors, hyperspec- ALOS/AVNIR-
indices, Forest tral imagery on 2, IKONOS .
Canopy Density airborne sensors EO-1/Hyperion,
model (FCD) EO-1/ALI, and
ALOS/AVNIR-2
Net Primary NPP (tDM/ha/yr) UAV or air- Sentinel, Land-
Production can be quantified borne remote sat, Modis
(NPP) using indices de- sensing spec-
rived from earth tral composites
observation data (i.e. vegetation
such as the Nor- indices)
malized Difference
Vegetation Index
(NDVI) or En-
hanced Vegetation
Index (EVI).

21
Biodiversity CBD Indicators LiDAR, UAV, Sentinel, Land-
Habitat rarity and optical re- sat, Modis,
(InVEST) Habitat mote sensing Copernicus, Big
Quality (InVEST) imagery, and Data
Habitat Connectiv- derived thematic
ity (FRAGSTATS) maps. Mapable
Habitat frag- data showing
mentation changes in the
(FRAGSTATS) distribution
Indicator Species. and/or richness
patterns of cer-
tain indicator
species under
monitoring pro-
grams can be
downloaded from
Big Data.
PFT/EllenbergLandsat seasonal UAV or air- DEMs, LULC,
composites and borne remote soil maps from
vegetation indices, sensing spec- various plat-
or from multi- tral composites forms.
source evidential (e.g. vegetation
reasoning (ER) indices), local
algorithm. calibration
Soil Erosion Compound To- Radar, LiDAR, DEMs, LULC,
pographic In- and Optical soil maps from
dex (CTI) or remote sens- various platforms
TWI, RUSLE ing. Thematic
or RUSLE2/GIS Maps for RUSLE
methodology. factors.
Water and Curve Number Radar and Opti- Thematic maps:
nutrient from the Soil Con- cal remote sens- DEMs, LULC
runoff servation Service of ing, local cali- maps, climate
the US (SCS-CN) bration from ter- maps, soil type
Nutrient runoff rain surveys, Em- maps. Empirical
estimates from pirical data, and equations and
models, e.g. Swat, equations. Hy- values (e.g. SCS-
AnnAGNPs. drological tools in CN equations,
GIS. EMCs values)

22
Soil Organic Multivariate re- Hyperspectral Updated satellite
Carbon gression modeling imagery from hyperspectral im-
such as Partial Orbital, UAV, or agery databases,
Least-Square Re- airborne remote drone imagery.
gression (PLSR) sensing
and multiple linear
regression based on
soil color attributes
(visible bands).
Indices based on
spectral reflectance
Regression krig-
ing of predictor
variables: Tasseled
cap brightness,
greenness and
wetness indices,
NDVI, Vegetation
Temp. Cond. In-
dex (VTCI), DEM,
slopes, Compound
Topography In-
dex (CTI)- Leaf
Area Indices for
grasslands
Surface wa- Turbidity, TSS, Spaceborne sen- Landsat TM,
ter quality Chl-a, CDOM sors, hyperspec- ALOS/AVNIR-
tral imagery on 2, IKONOS.
airborne sensors EO-1/Hyperion,
EO-1/ALI, and
ALOS/AVNIR-2

23
References
[ABP+ 18] Gregory P Asner, Philip G Brodrick, Christopher Philipson, Nico-
las R Vaughn, Roberta E Martin, David E Knapp, Joseph Heckler,
Luke J Evans, Tommaso Jucker, Benoit Goossens, et al. Mapped
aboveground carbon stocks to advance forest conservation and re-
covery in malaysian borneo. Biological Conservation, 217:289–310,
2018.
[And10] Linus Andersson. Soil loss estimation based on the usle/gis ap-
proach through small catchments-a minor field study in tunisia.
2010.
[BLMP17] Pasquale Borrelli, Emanuele Lugato, Luca Montanarella, and
Panos Panagos. A new assessment of soil loss due to wind erosion
in european agricultural soils using a quantitative spatially dis-
tributed modelling approach. Land Degradation & Development,
28(1):335–344, 2017.
[BSK+ 08] HM Bartholomeus, ME Schaepman, Lammert Kooistra, Antoine
Stevens, WB Hoogmoed, and OSP Spaargaren. Spectral re-
flectance based indices for soil organic carbon quantification. Geo-
derma, 145(1-2):28–36, 2008.
[EJBG+ 17] Aafaf El Jazouli, Ahmed Barakat, Abdessamad Ghafiri, Saida
El Moutaki, Abderrahim Ettaqy, and Rida Khellouk. Soil ero-
sion modeled with usle, gis, and remote sensing: a case study of
ikkour watershed in middle atlas (morocco). Geoscience Letters,
4(1):25, 2017.
[eur18] Agri-environmental indicator - soil erosion, 2018.
[Eyk17] Sarah Eykyn, January 2017.
[For03] Richard TT Forman. Road ecology: science and solutions. Island
Press, 2003.
[GMR16] Mohammad Haji Gholizadeh, Assefa M Melesse, and Lakshmi
Reddi. A comprehensive review on water quality parameters es-
timation using remote sensing techniques. Sensors, 16(8):1298,
2016.
[inv18] 2018.
[IUC17] IUCN. Summary statistics. http://www.iucnredlist.org/
about/summary-statistics, 2017.
[JFH12] Alison Curtin John Foster, Lisa Evans and Brydie Hill. The role
of wetlands in the carbon cycle. issues paper the role of wetlands
in the carbon cycle. www.environment.gov.au/water/wetlands/

24
 publications/issues-paper-wetlands-role-carbon-cycle,
2012.
[JSM+ 17] Rob HG Jongman, Andrew K Skidmore, CA Sander Mücher,
Robert GH Bunce, and Marc J Metzger. Global terrestrial ecosys-
tem observations: why, where, what and how? In The GEO hand-
book on biodiversity observation networks, pages 19–38. Springer,
2017.
[KM17] Lalit Kumar and Onisimo Mutanga. Remote sensing of above-
ground biomass, 2017.
[KVC+ 07] Alexandra-Maria Klein, Bernard E Vaissiere, James H Cane, In-
golf Steffan-Dewenter, Saul A Cunningham, Claire Kremen, and
Teja Tscharntke. Importance of pollinators in changing landscapes
for world crops. Proceedings of the Royal Society of London B: Bi-
ological Sciences, 274(1608):303–313, 2007.
[LWHT15] Yan Li, Yu-Gang Wang, RA Houghton, and Li-Song Tang. Hid-
den carbon sink beneath desert. Geophysical Research Letters,
42(14):5880–5887, 2015.
[McK02] Michael L McKinney. Urbanization, biodiversity, and conservation:
the impacts of urbanization on native species are poorly studied,
but educating a highly urbanized human population about these
impacts can greatly improve species conservation in all ecosystems.
BioScience, 52(10):883–890, 2002.
[MDBSM11] VL Mulder, S De Bruin, ME Schaepman, and TR Mayr. The use
of remote sensing in soil and terrain mapping?a review. Geoderma,
162(1-2):1–19, 2011.
[MKK+ 17] Arun Mondal, Deepak Khare, Sananda Kundu, Surajit Mondal,
Sandip Mukherjee, and Anirban Mukhopadhyay. Spatial soil or-
ganic carbon (soc) prediction by regression kriging using remote
sensing data. The Egyptian Journal of Remote Sensing and Space
Science, 20(1):61–70, 2017.
[NM08] Edward A Nater and Cheryl Miller. Terrestrial carbon sequestra-
tion monitoring networks and demonstration sites. part ii, report
to the minnesota department of natural resources from the min-
nesota terrestrial carbon sequestration initiative. 2008.

[NVJS01] Christian Nellemann, Ingunn Vistnes, Per Jordhøy, and Olav

Strand. Winter distribution of wild reindeer in relation to power
lines, roads and resorts. Biological Conservation, 101(3):351–360,
2001.

25
[PMS15] Zisis I Petrou, Ioannis Manakos, and Tania Stathaki. Remote sens-
ing for biodiversity monitoring: a review of methods for biodiver-
sity indicator extraction and assessment of progress towards inter-
national targets. Biodiversity and conservation, 24(10):2333–2363,
2015.
[QZT+ 17] Bingwen Qiu, Ke Zhang, Zhenghong Tang, Chongcheng Chen, and
Zhuangzhuang Wang. Developing soil indices based on brightness,
darkness, and greenness to improve land surface mapping accuracy.
GIScience & Remote Sensing, 54(5):759–777, 2017.
[RH00] DeAnne Rietz and Richard H Hawkins. Effects of land use on
runoff curve number. In Watershed Management and Operations
Management 2000, pages 1–11. 2000.
[RVWL+ 17] Pedro Rodrı́guez-Veiga, James Wheeler, Valentin Louis, Kevin
Tansey, and Heiko Balzter. Quantifying forest biomass carbon
stocks from space. Current Forestry Reports, 3(1):1–18, 2017.
[RZE03] Jerry C Ritchie, Paul V Zimba, and James H Everitt. Remote
sensing techniques to assess water quality. Photogrammetric Engi-
neering & Remote Sensing, 69(6):695–704, 2003.
[Sch05] Sebastian Schmidtlein. Imaging spectroscopy as a tool for mapping
ellenberg indicator values. Journal of Applied Ecology, 42(5):966–
974, 2005.

[SVN+ 18] Ninni Saarinen, Mikko Vastaranta, Roope Näsi, Tomi Rosnell,
Teemu Hakala, Eija Honkavaara, Michael A Wulder, Ville Luoma,
Antonio MG Tommaselli, Nilton N Imai, et al. Assessing biodiver-
sity in boreal forests with uav-based photogrammetric point clouds
and hyperspectral imaging. Remote Sensing, 10(2):338, 2018.
[UG10] Susan L Ustin and John A Gamon. Remote sensing of plant func-
tional types. New Phytologist, 186(4):795–816, 2010.

[UNA18] UNAVCO. Synthetic aperture radar (sar) satellites.

https://www.unavco.org/instrumentation/geophysical/
imaging/sar-satellites/sar-satellites.html, 2018.

[WMS+ 14] Hendrik Wulf, Titia Mulder, Michael E Schaepman, Armin Keller,
and P Jörg. Remote sensing of soils. University of Zurich: Zurich,
Switzerland, pages 1–71, 2014.

[WWLF] Nandika Tsendbazar WUR, Martin Herold WUR, Myroslava

Lesiv, and Steffen Fritz. Copernicus global land operations vege-
tation and energy.

26