Navigation öffnen

Uniting expert data from different disciplines

GRAS took strong efforts to carry together scientifically approved, reliable and reputable data from alll relevant sustainability sectors. Below please find all information about how the data was assessed and processed.

Biodiversity Methodology

1. Assessment and selection of databases

The biodiversity layers in the maps of the Countries show protected areas and biodiverse areas and are compiled from global, inter-regional (European Union), national and sub-national databases. GRAS researched and compiled the best available data sources for each country. The integration of complementary data sources ensures that a complete picture of biodiversity is painted.
The databases have been selected according to the criteria content, data source, data quality, data format and availability of data. Every database must fulfil the following minimum requirements:

  • Content: The database shows protected areas, i.e. areas designated by international or national law for nature protection purposes by a governmental authority, or it shows areas that have been identified as biodiverse and/or vulnerable by NGOs, Institutes or Organizations. The database is also eligible if it contains information about potential No Go areas and/or Risk Areas according to the EU Renewable Energy Directive (e.g. primary forests, highly biodiverse grassland).
  • Data source: The database is provided by the respective government authority responsible for protected areas. The database is also eligible if it is provided by renowned international NGO, research institutes, UN organization or if it was published in a scientific article.
  • Data quality: The database is up-to-date, accurate and it is updated regularly. Additionally the data is displayed in clearly defined polygons.
  • Data format: The data is provided in GIS format or can be transferred into GIS format with justifiable effort.
  • Availability of data: GRAS has the permission to implement and depict the data. The permission was either directly given to GRAS by the database provider or it is defined in the terms of use of the database.

2. Data structuring

All data files are maintained in the original format and structure provided in the respective database. On the one hand this is because of reasons concerning the terms of use and on the other hand this is to always ensure the correct reproduction of the data. However, the data has been filtered: Not all of the original attributes are displayed when klicking on a polygon but just the relevant ones. Furthermore, the names of the attribute categories have been harmonized and are all displayed in English.

3. Classification of No Go Areas and Risk Areas

The biodiversity data integrated in GRAS maps of the different Countries has a broad variety: It comprises protected areas with different designations and objectives, areas identified by NGOs as being biodiverse as well as areas with a certain land cover e.g. primary forests. In order to facilitate interpretation of the biodiversity data, the areas have been classified and compiled into so called „No Go Areas“ and „Risk Areas“. They are displayed as separate layers summarizing all the biodiversity and carbon stock information in the respective country.

The classification is based on the specifications made in article 17 of the EU Renewable Energy Directive (RED). The EU Renewable Energy Directive was originally set up for the biofuels and bioliquids market. Nevertheless, the land-related sustainability requirements are also applicable in the food, feed and chemical market. The sustainability requirements of the RED are the baseline for many certification systems and standards set up by companies. That is why they were used in GRAS as an orientation to classify the areas of various different biodiversity databases into No Go Areas and Risk Areas. The Renewable Energy Directive (RED) defines areas with high priority for nature conservation that are not suitable for sustainable biomass production:
Art. 17, Directive 2009/28/EC: "Biofuels and bioliquids (...) shall not be made from raw material obtained from land with high biodiversity value, namely (...) Primary Forests and other wooded land (...); highly biodiverse grassland (...).
Biofuels and bioliquids (...) shall not be made from raw material obtained from land with high carbon stock, namely (...) wetlands (...); continuously forested areas (...); land spanning more than one hectare with trees higher than five metres and a canopy cover of between 10 % and 30 % (...).
Biofuels and bioliquids (...) shall not be made from raw material obtained from land that was peatland in January 2008, unless evidence is provided that the cultivation and harvesting of that raw material does not involve drainage of previously undrained soil."

No Go Areas, as defined in GRAS, have a high conservation priority and are not suitable for a sustainable biomass production. They are identified using the following indicators:

  • Explicitly named in RED (e.g. Ramsar Sites)
  • Designation/definition of area is very similar to RED definition of No Go Area: e.g. definition of primary forests from Intact Forest Landscapes and from RED are very similar
  • IUCN category Ia, Ib, II or III
  • Protection goal of area is to perpetuate the area in its pristine state and allow for an undisturbed development of nature: e.g. National Parks

In contrast, Risk Areas are areas which do not clearly classify as No Go Areas but which are protected, potentially biodiverse and / or carbon stock intensive. Especially, areas identified in land cover maps are rather classified as risk areas because land cover maps can be biased. The following indicators are used to classify Risk Areas:

  • IUCN category, IV, V or VI
  • Area is grassland, peatland or forested area but the exact characteristics of these areas (e.g. biodiversity, management, canopy cover) are not specified
  • Area has a high carbon stock (carbon stock maps are based on land cover maps)

Wherever possible, the classification of protected areas was done with the help of the IUCN categories. IUCN protected area management categories classify protected areas according to their management objectives. The categories are recognized by the United Nations and by many national governments as the global standard for defining and recording protected areas. The IUCN categories are defined as follows:

  • IUCN categories Ia, Ib, II, III: Area is designated for nature protection purposes and is strictly protected: only scientific, educational, maintenance and military activities are allowed; agricultural and industrial activities are not allowed
  • IUCN categories IV, V, VI: Area is designated for nature protection purposes but certain regulated economic and agricultural activities are allowed by law or can be allowed by the responsible authority

Carbon Stock Methodology

1. Calculation

A complete carbon map shows the carbon stored in all of the biomass and soil. For selected Countries, GRAS includes its own carbon layers in displaying the full carbon stock in biomass and soil based on scientifically verified methodology and transparent data sources. For the calculation of a carbon stock associated with a particular land use, the carbon stock stored in biomass and the carbon stock stored in the soil need to be summarized. These two carbon stocks are calculated in separated calculation steps according to the IPCC 2006.

2. Carbon in biomass

The basis for determining the carbon stored in biomass is a land cover map. By associating a typical biomass carbon value with each land cover class in the land cover map, a biomass carbon map can be derived. The typical carbon values were taken from the relevant scientific literature and are transparent for each country.  The carbon values include carbon stored in above and below ground living vegetation.

3. Carbon in soil

Carbon stored in the soil concerns the stock of carbon in the soil that is not part of the living biomass, which means all carbon apart from living roots. The GRAS carbon layers are based on the Tier 1 approach of the IPCC 2006 which models soil carbon stocks influenced by climate, soil type, land use, management practices and inputs, which can increase or decrease the carbon content in the soil.
As a first step a suitable soil map needs to be chosen. For GRAS the FAO Harmonized World Soil Database (HWSD) generated by IIASA is used. It shows different soil types in a consistent categorization for the whole world. This first step includes the generation of a map of soil carbon as if the whole area were under natural land cover. This is done in the same fashion as for the biomass carbon map, namely, by associating carbon values to each soil category in the soil map. Carbon values were taken from the IPCC 2006 and show the carbon content in the first 30 centimeters of the soil. As most peatland soils possess carbon stocks in much thicker soil layers, they are not included in this map but shown in separate layers.
As a second step, to determine the actual carbon stock stored in the soil, the carbon stock under natural land cover must be adjusted with the soil use factors that correspond to the current land use. The factors indicate how much the land use type, the management practice and the inputs change the carbon stock stored in the soil compared to a natural land cover.
In order to determine which of these factors apply, the land cover map is used. Thus, for each land cover class, the land use factor, the typical management regime applied for a particular land use in the region and the corresponding typical input are defined. The corresponding values for the factors are taken from the IPCC 2006.
As a last step, to determine the actual carbon stock stored in the soil, the carbon values in the soil under natural vegetation are multiplied with the corresponding soil factors.

4. The total carbon map

The final carbon map is calculated by overlaying and summing the carbon stocks stored in total above and below ground living vegetation map with the actual carbon stock stored in the soil map. The result is a map of the total carbon content of a region.

Land Use Change Methodology

1. Processing of Remote Sensing images

Our methodology is based on detecting land use change (LUC) from MODIS (Moderate-resolution Imaging Spectroradiometer) greenness index time series (resolution 250m x 250m). The index is called Enhanced Vegetation Index (EVI) with values range from 0 to 1, the higher the values the more green the soil cover. From one EVI image we can differentiate between bare soil and green cover. From more than 300 EVI images since the year 2000 in GRAS we can differentiate among the type of green cover, see the history of the land, and indeed detect LUC. Tropical forest would look like a calm sea waves with minor changes around EVI value of 0.6. Deforestation would appear as a clear change in those waves with a drop of EVI to a value below 0.2. We can actually see what happened and when it happened.

GRAS uses that concept to detect land use change after the cut-of-date 1.1.2008. The following MODIS images are used in our analysis (from February 2000 till July 2014):

  • Enhanced vegetation index;
  • Composite day of the year;
  • and Vegetation index quality

2. Preprocessing and Reprojecting

Preprocessing is done on two steps:

  • Exporting relevant layers from MODIS Hierarchical Data Format (HDF) container
  • Reprojecting the layers’ geographic coordinates system from Sinusoidal, Lambert Azimuthal Equal-Area to WGS84

GRAS preprocess the images using the development framework R and GDAL library on Apple Mac Pro 2,7GHz Intel 12-Core Xeon E5; 68GB RAM; 1 TB GB Flash-Hard drive.

Time Series Smoothing

Smoothing is an important step to exclude outliers, realign EVI data (using composite day of the year layer) and to exclude bad values (using vegetation index quality layer). Time series smoothing is done using modified Whittaker smoothing algorithm. Visual land use change detection is possible from smoothed time series.

3. Automatic Land Use Change Detection

German Aerospace Center (DLR) experts have developed an algorithm for GRAS to detected land use change automatically from smoothed EVI time series. LUC detection algorithm is based on the following three assumptions: 

  • Agricultural crops reach maximum EVI value (corresponds to maximum leave area index) at certain time of the season;
  • The green cover falls down fast from the maximum in the following few weeks;
  • Agricultural season for annual crops season is relatively short.

These assumptions are translated into a fuzzy logic approach to detect whether the land was in fact cropped before 2008 or not. If a parcel GRAS algorithm found crop EVI peak after 2008 and no crop EVI peak before 2008, it will be considered that LUC took place.

The algorithm has been tested in several countries. In Chaco, Argentina, for instance, the agreement between visual and automatic LUC detection was 97%. Testing and improving GRAS algorithm will continue to reach a better output and more reliable information for our users.

Social Indices Methodology

1. Used Data

Working and general living, as well as health and education play a major role in the sustainable production of agriculture and forestry materials. The selection of the social indices to be introduced inside the GRAS Tool was made according to the social sustainability issues mentioned in the Renewable Energy Directive. These issues are the consequences of the increased production of biofuel on hunger, working conditions (e.g. loan, slavery, child work), human development and governance as an indicator for political stability and efficiency (e.g. land use rights, absence of violence, corruption). Further, the social consequences of environmental issues related to increasing agricultural production (water issues, use of pesticides, subsidies) were taken into consideration. In addition, the RED is specifically mentioning Conventions of the International Labour Organization, which shall be ratified and implemented by countries that are a significant source of raw material for biofuels consumed within the Community. According to the previously mentioned topics and RED specifications, social indices were selected and displayed inside the GRAS Tool. To mention one example, The Global Hunger Index is an indication for food security, which is a major issue in rural populations and farmers.

All social sources on national level get comprised in an overall GRAS Factor Social that is valid for the whole country. The overall GRAS Factor Social can be displayed as a global ranking of social sustainability risks in countries. For the assessment of sustainability Risk Areas in the GRAS-Index the according overall country GRAS Factor Social is integrated in the calculation. As a convention the overall index is issued on a scale from zero to 1 where a zero marks a perfect sustainability score (a “clean” record) and 1 the worst possible score. In the recent version of the GRAS Factor Social the following indexes were considered:

  • Global Hunger Index (GHI)
  • World Governance Indicators (WGI)
  • Human Development Index (HDI)
  • Global Slavery Index (GSI)
  • EPI Agricultural Subsidies (EPI AS)
  • EPI Pesticide Regulation (EPI PR)
  • EPI Water Resources (EPI WR)
  • UNICEF Access to Drinking Water (UNICEF WA)
  • UNICEF Access to Sanitation (UNICEF WS)

As an additional information the ratification of UN International Labor Organisation (ILO) Core Labor Standards is displayed within the maps tool. However this information is not considered for the overall social factor calculation.

2. Adjustment and weighting of the indices for the overall calculation

The first step in the creation of the GRAS Factor Social is to adjust each index on a 0 to 1 scale. Thus, 0 will be the best result and 1 the worst for all applied indices.

Additionally, a unifying stretching is applied to the adjusted Index values. The following is an example for the Global Hunger Index. The highest value scored by a country in 2014 is 35.6, resulting in a maximum GHI adj of 0.36. However, other indices imply a full minimum to maximum range from zero to one. Thus, GHI might be underrepresented in an overall comparison and implementation in the GRAS Factor Social. With regard to that matter the maximum value of each added adjusted index is stretched. In the case of the GHI example, the value of 35.6 for Country X is adjusted and then divided by its rounded up maximum value:

Afterwards, all indices are summed together for each country. For the calculation of the overall GRAS Factor Social, the used indices are weighted with regard to their impact on the sustainability issues mentioned in the Renewable Energy Directive. These issues are the effects of the increasing production of biomass for biofuels on hunger (represented in GRAS by the Global Hunger Index), working conditions (represented in GRAS by the Global Slavery Index), human development (represented in GRAS by the Human Development Index) and governance (represented in GRAS by the World Governance Indicators) as an indicator for political stability and efficiency (e.g. corruption, land use rights, etc.). In addition, the social consequences of environmental problems (represented in GRAS by the Environmental Performance Index), which are related to the increasing agricultural production (e.g. water problems, use of pesticides, agricultural subventions), were taken into account. For the calculation, all the following indices are weighted equally (0,2): Global Hunger Index (GHI), World Governance Indicators (WGI), Human Development Index (HDI), Global Slavery Index (GSI) and Environmental Performance Index (EPI). The EPI is based on a mix of data from international sources (e.g. UNICEF, which is displayed inside the GRAS Tool as source of the original data) and own indicators. Out of the 19 EPI indicators the five most important (EPI Agricultural Subsidies, EPI Pesticide Regulation, EPI Water Resources, UNICEF Access to Drinking Water and UNICEF Access to Sanitation) were included in the GRAS Factor Social and were weighted equally (0,04).

In the table below, the set weighting factors as well as the adjustments for each index is displayed:

Index Weighting Factor Adjustment
Global Hunger Index (GHI) 0.2 GHIadj = GHI * 0.01
World Governance Indicators (WGI) 0.2 WGIadj=1 - 0.2 * (WGI + 2.5)
Human Development Index (HDI) 0.2 HDIadj=1 - HDI
Global Slavery Index (GSI) 0.2 GSIadj=GSI * 0.01
EPI Agricultural Subsidies (EPI AS) 0.04 EPI ASadj=1 - EPI AS * 0.01
EPI Pesticide Regulation (EPI PR) 0.04 EPI PRadj = 1 - EPI PR * 0.01
EPI Water Resources (EPI WR) 0.04 EPI WRadj = 1 - EPI WR * 0.01
UNICEF Access to Drinking Water (UNICEF WA) 0.04 UNICEF WAadj = 1 - UNICEF WA * 0.01
UNICEF Access to Sanitation (UNICEF WS) 0.04 UNICEF WSadj = 1 - UNICEF WS * 0.01