Twenty per cent of the Earth’s total land area was degraded between 2000 and 2015. This assessment is based on the most extensive compilation of official data on this subject since world governments agreed to tackle the problem of land degradation in 1994, and then adopted a binding agreement – the United Nations Convention to Combat Desertification (UNCCD) – in 1996. In its capacity as custodian agency for SDG indicator 15.3.1 (the proportion of land that is degraded over total land area), the UNCCD acts as a catalyst for the national uptake of Earth Observation (EO) for land degradation monitoring and reporting.

Originally published by the UNCCD in 2017 and revised in 2021, the Good Practice Guidance (GPG) for SDG indicator 15.3.1 describes methods that can be used by all countries at any levels of capacity and technological development to measure the extent of degraded land in a globally-harmonized manner. As documented in the GPG, EO plays a critical role in the calculation of SDG Indicator 15.3.1 and its three sub-indicators. EO can remotely map land cover over time, identify changes in the trajectory and level of land productivity, and support the definition of spatially-explicit Soil Organic Carbon (SOC) baselines through the integration with data collected from in-situ monitoring systems and modelling. Major revisions implemented in version 2 of the GPG include: a greater focus on the identification of important degradation processes, and the use of an appropriate land cover legend to monitor them; an improvement of the statistical methods for assessment of the land productivity metrics (trends, state and performance) and their interpretation in terms of the severity of degradation; and a revision of the SOC sub-indicator in light of updates and new guidance from the IPCC.

Despite these advances, there are some limitations in the EO-based approach. Most notably while new high spatial resolution products, such as the ESA WorldCover global land cover product, are to be welcomed, they do not fulfill SDG reporting requirements in terms of time series length. For example, SDG indicator 15.3.1 uses a 16-year baseline period from 1 January 2000 to 31 December 2015. Thus, there is an urgent need to simulate high-resolution time series dating back to the year 2000, by harmonizing newly developed 10-30m datasets with moderate resolution (300m-1km) satellite data. In conclusion, operational production of contemporary very high-resolution land observations should be complemented by harmonization efforts with the more moderate resolution historical time series. The development of multi-decadal harmonized data products for the land surface would represent a significant contribution to improved reporting of SDG indicator 15.3.1 and would enable high spatial resolution baselines to assess progress towards Land Degradation Neutrality.

Background:
Human use of natural resources has not yet reached its peak, but everything points to a slowdown in the growth of agricultural productivity, a rapid depletion of ecosystem productive capacity and generation of environmental damage. Reversing ecosystem degradation through sustainable productive and inclusive growth contribute to Sustainable Development. The Sustainable Development Goals (SDGs) target 15.3 aims at protecting and restoring the territorial ecosystem to achieve a land degradation-neutral world by 2030. The indicator 15.3.1 (proportion of degraded land over total land area) has three sub-indicators: i) productivity, ii) land cover, iii) soil organic carbon. Earth observation datasets are the primary sources of data for deriving these sub-indicators. It requires selecting, querying and processing a substantial historical archive of data to derive the information for this indicator. To remove the complexities from this end, a module on the SEPAL platform (https://sepal.io) has been developed in compliance with the UNCCD Good Practice Guidance (GPG v2) to derive all the necessary statistics and maps for reporting by the relevant stakeholders.
Methods:
SEPAL is a cloud computing-based platform for geospatial data analysis with a focus on querying, processing and building advanced analytics using earth observation data. It contains all the major open-source geospatial tools and libraries. The SDG indicator module uses the python API of Google Earth Engine and SEPAL’s UI framework-based frontend. The module allows users to choose satellite data from Landsat 4, Landsat 5, Landsat 7, Landsat 8, Sentinel 2 along with the MODIS sensors to derive an annual time series of vegetation indices (NDVI, EVI2, MSVI2) as a measure of primary productivity. The inclusion of Landsat and Sentinel 2 data make it possible to investigate the local context at high resolution whereas MODIS is suitable for assessment at the national scale. The module also allows users to use in-country land cover data to compute the land cover change sub-indicator. To facilitate validation of the the results the module also provides functionality to extract time series data for a set of geographical locations.
Results:
The outputs from the module consist of maps (raster data at sensors’ resolution) of all the sub-indicators and productivity metrics; several useful diagrams e.g., Sankey plot to visualize land cover transition, time series plot with z-score to inspect the changes for a set of locations and bar diagrams to portray statistics by land cover categories. The module has been piloted for a set of area of interest like Angola, Bangladesh, Indonesia and Nigeria to monitor land degradation and restoration activities at local and national level. Comparing the results using different approaches and tools helps improving the results and their use for different applications, including at different spatial levels.
Conclusion:
One of the major advantages of the module is that a user does not need to set up a sophisticated computing environment to compute the indicator. It only requires a web browser and a decent internet connection to get the results. In addition, the land cover classification module of SEPAL, along with other Open Foris tools, can be used to prepare in-country land cover datasets, using country specific legends e.g. from the FAO Land Cover Legend Registry. The results can be used to report and monitor land degradation and restoration activities at local to national scales.

The indicator 15.3.1 was adopted by the United Nations Sustainable Development Goal 15 (Life on Land) to measure the Land Degradation Neutrality. It is based on three sub-indicators: (1) Trends in Land Cover, (2) Land Productivity and (3) Carbon Stocks. The second one, Land Productivity, refers to the total above-ground Net Primary Production and can capture changes in health and productive capacity of the land. It can be estimated using the Land Productivity Dynamics approach , which performs a combined assessment of the long term tendency of change of land productivity and its current level relative to homogeneous land areas. The Joint Research Centre (European Commision) has recently developed and published an open source tool, LPDynR , which implements the Land Productivity Dynamics approach for the calculation of the Land Productivity sub-indicator. LPDynR ingests vegetation-related indices derived from time series of remote sensed imagery and produces a final 5-class map, ranging from declining to increasing land productivity.
Here we present the Land Productivity Dynamics (LPD) indicator calculated for two very different areas of the globe, Europe and Sahel, and for two periods. The data sets integrate (a) a 15 years baseline observation period from 2000 to 2015 and (b) a 16 year reporting period from 2004 to 2019. The maps provide information on the direction, intensity and persistence of the trend and change of above-ground biomass – surface biomass - generated by photosynthetically active vegetation cover, widely equivalent to Gross Primary Production (GPP) of the global land surface.
The original Land Productivity Dynamics approach typically uses phenology-based products, such as standing biomass, season end date, season start date and season length. However, for reasons related to data availability, for these products we used the sum of the NDVI product for each year of the time-series (2000-2015 and 2004-2019) as the productivity variable. Additionally, the phenology variables were replaced by the Discrete Land Cover product in order to stratify the NDVI on the Land Cover Class. The data used for the products is the Copernicus Global Land long term Normalized Difference Vegetation Index , which is corrected for bidirectional effects, accumulated over a yearly period – annual SumNDVI. For the processing of the ‘status’ information, based on the local scaling approach, the Copernicus Global Land Cover (year 2019) has been used as underlying stratification.
For the European product, we can see that in Europe, compared to the baseline period, there is a general land productivity decreasing trend, maybe due to the drought conditions of the last few years. On the other hand, southern countries, such as Italy and Greece, are confirming a positive trend both in the baseline and reporting periods.
In the Sahel area the land productivity is very different between baseline and reporting periods and it is possible to see, in general, a worrying negative trend above all in the West Africa countries. In the eastern part of the Sahel it is possible to locate areas with positive trends, mostly related to local precipitation.
The Land Productivity Dynamics products are useful for identifying where vegetation, natural or managed, is affected by the stress of climate variability and land management. Where negative trends coincide with other issues, such as relevant land use change and decreasing carbon stocks, then a higher risk for land degradation is assigned under the SDG 15.3 indicator scheme. Adding other approaches that include more contextual data on land management and socio-economics, such as coincidence of decreasing land productivity with recurrent fire regimes, extensive forest loss, over-intensive or low-input agriculture, low income etc., a better estimation of land degradation is obtained.

The trends in carbon stocks (above and below ground) is one the three sub-indicators used to calculate the extent of land degradation for reporting on United Nations (UN) Sustainable Development Goal (SDG) Indicator 15.3.1: the proportion of land that is degraded over total land area. Currently, soil organic carbon (SOC) stock is the metric used to assess carbon stocks and its change in time is usually estimated using information on land cover change along with climate/land cover default change factors. However, the spatial modelling of SOC changes that do not necessarily involve land cover transitions is key to monitor progress towards LDN and to support the implementation of the most appropriate Sustainable Soil Management (SSM) practices. The rates of SOC sequestration under different land use and management practices can vary greatly depending on soil characteristics, topography and climate. Identifying which locations and agricultural systems have the most potential for raising SOC stocks is therefore important in the context of achieving LDN, particularly for informing decisions related to the mechanism for neutrality. This work describes the framework and model used to produce a SOC sequestration potential map for Turkey for forests, grasslands and croplands at 250 m resolution based on the national SOC map and Earth Observations and its integration in a nationally developed LDN Decision Support System (LDN DSS). Turkey is located at the intersection of Europe, Asia and Africa in an arid and semi-arid region with a rich biological diversity. It has been home to various civilisations since the first human settlements and agricultural practices applied throughout history have intensified degradation of Turkish soils. Due to its topography, water erosion is one of the primary issues, and the country is highly vulnerable to desertification and drought. Degradation seriously affects its forests, steppes and wetlands. However, Turkey has accumulated significant knowledge and experience on combating land degradation and has effectively applied soil conservation, afforestation, and rehabilitation activities, providing useful tools and methodologies scalable to the region. The country has also developed several systems and national products related to erosion and land degradation useful to effectively mainstream LDN into monitoring and planning processes, such as its national SOC model. Under the FAO-GEF project “Contributing to LDN Target Setting by Demonstrating the LDN Approach in the Upper Sakarya Basin for Scaling up at National Level” we estimated the potential SOC sequestration based on the process-based RothC model to analyze several scenarios of C inputs and estimate future SOC stocks, and the biophysical conditions of SOC saturation. RothC model also enabled the identification of the C inputs necessary for each land use to achieve a Carbon neutral state (ie. consonant with the LDN principle of maintaining or improving the delivery of ecosystem services). The maps provided information on how alternative C input scenarios related to soil sustainable management techniques may affect the soil C stocks. The resulting maps were integrated to other LDN and EO indicators in a national LDN DSS to facilitate its use by decision makers and the most suitable sites for different types of interventions to reverse, reduce and avoid land degradation were identified and prioritized via multicriteria analysis. Our results show that modelling and mapping SOC sequestration potential provides crucial information for identifying intervention areas and setting priorities for public and private policies that can advance land degradation neutrality at different levels of governance (catchment to national) and contribute to a carbon neutral future. It can also help identify options and pathways for interventions (e.g. to monitor changes in SOC that occur in areas with no land cover change, undergoing degradation processes due to unsustainable and management practices) and to evaluate success of interventions (e.g. area that are improving due to implementation of better management practices). Many countries are currently developing SOCseq maps within the Global Soil Partnership GSOCseq initiative to estimate the potentials of soils for sequestering SOC. The integration of Soil Organic Carbon Sequestration Potential maps within the Land Degradation Neutrality monitoring framework can contribute to achieve SDG 15.3.

Continental scale land degradation assessment supporting nature restoration targets under the EU Green Deal.



Eva Ivits (European Environment Agency), Jaume Fons Esteve, Mirko Gregor, Gundula Prokop, Manuel Loehnertz, Roger Milego, Emanuele Mancosu

Land is an essential natural resource, both for the survival and prosperity of people and for the maintenance of all terrestrial ecosystems. Current land use trends and land management focused on increasing production together with climate change create strong environmental pressures. Together they have become the key drivers for the decline of nature, that is pressured through land take, intensive forest management and agriculture, droughts, wildfires, etc. Land use change and the increasing impact of droughts, fires and floods leads to areas of less biodiversity value and of less carbon sink potential, including the loss of fertile soils. The limits on land resources are finite, while human demands on them are not. Increased demand, or pressure on land resources, result in declining crop production, degradation of land quality and quantity, and competition for land.

The EU Nature Restoration Plan, as part of the EU Biodiversity Strategy for 2030, highlights the urgent need for restoration of degraded ecosystems that are in a poor state, at the same time as reducing the pressures on habitats and species, and ensuring all use of ecosystems is sustainable. The Plan sets several priorities as bringing nature back to agricultural land, addressing land take and restoring soil ecosystems, increasing the quantity of forests and improving their health and resilience, or restoring freshwater ecosystems addressing riparian zones, among others. This work responds to the need to understand land processes better, that lead to land degradation, and corresponding worsening of ecosystem condition.

We use the concept of degraded land as “land in a state of persistent loss of biodiversity, ecosystem functions and services” according to the definition of IPBES. One of the key principles of the methodological concept is to use all available and relevant data, information and knowledge, both from Earth Observation or of statistical/administrative nature. Several land degradation processes and types are to be considered, e.g. land take, fragmentation, degraded floodplains, deforestation, drained wetlands and burnt areas. In that context, and over the last years, the EEA have processed and integrated a number of geospatial datasets in the EEA Integrated Data Platform (IDP), addressing land use change, land take, floodplains, fragmentation, droughts, and others. The platform uses all available data dimensions and data cubes in EEA´s IDP, in a series of interactive dashboards using technologies such as ArcGIS online (Operations Dashboard or Insights), Tableau, and others.

Unsustainable land management practices, which include land cover/use changes and over-exploitation, are combined with climate-induced impacts, such as wildfires or droughts and with soil degradation processes. The platform allows users to map, explore and plot land degradation processes pressuring specific regions (e.g. 10km grid and NUTS3 regions). Future work with concentrate on increasing the spatial resolution of the platforms towards Sentinels.

The GEO Land Degradation Neutrality Initiative (GEO-LDN) supports global efforts to reduce land degradation. 128 countries have set targets to achieve LDN, and each of these countries faces unique land degradation challenges. Differences in the geography, scale and drivers of land degradation mean that a diversity of datasets are required to meet the needs of these countries. Analytical tools to support the achievement of LDN must be able to accommodate these diverse datasets and yet be readily usable by countries with significant differences in their capacity to measure and monitor land degradation. To accommodate the wide variety of end users and use cases, GEO-LDN, in close collaboration with CEOS, is collating a federated suite of open-source and/or open licensed interoperable datasets and analytical tools. Informed by surveys of a wide cross section of end users, GEO-LDN has supported the development of decision trees to guide the identification of data suitability for LDN analysis, methods guidance for the use of these datasets to report on LDN and SDG Indicator 15.3.1, analytical tools that draw on these datasets and methods to measure and monitor land degradation , and planning tools to help users avoid land degradation while assessing their land degradation neutrality status. By engaging with reporting and land use planning authorities, GEO-LDN will continue to identify gaps and needs in the existing land degradation data and analytical offerings, and support the harmonisation of these tools or their development where required. This presentation will demonstrate the interoperability of key datasets and analytical tools currently supported by GEO-LDN, and their application in the land degradation assessment and planning process, including examples from pilot sites in several countries. This presentation will also describe plans for the continued engagement with end users and data providers, and how feedback from these communities will be used to guide the continued development of datasets and tools to improve the consistency, accuracy and simplicity of LDN assessment and planning across the globe. The datasets and tools developed under this strategy will be used in capacity building efforts, including an international postgraduate programme being implemented through the University of Energy and Natural Resources, Ghana, and their program partners.