Search

The global spatial extent of croplands is a crucial input to global and regional agricultural monitoring and modeling systems. Although many new remotely-sensed products are now appearing due to recent advances in the spatial and temporal resolution of satellite sensors, there are still issues with these products that are related to the definition of cropland used and the accuracies of these maps, particularly when examined spatially. To address the needs of the agricultural monitoring community, here we have created a hybrid map of global cropland extent at a 500 m resolution by fusing two of the latest high resolution remotely-sensed cropland products: the European Space Agency's WorldCereal and the cropland layer from the University of Maryland. We aggregated the two products to a common resolution of 500 m to produce percentage cropland and compared them spatially, calculating two kinds of disagreement: density disagreement, where the two maps differ by more than 80%, and absence-presence of cropland disagreement, where one map indicates the presence of cropland while the other does not. Based on these disagreements, we selected continuous areas of disagreement, referred to in the paper as hotspots of disagreement, for manual correction by experts using the Geo-Wiki land cover application. The hybrid map was then validated using a stratified random sample based on the disagreement layer, where the sample was visually interpreted by a different set of experts using Geo-Wiki. The results show that the hybrid product improves upon the overall accuracy statistics in the areas where the underlying cropland layer from the University of Maryland was improved with the WorldCereal product, but more importantly, it represents an improved spatially explicit cropland mask for early warning and food security assessment purposes.

The last few decades have seen an explosion in the availability of remotely sensed and geospatial big data, which are defined by the 3 Vs: a large volume of data; a variety of different forms of data; and the rapid velocity of data arrival. The term big data is particularly applicable to remote sensing. The opening of the Landsat archive, the spatially and temporally rich data now available from the Sentinel satellites, and the proliferation of small satellites photographing the Earth all provide new opportunities for characterizing and monitoring the Earth’s surface. New sources of geospatial big data (as well as regular geospatial data) can also benefit the mapping and monitoring of land cover and land use. These include data from authoritative sources, e.g., data from official censuses and surveys, as well as data generated by citizens, both actively and passively. Citizen science and volunteered geographic information can provide data on land cover and use through initiatives such as OpenStreetMap (OSM), Geo-Wiki, and many other projects that involve volunteers monitoring the environment or landscape features. Mobile phones and low-cost sensors can provide new streams of information through mobile apps that facilitate data collection or that collect information in the background, as well as a variety of different sensors that are being used for environmental monitoring. Data from social media, including geotagged photographs from sites such as Flickr or street-level photographs from providers such as Google Street View and Mapillary, can be processed using computer vision and segmentation to extract information related to land cover and land use. The logical progression of this field of study is the integration of remote sensing with these different sources of geospatial data using various machine learning and data fusion approaches to create new data sets on land cover and land use. Much of the previous integration work in this area has focused on urban areas b

Accurate and high-resolution land cover (LC) information is vital for addressing contemporary environmental challenges. With the advancement of satellite data acquisition, cloud-based processing, and deep learning technology, high-resolution Global Land Cover (GLC) map production has become increasingly feasible. With a growing number of available GLC maps, a comprehensive evaluation and comparison is necessary to assess their accuracy and suitability for diverse uses. This particularly applies to maps lacking statistically robust accuracy assessment or sufficient reported detail on the validation procedures. This study conducts a comparative independent validation of recent 10 m GLC maps, namely ESRI Land Use/Land Cover (LULC), ESA WorldCover, and Google and World Resources Institute (WRI)’s Dynamic World, examining their spatial detail representation and thematic accuracy at global, continental, and national (for 47 larger countries) levels. Since high-resolution map validation is impacted by reference data uncertainty owing to geolocation and labelling errors, five validation approaches dealing with reference data uncertainty were evaluated. Of the considered approaches, validation using the sample label supplemented by majority label within the neighborhood is found to produce more reasonable accuracy estimates compared to the overly optimistic approach of using any label within the neighborhood and the overly pessimistic approach of direct comparison between the map and reference labels. Overall global accuracies of the maps range between 73.4% ± 0.7% (95% confidence interval) to 83.8% ± 0.4% with WorldCover having the highest accuracy followed by Dynamic World and ESRI LULC. The quality of the maps varies across different LC classes, continents, and countries. The maps' spatial detail representation was assessed at various homogeneity levels within a 3 × 3 kernel. Although considered as high-resolution maps, this study reveals that ESRI LULC and Dynamic World

The challenge of global food security in the face of population growth, conflict and climate change requires a comprehensive understanding of cropped areas, irrigation practices and the distribution of major commodity crops like maize and wheat. However, such understanding should preferably be updated at seasonal intervals for each agricultural system rather than relying on a single annual assessment. Here we present the European Space Agency funded WorldCereal system, a global, seasonal, and reproducible crop and irrigation mapping system that addresses existing limitations in current global-scale crop and irrigation mapping. WorldCereal generates a range of global products, including temporary crop extent, seasonal maize and cereals maps, seasonal irrigation maps, seasonal active cropland maps, and confidence layers providing insights into expected product quality. The WorldCereal product suite for the year 2021 presented here serves as a global demonstration of the dynamic open-source WorldCereal system. The presented products are fully validated, e.g., global user's and producer's accuracies for the annual temporary crop product are 88.5 % and 92.1 %, respectively. The WorldCereal system provides a vital tool for policymakers, international organizations, and researchers to better understand global crop and irrigation patterns and inform decision-making related to food security and sustainable agriculture. Our findings highlight the need for continued community efforts such as additional reference data collection to support further development and push the boundaries for global agricultural mapping from space. The global products are available at https://doi.org/10.5281/zenodo.7875104 (Van Tricht et al., 2023).

Short-rotation woody plantations (SRWPs) play a major role in climate change mitigation and adaptation plans, because of their high yields of woody biomass and fast carbon storage. However, their benefits, trade-offs and growing-success are heavily location-dependent. Therefore, spatial data on the distribution of SRWPs are indispensable for assessing current distribution, trade-offs with other uses and potential contributions to climate mitigation. As current global datasets lack reliable information on SRWPs and full global mapping is difficult, we provide a consistent and systematic approach to estimate the spatial distribution of SRWPs in (sub-)tropical biomes under current and future climate. We combined three advanced methods (maximum entropy, random forest and multinomial regression) to evaluate spatially explicit probabilities of SRWPs. As inputs served a large empirical dataset on SRWP observations and 17 predictor variables, covering biophysical and socio-economic conditions. This dataset can help adding a more nuanced treatment of mitigation options and forest management in research on biodiversity and land use change.

The restoration of tree cover has been placed at the top of international policy agendas, yet often, the ‘type’ of restored forests can be widely different, with consequences for biodiversity and livelihoods. Here we used a map of forest management types to assess the extent of managed forests in recent tree cover gains globally. We call on policymakers to differentiate forest management as a distinct element of reforestation targets.

This dataset was created by using a new IIASA tool, called “Street Imagery validation” (https://svweb.cloud.geo-wiki.org/) where users could check street level images (e.g., Google Street Level images, Mapillary etc.) and identify the crop type where it is possible. The advantage of this tool is that there are plenty of georeferenced images with dates, going back in time. The disadvantage is that users need to check plenty of images where only few will clearly show cropland fields that are mature enough to be identified. To make the data collection more efficient, we provided our experts with preliminary maps of points in agricultural areas where street level images are available for the year 2021. Then, the experts checked those locations in an opportunistic way. The dataset is completely independent from all the existing maps and the reference datasets. There are 3 main data records uploaded: sv_croptype_poly.zip – an archive with a shapefile containing all the collected polygons with crop type information. Not all the polygons correspond to actual field boundaries. sv_croptype_validations.csv – a table with crop type observations with centroid coordinates in WGS84 sv_worldcereal_validation.csv – a table with a subset of crop type observations used in validation of WorldCereal crop type maps for 2021. Fields: "id" – unique observation identifier; "imgSource" – source of imagery used for visual inspection; "imgLoc" – image location; "svImgDate" – image date; "imageIdKey" – image unique identifier; "submitedAt" – date of submission of crop type observation; "cropType" - crop type observation; "irrType" – irrigation type; "x", "y" – centroids of submitted polygons in WGS84.

The European Space Agency (ESA) WorldCereal 10m 2021 product suite consist of global-scale annual and seasonal crop maps and (where applicable) their related confidence. Every file in this repository contains up to 106 agro-ecological zone (AEZ) products which were all processed with respect to their own regional seasonality and should be considered as independent products.

IIASA team has created a new validation data set, which is completely independent from all other existing maps or reference data sets, and which is in line with the cropland definitions and mapping period of the WorldCereal products. To decide if this is an active cropland in each period, the experts looked at very high-resolution Google historical imagery and Google Street level images, Microsoft Bing images, ESRI imagery, Planet historical data, Sentinel-2 time series, and Modis NDVI time series. The experts were asked to label 5 by 5 Sentinel pixels (each pixel 10m by 10m) either as winter crops, or as summer crops, or as maize (if this was possible), or as active crops (where it was not possible to confirm a growing season, e.g. overlap between seasons was too big or crop fields were too small in size), or as no crops, or as not sure where was too little information available for 2021. There was additional question on irrigation system, either circle, or other irrigation, or rainfed, or not sure.

This is a reference data set collected for validation of the monthly land cover maps at a 3m and at a 10m resolution produced in the WP5. The reference data set has been collected by using Geo-Wiki toolbox for visual interpretation of very high-resolution images, including Planet data and Google maps. The data set has been collected over 3 AOIs. Each reference sample site corresponds to a 30m-by-30m box and includes information about monthly land cover type over the period 2018-2020. Land cover legend is the same as in ESA WorldCover map at a 10m resolution (https://worldcover2021.esa.int/).

A prototype of monthly, 10 m resolution land surface categories, land use land cover (LULC) cover, and LULC change maps derived from Sentinel-2 data over three areas within Belgium, Portugal, and Sicily for the period 2018-2020. The LULC and LULC change maps were independently validated by IIASA. All products were generated within the framework of the RapidAI4EO project, funded from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 101004356. The data description can be found below. The validation report of the LULC and LULC change maps can be found in validation_LULC.pdf and validation_change.pdf, respectively, and the validation dataset can be found in Lesiv et al. (2023). Data description Increasing the cadence of the land cover updates from the typical (multi-)annual to monthly cadence poses several challenges. First, several land cover types are difficult to discriminate without any knowledge of temporal dynamics. For instance, croplands are characterized by a dynamic of vegetation growth and a harvest period (i.e. cycles of bare soil, sparsely vegetated and vegetated periods). This contrasts with grasslands that often lack the harvest period resulting in a bare soil cover. Without this temporal information, it is difficult to distinguish a vegetated cropland field from grassland. Second, phenological changes may introduce a large intra-class variability and thus also confusion between classes. For example, the shedding of leaves during autumn or wilting of herbaceous vegetation in dry summer periods introduces spectral variability within land cover classes. To overcome these challenges, we developed a workflow with two main phases. The first phase aims to map land surface categories (LSC) at a monthly resolution. The next phase uses the resulting monthly LSC probability time series to classify land cover.

Short-rotation woody plantations (SRWPs) play a major role in climate change mitigation and adaptation plans, because of their high yields of woody biomass and fast carbon storage. However, their benefits, trade-offs and growing-success are heavily location-dependent. Therefore, spatial data on the distribution of SRWPs are indispensable for assessing current distribution, trade-offs with other uses and potential contributions to climate mitigation. As current global datasets lack reliable information on SRWPs and full global mapping is difficult, we provide a consistent and systematic approach to estimate the spatial distribution of SRWPs in (sub-)tropical biomes under current and future climate. We combined three advanced methods (maximum entropy, random forest and multinomial regression) to evaluate spatially explicit probabilities of SRWPs. As inputs served a large empirical dataset on SRWP observations and 17 predictor variables, covering biophysical and socio-economic conditions. SRWP probabilities varied strongly between regions, and might not be feasible in major parts of (sub-)tropical biomes, challenging the feasibility of global mitigation plans that over-rely on tree plantations. Due to future climatic changes, SRWP probabilities decreased in many areas, particularly pronounced in higher emission scenarios. This indicates a negative feedback with higher emissions resulting in less mitigation potential. Less suitable land for SRWPs in the future could also result in fewer wood resources from these plantations, enhancing pressure on natural forests and hampering sustainability initiatives that use wood-based alternatives. Our results can help adding a more nuanced treatment of mitigation options and forest management in research on biodiversity and land use change.

It is becoming increasingly obvious that in order to address current global challenges and achieve the SDGs in the land-use sector, monitoring and evaluation using remote sensing technologies are essential. In particular, with the Copernicus program of the European Union, unprecedented free and open Earth observation data are becoming available. However, in order to improve our remotely sensed based machine learning models, training data in the form of in-situ or annotated land-use or land cover data which are based on the visual interpretation of aerial photographs or very high resolution satellite data are of utmost importance. Without sufficient training data, many land-use and land cover maps lack sufficient quality. The presentation will provide an overview of existing and open in-situ data in the field of land-use science. It will highlight what land-use data are currently available including data collected though crowdsourcing and the Geo-Wiki toolbox. In particular, it will provide insights into current gaps in land cover, land-use, livestock, forest as well as crop type information globally. It will draw on existing global data products such as those from the Copernicus global land monitoring service, and more recently generated products such as WorldCover and WorldCereal. Furthermore, tools to close those data gaps will be shown. The presentation will furthermore explore current obstacles and limitations to data sharing and debunk current arguments that are often put forth for not sharing in-situ data. These arguments include limited resources, quality issues, competition, as well as time constraints, etc. Specific attention will be given to the role of doners and funders in more clearly defining open and FAIR requirements for in-situ data. The presentation will close by making the audience aware of the LUCKINet consortium, which is trying to make more reference data openly accessible and to build a consistent global land-use change dataset as well as work

Aim Many vertebrate species globally are dependent on forests, most of which require active protection to safeguard global biodiversity. Forests, however, are increasingly either being disturbed, planted or managed in the form of timber or food plantations. Because of a lack of spatial data, forest management has commonly been ignored in previous conservation assessments. Location Global. Methods We combine a new global map of forest management types created solely from remote sensing imagery with spatially explicit information on the distribution of forest-associated vertebrate species and protected areas globally. Using Bayesian logistic regressions, we explore whether the amount of forested habitat available to a species as well as information on species-specific threats can explain differences in IUCN extinction risk categories. Results We show that disturbed and human-managed forests dominate the distributional ranges of most forest-associated species. Species considered as non-threatened had on average larger amounts of non-managed forests within their range. A greater amount of planted forests did not decrease the probability of species being threatened by extinction. Even more worrying, protected areas are increasingly being established in areas dominated by disturbed forests. Conclusion Our results imply that species extinction risk and habitat assessments might have been overly optimistic with forest management practices being largely ignored so far. With forest restoration being at the centre of climate and conservation policies in this decade, we caution that policy makers should explicitly consider forest management in global and regional assessments.

Citizen science has become an increasingly popular approach to scientific data collection, where classification tasks involving visual interpretation of images is one prominent area of application, e.g., to support the production of land cover and land-use maps. Achieving a minimum accuracy in these classification tasks at a minimum cost is the subject of this study. A Bayesian approach provides an intuitive and reasonably straightforward solution to achieve this objective. However, its application requires additional information, such as the relative frequency of the classes and the accuracy of each user. While the former is often available, the latter requires additional data collection. In this paper, we present a two-stage approach to gathering this additional information. We demonstrate its application using a hypothetical two-class example and then apply it to an actual crowdsourced dataset with five classes, which was taken from a previous Geo-Wiki crowdsourcing campaign on identifying the size of agricultural fields from very high-resolution satellite imagery. We also attach the R code for the implementation of the newly presented approach.

It is well established that nighttime radiance, measured from satellites, correlates with economic prosperity across the globe. In developing countries, areas with low levels of detected radiance generally indicate limited development – with unlit areas typically being disregarded. Here we combine satellite nighttime lights and the world settlement footprint for the year 2015 to show that 19% of the total settlement footprint of the planet had no detectable artificial radiance associated with it. The majority of unlit settlement footprints are found in Africa (39%), rising to 65% if we consider only rural settlement areas, along with numerous countries in the Middle East and Asia. Significant areas of unlit settlements are also located in some developed countries. For 49 countries spread across Africa, Asia and the Americas we are able to predict and map the wealth class obtained from ~2,400,000 geo-located households based upon the percent of unlit settlements, with an overall accuracy of 87%.

The global Forest Management data map for year 2015 (see DOI 10.5281/zenodo.4541512) was produced using a set of region-specific Random Forest Classifier models. These models are trained on and applied to each region defined in the Global Biome Cluster layer (see DOI 10.5281/zenodo.5848609). They can be run with Python's scikit-learn Random Forest Classifier and the Python joblib package. The model information is provided in three folders: training data (.csv files) for each model in the right Remote Sensing band order, including lat, lon of the location, and the class [coded as number] training parameters: random forest classifier parameters and used PROBA-V metrics bands (.ini files) to train the model with the given training data, after the 5folder cross-validation and optimization models (.joblib.z files) for each biome. The model names includes the identifier code from the global Biome Cluster layer.

ESA WorldCover 10 m 2021 v200 The European Space Agency (ESA) WorldCover 10 m 2021 product provides a global land cover map for 2021 at 10 m resolution based on Sentinel-1 and Sentinel-2 data. The WorldCover product comes with 11 land cover classes, aligned with UN-FAO's Land Cover Classification System, and has been generated in the framework of the ESA WorldCover project. The ESA WorldCover 10m 2021 v200 product updates the existing ESA WorldCover 10m 2020 v100 product to 2021 but is produced using an improved algorithm version (v200) compared to the 2020 map. Consequently, since the WorldCover maps for 2020 and 2021 were generated with different algorithm versions (v100 and v200, respectively), changes between the maps should be treated with caution, as they include both real changes in land cover and changes due to the algorithms used. The WorldCover 2021 v200 product is developed by a consortium lead by VITO Remote Sensing together with partners Brockmann Consult, Gamma Remote Sensing AG, IIASA and Wageningen University

Spatially explicit information on forest management at a global scale is critical for understanding the status of forests, for planning sustainable forest management and restoration, and conservation activities. Here, we produce the first reference data set and a prototype of a globally consistent forest management map with high spatial detail on the most prevalent forest management classes such as intact forests, managed forests with natural regeneration, planted forests, plantation forest (rotation up to 15 years), oil palm plantations, and agroforestry. We developed the reference dataset of 226 K unique locations through a series of expert and crowdsourcing campaigns using Geo-Wiki (https://www.geo-wiki.org/). We then combined the reference samples with time series from PROBA-V satellite imagery to create a global wall-to-wall map of forest management at a 100 m resolution for the year 2015, with forest management class accuracies ranging from 58% to 80%. The reference data set and the map present the status of forest ecosystems and can be used for investigating the value of forests for species, ecosystems and their services.

The development of remotely sensed products such as land cover requires large amounts of high-quality reference data, needed to train remote sensing classification algorithms and for validation. However, due to the lack of sharing and the high costs associated with data collection, particularly ground-based information, the amount of reference data available has not kept up with the vast increase in the availability of satellite imagery, e.g. from Landsat, Sentinel and Planet satellites. To fill this gap, the Geo-Wiki platform for the crowdsourcing of reference data was developed, involving visual interpretation of satellite and aerial imagery. Here we provide an overview of the crowdsourcing campaigns that have been run using Geo-Wiki over the last decade, including the amount of data collected, the research questions driving the campaigns and the outputs produced such as new data layers (e.g. a global map of forest management), new global estimates of areas or percentages of land cover/land use (e.g. the amount of extra land available for biofuels) and reference data sets, all openly shared. We demonstrate that the amount of data collected and the scientific advances in the field of land cover and land use would not have been possible without the participation of citizens. A relatively conservative estimate reveals that citizens have contributed more than 5.3 years of the data collection efforts of one person over short, intensive campaigns run over the last decade. We also provide key observations and lessons learned from these campaigns including the need for quality assurance mechanisms linked to incentives to participate, good communication, training and feedback, and appreciating the ingenuity of the participants.

Comparing the performance of different satellite sensors in global land cover change (LCC) monitoring is necessary to assess their potential and limitations for more accurate and operational LCC estimations. This paper aims to examine and compare the performance in LCC monitoring using three satellite sensors: PROBA-V, Landsat 8 OLI, and Sentinel-2 MSI. We utilized a unique set of global reference data containing four years of records (2015–2018) at 29,263 land cover change/no-change 100 × 100-m sites. The LCC monitoring was conducted using the BFAST(s)-Random Forest (BRF) change detection framework involving 15 global timeseries vegetation indices and three BFAST models. Due to the different spectral characteristics and data availability of the sensors, we designed 30 comparison scenarios to extensively evaluate their performance. The overall results were: 1) for global general LCC monitoring, Landsat 8 OLI slightly outperformed Sentinel-2, and PROBA-V performed the worst. The performance among the three sensors differed consistently despite different data availability and spectral observation regions. Sentinel-2 was more competitive with Landsat 8 when the red-edge 1 band was included; 2) Landsat 8 was more accurate in forest, herbaceous vegetation, and water monitoring. Sentinel-2 performed particularly well in wetland monitoring. In addition, we further observed: 3) missing data in time series decreased the accuracy in all sensors, but had little influence on the relative performance across sensors; 4) combining sensors would not necessarily improve the accuracy because the complementary effects enhanced the accuracy only when there was a large amount of data missing for all sensors; 5) the BRF framework maintained the performance gap among sensors, but obtained a higher and more balanced accuracy overall when compared with using BFAST methods alone, by involving ensemble learning with an embedded sample-balancing strategy; 6) among the random forest variables

Deforestation contributes to global greenhouse gas emissions and must be reduced if the 1.5° C limit to global warming is to be realized. Protected areas represent one intervention for decreasing forest loss and aiding conservation efforts, yet there is intense human pressure on at least one-third of protected areas globally (Jones et al., 2018). There have been numerous studies addressing the extent and identifying drivers of deforestation at the local, regional, and global level. Yet few have focused on drivers of deforestation in protected areas in high thematic detail. Here we use a new crowdsourced data set on drivers of tropical forest loss for the period 2008 to 2019, which has been collected using the Geo-Wiki crowdsourcing application for visual interpretation of very high-resolution imagery by volunteers. Extending on the published data on tree cover and forest loss from the Global Forest Change initiative (Hansen et al., 2013), we investigate the dominant drivers of deforestation in tropical protected areas situated within 30 degrees north and south of the equator. We find the deforestation rate in protected areas to be lower than the continental average for the Latin Americas (3.4% in protected areas compared to 5.4% in the Latin Americas) and Africa (3.3% compared to 3.9%), but it exceeds that of unprotected land in Asia (8.5% compared to 8.1%). Consistent with findings from foregoing studies, we also find that pastures and other subsistence agriculture are the dominant deforestation driver in the Latin Americas, while forest management, oil palm, shifting cultivation and other subsistence agriculture dominate in Asia, and shifting cultivation and other subsistence agriculture is the main driver in Africa. However, we find contrasting results in relation to the degree of protection, which indicate that the rate of deforestation in Latin America and Africa strictly protected areas might even exceed that of areas with no strict protection. This crucial fi

To meet the ambitious objectives of biodiversity and climate conventions, the international community requires clarity on how these objectives can be operationalized spatially and how multiple targets can be pursued concurrently. To support goal setting and the implementation of international strategies and action plans, spatial guidance is needed to identify which land areas have the potential to generate the greatest synergies between conserving biodiversity and nature’s contributions to people. Here we present results from a joint optimization that minimizes the number of threatened species, maximizes carbon retention and water quality regulation, and ranks terrestrial conservation priorities globally. We found that selecting the top-ranked 30% and 50% of terrestrial land area would conserve respectively 60.7% and 85.3% of the estimated total carbon stock and 66% and 89.8% of all clean water, in addition to meeting conservation targets for 57.9% and 79% of all species considered. Our data and prioritization further suggest that adequately conserving all species considered (vertebrates and plants) would require giving conservation attention to ~70% of the terrestrial land surface. If priority was given to biodiversity only, managing 30% of optimally located land area for conservation may be sufficient to meet conservation targets for 81.3% of the terrestrial plant and vertebrate species considered. Our results provide a global assessment of where land could be optimally managed for conservation. We discuss how such a spatial prioritization framework can support the implementation of the biodiversity and climate conventions.

This Validation Report describes in detail the quality of the satellite-based 100m Land Cover product of the global component of the Copernicus Land Service. It includes assessments of yearly global land cover layers (2015-2019), assessment of change as well as comparison with the previous version using an independent validation dataset. The related Product User Manual is the starting point for the reader and summarizes all aspects of the product (algorithm, quality, contents, format, etc).

Since the collapse of the Soviet Union and transition to a new forest inventory system, Russia has reported almost no change in growing stock (+ 1.8%) and biomass (+ 0.6%). Yet remote sensing products indicate increased vegetation productivity, tree cover and above-ground biomass. Here, we challenge these statistics with a combination of recent National Forest Inventory and remote sensing data to provide an alternative estimate of the growing stock of Russian forests and to assess the relative changes in post-Soviet Russia. Our estimate for the year 2014 is 111 ± 1.3 × 109 m3, or 39% higher than the value in the State Forest Register. Using the last Soviet Union report as a reference, Russian forests have accumulated 1163 × 106 m3 yr-1 of growing stock between 1988–2014, which balances the net forest stock losses in tropical countries. Our estimate of the growing stock of managed forests is 94.2 × 109 m3, which corresponds to sequestration of 354 Tg C yr-1 in live biomass over 1988–2014, or 47% higher than reported in the National Greenhouse Gases Inventory.

We provide four data records: 1.The reference data set as a comma-separated file ("reference_data_set.csv") with the following attributes: “ID” is a unique location identifier “Latitude, Longitude” are centroid coordinates of a 100m x 100m pixel. “Land_use_ID “is a land use class: 11 - Naturally regenerating forest without any signs of human activities, e.g., primary forests. 20 - Naturally regenerating forest with signs of human activities, e.g., logging, clear cuts etc. 31 - Planted forest. 32 - Short rotation plantations for timber. 40 - Oil palm plantations. 53 - Agroforestry. “Flag” identifies a data origin: 1- the crowdsourced locations, 2- the control data set, 0 – the additional experts' classifications following the opportunistic approach. 2. The 100 m forest management map in a geoTiff format with the classes presented - "FML_v3.2.tif ". 3. The predicted class probability from the Random Forest classification in a geoTiff format - "ProbaV_LC100_epoch2015_global_v2.0.3_forest-management--layer-proba_EPSG-4326.tif" 4. Validation data set as a comma-separated file ("validation_data_set.csv) with the following attributes: “ID” is a unique location identifier “pixel_center_x” , “pixel_center_y ” are centroid coordinates of a 100m x 100m pixel in lat/lon projection “first_landuse_class “is a land use class, as in (1). “second_landuse_class “is a second possible land use class, as in (1), identified in case it was difficult to assign one class with high confidence.

Currently most global land cover maps are produced with discrete classes, which express the dominant land cover class in each pixel, or a combination of several classes at a predetermined ratio. In contrast, land cover fraction mapping enables expressing the proportion of each pure class in each pixel, which increases precision and reduces legend complexity. To map land cover fractions, regression rather than classification algorithms are needed, and multiple approaches are available for this task. A major challenge for land cover fraction mapping models is data sparsity. Land cover fraction data is by its nature zero-inflated due to how common the 0% fraction is. As regression favours the mean, 0% and 100% fractions are difficult for regression models to predict accurately. We proposed a new solution by combining three models: a binary model determines whether a pixel is pure; if so, it is processed using a classification model; otherwise with a regression model. We compared multiple regression algorithms and implemented our proposed three-step model on the algorithm with the lowest RMSE. We further evaluated the spatial and per-class accuracy of the model and demonstrated a wall-to-wall prediction of seven land cover fractions over the globe. The models were trained on over 138,000 points and validated on a separate dataset of over 20,000 points, provided by the CGLS-LC100 project. Both datasets are global and aligned with the PROBA-V 100 m UTM grid. Results showed that the random forest regression model reached the lowest RMSE of 17.3%. Lowest MAE (7.9%) and highest overall accuracy (72% ± 2%) was achieved using random forest with our proposed three-model approach and median vote. This research proves that machine learning algorithms can be applied globally to map a wide variety of land cover fractions. Fraction mapping expresses land cover more precisely, and empowers users to create their own discrete maps using user-defined thresholds and rules, which enables

Annual global land cover maps (GLC) are being provided by several operational monitoring efforts. However, map validation is lagging, in the sense that the annual land cover maps are often not validated. Concurrently, users such as the climate and land management community require information on the temporal consistency of multi-date GLC maps and stability in their accuracy. In this study, we propose a framework for operational validation of annual global land cover maps using efficient means for updating validation datasets that allow timely map validation according to recommendations in the CEOS Stage-4 validation guidelines. The framework includes a regular update of a validation dataset and continuous map validation. For the regular update of a validation dataset, a partial revision of the validation dataset based on random and targeted rechecking (areas with a high probability of change) is proposed followed by additional validation data collection. For continuous map validation, an accuracy assessment of each map release is proposed including an assessment of stability in map accuracy addressing the user needs on the temporal consistency information of GLC map and map quality. This validation approach was applied to the validation of the Copernicus Global Land Service GLC product (CGLS-LC100). The CGLS-LC100 global validation dataset was updated from 2015 to 2019. The update was done through a partial revision of the validation dataset and an additional collection of sample validation sites. From the global validation dataset, a total of 40% (10% for each update year) was revisited, supplemented by a targeted revision focusing on validation sample locations that were identified as possibly changed using the BFAST time series algorithm. Additionally, 6720 sample sites were collected to represent possible land cover change areas within 2015 and 2019. Through this updating mechanism, we increased the sampling intensity of validation sample sites in possible land

ESA WorldCover 10 m 2020 v100 The European Space Agency (ESA) WorldCover 10 m 2020 product provides a global land cover map for 2020 at 10 m resolution based on Sentinel-1 and Sentinel-2 data. The WorldCover product comes with 11 land cover classes, aligned with UN-FAO's Land Cover Classification System, and has been generated in the framework of the ESA WorldCover project. The WorldCover product is developed by a consortium lead by VITO Remote Sensing together with partners Brockmann Consult, CS SI, Gamma Remote Sensing AG, IIASA and Wageningen University

This data repository contains the results of the NatureMap ( naturemap.earth/) conservation prioritization effort. The maps were created by jointly optimizing biodiversity and NCPs such as carbon and/or water. Maps are supplied at both 10km and 50km resolution and all maps that aim to find priority areas for all species considered in the analysis, utilize a series of representative sets. The ranks for each layer are area-specific and can be used to extract summary statistics by simple subsetting. For example, to obtain the top 30% of land area for biodiversity and carbon, one needs to create a mask of all areas lower than a value of 30 from the respective ranked layers. For convenience two files are supplied that contain the fraction of land area per grid cell times 1000. Multiplying those with the cell area (100km2, respectively 2500km2) gives the exact amount of land area in a given grid cell. These are labelled "globalgrid_mollweide_**km.tif " can be used to create masks for the priority maps. The geographic projection is World Mollweide Equal Area projection.

The data set is the result of the Drivers of Tropical Forest Loss crowdsourcing campaign. The campaign took place in December 2020. A total of 58 participants contributed validations of almost 120k locations worldwide. The locations were selected randomly from the Global Forest Watch tree loss layer (Hansen et al 2013), version 1.7. At each location the participants were asked to look at satellite imagery time series using a customized Geo-Wiki user interface and identify drivers of tropical forest loss during the years 2008 to 2019 following 3 steps: Step 1) Select the predominant driver of forest loss visible on a 1 km square (delimited by a blue bounding box); Step 2) Select any additional driver(s) of forest loss and; Step 3) Select if any roads, trails or buildings were visible in the 1 km bounding box. The Geo-Wiki campaign aims, rules and prizes offered to the participants in return for their work can be seen here: https://application.geo-wiki.org/Application/modules/drivers_forest_change/drivers_forest_change.html . The record contains 3 files: One “.csv” file with all the data collected by the participants during the crowdsourcing campaign (1158021 records); a second “.csv” file with the controls prepared by the experts at IIASA, used for scoring the participants (2001 unique locations, 6157 records) and a ”.docx” file describing all variables included in the two other files. A data descriptor paper explaining the mechanics of the campaign and describing in detail how the data was generated will be made available soon.

To meet the ambitious objectives of biodiversity and climate conventions, countries and the international community require clarity on how these objectives can be operationalized spatially, and multiple targets be pursued concurrently1. To support governments and political conventions, spatial guidance is needed to identify which areas should be managed for conservation to generate the greatest synergies between biodiversity and nature’s contribution to people (NCP). Here we present results from a joint optimization that maximizes improvements in species conservation status, carbon retention and water provisioning and rank terrestrial conservation priorities globally. We found that, selecting the top-ranked 30% (respectively 50%) of areas would conserve 62.4% (86.8%) of the estimated total carbon stock and 67.8% (90.7%) of all clean water provisioning, in addition to improving the conservation status for 69.7% (83.8%) of all species considered. If priority was given to biodiversity only, managing 30% of optimally located land area for conservation may be sufficient to improve the conservation status of 86.3% of plant and vertebrate species on Earth. Our results provide a global baseline on where land could be managed for conservation. We discuss how such a spatial prioritisation framework can support the implementation of the biodiversity and climate conventions.

The aim of the study was to study the status of cognitive functions in hypertensive patients, patients with hypothyroidism, and in patients with combination of these diseases, taking into account the state of disease compensation. 67 patients (36 men and 31 women), average age – 49.84±2.83 years were examined. The control group (CG) consisted of 18 practically healthy individuals (8 men and 10 women). Patients who received appropriate nosology treatment were divided into 3 groups: Group I – 21 patients with hypertension, systolic blood pressure (SBP) – 134.26±5.23 mm Hg, diastolic blood pressure (DBP) – 84.37±4.51 mm Hg; Group II – 18 patients with hypothyroidism, thyrotropic hormone (TSH) – 3.16±0.79 mIU/L, stage of hypothy­roidism compensation was diagnosed in 83.3%, subcompensation – in 16.7%; Group III – 28 patients with hyper­ten­sion (SBP 145.52±5.45 mm Hg; DBP 82.41±3.86 mm Hg) with concomitant hypothyroidism (TSH – 2.92±0.78 mIU/L, stage of compensation for hypothyroidism was diagnosed in 85.7%, subcompensation – in 14.3% of cases. Information about visits to the therapist/cardiologist/family doctor and endocrinologist was used to analyze the therapeutic correction of the disease: ambulatory medical records of patients with measurement of blood pressure (BP) and TSH during the disease were processed. To assess cognitive functions, Mini Mental State Examination (MMSE) and the Addenbrooke’s cognitive examination (ACE-R) were used. The relationship of average blood pressure data in patients with hypertension and TSH in patients with hypothyroidism during the disease period and the level of cognitive function was investigated. The average level of office BP (SBP/DBP) in Groups I and III compared to CG during the examination was: SBP 134.26±5.23 mm Hg (p=0.047), DBP 84.37±4.51 mm Hg (p=0.041) in Group I; SBP 145.52±5.45 mm Hg (p=0.031), DBP 82.41±3.86 mm Hg (p=0.050) in Group III. Analyzing the TSH levels it was found that at the time of the physical examination

We provide a global, spatially explicit characterization of 47 terrestrial habitat types, as defined in the International Union for Conservation of Nature (IUCN) habitat classification scheme, which is widely used in ecological analyses, including for quantifying species’ Area of Habitat. We produced this novel habitat map for the year 2015 by creating a global decision tree that intersects the best currently available global data on land cover, climate and land use. We independently validated the map using occurrence data for 828 species of vertebrates (35152 point plus 8181 polygonal occurrences) and 6026 sampling sites. Across datasets and mapped classes we found on average a balanced accuracy of 0.77 (+¯0.14 SD) at Level 1 and 0.71 (+¯0.15 SD) at Level 2, while noting potential issues of using occurrence records for validation. The maps broaden our understanding of habitats globally, assist in constructing area of habitat refinements and are relevant for broad-scale ecological studies and future IUCN Red List assessments. Periodic updates are planned as better or more recent data becomes available.

The first ever global map of forest management was generated based on remote sensing data. To collect training data, we launched a series of Geo-Wiki (https://www.geo-wiki.org/) campaigns involving forest experts from different world regions, to explore which information related to forest management could be collected by visual interpretation of very high-resolution images from Google Maps and Microsoft Bing, Sentinel time series and normalized difference vegetation index (NDVI) profiles derived from Google Earth Engine. A machine learning technique was then used with the visually interpreted sample (280K locations) as a training dataset to classify PROBA-V satellite imagery. Finally, we obtained a global wall-to-wall map of forest management at a 100m resolution for the year 2015. The map includes classes such as intact forests; forests with signs of management, including logging; planted forests; woody plantations with a rotation period up to 15 years; oil palm plantations; and agroforestry. The map can be used to deliver further information about forest ecosystems, protected and observed forest status changes, biodiversity assessments, and other ecosystem-related aspects.

Consolidated epoch 2017 from the Collection 3 of annual, global 100m land cover maps. Other available epochs: 2015 2016 2018 2019 Produced by the global component of the Copernicus Land Service, derived from PROBA-V satellite observations and ancillary datasets. The maps include a main discrete classification with 23 classes aligned with UN-FAO's Land Cover Classification System, a set of versatile cover fractions: percentage (%) of ground cover for the 10 main classes a forest type layer quality layers on input data density and on the confidence of the detected land cover change

Traditionally the accuracy assessment of a hard raster-based land use land cover (LULC) map uses a reference data set that contains one LULC class per pixel, which is the class that has the largest area in each pixel. However, when mixed pixels exist in the reference data, this is a simplification of reality that has implications for both the accuracy assessment and subsequent applications of LULC maps, such as area estimation. This paper demonstrates how the use of class proportions in the reference data set can be used easily within regular accuracy assessment procedures and how the use of class proportions can affect the final accuracy assessment. Using the CORINE land cover map (CLC) and the more detailed Urban Atlas (UA), two accuracy assessments of the raster version of CLC were undertaken using UA as the reference and considering for each pixel: (i) the class proportions retained from the UA; and (ii) the class with the majority area. The results show that for the study area and the classes considered here, all accuracy indices decrease when the class proportions are considered in the reference database, achieving a maximum difference of 16% between the two approaches. This demonstrates that if the UA is considered as representing reality, then the true accuracy of CLC is lower than the value obtained when using the reference data set that assigns only one class to each pixel. Arguments for and against using class proportions in reference data sets are then provided and discussed.

Almost 6 years ago, the now Center for Earth Observation and Citizen Science (EOCS) at the International Institute for Applied Systems Analysis (IIASA) pioneered a crowdsourcing mobile app that allowed citizens to report land use and land cover at specific locations across Austria. The app is called FotoQuest Austria (and FotoQuest Go Europe when extended outside of Austria) and uses the GPS capabilities of mobile phones to allow citizens to visit locations near to them and then provide information on various land-related characteristics. A subset of the locations in FotoQuest Austria matched those used in the three-yearly Land Use and Coverage Area frame Survey (LUCAS) from Eurostat. The interface was developed to mimic part of the same protocol that LUCAS surveyors use when visiting locations across Europe, but in this case allowing any citizen to record land use and land cover characteristics observed at these locations. Over a period of 4 years, the FotoQuest project continued to improve: In the 2015 FotoQuest Austria version, 76 citizens collected data at over 600 LUCAS locations, although only 300 were used for comparison, mostly due to quality reasons (Laso Bayas et al. 2016). In the 2018 FotoQuest Go Europe campaign, 140 users from 18 different countries visited 1600 locations, with almost 1400 being currently used for analysis. Apart from the increased number of countries and locations, the user interface, experience and interaction with the app was continuously enhanced. Although LUCAS happened only twice in this period (2015 and 2018), FotoQuest had 3 official campaigns, which allowed us to introduce improvements in each campaign, but it also enabled citizens to continue providing land use change information in between campaigns. In 2015, the agreement between the main land cover classes in LUCAS and FotoQuest Austria was 69% whereas in the 2018 FotoQuest Go Europe campaign, it was over 90%. Currently, data from all campaigns are being compiled and will b

Land cover maps are key elements for understanding global climate and land use. They are often created by automatically classifying satellite imagery. However, inconsistencies in classification may be introduced inadvertently. Experts can reconcile classification discrepancies by viewing satellite and high-resolution images taken on the ground. We present and evaluate a framework to filter relevant geo-tagged photos from social network sites for land cover classification tasks. Social network sites offer massive amounts of potentially relevant data, but its quality and fitness for research purposes must be verified. Our framework uses computer vision to analyse the content of geo-tagged photos on social network sites to generate descriptive tags. These are used to train artificial neural networks to predict a photo’s relevance for land cover classification. We apply our models to four African case studies and their neighbours. The framework has been implemented within Geo-Wiki to fetch relevant photos from Flickr.

In May 2019, Collection 2 of the Copernicus Global Land Cover layers was released. Next to a global discrete land cover map at 100 m resolution, a set of cover fraction layers is provided depicting the percentual cover of the main land cover types in a pixel. This additional continuous classification scheme represents areas of heterogeneous land cover better than the standard discrete classification scheme. Overall, 20 layers are provided which allow customization of land cover maps to specific user needs or applications (e.g., forest monitoring, crop monitoring, biodiversity and conservation, climate modeling, etc.). However, Collection 2 was not just a global up-scaling, but also includes major improvements in the map quality, reaching around 80% or more overall accuracy. The processing system went into operational status allowing annual updates on a global scale with an additional implemented training and validation data collection system. In this paper, we provide an overview of the major changes in the production of the land cover maps, that have led to this increased accuracy, including aligning with the Sentinel 2 satellite system in the grid and coordinate system, improving the metric extraction, adding better auxiliary data, improving the biome delineations, as well as enhancing the expert rules. An independent validation exercise confirmed the improved classification results. In addition to the methodological improvements, this paper also provides an overview of where the different resources can be found, including access channels to the product layer as well as the detailed peer-review product documentation.

There is increasing evidence that smallholder farms contribute substantially to food production globally yet spatially explicit data on agricultural field sizes are currently lacking. Automated field size delineation using remote sensing or the estimation of average farm size at subnational level using census data are two approaches that have been used. However, both have limitations, e.g. automatic field size delineation using remote sensing has not yet been implemented at a global scale while the spatial resolution is very coarse when using census data. This paper demonstrates a unique approach to quantifying and mapping agricultural field size globally using crowdsourcing. A campaign was run in June 2017 where participants were asked to visually interpret very high resolution satellite imagery from Google Maps and Bing using the Geo-Wiki application. During the campaign, participants collected field size data for 130K unique locations around the globe. Using this sample, we have produced the most accurate global field size map to date and estimated the percentage of different field sizes, ranging from very small to very large, in agricultural areas at global, continental and national levels. The results show that smallholder farms occupy up to 40% of agricultural areas globally, which means that, potentially, there are many more smallholder farms in comparison with the two different current global estimates of 12% and 24%. The global field size map and the crowdsourced data set are openly available and can be used for integrated assessment modelling, comparative studies of agricultural dynamics across different contexts, for training and validation of remote sensing field size delineation, and potential contributions to the Sustainable Development Goal of Ending hunger, achieve food security and improved nutrition and promote sustainable agriculture.