Forest cover is rapidly changing at the global scale as a result of land-use change (principally deforestation in many tropical regions and afforestation in many temperate regions) and climate change. However, a detailed map of global forest gain is still lacking at fine spatial and temporal resolutions. In this study, we developed a new automatic framework to map annual forest gain across the globe, based on Landsat time series, the LandTrendr algorithm and the Google Earth Engine (GEE) platform. First, samples of stable forest collected based on the Global Forest Change product (GFC) were used to determine annual Normalized Burn Ratio (NBR) thresholds for forest gain detection. Secondly, with the NBR time-series from 1982 to 2020 and LandTrendr algorithm, we produced dataset of global forest gain year from 1984 to 2020 based on a set of decision rules. Our results reveal that large areas of forest gain occurred in China, Russia, Brazil and North America, and the vast majority of the global forest gain has occurred since 2000. The new dataset was consistent in both spatial extent and years of forest gain with data from field inventories and alternative remote sensing products. Our dataset is valuable for policy-relevant research on the net impact of forest cover change on the global carbon cycle and provides an efficient and transferable approach for monitoring other types of land cover dynamics.

Tree canopy is defined as area of vegetation (including leaves, stems, branches, etc.) of woody plants above 5m in height. The dataset developers derived tree canopy cover estimates from the Global Forest Cover Change (GFCC) Surface Reflectance product (GFCC30SR), which is based on enhanced Global Land Survey (GLS) datasets. The GLS datasets are composed of high-resolution Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper Plus (ETM ) images at 30 meter resolution.CANUE staff retrieved tree canopy cover data from Google Earth Engine (GEE) for the year 2010 and 2015, extracted values (percent coverage) to postal codes and calculated summary measures (average percent coverage) within buffers of 100, 250, 500, and 1000 metres.

This data set, a collaboration between the GLAD (Global Land Analysis & Discovery) lab at the University of Maryland, Google, USGS, and NASA, measures areas of tree cover gain across all global land (except Antarctica and other Arctic islands) at 30 × 30 meter resolution, displayed as a 12-year cumulative layer. The data were generated using multispectral satellite imagery from the Landsat 7 thematic mapper plus (ETM+) sensor. Over 600,000 Landsat 7 images were compiled and analyzed using Google Earth Engine, a cloud platform for earth observation and data analysis. The clear land surface observations (30 × 30 meter pixels) in the satellite images were assembled and a supervised learning algorithm was then applied to identify per pixel tree cover gain.

Tree cover gain was defined as the establishment of tree canopy at the Landsat pixel scale in an area that previously had no tree cover. Tree cover gain may indicate a number of potential activities, including natural forest growth or the crop rotation cycle of tree plantations.

When zoomed out (< zoom level 13), pixels of gain are shaded according to the density of gain at the 30 x 30 meter scale. Pixels with darker shading represent areas with a higher concentration of tree cover gain, whereas pixels with lighter shading indicate a lower concentration of tree cover gain. There is no variation in pixel shading when the data is at full resolution (≥ zoom level 13).

The tree cover canopy density of the displayed data is >50%.

This data set, a collaboration between the GLAD (Global Land Analysis & Discovery) lab at the University of Maryland, Google, USGS, and NASA, measures areas of tree cover gain across all global land (except Antarctica and other Arctic islands) at 30 × 30 meter resolution, displayed as a 12-year cumulative layer. The data were generated using multispectral satellite imagery from the Landsat 7 thematic mapper plus (ETM+) sensor. Over 600,000 Landsat 7 images were compiled and analyzed using Google Earth Engine, a cloud platform for earth observation and data analysis. The clear land surface observations (30 × 30 meter pixels) in the satellite images were assembled and a supervised learning algorithm was then applied to identify per pixel tree cover gain.Tree cover gain was defined as the establishment of tree canopy at the Landsat pixel scale in an area that previously had no tree cover. Tree cover gain may indicate a number of potential activities, including natural forest growth or the crop rotation cycle of tree plantations.When zoomed out (< zoom level 13), pixels of gain are shaded according to the density of gain at the 30 x 30 meter scale. Pixels with darker shading represent areas with a higher concentration of tree cover gain, whereas pixels with lighter shading indicate a lower concentration of tree cover gain. There is no variation in pixel shading when the data is at full resolution (≥ zoom level 13).The tree cover canopy density of the displayed data is >50%.

This data set, a collaboration between the GLAD (Global Land Analysis & Discovery) lab at the University of Maryland, Google, USGS, and NASA, displays tree cover over all global land (except for Antarctica and a number of Arctic islands) for the year 2000 at 30 × 30 meter resolution. “Percent tree cover” is defined as the density of tree canopy coverage of the land surface and is color-coded by density bracket (see legend).

Data in this layer were generated using multispectral satellite imagery from the Landsat 7 thematic mapper plus (ETM+) sensor. The clear surface observations from over 600,000 images were analyzed using Google Earth Engine, a cloud platform for earth observation and data analysis, to determine per pixel tree cover using a supervised learning algorithm.

The tree cover canopy density of the displayed data varies according to the selection - use the legend on the map to change the minimum tree cover canopy density threshold.

'The RCMAP (Rangeland Condition Monitoring Assessment and Projection) dataset quantifies the percent cover of rangeland components across western North America using Landsat imagery from 1985-2023. The RCMAP product suite consists of ten fractional components: annual herbaceous, bare ground, herbaceous, litter, non-sagebrush shrub, perennial herbaceous, sagebrush, shrub, tree, and shrub height …

Tree canopy is defined as area of vegetation (including leaves, stems, branches, etc.) of woody plants above 5m in height. The dataset developers derived tree canopy cover estimates from the Global Forest Cover Change (GFCC) Surface Reflectance product (GFCC30SR), which is based on enhanced Global Land Survey (GLS) datasets. The GLS datasets are composed of high-resolution Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper Plus (ETM ) images at 30 meter resolution.CANUE staff retrieved tree canopy cover data from Google Earth Engine (GEE) for the year 2010 and 2015, extracted values (percent coverage) to postal codes and calculated summary measures (average percent coverage) within buffers of 100, 250, 500, and 1000 metres.

Tree canopy is defined as area of vegetation (including leaves, stems, branches, etc.) of woody plants above 5m in height. The dataset developers derived tree canopy cover estimates from the Global Forest Cover Change (GFCC) Surface Reflectance product (GFCC30SR), which is based on enhanced Global Land Survey (GLS) datasets. The GLS datasets are composed of high-resolution Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper Plus (ETM ) images at 30 meter resolution.CANUE staff retrieved tree canopy cover data from Google Earth Engine (GEE) for the year 2010 and 2015, extracted values (percent coverage) to postal codes and calculated summary measures (average percent coverage) within buffers of 100, 250, 500, and 1000 metres.

We used Landsat data to map and quantify land cover change and forest fragmentation in Liberia between 2000-2018. These are the land cover maps created by the methodology and with which we performed the forest fragmentation analysis.

The Lower Mekong Region (LMR) faces significant loss of mangrove forests, yet limited studies have explored this decline in the region. Here, we employ Google Earth Engine and Landsat satellite imagery to assess changes in mangrove forest cover across Myanmar, Thailand, Vietnam, and Cambodia between 1989 and 2020, with a five-year interval. Accordingly, we estimated carbon stock changes due to changes of forest cover. Our analysis yielded an overall average accuracy of 92.10% and an average kappa coefficient of 0.89 across the four countries. The findings reveal a 0.9% increase in mangrove area in Myanmar, 2.5% in Thailand, and 1.3% in Cambodia, while Vietnam experienced a 0.2% loss annually between 1989 and 2020. Carbon stocks in mangrove forests were estimated at 577.0 ​Tg of carbon or TgC, 250.0 TgC, 61.6 TgC, and 269.0 TgC in 1989 for Myanmar, Thailand, Cambodia, and Vietnam respectively, and increased to 736.0 TgC, 443.0 TgC, 86.7 TgC, and 254 TgC in 2020. Increase in mangrove areas resulted in carbon removals of 42.8 TgCO2 year−1 over the same period above. Depending on policies in these respective countries, such carbon removals could be used to claim for result-based payment under the REDD ​+ ​scheme of the United Nations Framework Convention on Climate Change.

The layers included in the code were from the study conducted by the research group of CNR-IBE (Institute of BioEconomy of the National Research Council of Italy) and ISPRA (Italian National Institute for Environmental Protection and Research), published by the Sustainability journal (https://doi.org/10.3390/su14148412).

Link to the Google Earth Engine (GEE) code (link: https://code.earthengine.google.com/715aa44e13b3640b5f6370165edd3002)

You can analyze and visualize the following spatial layers by accessing the GEE link:

Daytime summer land surface temperature (raster data, horizontal resolution 30 m, from Landsat-8 remote sensing data, years 2015-2019)

Surface thermal hot-spot (raster data, horizontal resolution 30 m) was obtained by using a statistical-spatial method based on the Getis-Ord Gi* approach through the ArcGIS Pro tool.

Surface albedo (raster data, horizontal resolution 10 m, Sentinel-2A remote sensing data, year 2017)

Impervious area (raster data, horizontal resolution 10 m, ISPRA data, year 2017)

Tree cover (raster data, horizontal resolution 10 m, ISPRA data, year 2018)

Grassland area (raster data, horizontal resolution 10 m, ISPRA data, year 2017)

Water bodies (raster data, horizontal resolution 2 m, Geoscopio Platform of Tuscany, year 2016)

Sky View Factor (raster data, horizontal resolution 1 m, lidar data from the OpenData platform of Florence, year 2016)

Buildings' units of Florence (shapefile from the OpenData platform of Florence) include data on the residential real estate value from the Real Estate Market Observatory (OMI) of the National Revenue Agency of Italy (source: https://www1.agenziaentrate.gov.it/servizi/Consultazione/ricerca.htm, accessed on 14 July 2022). Data on the characterization of the buffer area (50 m) surrounding the buildings are included in this shapefile [the names of table attributes are reported in the square brackets]: averaged values of the daytime summer land surface temperature [LST_media], thermal hot-spot pattern [Thermal_cl], mean values of sky view factor [SVF_medio], surface albedo [alb_medio], and average percentage areas of imperviousness [ImperArea%], tree cover [TreeArea%], grassland [GrassArea%] and water bodies [WaterArea%].

Here attached the .txt file of the GEE code.

E-mail

Giulia Guerri, CNR-IBE, giulia.guerri@ibe.cnr.it

Marco Morabito, CNR-IBE, marco.morabito@cnr.it

Alfonso Crisci, CNR-IBE, alfonso.crisci@ibe.cnr.it

30m resolution historically consistent land cover and cover fraction maps over the Sudano-Sahel for the period 1986-2015. These land cover / cover fraction maps are achieved based on the Landsat archive preprocessed on Google Earth Engine and a random forest classification / regression model, while historical consistency is achieved using the Hidden Markov Model.

Validated land cover / cover fraction maps covering the full Sudano-Sahel are provided for 2015 (2015_Sahel.zip), while historical maps are available for four focus areas. The extent of the areas are displayed in 11_study_area.jpeg

Each of the zip files contains 14 GeoTIFF files for the respective period and area:

• Landsat_LC30_epochYYYY_AREA_bare-coverfraction-layer.tif [0-100%]
• Landsat_LC30_epochYYYY_AREA_crops-coverfraction-layer.tif [0-100%]
• Landsat_LC30_epochYYYY_AREA_DataDensityIndicator.tif [# overpasses that are used as input for the creation of the maps for this region / epoch]
• Landsat_LC30_epochYYYY_AREA_discrete-classification-HMM.tif [temporally cleaned discrete classification map using the Hidden Markov Model; legend see below]
• Landsat_LC30_epochYYYY_AREA_discrete-classification.tif [original discrete classification map; legend see below]
• Landsat_LC30_epochYYYY_AREA_forest-type-layer.tif [legend see below]
• Landsat_LC30_epochYYYY_AREA_grass-coverfraction-layer.tif [0-100%]
• Landsat_LC30_epochYYYY_AREA_moss-coverfraction-layer.tif [0-100%]
• Landsat_LC30_epochYYYY_AREA_shrub-coverfraction-layer.tif [0-100%]
• Landsat_LC30_epochYYYY_AREA_snow-coverfraction-layer.tif [0-100%]
• Landsat_LC30_epochYYYY_AREA_tree-coverfraction-layer.tif [0-100%]
• Landsat_LC30_epochYYYY_AREA_urban-coverfraction-layer.tif [0-100%]
• Landsat_LC30_epochYYYY_AREA_water-permanent-coverfraction-layer.tif [0-100%]
• Landsat_LC30_epochYYYY_AREA_water-seasonal-coverfraction-layer.tif [0-100%]

Discrete classification legend:

• 0: Unknown. No or not enough satellite data available.
• 20: Shrubs. Woody perennial plants with persistent and woody stems and without any defined main stem being less than 5 m tall. The shrub foliage can be either evergreen or deciduous.
• 30: Herbaceous vegetation. Plants without persistent stem or shoots above ground and lacking definite firm structure. Tree and shrub cover is less than 10 %.
• 40: Cultivated and managed vegetation / agriculture. Lands covered with temporary crops followed by harvest and a bare soil period (e.g., single and multiple cropping systems). Note that perennial woody crops will be classified as the appropriate forest or shrub land cover type.
• 50: Urban / built up. Land covered by buildings and other man-made structures.
• 60: Bare / sparse vegetation. Lands with exposed soil, sand, or rocks and never has more than 10 % vegetated cover during any time of the year.
• 70: Snow and ice. Lands under snow or ice cover throughout the year.
• 80: Permanent water bodies. Lakes, reservoirs, and rivers. Can be either fresh or salt-water bodies.
• 90: Herbaceous wetland. Lands with a permanent mixture of water and herbaceous or woody vegetation. The vegetation can be present in either salt, brackish, or fresh water.
• 100: Moss and lichen.
• 111: Closed forest, evergreen needle leaf. Tree canopy >70 %, almost all needle leaf trees remain green all year. Canopy is never without green foliage.
• 112: Closed forest, evergreen broad leaf. Tree canopy >70 %, almost all broadleaf trees remain green year round. Canopy is never without green foliage.
• 113: Closed forest, deciduous needle leaf. Tree canopy >70 %, consists of seasonal needle leaf tree communities with an annual cycle of leaf-on and leaf-off periods.
• 114: Closed forest, deciduous broad leaf. Tree canopy >70 %, consists of seasonal broadleaf tree communities with an annual cycle of leaf-on and leaf-off periods.
• 115: Closed forest, mixed.
• 116: Closed forest, not matching any of the other definitions.
• 121: Open forest, evergreen needle leaf. Top layer- trees 15-70 % and second layer- mixed of shrubs and grassland, almost all needle leaf trees remain green all year. Canopy is never without green foliage.
• 122:Open forest, evergreen broad leaf. Top layer- trees 15-70 % and second layer- mixed of shrubs and grassland, almost all broadleaf trees remain green year round. Canopy is never without green foliage.
• 123: Open forest, deciduous needle leaf. Top layer- trees 15-70 % and second layer- mixed of shrubs and grassland, consists of seasonal needle leaf tree communities with an annual cycle of leaf-on and leaf-off periods.
• 124: Open forest, deciduous broad leaf. Top layer- trees 15-70 % and second layer- mixed of shrubs and grassland, consists of seasonal broadleaf tree communities with an annual cycle of leaf-on and leaf-off periods.
• 125: Open forest, mixed.
• 126: Open forest, not matching any of the other definitions.
• 200: Oceans, seas. Can be either fresh or salt-water bodies.

Forest type legend:

• 0: Unknown
• 1: Evergreen needle leaf
• 2: Evergreen broad leaf
• 3: Deciduous needle leaf
• 4: Deciduous broad leaf
• 5: Mix of forest types

More detail on the classification algorithm and the resulting maps can be found in the accompanying paper:

Souverijns, N.; Buchhorn, M.; Horion, S.; Fensholt, R.; Verbeeck, H.; Verbesselt, J.; Herold, M.; Tsendbazar, N.-E.; Bernardino, P.N.; Somers, B.; Van De Kerchove, R. Thirty Years of Land Cover and Fraction Cover Changes over the Sudano-Sahel Using Landsat Time Series. Remote Sens. 2020, 12, 3817. https://doi.org/10.3390/rs12223817

Please note that a quality layer is available for each of the historical areas / periods (Landsat_LC30_epochYYYY_AREA_DataDensityIndicator.tif). In case a value of 4 or lower is achieved here, the discrete land cover classification / cover fraction for this period / area is highly uncertain. Take this into account when analysing the maps. Furthermore, take note that there is a large difference between the temporally cleaned (Landsat_LC30_epochYYYY_AREA_discrete-classification-HMM.tif) and original discrete land cover classification (Landsat_LC30_epochYYYY_AREA_discrete-classification.tif). We recommend to use the temporally cleaned version in combination with the quality layer.

Mangroves provide crucial biodiversity values to the ecosystem and the carbon dioxide sequestration capacity of mangrove forests helps mitigate global climate change. Given the diverse ecosystem and economic values, mapping how mangroves change over time is important for mangrove forest management and protection. The traditional mangrove mapping methodology using remote sensing requires a comprehensive understanding of mangrove ecology, remote sensing knowledge, and programming skills. The lack of specialists may stop the mapping of mangrove forests in many areas. In 2020, the Google Earth Engine Mangrove Mapping Methodology was introduced to accessibly map and monitor mangroves with random forest classifier and Landsat satellite imagery. This study applied Google Earth Engine Mangrove Mapping Methodology in Saloum Delta, Senegal for land cover classification with an overall accuracy of 96.45% in 2013 and 97.51% in 2023. Mangrove forests in Saloum Delta experienced heavy harvesting since 1950, and conservation projects have been conducted since 2004 by international organizations and the local government. The results suggest that 93.9% of mangrove forests remained unchanged, while 3.1% were lost and 2.9% saw an increase over the last decade. The results highlighted the mangrove loss areas that need more conservation attention and provided a valuable mangrove forest dynamic map to help the local communities in Saloum Delta with mangrove forest management. For the future development of Google Earth Engine Mangrove Mapping Methodology, the addition of more classifiers for land cover classification and higher resolution Sentinel satellite imagery should be considered. With a user-friendly interface and detailed guidance, Google Earth Engine Mangrove Mapping Methodology has a bright potential to help people with mangrove forest mapping for sustainable management globally in the future.

Using 335,709 Landsat images on the Google Earth Engine, we built the first Landsat-derived annual land cover product of China (CLCD) from 1985 to 2019. We collected the training samples by combining stable samples extracted from China's Land-Use/Cover Datasets (CLUD), and visually-interpreted samples from satellite time-series data, Google Earth and Google Map. Several temporal metrics were constructed via all available Landsat data and fed to the random forest classifier to obtain classification results. A post-processing method incorporating spatial-temporal filtering and logical reasoning was further proposed to improve the spatial-temporal consistency of CLCD.

"*_albert.tif" are projected files via a proj4 string "+proj=aea +lat_1=25 +lat_2=47 +lat_0=0 +lon_0=105 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs".

CLCD in 2022 is now available.

1. Given that the USGS no longer maintains the Landsat Collection 1 data, we are now using the Collection 2 SR data to update the CLCD.

2. All files in this version have been exported as Cloud Optimized GeoTIFF for more efficient processing on the cloud. Please check here for more details.

3. Internal overviews and color tables are built into each file to speed up software loading and rendering.

This data set, a collaboration between the GLAD (Global Land Analysis & Discovery) lab at the University of Maryland, Google, USGS, and NASA, displays tree cover over all global land (except for Antarctica and a number of Arctic islands) for the year 2000 at 30 × 30 meter resolution. “Percent tree cover” is defined as the density of tree canopy coverage of the land surface and is color-coded by density bracket (see legend).

Data in this layer were generated using multispectral satellite imagery from the Landsat 7 thematic mapper plus (ETM+) sensor. The clear surface observations from over 600,000 images were analyzed using Google Earth Engine, a cloud platform for earth observation and data analysis, to determine per pixel tree cover using a supervised learning algorithm.

The tree cover canopy density of the displayed data varies according to the selection - use the legend on the map to change the minimum tree cover canopy density threshold.

The Xingu River basin is known for its high biodiversity. The state of the riparian ecosystems is directly linked to forest cover changes and has an impact on the unique biota. In order to assess the rapid changes that the Xingu river basin has experienced over the last four decades, we developed a forest/ non-forest classification for four time periods, from 1989 to 2018. These datasets were produced from Landsat TM5 and Landsat 8-OLI cloud free surface reflectance mosaics in Google Earth Engine and a supervised CART classification. The forest definition for the classifications uses a minimum mapping unit of 90 m x 90 m (~1 ha). We used around 3000 field photos to identify corresponding forest areas in the image. From the photos, our definition of forest corresponds with areas of approximately more than 30% tree canopy cover, which is consistent with the Brazilian definition of forest.

This dataset is a collection of data and results from a research project conducted by NASA SEES Interns. The research project aimed to study urban heat islands and their relationship with land cover observations. This dataset upload consists of 12 files. One file is a poster pdf that includes all the information needed about the project. The other 11 files are png images of heatmaps, bar graphs, scatter plots, and tables used in the analysis of our project. For quick reference, the abstract to this project is below:

The urban heat island (UHI) effect refers to the phenomenon in which urban areas experience higher temperatures compared to their rural counterparts. This research aims to quantify and examine the UHI effect within three areas of interest (AOIs) by utilizing LANDSAT imagery. In addition, this study seeks to explore the relationship between land cover classifications, which represent the most green (rural) and the most urban areas, and the intensity of the UHI effect. To achieve this, temperature data from local weather stations are analyzed, and statistical methods are employed to determine whether a correlation exists between the difference in land cover classifications and the intensity of the UHI effect, as determined by the average temperature difference between urban and rural areas. Google Earth Engine is used to visualize LANDSAT data from 2013 to 2022 in the months of July and August for each AOI. Subsequently, the data is compared with the land cover classifications from Collect Earth Online using statistical models in Microsoft Excel. These tools were used to take data from three pre-selected areas of interest in GLOBE Observer. The data findings from this analysis suggest that the more tree cover and rural an area is according to our classification method, the lower the UHI intensity. On the other hand, the higher the urban area, the higher the UHI intensity. By beginning this research, we have reinforced the validity of land cover classifications, and we now have the capability to generally predict the UHI intensity of locations based on their classifications. Overall, this investigation aims to contribute to a better understanding of the GLOBE land cover classifications and their potential indications of UHI intensity.

L'outil Hakai Google Earth Engine Kelp (outil GEEK) a été développé dans le cadre d'une collaboration entre l'Institut Hakai, l'Université de Victoria et le ministère des Pêches et des Océans pour tirer parti des capacités de cloud computing pour analyser l'imagerie satellite Landsat (30 m) afin d'extraire l'étendue de la canopée et du varech. La méthodologie originale est décrite dans Nijland et al. 2019*.

Remarque : Ce jeu de données est conçu comme une « lecture seule », car nous continuons à améliorer les résultats. Il vise à démontrer l'utilité de l'archive Landsat pour cartographier le varech. Ces données sont visibles sur la carte Web GEEK disponible ici.

Ce package de données contient deux jeux de données :

Etendue annuelle maximale estivale du varech formant la canopée (1984 - 2019) en tant que rasters. Etendue maximale décennale du varech formant la canopée (1984 - 1990, 1991 - 2000, 2001 - 2010, 2011 - 2020)

Ce jeu de données a été généré à la suite de modifications apportées aux méthodologies GEEK originales. Les paramètres utilisés pour générer les rasters étaient des scènes d'images avec :

Plage de mois Imagescene = 1er mai - 30 septembre Clouds maximum dans la scène = 80% Marée maximale = 3,2 m (+0,5 MWL des marées de la côte centrale selon les méthodes KIM-1) Marée minimale = 0 m Tampon de rivage appliqué au masque de terrain = 1 pixel (30 m) NDVI* minimum (pour qu'un pixel individuel soit classé comme varech) = -0,05 Nombre minimum de fois qu'un pixel de varech individuel doit être détecté en tant que varech au cours d'une seule année = 30 % de toutes les détections d'une année donnée K moyenne minimale (moyenne du NDVI pour tous les pixels à un emplacement donné détecté comme varech) = -0,05 * NDVI = indice de végétation de différence normalisée.

Ces paramètres ont été choisis sur la base d'une évaluation de la précision à l'aide d'une étendue de varech dérivée d'images WorldView-2 (2 m) de juillet 2014 et août 2014. Ces données ont été rééchantillonnées à 30 m. Bien que de nombreuses itérations exécutées pour l'outil aient donné des résultats très similaires, des paramètres ont été sélectionnés qui ont maximisé la précision du varech pour la comparaison de 2014.

Les résultats de l'évaluation de la précision ont été les suivants : Erreur de commission de 50 % Erreur d'omission de 25 %

En termes simples, les méthodes actuelles conduisent à un niveau élevé de « faux positifs », mais elles capturent avec précision l'étendue du varech par rapport au jeu de données de validation. Cette erreur peut être attribuée à la sensibilité de l'utilisation d'un seul NDVI pour détecter le varech. Nous observons des variations des seuils NDVI à la fois au sein d'une seule scène et entre les scènes.

L'objectif du jeu de données de séries chronologiques est censé prendre en compte une partie de cette erreur, car les pixels détectés seulement un par décennie sont supprimés.

Ce jeu de données fait partie du programme de cartographie de l'habitat de Hakai. L'objectif principal du programme de cartographie de l'habitat de Hakai est de générer des inventaires spatiaux des habitats côtiers, d'étudier comment ces habitats évoluent au fil du temps et les moteurs de ce changement.

*Nijland, W., Reshitnyk, L. et Rubidge, E. (2019). Télédétection par satellite de varech formant une canopée sur un littoral complexe : une nouvelle procédure utilisant les archives d'images Landsat. Télédétection de l'environnement, 220, 41-50. doi:10.1016/j.rse.2018.10.032

The objectives of this study are to map the past and present LULC accurately, ensure the precision of classification and prediction models, and estimate future LULC changes using Google Earth Engine. LULC classification from 2014-2024 was performed using robust SmileCART and future LULC prediction for 2024, 2030, 2040, 2050, 2060 and 2070 was conducted using the Smile Random Forest (RF) algorithm incorporating various socio-economic variables. The classification model SmileCART was trained by using the European Space Agency World Cover data and validated through Landsat 8 satellite imagery, achieving training and test accuracies of 83% and 84% respectively. SmileRF prediction model showed an accuracy of 87% with a kappa coefficient of 0.86. The results indicate a decline in vegetation cover, snow & ice and water bodies, and an increase in built-up areas and cropland, with other classes showing fluctuations in the Jhelum and Chenab River basins from 2014 to 2070. These insights contribute to a deeper understanding of these critical watersheds, informing sustainable land management, water resource planning, and decision-making for the future sustainable development of the Jhelum and Chenab River basins using GEE and remote sensing data.

What is SEPAL?

SEPAL (https://sepal.io/) is a free and open source cloud computing platform for geo-spatial data access and processing. It empowers users to quickly process large amounts of data on their computer or mobile device. Users can create custom analysis ready data using freely available satellite imagery, generate and improve land use maps, analyze time series, run change detection and perform accuracy assessment and area estimation, among many other functionalities in the platform. Data can be created and analyzed for any place on Earth using SEPAL.

https://data.apps.fao.org/catalog/dataset/9c4d7c45-7620-44c4-b653-fbe13eb34b65/resource/63a3efa0-08ab-4ad6-9d4a-96af7b6a99ec/download/cambodia_mosaic_2020.png" alt="alt text" title="Figure 1: Best pixel mosaic of Landsat 8 data for 2020 over Cambodia">

Figure 1: Best pixel mosaic of Landsat 8 data for 2020 over Cambodia

SEPAL reaches over 5000 users in 180 countries for the creation of custom data products from freely available satellite data. SEPAL was developed as a part of the Open Foris suite, a set of free and open source software platforms and tools that facilitate flexible and efficient data collection, analysis and reporting. SEPAL combines and integrates modern geospatial data infrastructures and supercomputing power available through Google Earth Engine and Amazon Web Services with powerful open-source data processing software, such as R, ORFEO, GDAL, Python and Jupiter Notebooks. Users can easily access the archive of satellite imagery from NASA, the European Space Agency (ESA) as well as high spatial and temporal resolution data from Planet Labs and turn such images into data that can be used for reporting and better decision making.

National Forest Monitoring Systems in many countries have been strengthened by SEPAL, which provides technical government staff with computing resources and cutting edge technology to accurately map and monitor their forests. The platform was originally developed for monitoring forest carbon stock and stock changes for reducing emissions from deforestation and forest degradation (REDD+). The application of the tools on the platform now reach far beyond forest monitoring by providing different stakeholders access to cloud based image processing tools, remote sensing and machine learning for any application. Presently, users work on SEPAL for various applications related to land monitoring, land cover/use, land productivity, ecological zoning, ecosystem restoration monitoring, forest monitoring, near real time alerts for forest disturbances and fire, flood mapping, mapping impact of disasters, peatland rewetting status, and many others.

The Hand-in-Hand initiative enables countries that generate data through SEPAL to disseminate their data widely through the platform and to combine their data with the numerous other datasets available through Hand-in-Hand.

https://data.apps.fao.org/catalog/dataset/9c4d7c45-7620-44c4-b653-fbe13eb34b65/resource/868e59da-47b9-4736-93a9-f8d83f5731aa/download/probability_classification_over_zambia.png" alt="alt text" title="Figure 2: Image classification module for land monitoring and mapping. Probability classification over Zambia">

Figure 2: Image classification module for land monitoring and mapping. Probability classification over Zambia