resource-watch · tamaramegan · Sep 7, 2021
diff --git a/...eWatch/dataset_sync_files/RW_prod_staging_match_13f4e065-b579-41f0-938d-b97c9dd54ce2.json b/...eWatch/dataset_sync_files/RW_prod_staging_match_13f4e065-b579-41f0-938d-b97c9dd54ce2.json
@@ -1 +1 @@
-[{"type": "dataset", "prodId": "13f4e065-b579-41f0-938d-b97c9dd54ce2", "stagingId": "e16d6b2d-7084-4891-878c-ae47fc49da52"}, {"type": "vocabulary", "prodId": null, "stagingId": [{"id": "knowledge_graph", "type": "vocabulary", "attributes": {"tags": ["geospatial", "global", "raster", "historical", "soil", "erosion", "water", "flood", "agriculture"], "name": "knowledge_graph", "application": "rw"}}]}, {"type": "layer", "prodId": "555d1d25-3112-430f-9fd5-5b00aa94b613", "stagingId": "6ab77e36-df9d-4047-8641-ca49922522b5"}, {"type": "layer", "prodId": "503f1001-e69c-4111-8ef6-5d34fa451e94", "stagingId": "6a7c59b9-2e79-42f4-9caf-6b1bf068ae09"}, {"type": "layer", "prodId": "065e6200-1670-4832-b0ae-70b7852a9875", "stagingId": "f37d8246-ddb6-4dbb-a9db-715fa889a88b"}, {"type": "layer", "prodId": "e6eafefd-bb28-429e-9fff-1d6205f5d5b2", "stagingId": "70b7071b-2c3a-425e-9ba2-51cab97fc681"}, {"type": "layer", "prodId": "5c79ccac-2c31-4399-9a54-ae964feb7419", "stagingId": "5ad82c6b-b5a4-47c5-8417-e768eff56790"}, {"type": "widget", "prodId": "fd292bc0-be35-4fc9-8dc1-8485ea0360b8", "stagingId": "6e26a922-3c1b-40cf-8fe2-8059b25e7261"}, {"type": "widget", "prodId": "a72e4432-dd8f-43b1-a33c-b0d61432d6f1", "stagingId": "7f18260c-aada-4dd3-baf1-a51af52947db"}, {"type": "metadata", "prodId": "60eca114b66636001aa44c11", "stagingId": [{"id": "611147d2cd40d0001afca550", "type": "metadata", "attributes": {"dataset": "e16d6b2d-7084-4891-878c-ae47fc49da52", "application": "rw", "resource": {"id": "7f18260c-aada-4dd3-baf1-a51af52947db", "type": "widget"}, "language": "en", "info": {"caption": "The Soil Erosion Prevalence dataset, produced by the World Agroforestry Centre, provides predictions for the percentage of area with soil erosion across the global tropics (between the parallels of 40\u00b0 south and 40\u00b0 north). The authors predicted soil erosion for the years 2002, 2007, 2012, 2017, and 2020 from satellite imagery at 500 m resolution using machine learning algorithms trained with field observations. ", "widgetLinks": [{"link": "https://www.mdpi.com/2072-4292/11/15/1800", "name": "Learn more"}]}, "createdAt": "2021-08-09T15:20:50.370Z", "updatedAt": "2021-08-09T15:52:27.938Z", "status": "published"}}]}, {"type": "metadata", "prodId": "60d64348173c43001ad6b8d1", "stagingId": [{"id": "60d645b3cb6c5c001a6546c3", "type": "metadata", "attributes": {"dataset": "e16d6b2d-7084-4891-878c-ae47fc49da52", "application": "rw", "resource": {"id": "e16d6b2d-7084-4891-878c-ae47fc49da52", "type": "dataset"}, "language": "en", "name": "Soil Erosion", "description": "### Overview  \n  \nThe Soil Erosion Prevalence dataset, produced by the World Agroforestry Centre (ICRAF), provides predictions for the percentage of area with soil erosion across the global tropics (between the parallels of 40\u00b0 south and 40\u00b0 north). The authors predicted soil erosion for the years 2002, 2007, 2012, 2017, and 2020 from satellite imagery at 500 m resolution using machine learning algorithms trained with field observations. \n\n  \n  \nSoil is eroding more quickly than its being formed, contributing to the degradation of millions of hectares of land globally. The nutrient rich layer of soil on the surface, called topsoil, is vulnerable to wind and water erosion and its loss is accelerated by modifications in land use. Erosion of this critical layer of soil comes at a [major economic and environmental cost](https://www.wri.org/insights/causes-and-effects-soil-erosion-and-how-prevent-it). It causes [billions](https://www.sciencedirect.com/science/article/pii/S0264837718319343) of dollars of losses due to decreased soil fertility, reduced crop yields, and increased water usage. The eroded soil can be carried into rivers and streams. This creates a heavy layer of sediment which is carried downstream. This process clogs waterways, prevents smooth water flow, and may eventually lead to flooding. Other environmental costs include loss of productivity and biodiversity, decreased resilience of marine and terrestrial ecosystems, and increased vulnerability to climate change and food insecurity. \n\n  \n  \nEstimates of both spatial and temporal dynamics of soil erosion are needed to better track the occurrence and severity of erosion in landscapes over time. The main objectives of this dataset are to provide rapid assessments of soil erosion for spatially distributed monitoring as well as assess changes in soil erosion prevalence over time. The spatial assessments of erosion provide estimates of land degradation hotspots and can be combined with other indicators of ecosystem health, including social factors, to better assess and identify drivers of land degradation and target land management interventions to reverse degradation.\n  \n  \n### Methodology  \n  \nThe authors used a combination of systematically collected field observations of erosion and remote sensing to predict the spatial and temporal distribution of soil erosion prevalence at moderate spatial resolution (500 m) for the years 2002, 2007, 2012, 2017, and 2020. The model was trained and tested using field data collected via [the Land Degradation Surveillance Framework](http://landscapeportal.org/blog/2015/03/25/the-land-degradation-surveillance-framework-ldsf) (LDSF). As part of the LDSF, sample sites are divided into plots and subplots. Erosion prevalence was scored at the plot level by summing up the number of subplots with visible signs of erosion, with 0 being no observed erosion and 4 being erosion observed in all four subplots. The authors took a cut-off at three subplots or more (>50%) to represent \u201csevere\u201d soil erosion. Erosion prevalence was modeled using a decision-tree approach known as [Random Forests](https://www.jstatsoft.org/article/view/v077i01/v77i01.pdf) to generate a classification model for soil erosion prevalence using imagery from the [Moderate Resolution Imaging Spectroradiomer](https://modis.gsfc.nasa.gov/data/) (MODIS) platform. Annual composites were matched to the years of LDSF field data collection and fitted to annual composite reflectance data for 2002, 2007, 2012, and 2017. Soil erosion prevalence is displayed on the maps as the percentage of area (of a pixel) predicted to be eroded. Model accuracy for the detection of erosion was about 86% and accuracy for non-detection was about 91%.\n\n  \n  \nFor the full documentation, please see the source [methodology](https://www.mdpi.com/2072-4292/11/15/1800).\n  \n  \n### Additional Information  \n  \nFor access to additional information, click on the \u201cLearn more\u201d button.  \n  \n### Visualizing the Data  \n  \nOur team reformatted this dataset before displaying it on Resource Watch. See the documentation on how Resource Watch retrieved the data on [Github](https://github.com/resource-watch/data-pre-processing/tree/master/wat_070_rw0_soil_erosion).  \n  \n### Disclaimer  \n  \nExcerpts of this description page were taken from the source metadata. Resource Watch shows only a subset of the dataset. For access to the full dataset and additional information, click on the \u201cLearn more\u201d button.", "source": "ICRAF", "info": {"rwId": "wat.070.rw0", "data_type": "Raster", "name": "Soil Erosion", "sources": [{"source-name": "", "id": 0, "source-description": "World Agroforestry Centre (ICRAF)\n"}], "technical_title": "Predicted Soil Erosion Prevalence", "functions": "Predicted percentage of eroded area for the years 2002, 2007, 2012, 2017, and 2020", "cautions": "- There is no calibration/validation data for the US, Europe and Australia, although these regions are included in the maps.\n  \n  \n- Hyperarid areas have been masked out (excluded) based on TRMM rainfall data.\n  \n  \n- The drivers of soil erosion, which include both social and ecological factors can be challenging to detect remotely.\n", "citation": "V\u00e5gen T-G, Winowiecki LA. Predicting the Spatial Distribution and Severity of Soil Erosion in the Global Tropics using Satellite Remote Sensing. Remote Sensing. 2019; 11(15):1800. https://doi.org/10.3390/rs11151800. Accessed through Resource Watch, (date). [www.resourcewatch.org](https://www.resourcewatch.org).", "license": "Restrictions Apply", "license_link": null, "geographic_coverage": "40\u00b0S\u2013 40\u00b0N", "spatial_resolution": "500 m", "date_of_content": "2002, 2007, 2012, 2017, 2020", "frequency_of_updates": "Unknown", "learn_more_link": "https://www.mdpi.com/2072-4292/11/15/1800", "data_download_link": "http://wri-public-data.s3.amazonaws.com/resourcewatch/raster/wat_070_rw0_soil_erosion.zip", "data_download_original_link": "https://www.mdpi.com/2072-4292/11/15/1800"}, "createdAt": "2021-06-25T21:08:03.264Z", "updatedAt": "2021-08-09T15:52:29.537Z", "status": "published"}}]}]
+[{"prodId": "13f4e065-b579-41f0-938d-b97c9dd54ce2", "stagingId": "e16d6b2d-7084-4891-878c-ae47fc49da52", "type": "dataset"}, {"prodId": null, "stagingId": [{"attributes": {"application": "rw", "name": "knowledge_graph", "tags": ["geospatial", "global", "raster", "historical", "soil", "erosion", "water", "flood", "agriculture"]}, "id": "knowledge_graph", "type": "vocabulary"}], "type": "vocabulary"}, {"prodId": "555d1d25-3112-430f-9fd5-5b00aa94b613", "stagingId": "6ab77e36-df9d-4047-8641-ca49922522b5", "type": "layer"}, {"prodId": "503f1001-e69c-4111-8ef6-5d34fa451e94", "stagingId": "6a7c59b9-2e79-42f4-9caf-6b1bf068ae09", "type": "layer"}, {"prodId": "065e6200-1670-4832-b0ae-70b7852a9875", "stagingId": "f37d8246-ddb6-4dbb-a9db-715fa889a88b", "type": "layer"}, {"prodId": "e6eafefd-bb28-429e-9fff-1d6205f5d5b2", "stagingId": "70b7071b-2c3a-425e-9ba2-51cab97fc681", "type": "layer"}, {"prodId": "5c79ccac-2c31-4399-9a54-ae964feb7419", "stagingId": "5ad82c6b-b5a4-47c5-8417-e768eff56790", "type": "layer"}, {"prodId": "fd292bc0-be35-4fc9-8dc1-8485ea0360b8", "stagingId": "6e26a922-3c1b-40cf-8fe2-8059b25e7261", "type": "widget"}, {"prodId": "dbd066c0-a1c0-4e0c-9eec-9294b96b49ef", "stagingId": "c5237807-c4f4-4ec1-8d6a-3b5931679eba", "type": "widget"}, {"prodId": "a72e4432-dd8f-43b1-a33c-b0d61432d6f1", "stagingId": "7f18260c-aada-4dd3-baf1-a51af52947db", "type": "widget"}, {"prodId": "60eca114b66636001aa44c11", "stagingId": [{"attributes": {"application": "rw", "createdAt": "2021-08-09T15:20:50.370Z", "dataset": "e16d6b2d-7084-4891-878c-ae47fc49da52", "info": {"caption": "The Soil Erosion Prevalence dataset, produced by the World Agroforestry Centre, provides predictions for the percentage of area with soil erosion across the global tropics (between the parallels of 40\u00b0 south and 40\u00b0 north). The authors predicted soil erosion for the years 2002, 2007, 2012, 2017, and 2020 from satellite imagery at 500 m resolution using machine learning algorithms trained with field observations. ", "widgetLinks": [{"link": "https://www.mdpi.com/2072-4292/11/15/1800", "name": "Learn more"}]}, "language": "en", "resource": {"id": "7f18260c-aada-4dd3-baf1-a51af52947db", "type": "widget"}, "status": "published", "updatedAt": "2021-09-07T09:11:07.303Z"}, "id": "611147d2cd40d0001afca550", "type": "metadata"}], "type": "metadata"}, {"prodId": "60d64348173c43001ad6b8d1", "stagingId": [{"attributes": {"application": "rw", "createdAt": "2021-06-25T21:08:03.264Z", "dataset": "e16d6b2d-7084-4891-878c-ae47fc49da52", "description": "### Overview  \n  \nThe Soil Erosion Prevalence dataset, produced by the World Agroforestry Centre (ICRAF), provides predictions for the percentage of area with soil erosion across the global tropics (between the parallels of 40\u00b0 south and 40\u00b0 north). The authors predicted soil erosion for the years 2002, 2007, 2012, 2017, and 2020 from satellite imagery at 500 m resolution using machine learning algorithms trained with field observations. \n\n  \n  \nSoil is eroding more quickly than its being formed, contributing to the degradation of millions of hectares of land globally. The nutrient rich layer of soil on the surface, called topsoil, is vulnerable to wind and water erosion and its loss is accelerated by modifications in land use. Erosion of this critical layer of soil comes at a [major economic and environmental cost](https://www.wri.org/insights/causes-and-effects-soil-erosion-and-how-prevent-it). It causes [billions](https://www.sciencedirect.com/science/article/pii/S0264837718319343) of dollars of losses due to decreased soil fertility, reduced crop yields, and increased water usage. The eroded soil can be carried into rivers and streams. This creates a heavy layer of sediment which is carried downstream. This process clogs waterways, prevents smooth water flow, and may eventually lead to flooding. Other environmental costs include loss of productivity and biodiversity, decreased resilience of marine and terrestrial ecosystems, and increased vulnerability to climate change and food insecurity. \n\n  \n  \nEstimates of both spatial and temporal dynamics of soil erosion are needed to better track the occurrence and severity of erosion in landscapes over time. The main objectives of this dataset are to provide rapid assessments of soil erosion for spatially distributed monitoring as well as assess changes in soil erosion prevalence over time. The spatial assessments of erosion provide estimates of land degradation hotspots and can be combined with other indicators of ecosystem health, including social factors, to better assess and identify drivers of land degradation and target land management interventions to reverse degradation.\n  \n  \n### Methodology  \n  \nThe authors used a combination of systematically collected field observations of erosion and remote sensing to predict the spatial and temporal distribution of soil erosion prevalence at moderate spatial resolution (500 m) for the years 2002, 2007, 2012, 2017, and 2020. The model was trained and tested using field data collected via [the Land Degradation Surveillance Framework](http://landscapeportal.org/blog/2015/03/25/the-land-degradation-surveillance-framework-ldsf) (LDSF). As part of the LDSF, sample sites are divided into plots and subplots. Erosion prevalence was scored at the plot level by summing up the number of subplots with visible signs of erosion, with 0 being no observed erosion and 4 being erosion observed in all four subplots. The authors took a cut-off at three subplots or more (>50%) to represent \u201csevere\u201d soil erosion. Erosion prevalence was modeled using a decision-tree approach known as [Random Forests](https://www.jstatsoft.org/article/view/v077i01/v77i01.pdf) to generate a classification model for soil erosion prevalence using imagery from the [Moderate Resolution Imaging Spectroradiomer](https://modis.gsfc.nasa.gov/data/) (MODIS) platform. Annual composites were matched to the years of LDSF field data collection and fitted to annual composite reflectance data for 2002, 2007, 2012, and 2017. Soil erosion prevalence is displayed on the maps as the percentage of area (of a pixel) predicted to be eroded. Model accuracy for the detection of erosion was about 86% and accuracy for non-detection was about 91%.\n\n  \n  \nFor the full documentation, please see the source [methodology](https://www.mdpi.com/2072-4292/11/15/1800).\n  \n  \n### Additional Information  \n  \nFor access to additional information, click on the \u201cLearn more\u201d button.  \n  \n### Visualizing the Data  \n  \nOur team reformatted this dataset before displaying it on Resource Watch. See the documentation on how Resource Watch retrieved the data on [Github](https://github.com/resource-watch/data-pre-processing/tree/master/wat_070_rw0_soil_erosion).  \n  \n### Disclaimer  \n  \nExcerpts of this description page were taken from the source metadata. Resource Watch shows only a subset of the dataset. For access to the full dataset and additional information, click on the \u201cLearn more\u201d button.", "info": {"cautions": "- There is no calibration/validation data for the US, Europe and Australia, although these regions are included in the maps.\n  \n  \n- Hyperarid areas have been masked out (excluded) based on TRMM rainfall data.\n  \n  \n- The drivers of soil erosion, which include both social and ecological factors can be challenging to detect remotely.\n", "citation": "V\u00e5gen T-G, Winowiecki LA. Predicting the Spatial Distribution and Severity of Soil Erosion in the Global Tropics using Satellite Remote Sensing. Remote Sensing. 2019; 11(15):1800. https://doi.org/10.3390/rs11151800. Accessed through Resource Watch, (date). [www.resourcewatch.org](https://www.resourcewatch.org).", "data_download_link": "http://wri-public-data.s3.amazonaws.com/resourcewatch/raster/wat_070_rw0_soil_erosion.zip", "data_download_original_link": "https://www.mdpi.com/2072-4292/11/15/1800", "data_type": "Raster", "date_of_content": "2002, 2007, 2012, 2017, 2020", "frequency_of_updates": "Unknown", "functions": "Predicted percentage of eroded area for the years 2002, 2007, 2012, 2017, and 2020", "geographic_coverage": "40\u00b0S\u2013 40\u00b0N", "learn_more_link": "https://www.mdpi.com/2072-4292/11/15/1800", "license": "Restrictions Apply", "license_link": null, "name": "Soil Erosion", "rwId": "wat.070a.rw0", "sources": [{"id": 0, "source-description": "World Agroforestry Centre (ICRAF)\n", "source-name": ""}], "spatial_resolution": "500 m", "technical_title": "Predicted Soil Erosion Prevalence"}, "language": "en", "name": "Soil Erosion", "resource": {"id": "e16d6b2d-7084-4891-878c-ae47fc49da52", "type": "dataset"}, "source": "ICRAF", "status": "published", "updatedAt": "2021-09-07T09:11:08.733Z"}, "id": "60d645b3cb6c5c001a6546c3", "type": "metadata"}], "type": "metadata"}]
Original file line number	Diff line number	Diff line change
		@@ -1 +1 @@
		[{"type": "dataset", "prodId": "13f4e065-b579-41f0-938d-b97c9dd54ce2", "stagingId": "e16d6b2d-7084-4891-878c-ae47fc49da52"}, {"type": "vocabulary", "prodId": null, "stagingId": [{"id": "knowledge_graph", "type": "vocabulary", "attributes": {"tags": ["geospatial", "global", "raster", "historical", "soil", "erosion", "water", "flood", "agriculture"], "name": "knowledge_graph", "application": "rw"}}]}, {"type": "layer", "prodId": "555d1d25-3112-430f-9fd5-5b00aa94b613", "stagingId": "6ab77e36-df9d-4047-8641-ca49922522b5"}, {"type": "layer", "prodId": "503f1001-e69c-4111-8ef6-5d34fa451e94", "stagingId": "6a7c59b9-2e79-42f4-9caf-6b1bf068ae09"}, {"type": "layer", "prodId": "065e6200-1670-4832-b0ae-70b7852a9875", "stagingId": "f37d8246-ddb6-4dbb-a9db-715fa889a88b"}, {"type": "layer", "prodId": "e6eafefd-bb28-429e-9fff-1d6205f5d5b2", "stagingId": "70b7071b-2c3a-425e-9ba2-51cab97fc681"}, {"type": "layer", "prodId": "5c79ccac-2c31-4399-9a54-ae964feb7419", "stagingId": "5ad82c6b-b5a4-47c5-8417-e768eff56790"}, {"type": "widget", "prodId": "fd292bc0-be35-4fc9-8dc1-8485ea0360b8", "stagingId": "6e26a922-3c1b-40cf-8fe2-8059b25e7261"}, {"type": "widget", "prodId": "a72e4432-dd8f-43b1-a33c-b0d61432d6f1", "stagingId": "7f18260c-aada-4dd3-baf1-a51af52947db"}, {"type": "metadata", "prodId": "60eca114b66636001aa44c11", "stagingId": [{"id": "611147d2cd40d0001afca550", "type": "metadata", "attributes": {"dataset": "e16d6b2d-7084-4891-878c-ae47fc49da52", "application": "rw", "resource": {"id": "7f18260c-aada-4dd3-baf1-a51af52947db", "type": "widget"}, "language": "en", "info": {"caption": "The Soil Erosion Prevalence dataset, produced by the World Agroforestry Centre, provides predictions for the percentage of area with soil erosion across the global tropics (between the parallels of 40\u00b0 south and 40\u00b0 north). The authors predicted soil erosion for the years 2002, 2007, 2012, 2017, and 2020 from satellite imagery at 500 m resolution using machine learning algorithms trained with field observations. ", "widgetLinks": [{"link": "https://www.mdpi.com/2072-4292/11/15/1800", "name": "Learn more"}]}, "createdAt": "2021-08-09T15:20:50.370Z", "updatedAt": "2021-08-09T15:52:27.938Z", "status": "published"}}]}, {"type": "metadata", "prodId": "60d64348173c43001ad6b8d1", "stagingId": [{"id": "60d645b3cb6c5c001a6546c3", "type": "metadata", "attributes": {"dataset": "e16d6b2d-7084-4891-878c-ae47fc49da52", "application": "rw", "resource": {"id": "e16d6b2d-7084-4891-878c-ae47fc49da52", "type": "dataset"}, "language": "en", "name": "Soil Erosion", "description": "### Overview \n \nThe Soil Erosion Prevalence dataset, produced by the World Agroforestry Centre (ICRAF), provides predictions for the percentage of area with soil erosion across the global tropics (between the parallels of 40\u00b0 south and 40\u00b0 north). The authors predicted soil erosion for the years 2002, 2007, 2012, 2017, and 2020 from satellite imagery at 500 m resolution using machine learning algorithms trained with field observations. \n\n \n \nSoil is eroding more quickly than its being formed, contributing to the degradation of millions of hectares of land globally. The nutrient rich layer of soil on the surface, called topsoil, is vulnerable to wind and water erosion and its loss is accelerated by modifications in land use. Erosion of this critical layer of soil comes at a [major economic and environmental cost](https://www.wri.org/insights/causes-and-effects-soil-erosion-and-how-prevent-it). It causes [billions](https://www.sciencedirect.com/science/article/pii/S0264837718319343) of dollars of losses due to decreased soil fertility, reduced crop yields, and increased water usage. The eroded soil can be carried into rivers and streams. This creates a heavy layer of sediment which is carried downstream. This process clogs waterways, prevents smooth water flow, and may eventually lead to flooding. Other environmental costs include loss of productivity and biodiversity, decreased resilience of marine and terrestrial ecosystems, and increased vulnerability to climate change and food insecurity. \n\n \n \nEstimates of both spatial and temporal dynamics of soil erosion are needed to better track the occurrence and severity of erosion in landscapes over time. The main objectives of this dataset are to provide rapid assessments of soil erosion for spatially distributed monitoring as well as assess changes in soil erosion prevalence over time. The spatial assessments of erosion provide estimates of land degradation hotspots and can be combined with other indicators of ecosystem health, including social factors, to better assess and identify drivers of land degradation and target land management interventions to reverse degradation.\n \n \n### Methodology \n \nThe authors used a combination of systematically collected field observations of erosion and remote sensing to predict the spatial and temporal distribution of soil erosion prevalence at moderate spatial resolution (500 m) for the years 2002, 2007, 2012, 2017, and 2020. The model was trained and tested using field data collected via [the Land Degradation Surveillance Framework](http://landscapeportal.org/blog/2015/03/25/the-land-degradation-surveillance-framework-ldsf) (LDSF). As part of the LDSF, sample sites are divided into plots and subplots. Erosion prevalence was scored at the plot level by summing up the number of subplots with visible signs of erosion, with 0 being no observed erosion and 4 being erosion observed in all four subplots. The authors took a cut-off at three subplots or more (>50%) to represent \u201csevere\u201d soil erosion. Erosion prevalence was modeled using a decision-tree approach known as [Random Forests](https://www.jstatsoft.org/article/view/v077i01/v77i01.pdf) to generate a classification model for soil erosion prevalence using imagery from the [Moderate Resolution Imaging Spectroradiomer](https://modis.gsfc.nasa.gov/data/) (MODIS) platform. Annual composites were matched to the years of LDSF field data collection and fitted to annual composite reflectance data for 2002, 2007, 2012, and 2017. Soil erosion prevalence is displayed on the maps as the percentage of area (of a pixel) predicted to be eroded. Model accuracy for the detection of erosion was about 86% and accuracy for non-detection was about 91%.\n\n \n \nFor the full documentation, please see the source [methodology](https://www.mdpi.com/2072-4292/11/15/1800).\n \n \n### Additional Information \n \nFor access to additional information, click on the \u201cLearn more\u201d button. \n \n### Visualizing the Data \n \nOur team reformatted this dataset before displaying it on Resource Watch. See the documentation on how Resource Watch retrieved the data on [Github](https://github.com/resource-watch/data-pre-processing/tree/master/wat_070_rw0_soil_erosion). \n \n### Disclaimer \n \nExcerpts of this description page were taken from the source metadata. Resource Watch shows only a subset of the dataset. For access to the full dataset and additional information, click on the \u201cLearn more\u201d button.", "source": "ICRAF", "info": {"rwId": "wat.070.rw0", "data_type": "Raster", "name": "Soil Erosion", "sources": [{"source-name": "", "id": 0, "source-description": "World Agroforestry Centre (ICRAF)\n"}], "technical_title": "Predicted Soil Erosion Prevalence", "functions": "Predicted percentage of eroded area for the years 2002, 2007, 2012, 2017, and 2020", "cautions": "- There is no calibration/validation data for the US, Europe and Australia, although these regions are included in the maps.\n \n \n- Hyperarid areas have been masked out (excluded) based on TRMM rainfall data.\n \n \n- The drivers of soil erosion, which include both social and ecological factors can be challenging to detect remotely.\n", "citation": "V\u00e5gen T-G, Winowiecki LA. Predicting the Spatial Distribution and Severity of Soil Erosion in the Global Tropics using Satellite Remote Sensing. Remote Sensing. 2019; 11(15):1800. https://doi.org/10.3390/rs11151800. Accessed through Resource Watch, (date). [www.resourcewatch.org](https://www.resourcewatch.org).", "license": "Restrictions Apply", "license_link": null, "geographic_coverage": "40\u00b0S\u2013 40\u00b0N", "spatial_resolution": "500 m", "date_of_content": "2002, 2007, 2012, 2017, 2020", "frequency_of_updates": "Unknown", "learn_more_link": "https://www.mdpi.com/2072-4292/11/15/1800", "data_download_link": "http://wri-public-data.s3.amazonaws.com/resourcewatch/raster/wat_070_rw0_soil_erosion.zip", "data_download_original_link": "https://www.mdpi.com/2072-4292/11/15/1800"}, "createdAt": "2021-06-25T21:08:03.264Z", "updatedAt": "2021-08-09T15:52:29.537Z", "status": "published"}}]}]
		[{"prodId": "13f4e065-b579-41f0-938d-b97c9dd54ce2", "stagingId": "e16d6b2d-7084-4891-878c-ae47fc49da52", "type": "dataset"}, {"prodId": null, "stagingId": [{"attributes": {"application": "rw", "name": "knowledge_graph", "tags": ["geospatial", "global", "raster", "historical", "soil", "erosion", "water", "flood", "agriculture"]}, "id": "knowledge_graph", "type": "vocabulary"}], "type": "vocabulary"}, {"prodId": "555d1d25-3112-430f-9fd5-5b00aa94b613", "stagingId": "6ab77e36-df9d-4047-8641-ca49922522b5", "type": "layer"}, {"prodId": "503f1001-e69c-4111-8ef6-5d34fa451e94", "stagingId": "6a7c59b9-2e79-42f4-9caf-6b1bf068ae09", "type": "layer"}, {"prodId": "065e6200-1670-4832-b0ae-70b7852a9875", "stagingId": "f37d8246-ddb6-4dbb-a9db-715fa889a88b", "type": "layer"}, {"prodId": "e6eafefd-bb28-429e-9fff-1d6205f5d5b2", "stagingId": "70b7071b-2c3a-425e-9ba2-51cab97fc681", "type": "layer"}, {"prodId": "5c79ccac-2c31-4399-9a54-ae964feb7419", "stagingId": "5ad82c6b-b5a4-47c5-8417-e768eff56790", "type": "layer"}, {"prodId": "fd292bc0-be35-4fc9-8dc1-8485ea0360b8", "stagingId": "6e26a922-3c1b-40cf-8fe2-8059b25e7261", "type": "widget"}, {"prodId": "dbd066c0-a1c0-4e0c-9eec-9294b96b49ef", "stagingId": "c5237807-c4f4-4ec1-8d6a-3b5931679eba", "type": "widget"}, {"prodId": "a72e4432-dd8f-43b1-a33c-b0d61432d6f1", "stagingId": "7f18260c-aada-4dd3-baf1-a51af52947db", "type": "widget"}, {"prodId": "60eca114b66636001aa44c11", "stagingId": [{"attributes": {"application": "rw", "createdAt": "2021-08-09T15:20:50.370Z", "dataset": "e16d6b2d-7084-4891-878c-ae47fc49da52", "info": {"caption": "The Soil Erosion Prevalence dataset, produced by the World Agroforestry Centre, provides predictions for the percentage of area with soil erosion across the global tropics (between the parallels of 40\u00b0 south and 40\u00b0 north). The authors predicted soil erosion for the years 2002, 2007, 2012, 2017, and 2020 from satellite imagery at 500 m resolution using machine learning algorithms trained with field observations. ", "widgetLinks": [{"link": "https://www.mdpi.com/2072-4292/11/15/1800", "name": "Learn more"}]}, "language": "en", "resource": {"id": "7f18260c-aada-4dd3-baf1-a51af52947db", "type": "widget"}, "status": "published", "updatedAt": "2021-09-07T09:11:07.303Z"}, "id": "611147d2cd40d0001afca550", "type": "metadata"}], "type": "metadata"}, {"prodId": "60d64348173c43001ad6b8d1", "stagingId": [{"attributes": {"application": "rw", "createdAt": "2021-06-25T21:08:03.264Z", "dataset": "e16d6b2d-7084-4891-878c-ae47fc49da52", "description": "### Overview \n \nThe Soil Erosion Prevalence dataset, produced by the World Agroforestry Centre (ICRAF), provides predictions for the percentage of area with soil erosion across the global tropics (between the parallels of 40\u00b0 south and 40\u00b0 north). The authors predicted soil erosion for the years 2002, 2007, 2012, 2017, and 2020 from satellite imagery at 500 m resolution using machine learning algorithms trained with field observations. \n\n \n \nSoil is eroding more quickly than its being formed, contributing to the degradation of millions of hectares of land globally. The nutrient rich layer of soil on the surface, called topsoil, is vulnerable to wind and water erosion and its loss is accelerated by modifications in land use. Erosion of this critical layer of soil comes at a [major economic and environmental cost](https://www.wri.org/insights/causes-and-effects-soil-erosion-and-how-prevent-it). It causes [billions](https://www.sciencedirect.com/science/article/pii/S0264837718319343) of dollars of losses due to decreased soil fertility, reduced crop yields, and increased water usage. The eroded soil can be carried into rivers and streams. This creates a heavy layer of sediment which is carried downstream. This process clogs waterways, prevents smooth water flow, and may eventually lead to flooding. Other environmental costs include loss of productivity and biodiversity, decreased resilience of marine and terrestrial ecosystems, and increased vulnerability to climate change and food insecurity. \n\n \n \nEstimates of both spatial and temporal dynamics of soil erosion are needed to better track the occurrence and severity of erosion in landscapes over time. The main objectives of this dataset are to provide rapid assessments of soil erosion for spatially distributed monitoring as well as assess changes in soil erosion prevalence over time. The spatial assessments of erosion provide estimates of land degradation hotspots and can be combined with other indicators of ecosystem health, including social factors, to better assess and identify drivers of land degradation and target land management interventions to reverse degradation.\n \n \n### Methodology \n \nThe authors used a combination of systematically collected field observations of erosion and remote sensing to predict the spatial and temporal distribution of soil erosion prevalence at moderate spatial resolution (500 m) for the years 2002, 2007, 2012, 2017, and 2020. The model was trained and tested using field data collected via [the Land Degradation Surveillance Framework](http://landscapeportal.org/blog/2015/03/25/the-land-degradation-surveillance-framework-ldsf) (LDSF). As part of the LDSF, sample sites are divided into plots and subplots. Erosion prevalence was scored at the plot level by summing up the number of subplots with visible signs of erosion, with 0 being no observed erosion and 4 being erosion observed in all four subplots. The authors took a cut-off at three subplots or more (>50%) to represent \u201csevere\u201d soil erosion. Erosion prevalence was modeled using a decision-tree approach known as [Random Forests](https://www.jstatsoft.org/article/view/v077i01/v77i01.pdf) to generate a classification model for soil erosion prevalence using imagery from the [Moderate Resolution Imaging Spectroradiomer](https://modis.gsfc.nasa.gov/data/) (MODIS) platform. Annual composites were matched to the years of LDSF field data collection and fitted to annual composite reflectance data for 2002, 2007, 2012, and 2017. Soil erosion prevalence is displayed on the maps as the percentage of area (of a pixel) predicted to be eroded. Model accuracy for the detection of erosion was about 86% and accuracy for non-detection was about 91%.\n\n \n \nFor the full documentation, please see the source [methodology](https://www.mdpi.com/2072-4292/11/15/1800).\n \n \n### Additional Information \n \nFor access to additional information, click on the \u201cLearn more\u201d button. \n \n### Visualizing the Data \n \nOur team reformatted this dataset before displaying it on Resource Watch. See the documentation on how Resource Watch retrieved the data on [Github](https://github.com/resource-watch/data-pre-processing/tree/master/wat_070_rw0_soil_erosion). \n \n### Disclaimer \n \nExcerpts of this description page were taken from the source metadata. Resource Watch shows only a subset of the dataset. For access to the full dataset and additional information, click on the \u201cLearn more\u201d button.", "info": {"cautions": "- There is no calibration/validation data for the US, Europe and Australia, although these regions are included in the maps.\n \n \n- Hyperarid areas have been masked out (excluded) based on TRMM rainfall data.\n \n \n- The drivers of soil erosion, which include both social and ecological factors can be challenging to detect remotely.\n", "citation": "V\u00e5gen T-G, Winowiecki LA. Predicting the Spatial Distribution and Severity of Soil Erosion in the Global Tropics using Satellite Remote Sensing. Remote Sensing. 2019; 11(15):1800. https://doi.org/10.3390/rs11151800. Accessed through Resource Watch, (date). [www.resourcewatch.org](https://www.resourcewatch.org).", "data_download_link": "http://wri-public-data.s3.amazonaws.com/resourcewatch/raster/wat_070_rw0_soil_erosion.zip", "data_download_original_link": "https://www.mdpi.com/2072-4292/11/15/1800", "data_type": "Raster", "date_of_content": "2002, 2007, 2012, 2017, 2020", "frequency_of_updates": "Unknown", "functions": "Predicted percentage of eroded area for the years 2002, 2007, 2012, 2017, and 2020", "geographic_coverage": "40\u00b0S\u2013 40\u00b0N", "learn_more_link": "https://www.mdpi.com/2072-4292/11/15/1800", "license": "Restrictions Apply", "license_link": null, "name": "Soil Erosion", "rwId": "wat.070a.rw0", "sources": [{"id": 0, "source-description": "World Agroforestry Centre (ICRAF)\n", "source-name": ""}], "spatial_resolution": "500 m", "technical_title": "Predicted Soil Erosion Prevalence"}, "language": "en", "name": "Soil Erosion", "resource": {"id": "e16d6b2d-7084-4891-878c-ae47fc49da52", "type": "dataset"}, "source": "ICRAF", "status": "published", "updatedAt": "2021-09-07T09:11:08.733Z"}, "id": "60d645b3cb6c5c001a6546c3", "type": "metadata"}], "type": "metadata"}]