The CYGNSS Level 3 UC Berkeley Watermask Record Version 3.1 was developed by CYGNSS investigators in the Department of Civil and Environmental Engineering at the University of California, Berkeley. CYGNSS was launched on 15 December 2016, it is a NASA Earth System Science Pathfinder Mission that was launched with the purpose of collecting the first frequent space‐based measurements of surface wind speeds in the inner core of tropical cyclones. Originally made up of a constellation of eight micro-satellites, the observatories provide nearly gap-free Earth coverage using an orbital inclination of approximately 35° from the equator, with a mean (i.e., average) revisit time of seven hours and a median revisit time of three hours. This dataset is derived from version 3.1 of the CYGNSS L1 SDR dataset (https://doi.org/10.5067/CYGNS-L1X31), and provides monthly binary inland surface water classification data at a 0.01-degree (~1x1 kilometer) resolution with a 1-month latency. This product, known as the UC Berkeley Random Walk Algorithm WaterMask from CYGNSS (Berkeley-RWAWC), generates water classification for a given _location based on CYGNSS observations combined with a random walker algorithm. The watermask variable includes binary values indicating land (0), surface water (1), and no data/ocean (-99). The data product is archived in monthly files in netCDF-4 format and covers the period from August 2018 to present.

The 250m water mask product utilizes the SWBD (SRTM Water Body Data) and complement it with information from 250m MODIS data to create a complete representation of global surface water. The original intent of this product is not for hydrologic modeling, rather for masking water in products.

Auxiliary information for pixel cloud product indicating to which water bodies the pixels are assigned in river and lake products. Also includes height-constrained pixel geolocation after reach- or lake-scale averaging. Point cloud over tile (approx 64x64 km2); half swath (left or right side of full swath). Available in netCDF-4 file format.

The SWOT Level 2 Water Mask Raster Image 100m Data Product from the Surface Water Ocean Topography (SWOT) mission provides global surface water elevation and inundation extent derived from high rate (HR) measurements from the Ka-band Radar Interferometer (KaRIn) on SWOT. SWOT launched on December 16, 2022 from Vandenberg Air Force Base in California into a 1-day repeat orbit for the "calibration" or "fast-sampling" phase of the mission, which completed in early July 2023. After the calibration phase, SWOT entered a 21-day repeat orbit in August 2023 to start the "science" phase of the mission, which is expected to continue through 2025.\r Water surface elevation, area, water fraction, backscatter, geophysical information are provided in geographically fixed scenes at 100 meter horizontal resolution in Universal Transverse Mercator (UTM) projection. Available in netCDF-4 file format. On-demand processing available to users for different resolutions, sampling grids, scene sizes, and file formats.Please note that this collection contains SWOT Version C science data products.This collection is a sub-collection of its parent: https://podaac.jpl.nasa.gov/dataset/SWOT_L2_HR_Raster_2.0

The MOD44W Version 6 data product was decommissioned on July 18, 2024. Users are encouraged to use the MOD44W Version 6.1 data product (https://doi.org/10.5067/MODIS/MOD44W.061).The Terra Moderate Resolution Imaging Spectroradiometer (MODIS) Land Water Mask (MOD44W) Version 6 data product provides a global map of surface water at 250 meter (m) spatial resolution. The data are available annually from 2000 to 2015. MOD44W Version 6 is derived using a decision tree classifier trained with MODIS data and validated with the Version 5 MOD44W data product. A series of masks are applied to address known issues caused by terrain shadow, burn scars, cloudiness, or ice cover in oceans. A primary improvement in Version 6 is the generation of time series data rather than a simple static representation of water, given that water bodies fluctuate in size and location over time due to both natural and anthropogenic causes. Provided in each MOD44W Version 6 Hierarchical Data Format 4 (HDF4) file are layers for land, water, no data, and an associated per pixel quality assurance (QA) layer that provides users with information on the determination of water. Known Issues The mask area for pixels outside of the Sinusoidal Grid for tile h27v14 should be labeled as 250 (Outside of Projection) instead of 0 (Land) and the QA values should be 10 (Fill) instead of 3 (Lower confidence land) for years 2000 through 2015. Additional known issues are described in the User Guide and ATBD. For complete information about known issues please refer to the MODIS/VIIRS Land Quality Assessment website.Improvements/Changes from Previous Versions Corrected for errors of many misplaced rivers in South America due to limited spatial resolution of previously used input data. Improved data in the far northern latitudes. Water extent provided as a time series that can be used to detect changes between 2000 and 2015. Increased spatial resolution of 250 m. The Version 5 product was 500 m. Improved terrain shadows with slope and elevation masking.* Burned areas (scars) delineated.

Seg Watermark

## Overview

Seg Watermark is a dataset for instance segmentation tasks - it contains Watermark annotations for 5,000 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

The MODIS Land Water Mask is created from MODIS 250 m imagery incombination with Shuttle Radar Topography Mission (SRTM) Water Body Data (SWBD) tocreate a global map of surface water at 250 m spatial resolution. Currently, only one mapexists, created in 2009 by Carroll et al. (2009). Because only one MODIS-based map exists,an analysis of surface water change is not possible at this time.

The MOD44W V6 land/water mask 250m product is derived using a decision tree classifier trained with MODIS data and validated with the MOD44W V5 product. A series of masks are applied to address known issues caused by terrain shadow, burn scars, cloudiness, or ice cover in oceans.

This is the updated Abstact. Watermasks were created for Chazuta, Peru using SAR imagery. VV+VH polarizations were used to get the layer and other source code provided by the Alaska Satellite Facility (ASF) group. Agriculture data was retrieved from croplands.org and other data was found at openstreetmaps.org

Rasterized water surface elevation and inundation extent in geographically fixed tiles at resolutions of 100 m and 250 m in a Universal Transverse Mercator projection grid. Provides rasters with water surface elevation, area, water fraction, backscatter, geophysical information. On-demand processing available to users for different resolutions, sampling grids, scene sizes, and file formats. Gridded scene (approx 128x128 km2, georeferenced); full swath. Available in netCDF-4 file format.

The SWOT Level 2 KaRIn High Rate Raster Product (SWOT_L2_HR_Raster_D) provides rasterized estimates of water surface elevation, inundation extent, and radar backscatter derived from high-resolution radar observations by the Ka-band Radar Interferometer (KaRIn) on the SWOT satellite. This product aggregates the irregularly spaced pixel cloud data from the PIXC and PIXCVec products onto a uniform geographic grid to facilitate spatial analysis of water surface features across inland, estuarine, and coastal domains.Standard granules cover non-overlapping 128 × 128 km² scenes in the UTM projection at 100 m and 250 m resolution, stored in NetCDF-4 format. Each file contains 2D image layers representing water surface elevation (corrected for geoid, solid Earth, load, and pole tides, as well as atmospheric and ionospheric path delays), surface area, water fraction, and sigma0, along with quality flags and uncertainty estimates. On-demand versions are available at user-specified resolutions and projections, with optional overlapping granules and GeoTIFF output via SWODLR: https://swodlr.podaac.earthdatacloud.nasa.gov/The raster product offers a gridded alternative to the unstructured pixel cloud, supporting hydrologic and geomorphic analyses in complex flow environments such as braided rivers, floodplains, wetlands, and coastal zones. It enables consistent spatiotemporal sampling while reducing noise through spatial aggregation, making it especially suitable for applications that require map-like continuity or integration with geospatial models.This collection is a sub-collection of its parent: https://podaac.jpl.nasa.gov/dataset/SWOT_L2_HR_Raster_D

Auxiliary information for pixel cloud product indicating to which water bodies the pixels are assigned in river and lake products. Also includes height-constrained pixel geolocation after reach- or lake-scale averaging. Point cloud over tile (approx 64x64 km2); half swath (left or right side of full swath). Available in netCDF-4 file format.

This personal geodatabase contains land and water masks (as rasters and polygons) for the remotely sensed data. It also contains a polygon feature class named: Spatial_Extent_Remote_Sensing_Data, which denotes the outer boundaries of all of the remote sensing data. All of these masks were derived directly from the remotely sensed imagery using geoprocessing functionality in ArcGIS 9.1.

Watermark-or-Not-20K Dataset

  Overview

The Watermark-or-Not-20K dataset consists of 20,000 images annotated with binary labels indicating the presence or absence of a watermark. It is designed to support training and evaluation of models focused on watermark detection, which is useful for content filtering, copyright protection, and image moderation tasks.

  Dataset Structure

Split: train Number of samples: 20,000 Label Type: Categorical (2 classes) Image… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/Watermark-or-Not-20K.

The SWOT Level 2 Water Mask Pixel Cloud Product (SWOT_L2_HR_PIXC_D) provides high-resolution, geolocated observations of terrestrial surface water pixels detected by the Ka-band Radar Interferometer (KaRIn) onboard the Surface Water and Ocean Topography (SWOT) satellite. This product contains the foundational “pixel cloud” data from which hydrologic features such as rivers, lakes, floodplains, and wetlands are later identified and analyzed in higher-level products.The PIXC product provides the unstructured point cloud of water detections at full instrument resolution (~15–25 m cross-track, ~5–10 m along-track), supporting fine-scale hydrologic analyses and applications such as water body delineation, flood mapping, and change detection. Each granule represents one KaRIn high-rate tile (~64 × 64 km²), corresponding to either the left or right half of the instrument swath. It includes individual water pixel detections with associated geodetic coordinates, ellipsoidal heights, normalized radar backscatter, and classification flags indicating surface type (e.g., land, water, ice, or layover). Pixel heights are referenced to the WGS84 ellipsoid and are not corrected for solid Earth tides, ocean tides, or pole tides like KaRIn low-rate oceanography products. Users may apply additional processing or corrections depending on the intended scientific use. Data are distributed in NetCDF-4 format and follow Climate and Forecast (CF) metadata conventions.The PIXC product is complemented by an auxiliary pixel cloud product (PIXCVec) that provides a less noisy, height-constrained geolocation of the observed water pixels and identifies associated river and lake features in the SWOT Prior River & Lake Databases (PRD & PLD). Together they support the derivation of higher-level, featured based River and Lake hydrology products from SWOT.

Land Water Mask Derived from MODIS and SRTM L3 Global 250m SIN Grid MOD44W The new MODIS 250 m land-water mask (Short Name: MOD44W) is an improvement over the existing MODIS Nadir Bidirectional Reflectance Distribution Function (BRDF)-Adjusted Reflectance (NBAR) and MODIS land cover-based global land-water mask (Salomon et al., 2004). The MODIS NBAR and land cover based mask was itself an improved version of the EOS DEM-based land-water mask (Logan et al., 1999). The new MODIS 250 m mask is primarily created with three different data inputs: The Shuttle Radar Topography Mission's (SRTM) Water Body Dataset (SWBD) (Areas between 60? S to 60 N); The MOD44C, a non-public, 250 m global 16-day composite collection based on 8+ years of Terra MODIS data, and 6+ years of Aqua MODIS data. This data set originally provided the input to produce the Vegetative Cover Conversion, and Vegetative Continuous Fields products (Areas between 60 N to 90 N); and the MODIS based Mosaic of Antarctica (MOA), which is a 250 m MODIS level-1b mosaic for Antarctica (Areas within Antarctica between 60 S and 90 S). Other appropriate and publicly available data sets were also used to supplement the production of the MODIS 250 m land water mask. Additional details regarding the methodology are available in the User Guide. This marks the first time that such a global MODIS-SRTM land-water mask is offered publicly to end-users. This data set is provided in the same gridded tile structure that is common to several higher-level MODIS land products.

The Terra Moderate Resolution Imaging Spectroradiometer (MODIS) Land Water Mask (MOD44W) Version 6.1 data product provides a global map of surface water at 250 meter (m) spatial resolution. The data are available annually from 2000 to present. MOD44W Version 6.1 is derived using a decision tree classifier trained with MODIS data and validated with the Version 5 MOD44W data product. A series of masks are applied to address known issues caused by terrain shadow, burn scars, cloudiness, or ice cover in oceans. Version 6.1 is the generation of a time series rather than a simple static representation of water, given that water bodies fluctuate in size and location over time due to both natural and anthropogenic causes. Provided in each MOD44W Version 6.1 Hierarchical Data Format 4 (HDF4) file are layers for land water mask and water body classification. A quality assurance (QA) layer provides users with information on the determination of water. The new seven class water classification layer provides values for shallow ocean, land, shoreline, inland water, ephemeral water, deep inland water, moderate ocean, deep ocean, and a classification deemed to fall outside of the projection.Known Issues The seven_class SDS layer legend for classes 4 and 5 are incorrectly listed as Deep Inland Water and Ephemeral Water in the data for the entire mission period. Users are requested to follow the legend provided in the User Guide. For complete information about known issues please refer to the MODIS/VIIRS Land Quality Assessment website.Improvements/Changes from Previous Versions Additional data files will be produced annually and are scheduled for distribution in the first quarter of the calendar year. * Seven-class water body classification layer produced at 250 meter resolution is provided in tiled format. For Collection 6.1 and beyond, Antarctica data are now being produced.

Point cloud of water mask pixels (“pixel cloud”) with geolocated heights, backscatter, geophysical fields, and flags. Point cloud over tile (approx 64x64 km2); half swath (left or right side of full swath). Available in netCDF-4 file format.Please note that this collection contains SWOT Version C science data products.

This is a multi-date water mask image. It represents the number of images in the time series classified as water using the Landsat based water index. This index was developed using Canonical Variates Analysis (CVA) of visually identified water and non-water signatures in radiometrically calibrated Queensland wide Landsat imagery. The index is a linear combination of bands, Log transformations of bands and interactive band terms.; ; Unsigned 8 bit; Band 1 = number of dates classified as water; ; Reference:; Danaher, TJ & Collett, L (2006). Development, optimisation, and multi-temporal application of a simple Landsat based Water Index. In: Proceedings of the 13th Australasian Remote Sensing and Photogrammetry Conference, Canberra, Australia.

Watermark Detect

## Overview

Watermark Detect is a dataset for classification tasks - it contains Images annotations for 6,336 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).