Global Surface Water analysis workflow: global water-occurrence data for GIS and Earth observation

Global Surface Water is a global Earth-observation dataset from the European Commission Joint Research Centre that describes where surface water has appeared and how its presence has changed over time. The current download release provides 1984–2024 products, while the widely used JRC/GSW1_4 Earth Engine asset covers 1984–2021. At 30-metre pixel size, its layers support water-occurrence mapping, change analysis, seasonality studies, regional statistics, floodplain screening, wetland research, and combinations with roads, buildings, infrastructure, administrative boundaries, and other geographic data. The dataset is not a live flood-monitoring service, a hydrologic model, a legal wetland inventory, or a substitute for locally authoritative water data.

Get Started Browse datasets

The Global Surface Water analysis workflow turns a long satellite-derived record into geographic evidence about surface-water presence, persistence, seasonality, and change. It is especially useful when the question is not simply “Where is water now?” but “Where has water appeared repeatedly, how stable is it, and how has its distribution changed across a place?”

How EDK fits

For: GIS analysts, Earth-observation researchers, environmental consultants, hydrologists, conservation teams, planners, cartographers, and developers building location-based analysis workflows.

Global Surface Water is a focused environmental layer for understanding one important class of physical-world feature: visible surface water. Its value within a broader world-model perspective comes from making water behavior computable alongside the places and infrastructure people care about. A water-occurrence raster can be summarized within neighborhoods, watersheds, parcels, road buffers, building footprints, protected areas, utility corridors, or administrative regions. That allows an analyst to move from a satellite-derived observation to a geographic question such as which roads intersect persistent water, which buildings are near areas with recurring water detection, or how water patterns differ across regions. E2’s relevance here is editorial and conceptual: Global Surface Water supplies an environmental measurement layer, while a world model provides the surrounding geographic context and computation needed to find, compare, filter, aggregate, and relate that layer to other physical features. This page does not imply that E2 publishes, hosts, updates, or reproduces Global Surface Water.

•Use Global Surface Water as an environmental evidence layer when ranking or comparing physical locations by historical water occurrence, recurrence, seasonality, or detected change.
•Combine water-history rasters with roads, buildings, drainage features, public assets, land use, or administrative areas to compute proximity, intersection, coverage, and regional summaries around real-world places.
•Use the dataset as a temporal context layer for investigating whether a physical feature lies near persistent, seasonal, newly appearing, or historically lost surface water; treat the result as screening evidence rather than a regulatory determination.

Producer

European Commission Joint Research Centre, with the Earth Engine catalog identifying EC JRC / Google.

Geography

Global; Earth Engine metadata reports approximately 90°S to 90°N and 180°W to 180°E.

Spatial resolution

30 metres for the documented mapping layers.

Historical coverage

GSW1_4: 1984–2021; current JRC download release: 1984–2024.

Primary data type

Satellite-derived raster layers and temporal water-history products.

Current update note

The JRC download page was marked updated August 26, 2026; no guaranteed recurring release schedule is stated there.

Dataset profile

license

Free of charge without restriction of use under the Copernicus Programme, with required acknowledgement, citation, and map attribution.

geography

Global surface-water mapping domain, approximately 90°S–90°N and 180°W–180°E in the Earth Engine catalog.

resolution

30-metre raster pixels for the documented mapping layers.

update cadence

No guaranteed recurring cadence is stated on the current access page. The current page identifies a 1984–2024 release and was marked updated August 26, 2026.

temporal coverage

GSW1_4 covers 1984–2021. The current JRC download release extends selected products through 2024; layer-specific temporal coverage varies in V1.5.

Formats

TIFF raster files delivered as 10° by 10° tiles.

Full-dataset download through an official Python 3 script.

Google Earth Engine image and image-collection assets.

Web map services and tiled web imagery for visualization.

Supporting QGIS symbology files and ISO 19139 metadata files for selected downloadable layers.

Access methods

Global Surface Water Explorer for visual inspection and historical map navigation.

Official JRC data-download page for tiled TIFF files.

Official Python 3 script for full V1.4 or V1.5 downloads.

Google Earth Engine catalog and project assets for cloud-based raster computation.

WMS or WMTS-style services for cartographic display in QGIS, ArcGIS, and web maps.

Quality and limits

The dataset is a surface-water detection product, not a live monitoring feed, hydrologic model, water-depth product, groundwater dataset, legal wetland inventory, or definitive floodplain boundary.

Thirty-metre pixels can omit or mix narrow channels, small ponds, shaded water, vegetation-covered water, and urban drainage features.

The 2024 extension combines Landsat Collection 1 and Collection 2 products. The JRC warns of residual co-registration offsets that can be spatially variable and may reach or exceed one 30-metre pixel in some areas.

Seasonality, monthly history, and yearly history in the 2024 release may require merging separate historical and recent assets to obtain a complete timeline.

The JRC identifies an unresolved issue in the monthly-recurrence layer that can lower long-term monthly values where observations are sparse.

The downloadable files do not contain all metadata internally; obtain and retain the associated metadata and user-guide materials.

The Global Surface Water Explorer and web map services are for visualization and are not suitable substitutes for analytical raster access.

Analysis workflows

Select the product version and time range before analysis; record the asset identifier, release, download date, and source documentation.

Choose occurrence, recurrence, seasonality, change, transitions, maximum extent, monthly history, or yearly history according to the question being studied.

Clip or mask the raster to a study area while preserving the source grid where practical.

Calculate zonal statistics for administrative areas, watersheds, protected areas, parcels, neighborhoods, or other geographic units.

Intersect or buffer roads, buildings, utilities, public assets, or environmental features and summarize water occurrence or maximum extent around them.

Compare regions using consistent thresholds, projections, masks, observation periods, and missing-data rules.

Inspect Collection 1 and Collection 2 transition areas when interpreting fine-scale change in the 2024 release.

Validate consequential results against current imagery, field observations, local hydrography, flood products, or other authoritative sources.

Use the required JRC/Copernicus acknowledgement and cite the Pekel et al. publication in derived outputs.

Value inside EDK

What the Global Surface Water analysis workflow contains

Global Surface Water is derived from Landsat observations and classifies individual pixels as water or non-water before aggregating those detections into spatial and temporal summary layers. The v1.4 Earth Engine documentation describes a 1984–2021 record generated from more than 4.7 million Landsat 5, 7, and 8 scenes. The newer access page adds products through 2024 and identifies the extended release as V1.5. (global-surface-water.appspot.com)

The principal summary layers are designed for different questions. Occurrence expresses how often water was detected during the observation period. Seasonality records the number of months in which water appears. Recurrence describes how regularly water returns from year to year. Change layers compare historical periods, transitions classify changes between early and late conditions, and maximum water extent records locations where water was detected at least once. The downloadable release also includes monthly and yearly history assets, although the 2024 documentation notes that some historical products must be combined from separate periods. (global-surface-water.appspot.com)

•Global geographic scope, represented as raster imagery and derived summary layers.
•Nominal 30-metre pixel size for the documented mapping layers.
•Long-term water-occurrence and change information rather than a real-time observation feed.
•Layer families for occurrence, change, seasonality, recurrence, transitions, maximum extent, monthly history, yearly history, and observation metadata.

Geography, temporal coverage, and update cadence

The dataset is global, spanning approximately 90° south to 90° north and 180° west to 180° east in the Earth Engine catalog. The established JRC/GSW1_4 asset covers observations from March 16, 1984 through December 31, 2021. The JRC download page now identifies a newer dataset extending the principal record through 2024, with the access page marked as updated August 26, 2026. (global-surface-water.appspot.com)

There is an important version distinction for reproducible work. GSW1_4 remains the documented 1984–2021 Earth Engine asset, while the newer V1.5 release adds 2022–2024 products and also reprocesses selected earlier layers. The JRC states that long-term summary products such as occurrence, recurrence, and transitions synthesize 1984–2024, but seasonality, monthly history, and yearly history may be split into separate historical and recent assets that need to be merged. (global-surface-water.appspot.com)

•Geography: global land and surface-water mapping domain.
•GSW1_4 temporal range: 1984–2021.
•Current download release identified by JRC: 1984–2024.
•Update cadence: no guaranteed schedule stated on the current access page; verify the release and asset version at the time of use.

Formats and access methods

For desktop GIS and scripted raster processing, the JRC download service provides individual 10° by 10° files as TIFF files. It also provides a Python 3 script for downloading the full V1.4 or V1.5 collections. Analysts should retain the accompanying product documentation and metadata because the JRC notes that downloadable files do not contain all metadata internally. (global-surface-water.appspot.com)

Google Earth Engine provides the widely used JRC/GSW1_4 assets, including the main mapping layers, yearly history, monthly history, monthly recurrence, and metadata. The newer V1.5 assets are published under project-specific Earth Engine asset paths, and the JRC supplies examples for merging historical and recent image collections. (global-surface-water.appspot.com)

•TIFF raster files in 10° by 10° download tiles.
•Python 3 download script for full V1.4 or V1.5 collections.
•Google Earth Engine image and image-collection assets.
•Global Surface Water Explorer for visual inspection.
•WMS or WMTS-style services for cartographic display, not analytical raster access.
•QGIS symbology files for several downloadable layers.

A defensible GIS analysis workflow

Start by defining the geographic question and selecting the layer that measures it. Use occurrence or recurrence for persistence, seasonality for within-year behavior, transitions or change for historical comparison, and maximum extent for a broad “detected at least once” mask. Do not use the visually convenient web map service when the analysis depends on numeric pixel values.

Next, choose a consistent product vintage. For a study that must reproduce published or established Earth Engine work, document the specific GSW1_4 asset. For work requiring the latest available period, use the V1.5 1984–2024 release and explicitly record whether a layer is a complete 1984–2024 synthesis or a 2022–2024 component that must be merged with earlier assets. The JRC identifies residual co-registration offsets between Landsat Collection 1 and Collection 2 and says that complete backprocessing is planned, so change analysis near boundaries or narrow features deserves extra inspection. (global-surface-water.appspot.com)

•Record product version, asset ID, temporal range, download date, and processing steps.
•Use numeric raster downloads or Earth Engine assets for analysis.
•Keep Collection 1 and Collection 2 transition effects in mind when interpreting change.
•Use consistent projections, thresholds, masks, and zonal-statistics rules across regions.
•Validate important conclusions with current or locally authoritative evidence.

License, strengths, and tradeoffs

The JRC states that Global Surface Water data are produced under the Copernicus Programme and provided free of charge without restriction of use, subject to the applicable Copernicus rules. Publications, models, and data products must include proper acknowledgement and citation. Published maps should include the attribution “Source: EC JRC/Google.” (global-surface-water.appspot.com)

The dataset’s strongest characteristic is its consistent global time series. It offers a common raster framework for comparing surface-water behavior across countries and regions, which is valuable when local datasets use incompatible classifications or inconsistent historical coverage. Its multiple derived layers also reduce the need to calculate every water-history statistic from raw imagery.

•Strength: global coverage with a multi-decade historical record.
•Strength: multiple ready-made layers for occurrence, recurrence, seasonality, change, and transitions.
•Strength: open access with downloadable rasters, Earth Engine assets, and visualization services.
•Tradeoff: 30-metre pixels can miss or mix narrow, small, shaded, or partly obscured water features.
•Tradeoff: the product does not provide real-time conditions, water depth, discharge, or legal wetland determinations.
•Tradeoff: V1.5 combines processing generations and requires care when merging or comparing historical and recent assets.
•Tradeoff: the JRC documents an unresolved monthly-recurrence issue in the newer release.

Questions teams ask

Is Global Surface Water a live flood-monitoring dataset?

No. It is a historical satellite-derived surface-water dataset. Its occurrence, recurrence, seasonality, and change layers describe water detections over defined observation periods. Use a dedicated near-real-time Earth-observation or flood-monitoring source for current inundation conditions.

What is the difference between Global Surface Water v1.4 and v1.5?

The established JRC/GSW1_4 Earth Engine asset covers 1984–2021. The current JRC download page identifies V1.5 products extending through 2024, including reprocessed and newly added assets. The newer release combines Landsat Collection 1 and Collection 2 components, so analysts should document the version and avoid treating the products as identical. (global-surface-water.appspot.com)

Can I use Global Surface Water in QGIS or ArcGIS?

Yes. The JRC provides downloadable TIFF tiles and supporting symbology files, and it documents web map service connections for desktop GIS. Use downloaded rasters when you need pixel values or statistics; use WMS or WMTS-style services primarily for visualization. (global-surface-water.appspot.com)

What does the occurrence layer measure?

Occurrence measures how frequently water was detected at a pixel during the relevant observation period. It is not water depth, average water level, flood probability, or a guarantee that the pixel is wet on a particular date. Interpret it together with the temporal range, observation metadata, and the needs of the analysis.

Is Global Surface Water suitable for legal wetland or floodplain decisions?

It can be useful as screening or historical context, but it should not be treated as a legal wetland inventory, definitive floodplain boundary, or locally authoritative regulatory dataset. Important decisions should be checked against the applicable agency data, current imagery, survey information, and governing standards.

Keep exploring