Marketplace

Sentinel-2 Reservoir and Lake Surface Area Monitoring

Generates water extent maps for lakes and reservoirs within the given bounding box and time range.

Description

Sentinel-2 Reservoir and Lake Surface Area Monitoring provides a more robust surface water extent mapping for lakes and reservoirs using a deep neural network. Surface extent maps are significantly less impacted by atmospheric disturbances compared to traditional index-based thresholding methods.

The block uses an existing dataset to identify water bodies within the AOI so that the users only pay for analyzing water bodies and avoid unnecessary computations in regions without water bodies. Please visit Terra Cover's free data block Reservoirs and Lakes Surface Area Timeseries Dataset to learn more about this dataset. The current version provides information for water bodies that are below 50 degree North and are of size between 0.1 and 100 sq. kms. Both size and geographical coverage will be increased soon. If water bodies of your interest are currently not available, please contact [email protected] to include them in the dataset on request.

Supported Workflows

  • Sentinel-2 L1C (SAFE) --> Sentinel-2 Reservoir and Lake Surface Area Monitoring

Technical Information

The surface extent mapping process uses a semantic segmentation based deep neural network that was trained using Sentinel-2 images taken from different parts of the world.

FeatureInfo
Data DependencyThe block must be used together with Sentinel-2 L1C (SAFE) data block.
Cloud Masks2cloudless package by Sentinel Hub was used without any modifications
Input ParametersThe processing block uses the same AOI used by the data block. The AOI is used to select water bodies within the Sentinel-2 granules returned by the data block.
Output FormatThe surface extent maps are provided as an 8-bit single band raster in GeoTIFF format. Pixel values: 0 (land), 1 (water), 2 (cloud).
Spatial Coverage180 W to 180 E, 50 N to 60 S
Usage NotesThis block works directly with the Sentinel-2 LIC (SAFE) data block. Copy the AOI with the geometry filter (bbox, intersects, contains) into the appropriate job parameter for this block. Only water bodies are processed that are fully covered by the AOI.

Performance

Most cloud detection algorithms focus on masking only clouds which leads to large errors of commission (false positives) in surface extent maps. The current version of the deep learning model is significantly more robust to cloud shadows.

Use Cases

Reservoirs play a crucial role for human sustenance as they provide freshwater for agriculture, power generation, human consumption, and recreation. The block enables regular monitoring of surface water availability at global scale. As an example, this case study quantifies the impact of recent droughts on surface water availability in Uruguay.

Pricing

The price of 10 credits / sqkm is charged on the output size of the lake areas (bounding box) and is not based on the input AOI.

More Information

For more information, please see the provider website.

Example Job Parameters

{
  "sobloo-s2-l1c-fullscene:1": {
    "ids": null,
    "time": "2018-01-01T00:00:00+00:00/2020-12-31T23:59:59+00:00",
    "limit": 2,
    "time_series": null,
    "max_cloud_cover": 100,
    "bbox": [
      -56.195148,
      -33.032783,
      -56.169135,
      -33.012469
    ]
  },
  "terracover-realsat:1": {
    "bbox": [
      -56.195148,
      -33.032783,
      -56.169135,
      -33.012469
    ],
    "contains": null,
    "intersects": null
  }
}

Capabilities

Input Capabilities

raster
up42_standard
bands
[
  "coastal",
  "blue",
  "green",
  "red",
  "rededge",
  "rededge2",
  "rededge3",
  "nir",
  "nir2",
  "watervapour",
  "swir",
  "swir2",
  "swir3"
]
dtype
uint16
format
SAFE
sensor
Sentinel2
resolution
10
processing_level
l1

Output Capabilities

raster
custom
bands
[
  "water_extent"
]
dtype
uint8
format
GTiff
resolution
10

End User License Agreement

View the End User License Agreement conditions in a new tab.