Skip to content

Open-ET/openet-ssebop

Repository files navigation

OpenET - SSEBop

Latest version on PyPI Build status

WARNING: This code is in development, is being provided without support, and is subject to change at any time without notification

This repository provides Google Earth Engine Python API based implementation of the SSEBop ET model.

The Operational Simplified Surface Energy Balance (SSEBop) model computes daily total actual evapotranspiration (ETa) using land surface temperature (Ts), maximum air temperature (Ta) and reference ET (ETr or ETo). The SSEBop model does not solve all the energy balance terms explicitly; rather, it defines the limiting conditions based on "gray-sky" net radiation balance principles and an air temperature parameter. This approach predefines unique sets of "hot/dry" and "cold/wet" limiting values for each pixel, allowing an operational model setup and a relatively shorter compute time. More information on the GEE implementation of SSEBop is published in Senay2022 and Senay2023 with additional details and model assessment.

Basic SSEBop model implementation in Earth Engine:

docs/SSEBop_GEE_diagram.jpg

Model Design

The primary component of the SSEBop model is the Image() class. The Image class can be used to compute a single fraction of reference ET (ETf) image from a single input image. The Image class should generally be instantiated from an Earth Engine Landsat image using the collection specific methods listed below. ET image collections can be built by computing ET in a function that is mapped over a collection of input images. Please see the Example Notebooks for more details.

Input Collections

SSEBop ET can currently be computed for Landsat Collection 2 Level 2 (SR/ST) images from the following Earth Engine image collections:

  • LANDSAT/LC09/C02/T1_L2
  • LANDSAT/LC08/C02/T1_L2
  • LANDSAT/LE07/C02/T1_L2
  • LANDSAT/LT05/C02/T1_L2

Note: Users are encouraged to prioritize use of Collection 2 data where available. Collection 1 was produced by USGS until 2022-01-01, and maintained by Earth Engine until 2023-01-01. [More Information]

Landsat Collection 2 SR/ST Input Image

To instantiate the class for a Landsat Collection 2 SR/ST image, use the Image.from_landsat_c2_sr method.

The input Landsat image must have the following bands and properties:

SPACECRAFT_ID Band Names
LANDSAT_5 SR_B1, SR_B2, SR_B3, SR_B4, SR_B5, SR_B7, ST_B6, QA_PIXEL
LANDSAT_7 SR_B1, SR_B2, SR_B3, SR_B4, SR_B5, SR_B7, ST_B6, QA_PIXEL
LANDSAT_8 SR_B1, SR_B2, SR_B3, SR_B4, SR_B5, SR_B6, SR_B7, ST_B10, QA_PIXEL
LANDSAT_9 SR_B1, SR_B2, SR_B3, SR_B4, SR_B5, SR_B6, SR_B7, ST_B10, QA_PIXEL

Model Output

The primary output of the SSEBop model is the fraction of reference ET (ETf). The actual ET (ETa) can then be computed by multiplying the Landsat-based ETf image with the reference ET (e.g. ETr from GRIDMET).

Example SSEBop ETa from Landsat:

docs/ET_example.PNG

Example

import openet.ssebop as ssebop

landsat_img = ee.Image('LANDSAT/LC08/C02/T1_L2/LC08_044033_20170716')
et_fraction = ssebop.Image.from_landsat_c2_sr(landsat_img).et_fraction
et_reference = ee.Image('IDAHO_EPSCOR/GRIDMET/20170716').select('etr')
et_actual = et_fraction.multiply(et_reference)

Custom Input Image

SSEBop images can also be built manually by instantiating the class with an ee.Image with the following bands: 'lst' (land surface temperature [K]) and 'ndvi' (normalized difference vegetation index). The input image must have 'system:index' and 'system:time_start' properties (described above).

import openet.ssebop as ssebop

input_img = (
    ee.Image([ee.Image(lst), ee.Image(ndvi)])
    .rename(['lst', 'ndvi'])
    .set({
        'system:index': 'LC08_044033_20170716',
        'system:time_start': ee.Date.fromYMD(2017, 7, 16).millis()})
)
et_fraction = ssebop.Image(input_img).et_fraction

Example Notebooks

Detailed Jupyter Notebooks of the various approaches for calling the OpenET SSEBop model are provided in the "examples" folder.

Ancillary Datasets

Maximum Daily Air Temperature (Tmax)

The daily maximum air temperature (Tmax) is essential for establishing the maximum ET limit (cold boundary) as explained in Senay2017. Support for source options includes CIMIS, GRIDMET, DAYMET, and other custom Image Collections. See the model Image class docstrings for more information.

Default Asset ID: projects/usgs-ssebop/tmax/daymet_v4_mean_1981_2010 (Daily median from 1981-2010)

Land Surface Temperature (LST)

Land Surface Temperature is currently calculated in the SSEBop approach two ways:

Temperature Difference (dT)

The SSEBop ET model uses dT as a predefined temperature difference between Thot and Tcold for each pixel. In SSEBop formulation, hot and cold limits are defined on the same pixel; therefore, dT actually represents the vertical temperature difference between the surface temperature of a theoretical bare/dry condition of a given pixel and the air temperature at the canopy level of the same pixel as explained in Senay2018. The input dT is calculated under "gray-sky" conditions and assumed not to change from year to year, but is unique for each day and location.

Default Asset ID: projects/usgs-ssebop/dt/daymet_median_v6

Temperature Correction (c factor)

In order to correspond the maximum air temperature with cold/wet limiting environmental conditions, the SSEBop model uses a temperature correction coefficient (c factor, sometimes labeled interchangeably as Tcorr) uniquely calculated for each Landsat scene. This temperature correction component is uniquely developed for SSEBop using a Forcing and Normalizing Operation (FANO) method featuring a linear relation between a normalized land surface temperature difference and NDVI difference using the dT parameter and a proportionality constant.

Note: Tcorr refers to the pixel-based ratio of LST_cold and Tmax while c factor is a statistical value that represents a region such as a 5-km grid size (or larger) value.

More information on parameter design and model improvements using the FANO method can be found in Senay2023. Additional SSEBop model algorithm theoretical basis documentation can be found here.

The 'FANO' parameter (default) can be implemented dynamically for each Landsat scene within the SSEBop Image object using the following Tcorr source:

model_obj = model.Image.from_landsat_c2_sr(
    ee.Image('LANDSAT/LC08/C02/T1_L2/LC08_044033_20170716'),
    tcorr_source='FANO',
    )

The FANO parameterization allows the establishment of the cold boundary condition regardless of vegetation cover density, improving the performance and operational implementation of the SSEBop ET model in sparsely vegetated landscapes, dynamic growing seasons, and varying locations around the world.

Installation

The OpenET SSEBop python module can be installed via pip:

pip install openet-ssebop

Dependencies

OpenET Namespace Package

Each OpenET model is stored in the "openet" folder (namespace). The model can then be imported as a "dot" submodule of the main openet module.

import openet.ssebop as ssebop

Development and Testing

Please see the CONTRIBUTING.rst.

References

[Senay2013]
Senay, G., Bohms, S., Singh, R., Gowda, P., Velpuri, N., Alemu, H., Verdin, J. (2013). Operational Evapotranspiration Mapping Using Remote Sensing and Weather Datasets: A New Parameterization for the SSEB Approach. Journal of the American Water Resources Association, 49(3).
[Senay2016]
Senay, G., Friedrichs, M., Singh, R., Velpui, N. (2016). Evaluating Landsat 8 evapotranspiration for water use mapping in the Colorado River Basin. Remote Sensing of Environment, 185.
[Senay2017]
Senay, G., Schauer, M., Friedrichs, M., Manohar, V., Singh, R. (2017). Satellite-based water use dynamics using historical Landsat data (1984-2014) in the southwestern United States. Remote Sensing of Environment, 202.
[Senay2018]
Senay, G. (2018). Satellite Psychrometric Formulation of the Operational Simplified Surface Energy Balance (SSEBop) Model for Quantifying and Mapping Evapotranspiration. Applied Engineering in Agriculture, 34(3).
[Senay2019]
Senay, G., Schauer, M., Velpuri, N.M., Singh, R.K., Kagone, S., Friedrichs, M., Litvak, M.E., Douglas-Mankin, K.R. (2019). Long-Term (1986–2015) Crop Water Use Characterization over the Upper Rio Grande Basin of United States and Mexico Using Landsat-Based Evapotranspiration. Remote Sensing, 11(13):1587.
[Schauer2019]
Schauer, M., Senay, G. (2019). Characterizing Crop Water Use Dynamics in the Central Valley of California Using Landsat-Derived Evapotranspiration. Remote Sensing, 11(15):1782.
[Senay2022]
Senay, G.B., Friedrichs, M., Morton, C., Parrish, G. E., Schauer, M., Khand, K., ... & Huntington, J. (2022). Mapping actual evapotranspiration using Landsat for the conterminous United States: Google Earth Engine implementation and assessment of the SSEBop model. Remote Sensing of Environment, 275, 113011
[Senay2023]
Senay, G.B., Parrish, G. E., Schauer, M., Friedrichs, M., Khand, K., Boiko, O., Kagone, S., Dittmeier, R., Arab, S., Ji, L. (2023). Improving the Operational Simplified Surface Energy Balance evapotranspiration model using the Forcing and Normalizing Operation. Remote Sensing, 15(1):260.