OpenET - Google Earth Engine ASCE Standardized Reference Evapotranspiration (ET)

This repository provides Google Earth Engine Python API based implementation of the ASCE Standardized Reference Evapotranspiration Equations (ASCE2005) for computing daily and hourly reference ET.

Usage

Daily

The following demonstrates how to compute a single daily ETr value using weather data for 2015-07-01 from the Fallon, NV AgriMet station. The necessary unit conversions are shown on the input values. The raw input data is available here.

import math
import ee
import openet.refetgee

# Unit conversions
tmin_c = (66.65 - 32) * (5.0 / 9)                          # F -> C
tmax_c = (102.80 - 32) * (5.0 / 9)                         # F -> C
tdew_c = (57.26 - 32) * (5.0 / 9)                          # F -> C
ea = 0.6108 * math.exp(17.27 * tdew_c / (tdew_c + 237.3))  # kPa
rs = (674.07 * 0.041868)                                   # Langleys -> MJ m-2 d-1
uz = 4.80 * 0.44704                                        # mpg -> m s-1
lat = 39.4575                                              # degrees

etr = openet.refetgee.Daily(
    tmin=tmin_c, tmax=tmax_c, ea=ea, rs=rs, uz=uz, zw=3, elev=1208.5,
    lat=lat, doy=182).etr.getInfo()

print('ETr: {:.2f} mm'.format(float(etr)))

Hourly

The following demonstrates how to compute a single hourly ETr value using weather data for 18:00 UTC (11:00 AM PDT) on 2015-07-01 from the Fallon, NV AgriMet station. The necessary unit conversions are shown on the input values. The raw input data is available here

import math
import ee
import openet.refetgee

# Unit conversions
tmean_c = (91.80 - 32) * (5.0 / 9)           # F -> C
ea = 1.20                                    # kPa
rs = (61.16 * 0.041868)                      # Langleys -> MJ m-2 h-1
uz = 3.33 * 0.44704                          # mph -> m s-1
lat = 39.4575                                # degrees
lon = -118.77388                             # degrees

etr = openet.refetgee.Hourly(
    tmean=tmean_c, ea=ea, rs=rs, uz=uz, zw=3, elev=1208.5,
    lat=lat, lon=lon, doy=182, time=18).etr.getInfo()

print('ETr: {:.2f} mm'.format(float(etr)))

GRIDMET

A helper function for computing daily ETo and ETr for GRIDMET images is available.

import ee
import openet.refetgee

source_img = ee.Image(ee.ImageCollection('IDAHO_EPSCOR/GRIDMET').first())
etr = openet.refetgee.Daily.gridmet(source_img).etr\
    .reduceRegion(reducer=ee.Reducer.first(),
                  geometry=ee.Geometry.Point(-118.77388, 39.4575),
                  scale=1000)\
    .getInfo()

print('ETr: {:.2f} mm'.format(float(etr['etr'])))

NLDAS

Helper functions for computing daily/hourly ETo/ETr for NLDAS images are available.

For the daily function, the NLDAS collection must be filtered to a single 24 hour period.

import ee
import openet.refetgee

source_coll = ee.ImageCollection('NASA/NLDAS/FORA0125_H002')\
    .filterDate('2015-07-01', '2015-07-02')
etr = openet.refetgee.Daily.nldas(source_coll).etr\
    .reduceRegion(reducer=ee.Reducer.first(),
                  geometry=ee.Geometry.Point(-118.77388, 39.4575),
                  scale=1000)\
    .getInfo()

print('ETr: {:.2f} mm'.format(float(etr['etr'])))

import ee
import openet.refetgee

source_img = ee.Image('NASA/NLDAS/FORA0125_H002/A20150701_2000')
etr = openet.refetgee.Hourly.nldas(source_img).etr\
    .reduceRegion(reducer=ee.Reducer.first(),
                  geometry=ee.Geometry.Point(-118.77388, 39.4575),
                  scale=1000)\
    .getInfo()

print('ETr: {:.2f} mm'.format(float(etr['etr'])))

CFSv2

A helper function for computing daily ETo and ETr for CFSv2 images is available.

For the daily function, the CFSv2 collection must be filtered to a single 24 hour period.

import ee
import openet.refetgee

source_coll = ee.ImageCollection('NOAA/CFSV2/FOR6H')\
    .filterDate('2015-07-01', '2015-07-02')
etr = openet.refetgee.Daily.cfsv2(source_coll).etr\
    .reduceRegion(reducer=ee.Reducer.first(),
                  geometry=ee.Geometry.Point(-118.77388, 39.4575),
                  scale=1000)\
    .getInfo()

print('ETr: {:.2f} mm'.format(float(etr['etr'])))

RTMA

Helper functions for computing daily/hourly ETo/ETr for RTMA images are available.

For the daily function, the RTMA collection must be filtered to a single 24 hour period.

import ee
import openet.refetgee

source_coll = ee.ImageCollection('NOAA/NWS/RTMA')\
    .filterDate('2015-07-01', '2015-07-02')
etr = openet.refetgee.Daily.rtma(source_coll).etr\
    .reduceRegion(reducer=ee.Reducer.first(),
                  geometry=ee.Geometry.Point(-118.77388, 39.4575),
                  scale=1000)\
    .getInfo()

print('ETr: {:.2f} mm'.format(float(etr['etr'])))

import ee
import openet.refetgee

source_img = ee.Image('NOAA/NWS/RTMA/2015070120')
etr = openet.refetgee.Hourly.nldas(source_img).etr\
    .reduceRegion(reducer=ee.Reducer.first(),
                  geometry=ee.Geometry.Point(-118.77388, 39.4575),
                  scale=1000)\
    .getInfo()

print('ETr: {:.2f} mm'.format(float(etr['etr'])))

ERA5-Land

Helper functions for computing daily/hourly ETo/ETr for ERA5-Land images are available.

For the daily function, the ERA5-Land collection must be filtered to a single 24 hour period.

import ee
import openet.refetgee

source_coll = ee.ImageCollection('ECMWF/ERA5_LAND/HOURLY')\
    .filterDate('2015-07-01', '2015-07-02')
etr = openet.refetgee.Daily.era5_land(source_coll).etr\
    .reduceRegion(reducer=ee.Reducer.first(),
                  geometry=ee.Geometry.Point(-118.77388, 39.4575),
                  scale=1000)\
    .getInfo()

print('ETr: {:.2f} mm'.format(float(etr['etr'])))

import ee
import openet.refetgee

source_img = ee.Image('ECMWF/ERA5_LAND/HOURLY/20150701T20')
etr = openet.refetgee.Hourly.era5_land(source_img).etr\
    .reduceRegion(reducer=ee.Reducer.first(),
                  geometry=ee.Geometry.Point(-118.77388, 39.4575),
                  scale=1000)\
    .getInfo()

print('ETr: {:.2f} mm'.format(float(etr['etr'])))

Input Parameters

Required Parameters (hourly & daily)

Variable	Type	Description [units]
ea	ee.Image	Actual vapor pressure [kPa]
rs	ee.Image	Incoming shortwave solar radiation [MJ m-2 day-1]
uz	ee.Image	Wind speed [m s-1]
zw	ee.Number	Wind speed height [m]
elev	ee.Image, ee.Number	Elevation [m]
lat	ee.Image, ee.Number	Latitude [degrees]
doy	ee.Image, ee.Number	Day of year

Required Daily Parameters

Variable	Type	Description [units]
tmin	ee.Image	Minimum daily temperature [C]
tmax	ee.Image	Maximum daily temperature [C]

Required Hourly Parameters

Variable	Type	Description [units]
tmean	ee.Image	Average hourly temperature [C]
lon	ee.Image, ee.Number	Longitude [degrees]
time	ee.Number	UTC hour at start of time period

Optional Parameters

Variable	Type	Description [units]
method	str	Calculation method 'asce' -- Calculations will follow ASCE-EWRI 2005 (default) 'refet' -- Calculations will follow RefET software
rso_type	str	Override default clear sky solar radiation (Rso) calculation Defaults to None if not set 'full' -- Full clear sky solar formulation (default) 'simple' -- Simplified clear sky solar formulation (Eq. 19) 'array' -- Read Rso values from "rso" function parameter
rso	ee.Image, ee.Number	Clear sky solar radiation [MJ m-2 day-1] Only needed if rso_type is 'array' Defaults to None if not set

Issues

Currently the user must handle all of the file I/O and unit conversions.

Cloudiness Fraction (hourly)

The cloudiness fraction (fcd) is computed as the ratio of the measured solar radiation (Rs) to the theoretical clear sky solar radiation (Rso). This ratio cannot be computed directly at night since Rso is 0. ASCE-EWRI 2005 suggests computing a representative nighttime fcd based on the fcd at sunset and/or sunrise.

In the RefET module fcd is hard coded to 1 for all time steps with very low sun angles since the hourly reference ET is computed independently for each time step.

Calculation Method - ASCE vs. RefET

The main difference between the two "methods" is that the "asce" method attempts to follow the equations in ASCE2005, whereas the "refet" method attempts to follow the calculations of the RefET Software as closely as possible. The difference in output between these methods is generally negligible (if not identical for realistic numbers of significant digits). Note that the default is set to "asce" to best match the calculations a user would expect to have happen. The "refet" method was added in order to help validate this code to the RefET Software.

Installation

The OpenET RefET GEE python module can be installed via pip:

pip install openet-refet-gee

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.refetgee as refetgee

Validation

Please see the validation document for additional details on the source of the test values and the comparison of the functions to the Ref-ET software.

Dependencies

• earthengine-api

Modules needed to run the test suite:

• pandas
• pytest
• pytz

References

[ASCE2005]

ASCE-EWRI (2005). The ASCE standardized reference evapotranspiration equation. https://ascelibrary.org/doi/book/10.1061/9780784408056