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.
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)))
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)))
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'])))
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'])))
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'])))
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'])))
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 |
Variable | Type | Description [units] |
---|---|---|
tmin | ee.Image | Minimum daily temperature [C] |
tmax | ee.Image | Maximum daily temperature [C] |
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 |
Variable | Type | Description [units] |
---|---|---|
method | str | Calculation method
|
rso_type | str | Override default clear sky solar radiation (Rso) calculation
Defaults to None if not set
|
rso | ee.Image, ee.Number | Clear sky solar radiation [MJ m-2 day-1]
|
Currently the user must handle all of the file I/O and unit conversions.
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.
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.
The OpenET RefET GEE python module can be installed via pip:
pip install openet-refet-gee
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
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.
Modules needed to run the test suite:
[ASCE2005] | ASCE-EWRI (2005). The ASCE standardized reference evapotranspiration equation.
|