Quicklinks: Data Preparation --- Project Setup --- Running METRIC
The following scripts should be ran in the following sequence in order to produce ET estimates. These scripts are all located in the "code/local/" directory.
Prepare Landsat path/row data and populates input files to be used later in the PyMETRIC process.
Prepares Landsat scenes for processing. This step inlcudes the creation of rasters which are a subset of those data downloaded during the Data Preperation step of pymetric.
#####Expected Raster Output:
- Common Area Raster
- Vapor Pressure
- Wind
- fmask
- Reference ETr
- Modified Cropland Data Layer
Runs METRIC Model 1 for all images.
#####Expected Raster Output:
- Cosine of Theta (cos_theta.img)
- Albedo
- Top of Atmosphere Reflectance
- Delapsed Temperature
- Ts Brightness ('ts_bt.img')
- NDVI (Normalized Difference Vegetation Index)
- NDWI (Normalized Difference Water Index)
- SAVI (Soil Adjusted Vegetation Index)
- LAI (Leaf Area Index)
Runs METRIC pixel rating function for all images, identifying potential calibration points.
#####Expected Raster Output:
- Region Mask
- Cold Pixel Rating
- Cold Pixel Suggestion
- Hot Pixel Rating
- Hot Pixel Suggestion
Runs METRIC pixel points function for all images, selecting initial calibration points for each Landsat image.
Runs METRIC Model 2 for all images.
Expected Raster Output:
- Fraction of Reference ET
Interpolates seasonal ET data from individual METRIC scenes
This workflow is setup to run with the example input file (C:\pymetric\example\landsat_2015.ini). Use this workflow as a starting point when using pyMETRIC for your data.
python C:\pymetric\code\local\landsat_prep_path_row.py -i C:\pymetric\example\landsat_2015.ini
python C:\pymetric\code\local\landsat_prep_ini.py -i C:\pymetric\example\landsat_2015.ini
python C:\pymetric\code\local\landsat_prep_scene.py -i C:\pymetric\example\landsat_2015.ini
python C:\pymetric\code\local\metric_model1.py -i C:\pymetric\example\landsat_2015.ini
python C:\pymetric\code\local\metric_pixel_rating.py -i C:\pymetric\example\landsat_2015.ini
python C:\pymetric\code\local\metric_pixel_points.py -i C:\pymetric\example\landsat_2015.ini
Prior to the running of METRIC model 2, calibration pixels must be adjusted manually with ArcGIS. At this point in the workflow, the software has automatically chose sample pixels, however the location of the calibration pixels must be changed for best results. The METRIC Manual should be consulted during the calibration process in order to provide the best possible estimates of ET. If pixels are left un-modified, ETrF rasters will still be produced, however the validity of the ETrF data will be significantly degraded.
python C:\pymetric\code\local\metric_model2.py -i C:\pymetric\example\landsat_2015.ini
python C:\pymetric\code\local\landsat_interpolate.py -i C:\pymetric\example\landsat_2015.ini --tables
python C:\pymetric\code\local\landsat_interpolate.py -i C:\pymetric\example\landsat_2015.ini --rasters
The following will run one iteration of the Monte Carlo tool with fixed tail sizes of 1% (cold) and 4% (hot). This value is the percent of agricultural pixels with ETrFs greater than the cold calibration ETrF value (for the cold calibration point).
python C:\pymetric\code\local\metric_monte_carlo.py -i C:\pymetric\example\landsat_2015.ini -mc 0 --tails 1 4
The following will run ten different iterations of the Monte Carlo tool with varying tail sizes (developed from the training data in 'misc/etrf_training_test.csv'). The "mc" parameter specifies which iterations to run.
python C:\pymetric\code\local\metric_monte_carlo.py -i C:\pymetric\example\landsat_2015.ini -mc 1-10