Pedotransfer functions by Schaap et al., 2001 and Zhang and Schaap, 2016.

Rosetta predicts van Genuchten soil water retention curve parameters utilizing a weighted recalibration of the Rosetta pedotransfer model with improved estimates of hydraulic parameter distributions and summary statistics (Rosetta3). Journal of Hydrology.

This version of Rosetta has been refactored by Roger Lew (2021) for use as a Python3 package under GNU GPL V2.

For more information regrading Rosetta see https://cals.arizona.edu/research/rosetta/

Easiest method of installation is to clone to your dist-packages folder.

e.g. homebrew Python install on MacOS

% cd /opt/homebrew/lib/python3.9/site-packages/rosetta/
% git clone https://github.com/rogerlew/rosetta
% cd ~

> from rosetta import Rosetta
> import numpy as np
> ros = Rosetta(model_no=3, debug=False)
> data = np.array([[35.0, 50.0, 15.0, 0.2]]) # sand, silt, clay, bulk density
> ros.predict(data)
{'theta_r': array([0.14238712]), 
 'theta_s': array([0.7368426]), 
 'alpha': array([0.00368141]), 
 'npar': array([1.44084392]), 
 'ks': array([1047.97575913])}

data should be a np.array with shape of nsamp, ninput. The ninput parameters should coorespond to the model.

The model_no parameter specifies the Rosetta model.

The expected units are:
SSC in weight %
BD in g/cm3
TH33 and T1500 as cm3/cm3

>>> ros.predict(data, calc_wilting_point=True, calc_field_capacity=True)
{'theta_r': array([0.14238712]), 
 'theta_s': array([0.7368426]), 
 'alpha': array([0.00368141]), 
 'npar': array([1.44084392]), 
 'ks': array([1047.97575913]), 
 'wp': array([0.2437112]), 
 'fc': array([0.34023513])}

The Rosetta class is intended to be backwards compatible with the original library. A more pythonic interface is provided by Rosetta3() (Rosetta(model_no=3) (also Rosetta2, Rosetta4 and Rosetta5). These provide a .predict_kwargs() method that accepts parameters as arguments and return scalars. By default these will also estimate wilting point and field capacity.

> from rosetta import Rosetta3
> ros = Rosetta3()
> ros.predict_kwargs(sand=35.0, silt=50.0, clay=15.0, bd=0.2)
{'theta_r': 0.14238712, 
 'theta_s': 0.7368426, 
 'alpha': 0.00368141, 
 'npar': 1.44084392, 
 'ks': 1047.97575913,
 'wp': 0.2437112, 
 'fc': 0.34023513}

%  python3 -m rosetta -h
usage: __main__.py [-h] [--raw] [--calc_wilting_point] [--calc_field_capacity] [-i INPUT] [-o OUTPUT] model_n

Rosetta 3 pedotransfer function CLI.

positional arguments:
  model_n

optional arguments:
  -h, --help            show this help message and exit
  --raw                 get raw data
  --calc_wilting_point  calculate wilting point
  --calc_field_capacity
                        calculate field capacity
  -i INPUT, --input INPUT
                        input from file
  -o OUTPUT, --output OUTPUT
                        store predicted data

% python3 -m rosetta 3 -i test_input_H3w.txt -o test_output_H3w.txt