Getting started#
Here we compare three different temperature functions with different cardinal temperature combinations with and without VPD stress function for the period from heading to maturity (grain filling period).
Creating application directory and installing dependencies#
Launch your terminal and navigate to your desired location to create the project directory. Run the following commands to create a directory for your project and initialize a grain yield model inside it:
Create a virtual environment with:
Note
Please follow the installation instructions to complete the steps above according to your local system.
Setup a folder#
Enter a folder or subfolder of your project root where you created the python environment and activate it.
If you didn't install thevenv yet, please review the detailed installation instructions
Install library#
Setting up temperature response functions is as simple as using the familiar pip install command. By executing the following line, you will have the library installed and ready to use:
Verify install#
If a version similar to tfunct version 1.0.0 is returned, you've installed the package successfully.
Success: Temperature Response Functions package is now installed.
View Sample quickstart for beginners
To get started with temperature response functions, we need to create a project folder and install a python virtual environment for our packages and further analysis as follows:
Google Colab
An easy way to learn and use temperature response functions package#
No install necessary, run the temperature response functions tutorials directly in the browser with Colaboratory, a Google research project created to help disseminate machine learning education and research. It's a Jupyter notebook environment that requires no setup to use and runs entirely in the cloud.
Note: this tutorial is also available as a single python notebook. You can download it on Github below:
Once tfunct library is successfully installed, you can import the essential modules for dataset creation and data pre-processing.
Load a dataset recommended for beginners#
Load and prepare the trial dataset.
Loading the data#
Example dataset can be loaded directly from the library, as follows:
import tfunct
from tfunct.data import load_dataset # Function to load existing dataset
# Load example dataset (Phenology, NDVI and Weather data for each site)
data = load_dataset()
# Display available datasets
print(data.keys()) # ['Pheno', 'NDVI', 'Weather']
# Display Phenology
data['Pheno'].head()
Creating a model#
The model contains all of the required functions to analyse the data
# Load module to create a model
from tfunct.model import Model
# ------------------------
# MODEL CONFIGURATION
# ------------------------
# Define the Path where the output data will be stored
PATH_PRJ = '/Users/ernestogiron/Desktop/TemperatureFunctions/'
RESULTS_PATH = PATH_PRJ + 'results/'
config = {
"PROJECT_PATH": PATH_PRJ,
"RESULTS_PATH": RESULTS_PATH, #'./', # Results will be put in the same folder where the model is running
}
# Parameters used by default
parameters = dict(
RUE = 3,
DRYMATTER = 0.8,
FACTOR_TON_HA = 0.01,
YIELD_FACTOR = 0.8 * 0.01,
TMIN_PERC_FACTOR = 0.25,
NDVI_lowerThreshold = 0.16,
Toptmin = 15,
Topt = 18,
Toptmax = 25,
Tmin = 9,
Tmax = 34,
Lvpd = 1,
Uvpd = 4,
SFvpd_Lthres = 0.2,
SFvpd_Uthres = 1,
)
# create model to estimate grain yield
# model = Model(config, parameters) # Use this if you change any parameter above
model = Model(config)
# Preprocess datasets
model.preprocess_raw_datasets(data)
Preparing locations#
Prepare dataset to run all process in parallel using NDVIA GPU if available
{'country': 'Nepal',
'location': 'Bhairahawa',
'loc_code': 'BHR',
'lat': 27.5,
'lon': 83.45,
'cycle': 2019,
'Days_To_Heading': 89,
'Days_To_Maturity': 122,
'ObsYield': 2.96685,
'Sowing_date': '2018-11-26',
'Heading_date': '2019-02-23',
'Maturity_date': '2019-03-28',
'UID': 1,
'ndays_tmn_lt9': 1,
'ndays_tmx_gt34': 0,
'avg_Tdaymax': 24.578,
'avg_NDVI': 0.447,
'avg_iPAR': 0.369}
Using one of the Temperature functions#
Calculating grain yield using Ritchie's Temperature-based function affecting Photosynthetic Reduction Factor (PRFT).
No stress conditions#
Stressed VPD#
For further information about model.getYield visit the API reference.
Displaying Grain Yield#
Create a figure to compare simulated grain yield against observed.
from tfunct.util import figures
dirname=os.path.join(config['RESULTS_PATH'], 'PRFT', 'Figures')
figures.chart_compareResults(df_result=PRFT_noStress, fld1="ObsYield", fld2="SimYield", alpha=.75, s=45, xy_lim=2,
hue='loc_code', loc_leg=2, ncol=2, ha='left', va='top',
title='PRFT\nNo streess condition', dirname=dirname, fname='PRFT_noStress',
dispScore=True, dispLegend=True, saveFig=True, showFig=True, fmt='jpg')
Changing the parameter df_result to PRFT_SFvpd, you can see the results for VPD stress conditions
figures.chart_compareResults(df_result=PRFT_SFvpd, fld1="ObsYield", fld2="SimYield", alpha=.75, s=45, xy_lim=2,
hue='loc_code', loc_leg=2, ncol=2, ha='left', va='top',
title='PRFT\nVPD streess condition', dirname=dirname, fname='PRFT_SFvpd',
dispScore=True, dispLegend=True, saveFig=True, showFig=True, fmt='jpg')
If you prefer see both figures sid-by-side in a single one to compare results, you can use the plot_corrTempFunct function as follows:
figures.plot_corrTempFunct(cmb_noStress=PRFT_noStress, cmb_noStress_filtered=None,
cmb_SFvpd=PRFT_SFvpd, cmb_SFvpd_filtered=None,
functype='PRFT',fld1='ObsYield',fld2='SimYield',hue='location',
ncol=6, s=80, alpha=0.95, xy_lim=1, fonts_axes=10, fonts_titles=12, dispScore=True, errorbar=True, saveFig=True, showFig=True, path_to_save_results=path_to_save_results, dirname='Figures', fname='Fig_1_errorbar', fmt='jpg')
Conclusion#
Congratulations
You have run a simulation using a prebuilt dataset and the Temperature Functions API.