Knowledge-Guided Machine Learning Models to Upscale Evapotranspiration in the U.S. Midwest

This repository contains the code that accompanies the paper
Knowledge-Guided Machine Learning Models to Upscale Evapotranspiration in the U.S. Midwest.
Aleksei Rozanov, Samikshya Subedi, Vasudha Sharma, Bryan Runck

Citation

Resource	DOI
Paper	(under review)
Dataset	10.13020/37XE-QQ18
Repository (code)

Abstract

Evapotranspiration (ET) plays a critical role in the land-atmosphere interactions, yet its accurate quantification across various spatiotemporal scales remains a challenge. In situ measurement approaches, like eddy covariance (EC) or weather station-based ET estimation, allow for measuring ET at a single location. Agricultural uses of ET require estimates for each field over broad areas, making it infeasible to deploy sensing systems at each location. This study integrates tree-based and knowledge-guided machine learning (ML) techniques with multispectral remote sensing data, griddled meteorology and EC data to upscale ET across the Midwest United States. We compare four tree-based models - Random Forest, CatBoost, XGBoost, LightGBM - and a simple feed-forward artificial neural network in combination with features engineered using knowledge-guided ML principles. Models were trained and tested on EC towers located in the Midwest of the United States using k-fold cross validation with k=5 and site-year, biome stratified train-test split to avoid data leakage. Results show that LightGBM with knowledge-guided features outperformed other methods with an R2=0.86, MSE=14.99 W·m^-2 and MAE = 8.82 W·m^-2 according to grouped k-fold validation (k=5). Feature importance analysis shows that knowledge-guided features were most important for predicting evapotranspiration. Using the best performing model, we provide a data product at 500 m spatial and one-day temporal resolution for gridded ET for the period of 2019-2024. Intercomparison between the new gridded product and state-level weather station-based ET estimates show best-in-class correspondence.

Content

ET_LCCMR/ ├─ LICENSE ├─ README.md ├─ requirements.txt # Core deps for MSI usage ├─ requirements_local.txt # Extras for local/dev use ├─ .gitignore ├─ paper/ # Drafts, manuscript assets ├─ fig/ # Exported figures ├─ models/ │ └─ ligthgbm_model.txt ├─ preprocessing/ # One-off data prep notebooks (site/stack/mesonet) │ ├─ 1. Carbon_Data.ipynb │ ├─ 2. Site_Selection.ipynb │ ├─ 3. Stack_Data.ipynb │ └─ 4. Mesonet_ValidationData.ipynb ├─ notebooks/ │ ├─ src/ │ │ ├─ PM_eq.py # PM helpers (LE↔ET conversion, etc.) │ │ ├─ train_ann.py # Simple ANN training │ │ └─ train.py # Tree-models training / utils │ ├─ 1. Final_Model_Training_and_Validation.ipynb │ └─ 2. Model_Mesonet_Testing.ipynb └─ pipeline/ # Data pipelines (ERA, MODIS) + utilities ├─ src/ │ ├─ config.py # Centralized config │ ├─ PM_eq.py # PM helpers (duplicated with notebooks/src/) │ ├─ utils_era.py # ERA5-Land acquisition and transforms │ └─ utils_modis.py # MODIS acquisition and transforms ├─ driveDerive.py ├─ 1. ERA_Pipeline.ipynb ├─ 2. MODIS_Pipeline.ipynb └─ 3. DataStacking.ipynb

Getting started

To replicate the analysis presented in this repository, follow these steps:

1. Clone the Repository

First, clone the repository to your local machine:

git clone https://github.com/RTGS-Lab/ET_LCCMR.git
cd ET_LCCMR

2. Set Up the Python Environment

It is highly recommended to use a virtual environment to manage dependencies.

  python -m venv .venv
  source .venv/bin/activate  # On Windows, use .venv\Scripts\activate
  pip install -r requirements_local.txt # Or requirements.txt to run on MSI

3. Use the notebooks

Navigate to the preprocessing/ directory and run the Jupyter notebooks in the specified order to prepare the data for model training:

1. Carbon_Data.ipynb

2. Site_Selection.ipynb

3. Stack_Data.ipynb

4. Mesonet_Validation_Data.ipynb