README – ET Evaluation HydroShare Resource

This repository supports the study:
“Evaluating the U.S. National Water Model Retrospective Evapotranspiration Simulation using Eddy-Covariance Flux Tower Measurements”
by Chapagain, Maghami, and Ames (2025).

Getting Started

Before running the notebooks, please unzip both archives into the same working directory:

• Notebooks_and_Data.zip – contains all Jupyter notebooks (JN1–JN5) and their input/output subdirectories.
• Figure1.zip – contains the ArcGIS Pro project and layer files used to create the site-location map (Figure 1).

Once unzipped, open and execute the notebooks sequentially (JN1 → JN5) to reproduce the results, figures, and tables presented in the paper.

Folder Structure and Workflow

1. Notebooks_and_Data.zip

Contains all analysis and visualization notebooks (JN1–JN5) and their associated data folders.

JN1_Input

Contains the primary input datasets and preprocessing outputs.

Preprocessing Overview

This study involved two main preprocessing stages prior to analysis.

(1) ET Preprocessing (NWM + AmeriFlux)
- Downloaded and aggregated NWM evapotranspiration (ET) data.
- Processed AmeriFlux latent heat flux (LE) data to compute monthly ET.
- Integrated site metadata (RFC, Köppen–Geiger climate zone, land use, and elevation).
- Generated a combined dataset of monthly ET from both NWM and tower observations across 72 sites.
- Note: Raw AmeriFlux and NWM data, along with intermediate scripts, are not included due to size and access constraints.
- The summarized file Monthly_NWM_Tower_ET_Siteinfo.csv serves as the starting point for all subsequent analyses.
- Refer to the published paper for additional details on preprocessing and data sources.

(2) PET and P/PET Ratio Preprocessing
- Calculates the precipitation-to-potential evapotranspiration ratio (P/PET) used to classify each site-month as water-limited or energy-limited.
- Corresponding scripts and derived data are included inside
JN1_Input/PET_and_P_PET_ratio_Preprocess/.
- Outputs are used in downstream notebooks (e.g., JN5) for regime-based ET performance evaluation.

JN4_Input

Contains merged temperature and ET datasets from NWM (AORC forcing) and AmeriFlux, used in JN4 to evaluate temperature bias and its relationship with ET performance.

Output Folders

Each notebook (JN1 through JN5) automatically generates its own output directory (e.g., JN2_Output, JN3_Output, etc.) upon execution.
These folders store all intermediate results, metrics, and figures, ensuring full reproducibility without manual setup.

Analysis and Visualization

• JN1 – Integrates NWM and AmeriFlux ET data, computes P/PET ratios, and prepares summarized input files.
• JN2 – Computes site-level ET performance metrics (Scaled KGE, R², RMSE, PBIAS).
• JN3 – Visualizes and ranks ET performance across sites, RFCs, and climate zones.
• JN4 – Evaluates AORC temperature forcing accuracy and examines its relationship with ET biases.
• JN5 – Analyzes and visualizes water-limited vs. energy-limited ET performance using P/PET-based regime classification.

2. Figure1.zip

Contains the ArcGIS Pro project and layer files used to create Figure 1 of the paper.
The map displays all 72 AmeriFlux tower sites, color-coded by RFC, land use, and Köppen–Geiger climate zone.
Clustered site symbols are offset slightly for clarity.
ArcGIS Pro was used for both spatial attribute assignment (e.g., RFC, climate, land use, elevation) and visualization.

Notes

• Most notebooks (JN1–JN5) run locally in Python using standard scientific libraries such as pandas, numpy, and matplotlib.
• ArcGIS Pro was used only to generate the site-location and hydroclimatic maps included in Figure 1.
• All analytical results, figures, and tables in the manuscript can be reproduced using the provided Jupyter notebooks and summarized input files.