Data and code for: "Meteorological drivers of three sequential honeybee pollination behaviors in gold kiwifruit: species-level thresholds, prediction hierarchy, and cross-site generalizability"

This repository contains the complete reproducibility package — raw data and analysis code — supporting the manuscript: Min, S.-H., Lee, Y.-B., Lee, K.-Y. (2026). Meteorological drivers of three sequential honeybee pollination behaviors in gold kiwifruit: species-level thresholds, prediction hierarchy, and cross-site generalizability. Submitted to Agriculture, Ecosystems & Environment. CONTENTS - data_wando_hourly.csv — Hourly meteorological and honeybee foraging activity records from the Wando development site (2018–2025; 11,546 hourly records, 10-minute observation intervals aggregated to hourly). Variables: air temperature, relative humidity, rainfall, wind speed, solar radiation, hive exit frequency, flower visitation count, flower dwell time. - data_boseong_hourly.csv — Hourly meteorological and honeybee foraging activity records from the Boseong validation site (2018–2025; 5,856 hourly records, 30-minute observation intervals aggregated to hourly). Same variable structure as Wando. - Ch1_reproduce.py — Python reproduction script for Tables 1–4, Table S1 (meteorological distribution overlap), Table S6 (raster variable compatibility), BPEI weight sensitivity analysis (171 weight combinations), and optional bootstrap 95% confidence intervals for temperature optima (Table S7). - Ch1_supplementary_figures.py — Python script reproducing Figure S2 (solar radiation response curves) and Figure S3 (variable importance rankings). - README.md — Detailed documentation including column definitions, column-by-column data descriptions, execution instructions, and a reproduction verification table comparing manuscript values to code output. SCOPE The study establishes species-level meteorological thresholds for three sequential honeybee pollination behaviors — hive exit, flower visitation, and flower dwell time — in open-field gold kiwifruit (Actinidia chinensis cv. 'Haegeum'). Using eight years of field observations from two sites on the southern coast of the Korean Peninsula, the analysis identifies temperature optima, quantifies cross-site transferability, and evaluates the meteorological predictability of each behavioral component. REQUIREMENTS Python 3.8 or later, with the following packages: pandas, numpy, scipy, scikit-learn, matplotlib Installation: pip install pandas numpy scipy scikit-learn matplotlib Expected runtime: approximately 3 minutes on a standard desktop computer (bootstrap CI analysis adds ~9 minutes when enabled). USAGE 1. Place all CSV files and Python scripts in the same directory. 2. Execute: python Ch1_reproduce.py 3. For supplementary figures: python Ch1_supplementary_figures.py 4. Reproduction outputs are compared to manuscript values in README.md. FUNDING This work was supported by the Cooperative Research Program for Agriculture Science and Technology Development (Project No. PJ015878), Rural Development Administration, Republic of Korea. LICENSE Released under Creative Commons Attribution 4.0 International (CC BY 4.0). Please cite both this dataset and the associated manuscript when using these materials. KEYWORDS honeybee foraging, Apis mellifera, kiwifruit pollination, Actinidia chinensis, asymmetric Gaussian response, temperature optimum, cross-site validation, meteorological prediction, machine learning, agricultural