Flash droughts threaten global managed forests

Jianzhuang Pang ^{1, 2, 3}, Hang Xu ^{1, 2, 3, *}, Yang Xu ^{1, 2, 3}, Yifan Zhang ^{1, 2, 3}, Xiaoyun Wu ^{1, 2, 3}, Kexin Li ^{1, 2, 3}, Zhiqiang Zhang ^{1, 2, 3, *}

¹ Jixian National Forest Ecosystem Observation and Research Station, CNERN, Beijing Forestry University, Beijing 100083, P.R. China
² Key Laboratory of Soil and Water Conservation and Desertification Combating, State Forestry and Grassland Administration, Beijing 100083, P.R. China
³ School of Soil and Water Conservation, Beijing Forestry University, Beijing 100083, P.R. China

*Correspondence
Dr. Hang Xu, School of Soil and Water Conservation, Beijing Forestry University, Beijing, China. Email: hangxu@bjfu.edu.cn
Dr. Zhiqiang Zhang, School of Soil and Water Conservation, Beijing Forestry University, Beijing, China. Tel: 0086-10-62338828 Email: zhqzhang@bjfu.edu.cn

Description

The code for plotting and data analysis used in the manuscript.

• Developed by: Jianzhuang Pang
• Language(s) (Programming): Python
• Language(s) (NLP): English
• License: MIT License

How to Get Start

The results of the paper can be reproduced by using the following code in main.py.
Before using the code, you should download the data for analysis and replace manuscript_data with the folder path where the data is stored, and set figure_path to the folder path where you wish to save the results.

• Fig. 1 Spatial patterns and latitudinal variations in flash drought (FD) characteristics across global forests.

  figure_1abcd()  # Subplot a, b, c, and d
  figure_1e()  # Subplot e

• Fig. 2 Impacts of flash drought events (FDs) on global forest.

  figure_2a()  # Subplot a
  figure_2b()  # Subplot b
  figure_2c()  # Subplot c

• Fig. 3 Fig. 3 Interactive effects of regulating factors on forest responses (ΔNDVI) to flash drought events between intact (IF) and managed forests (MF).

  figure_3()

• Supplementary Figure 1. Spatial distribution of temporal trends in flash drought event characteristics.

  supplementary_figure_1_ace()  # Subplot a, c, and e
  supplementary_figure_1_bdf()  # Subplot b, d, and f
  supplementary_figure_1_g()  # Subplot g

• Supplementary Figure 2. Spatial distribution of (a) normalized temperature anomalies (ΔTa), (b) normalized precipitation anomalies (ΔP), and (c) consistency between ΔP and ΔTa during flash drought events.

  supplementary_figure_2_ab()  # Subplot a and b
  supplementary_figure_2_c()  # Subplot c

• Supplementary Figure 3. Spatial patterns and latitudinal variations in flash drought (FD) characteristics across global forests (soil moisture-based).

  supplementary_figure_3abcd()  # Subplot a, b, c, and d
  supplementary_figure_3_e()  # Subplot e

• Supplementary Figure 4. The spatial distribution of flash drought events frequency.

  supplementary_figure_4()

• Supplementary Figure 5. Spatial patterns of mean anomalies during flash drought events.

  supplementary_figure_5_abc()  # Subplot a, b, and c
  supplementary_figure_5_de()  # Subplot d and e

• Supplementary Figure 6. Impacts of flash drought events (FDs) on global forest (soil moisture-based).

  supplementary_figure_6_a() # Subplot a
  supplementary_figure_6_b() # Subplot b
  supplementary_figure_6_c() # Subplot c

• Supplementary Figure 7. Normalized Difference Vegetation Index anomalies (ΔNDVI) during (a) Standardized Precipitation-Evapotranspiration Index-base, (b) soil moisture-base flash drought events relative to concurrent normal conditions under different forest types.

  supplementary_figure_7_a() # Subplot a
  supplementary_figure_7_b() # Subplot b

• Supplementary Figure 8. The individual effects of regulating factors on forest response (ΔNDVI) to flash drought events.

  supplementary_figure_8()

• Supplementary Figure 9. Normalized Difference Vegetation Index anomalies (ΔNDVI) during flash drought events relative to concurrent normal conditions under different (a) Forest management; (b) Forest management practices.

  supplementary_figure_9_a()
  supplementary_figure_9_b()

• Supplementary Figure 11. Performance evaluation of XGBoost model with (a) training dataset (n=808,407); (b) validation datasets (n=538,939).

  supplementary_figure_11()

Technical Specifications

Hardware

• CPU: Intel Core i9-13900
• GPU: NVIDIA GeForce RTX 4090
• RAM: 128 GB
• Storage: 10 TB

Software

• IDE: PyCharm Community Edition 2023.3.5
• Operating System: Windows 11
• Python Version: Python 3.12.2
• Libraries/Frameworks:
  • pandas=2.2.1
  • numpy=1.26.4
  • scipy=1.12.0
  • statsmodels=0.14.0
  • joblib=1.2.0
  • seaborn=0.12.2
  • matplotlib=3.8.4
  • xarray=2025.1.1
  • rasterio=1.3.9
  • cartopy=0.22.0
  • dask=2023.11.0
  • xgboost=2.0.3
  • scikit-learn=1.3.0

Citation

Pang JZ, Xu H, Xu Y, Zhang YF, Wu XY, Li KX, and Zhang ZZ. 2025. Flash droughts threaten global managed forests. Nature Communications, under peer review.