Jing et al. (2026) Evaluating composite drought index performance in capturing soil moisture dynamics and vegetation response across heterogeneous dryland environments
Identification
- Journal: Ecological Indicators
- Year: 2026
- Date: 2026-11-16
- Authors: Qiqi Jing, Akinwale Temitope Ogunrinde, Xian Xue, Paul Adigun
- DOI: 10.1016/j.ecolind.2026.115831
Research Groups
- State Key Laboratory of Ecological Safety and Sustainable Development in Arid Lands, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Lanzhou, Gansu, China
- University of Chinese Academy of Sciences, Beijing 101408, China
- Department of Engineering Mechanics and Energy, University of Tsukuba, Tsukuba, Japan
Short Summary
This study develops and evaluates the Multi-source Integrated Drought Index (MIDI) using principal component analysis (MIDIP) and Euclidean distance (MIDIE) to better characterize coupled hydroclimatic–ecological drought processes across heterogeneous dryland environments. The MIDI framework, particularly MIDI_P, significantly improves drought characterization by integrating atmospheric forcing, terrestrial water dynamics, and ecosystem responses.
Objective
- Develop and systematically evaluate the Multi-source Integrated Drought Index (MIDI), integrating precipitation, root-zone soil moisture, surface moisture conditions, and vegetation stress using Principal Component Analysis (MIDIP) and Euclidean distance (MIDIE) approaches, to improve drought characterization across heterogeneous dryland environments.
Study Configuration
- Spatial Scale: Arid and semi-arid regions of North Africa, West Asia, and Central Asia (approximately 17°W to 126°E and 7.5°N to 55°N), defined by an aridity index less than 0.65. The domain is subdivided into 9 IPCC climate reference regions.
- Temporal Scale: 1982–2022 (41-year period), with monthly time series analysis.
Methodology and Data
- Models used:
- Multi-source Integrated Drought Index (MIDI), with two versions: MIDIP (Principal Component Analysis) and MIDIE (Euclidean distance).
- Component indices: Standardized Precipitation Index (SPI-1), Soil Moisture Index (SMI), Temperature Vegetation Dryness Index (TVDI), Evapotranspiration Condition Index (ETCI).
- Statistical methods: Pearson's correlation coefficient, Mann–Kendall (MK) test, ordinary least-squares regression, run-theory-based analysis (Frequency Rate, Mean Duration, Mean Intensity, Persistence Index), Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Hurst exponent (R/S analysis).
- Data sources:
- ERA5-Land reanalysis: Monthly precipitation, land-surface temperature, potential evapotranspiration, actual evapotranspiration (0.1° spatial resolution).
- Global Land Evaporation Amsterdam Model (GLEAM): Root-zone soil moisture.
- GIMMS-3G+ NDVI dataset: Global Normalized Difference Vegetation Index (approximately 0.0833° spatial resolution).
- rEC-LUE model: Global monthly gross primary productivity (0.05° spatial resolution).
- SPEIbase v2.8: Standardized Precipitation–Evapotranspiration Index (SPEI).
- CRU TS data: Self-calibrating Palmer Drought Severity Index (sc-PDSI) (0.5° × 0.5° resolution).
- MODIS MCD12Q1 product: Land-cover information.
- Emergency Events Database (EM-DAT): Documented historical drought events.
- Global digital elevation data.
- Global Aridity Index and Potential Evapotranspiration (ET0) Database, Version 3.
Main Results
- MIDIP exhibited stronger correlations with gross primary productivity (GPP) than MIDIE across approximately 62% of the study area.
- The first principal component (PC1) of MIDI_P explained 75.56% of the total variance across meteorological, hydrological, and ecological variables, indicating strong hydroclimatic coherence.
- MIDI_P showed the highest correlation with GPP and NDVI variations, superior characterization of drought frequency, duration, severity, and persistence metrics, and stronger agreement with soil moisture and evaporative stress observations compared with conventional indices.
- CEEMDAN analysis of MIDI revealed dominant periodicities at 0.5, 1.0, 1.7, and 5.1 years, superimposed on a significant accelerating drying trend.
- Over 85% of the study area exhibited a Hurst exponent (H) > 0.5, indicating persistent drought behavior.
- Regional analysis showed contrasting drought characteristics: North Africa and the Arabian Peninsula experienced persistent severe drought conditions, while Central Asia exhibited relatively frequent but shorter drought events.
- Both MIDIE and MIDIP showed strong negative relationships with drought-affected area (DAA) and deep-drought area (DDA) (p < 0.001). MIDIE had a slightly stronger association with DAA (R² = 0.88, r = -0.94), while MIDIP showed a stronger association with DDA (R² = 0.95, r = -0.97).
Contributions
- Development of the Multi-source Integrated Drought Index (MIDI) as a physically interpretable framework for dryland drought monitoring, explicitly linking climate drivers to hydrological and ecological consequences.
- Systematic comparison of two contrasting index-construction strategies (Euclidean distance and Principal Component Analysis) within a common analytical and evaluation framework, examining how integration methods influence drought characterization and consistency with climatic references, ecosystem productivity, and documented drought events.
- Advancement of large-scale drought assessment, early warning capabilities, and climate adaptation planning for dryland populations.
- Provides an operational, physically transparent monitoring system that improves drought characterization across heterogeneous dryland systems by coupling atmospheric forcing, terrestrial water dynamics, and ecosystem responses.
Funding
- Northwest Institute of Ecological Environment and Resources, Chinese Academy of Science (grant number: E429020101).
Citation
@article{Jing2026Evaluating,
author = {Jing, Qiqi and Ogunrinde, Akinwale Temitope and Xue, Xian and Adigun, Paul},
title = {Evaluating composite drought index performance in capturing soil moisture dynamics and vegetation response across heterogeneous dryland environments},
journal = {Ecological Indicators},
year = {2026},
doi = {10.1016/j.ecolind.2026.115831},
url = {https://doi.org/10.1016/j.ecolind.2026.115831}
}
Original Source: https://doi.org/10.1016/j.ecolind.2026.115831