Akbar et al. (2026) Unveiling kazakhstan's ecosystem service puzzle: Spatiotemporal shifts and drivers of supply and demand through multi-model integration and machine learning methods

Identification

• Journal: Ecological Indicators
• Year: 2026
• Date: 2026-01-01
• Authors: Adila Akbar, Alim Samat, Jilili Abuduwaili, Lunche WANG, Peijun Du, Dana Shokparova, Galymzhan Saparov, Sanim Bissenbayeva
• DOI: 10.1016/j.ecolind.2025.114569

Research Groups

• State Key Laboratory of Ecological Safety and Sustainable Development in Arid Lands, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi, China
• China-Kazakhstan Joint Laboratory for Remote Sensing Technology and Application, Al-Farabi Kazakh National University, Almaty, Kazakhstan
• Research Center for Ecology and Environment of Central Asia, Chinese Academy of Sciences, Urumqi, China
• Hubei Key Laboratory of Regional Ecology and Environmental Change, School of Geography and Information Engineering, China University of Geosciences, Wuhan, China
• School of Geography and Ocean Science, Nanjing University, Nanjing, China

Short Summary

This study quantifies the spatiotemporal dynamics of water yield, carbon storage, and food production in Kazakhstan from 1995 to 2022 using multi-model integration and machine learning, revealing growing ecosystem service supply-demand mismatches driven by population growth, climate variability, and energy consumption.

Objective

• To quantify the spatiotemporal dynamics of water yield (WY), carbon storage (CS), and food production (FP) in Kazakhstan from 1995 to 2022.
• To identify the dominant drivers of the ecosystem service supply–demand ratio (ESDR) and understand their nonlinear responses and spatial heterogeneity.
• To provide actionable, quantitative guidance for region-specific resource management and sustainable development in arid Central Asia.

Study Configuration

• Spatial Scale: Kazakhstan, 1 km grid resolution.
• Temporal Scale: 1995 to 2022 (28 years).

Methodology and Data

• Models used: InVEST (Annual WY module, Carbon Storage module), Google Earth Engine (GEE), Random Forest (RF), SHAP (SHapley Additive exPlanations) analysis, Normalized Partial Dependence Plots (PDPs), Geographically and Temporally Weighted Regression (GTWR).
• Data sources:
  • Satellite: GLCFCS30 Global 30 m Resolution Dataset (Land use/Land cover), AVHRR (NDVI 1995), MODIS (NDVI 1996–2022, Evapotranspiration 1996–2022, GPP 2000–2022), DMSP-OLS (Night light 1992–2012), VIIRS-DNB (Night light 2013–2022).
  • Reanalysis: ERA5LAND (PAR, Evapotranspiration 1995, Precipitation 1995).
  • Observation/Other: CHIRPS dataset (Precipitation 1996–2022), USGS STRM data (DEM), ISRIC - SoilGrids dataset (Soil root depth), GHSL-POP dataset (Population density), Figshare Electricity Consumption dataset (Electricity consumption 1995–2019), Landscan global population data (Electricity consumption 2022), Natural Earth data sets (Administrative boundary data).
  • Guidelines: IPCC National Greenhouse Gas Inventory Guidelines (Carbon storage parameters).

Main Results

• Temporal changes (1995-2022): Water yield (WY) fluctuated over 25% (national decline by approximately 14%), carbon storage (CS) declined by 18.6% (national decrease by approximately 9%), and food production (FP) increased by 42.3% (national increase by approximately 35%).
• Spatial patterns: WY was concentrated in eastern and northern regions; CS in central and southern regions; FP in northern and western regions. Natural barriers limited ES supply in the south and east.
• ESDR dynamics:
  • WY: Surplus or balanced areas reached 64.8%; severe deficits decreased from 20.7% to 16.5%.
  • CS: High-capacity areas declined from 19.3% to 14.2%; low-capacity areas rose from 44.7% to 51.1%, indicating reduced carbon sequestration.
  • FP: High-capacity areas grew from 13.3% to 20.8%; deficit areas also increased from 17.3% to 21.7%, highlighting emerging food security risks.
• Mechanistic drivers: Population growth, climate variability (precipitation), vegetation cover (NDVI), and energy consumption (electricity consumption, night light) collectively explained over 60% of ESDR variance. These drivers exhibited threshold-dependent and nonlinear responses, with effects plateauing at higher intensities.
• Spatial heterogeneity of drivers (GTWR): Topography and precipitation strongly influenced WY in mountainous south/east; vegetation conditions and anthropogenic emissions drove CS heterogeneity; intensive cropland management and technological input influenced FP in northern agricultural belt and southern oases.

Contributions

• Provides a long-term (1995–2022), high-resolution assessment of key ecosystem services (water yield, carbon storage, food production) in Kazakhstan using the InVEST model and Google Earth Engine.
• Develops a novel methodological framework integrating Random Forest-SHAP analysis, Partial Dependence Plots, and Geographically and Temporally Weighted Regression to identify nonlinear drivers, quantify their importance, and reveal spatiotemporal heterogeneity of their effects.
• Evaluates ecosystem service supply-demand mismatches, offering actionable insights for balancing conservation and socio-economic needs and supporting targeted ecological protection and sustainable development in arid Central Asia.

Funding

• Western Young Scholars Project of the Chinese Academy of Sciences (grant number 2022-XBQNXZ-001)
• National Natural Science Foundation of China (grant number 42371389)

Citation

@article{Akbar2026Unveiling,
  author = {Akbar, Adila and Samat, Alim and Abuduwaili, Jilili and WANG, Lunche and Du, Peijun and Shokparova, Dana and Saparov, Galymzhan and Bissenbayeva, Sanim},
  title = {Unveiling kazakhstan's ecosystem service puzzle: Spatiotemporal shifts and drivers of supply and demand through multi-model integration and machine learning methods},
  journal = {Ecological Indicators},
  year = {2026},
  doi = {10.1016/j.ecolind.2025.114569},
  url = {https://doi.org/10.1016/j.ecolind.2025.114569}
}