Watershed Drought and Ecosystem Services: Spatiotemporal Characteristics and Gray Relational Analysis

Jizhou Bai, Zixiang Zhou, Yufeng Zou, Bakhtiyor Pulatov, Kadambot H. M. SiddiqueISPRS Int. J. Geo Inf 2021 · DOI

TL;DR

This study investigates the complex interplay between drought and ecosystem services, specifically soil conservation, in China's YanHe Watershed. By integrating remote sensing data with hydrological modeling and employing a novel Gray Relational Analysis (GRA) approach, the research reveals that while vegetation restoration enhances soil conservation, it also increases water consumption and can exacerbate drought conditions.

What Problem Does This Paper Address?

The YanHe Watershed, located in the Loess Plateau, is a region characterized by both severe soil erosion and water scarcity. The Chinese government's "Grain for Green" Project (GFGP) has led to extensive vegetation restoration efforts aimed at combating erosion and improving the environment. However, the ecological consequences of this restoration are not fully understood, particularly the trade-offs between improved soil conservation and increased water demand. Previous studies have often analyzed these factors in isolation or at a macro scale. This paper aims to bridge this gap by providing a detailed spatiotemporal analysis of drought, soil conservation services, and their interrelationships, considering the impact of vegetation restoration.

Methods

The study employs a multi-faceted approach:

  • Hydrological Modeling (SWAT): The Soil and Water Assessment Tool (SWAT) was used to simulate soil conservation services. A key innovation here is the modification of the SWAT model to incorporate remotely sensed Normalized Difference Vegetation Index (NDVI) data. This allows for a more accurate representation of vegetation cover changes, which directly influences the C-factor in soil erosion calculations, leading to improved simulation accuracy.
  • Drought Monitoring (ESI): The Evaporative Stress Index (ESI) was chosen to assess drought conditions. ESI is derived from actual evapotranspiration (ET) and potential ET (PET) and is considered more responsive to drought caused by multiple factors, including vegetation and soil moisture, compared to precipitation-based indices like SPI.
  • Gray Relational Analysis (GRA): This multivariate statistical method was used to analyze the complex relationships between drought, soil conservation services, and other influencing factors (meteorological data, vegetation, ET). GRA is particularly useful for systems with intricate interdependencies and unclear internal mechanisms, such as ecosystems, and does not require data to follow a specific distribution.
  • Trend Analysis (Mann-Kendall Test): The Mann-Kendall test was applied to assess the statistical significance of trends in vegetation cover (NDVI) and evapotranspiration (ET) over time.

Datasets

The study utilized a comprehensive set of datasets:

  • Digital Elevation Model (DEM): 30m resolution, from Geospatial Data Cloud.
  • Soil Data: China Soil Dataset based on the World Soil Database (HWSD), from the National Tibetan Plateau Data Center.
  • Land Use Data: 30m resolution, reclassified into six types, from the Resource and Environmental Science Data Center of the Chinese Academy of Sciences.
  • Meteorological Data: Daily precipitation, relative humidity, wind speed, maximum and minimum temperature (1999-2015), from the China Meteorological Data Sharing Network.
  • Hydrological Data: Daily runoff and sediment transport (2000-2015), from the Ganguyi Hydrological Station.
  • Vegetation and ET Data: MODIS NDVI (MOD13, 16 days/250m) and ET (MOD16, 8 days/500m) for 2000-2015, obtained from NASA. These were processed to obtain monthly NDVI and ET levels.

Key Insights and Ideas

  • Vegetation Restoration's Double-Edged Sword: While the GFGP has successfully increased vegetation cover and significantly improved soil conservation services, this comes at the cost of increased evapotranspiration (ET) and, consequently, a potential exacerbation of drought conditions. This highlights a critical trade-off in ecological restoration efforts.
  • GRA as a Powerful Tool for Ecosystem Analysis: The study demonstrates the efficacy of GRA in untangling complex, non-linear relationships within ecosystems, providing a robust method for understanding the drivers of ecosystem services and their interactions with environmental stressors like drought.
  • Remote Sensing Integration Enhances Hydrological Models: Modifying the SWAT model to incorporate NDVI data significantly improved its accuracy in simulating sediment yield, underscoring the value of integrating high-resolution remote sensing data into traditional hydrological models.
  • Spatial Heterogeneity is Crucial: The relationships between drought and soil conservation are not uniform across the watershed. Negative correlations (where increased soil conservation leads to increased drought) are more pronounced in areas with high vegetation coverage, particularly in the middle and lower reaches, suggesting that the type and density of vegetation are critical factors.

Key Results

  • Vegetation Growth: The YanHe Watershed experienced significant vegetation cover increase, with an average annual NDVI growth rate of 0.0245 × 10⁻¹ and a substantial increase in ET by 14.35 mm·a⁻¹.
  • Drought Extent: Approximately 63.86% of the watershed is classified as arid, with drought generally worsening from south to north.
  • Soil Conservation Improvement: Soil conservation modulus (SCM) increased significantly from 116.87 t·ha⁻¹·a⁻¹ in 2000 to 412.58 t·ha⁻¹·a⁻¹ in 2015, with a multi-year average of 235.69 t·ha⁻¹·a⁻¹.
  • Model Accuracy: Integrating NDVI into SWAT improved simulation accuracy, with R² increasing from 0.63 to 0.76 and NSE from 0.46 to 0.51 during calibration, and R² from 0.82 to 0.93 and NSE from 0.57 to 0.61 during validation.
  • GRA Findings: Vegetation cover and drought were identified as the primary drivers affecting ET and soil conservation services. A strong negative correlation (0.9779) was found between soil conservation services (SCM) and drought (ESI), particularly in vegetated areas.

Limitations and Future Work

  • Remote Sensing Data Resolution: The study acknowledges limitations due to the availability of remote sensing data, particularly the lack of high temporal and spatial resolution NDVI data for the modified SWAT model.
  • Value Accounting of ESs: The paper notes a lack of research on the economic valuation of ecosystem services (ESs), which limits the guidance for eco-compensation policies.
  • Future Research Directions: The authors suggest further research should consider the impact of climate change and human activities, analyze drought-ES relationships from a physical mechanism perspective, and evaluate the monetary value of ESs to inform eco-compensation strategies. They also recommend incorporating low-water-consumption tree species and restricting afforestation density in future restoration efforts.

Summary updated

BibTeX
@article{bai2021watershed,
  author = {Jizhou Bai and Zixiang Zhou and Yufeng Zou and Bakhtiyor Pulatov and Kadambot H. M. Siddique},
  title = {{Watershed Drought and Ecosystem Services: Spatiotemporal Characteristics and Gray Relational Analysis}},
  journal = {ISPRS Int. J. Geo Inf.},
  year = {2021},
  doi = {10.3390/ijgi10020043},
}
Figure