TL;DR
This paper introduces a novel "Runoff Observation Method" (ROM) to more accurately assess regional water conservation services by integrating actual hydrological station data with ecosystem and soil type information. This method overcomes the limitations of traditional "lookup table" approaches, which often overestimate water conservation due to their reliance on generalized coefficients and lack of real-world runoff data.
What Problem Does This Paper Address?
Assessing the water conservation capacity of terrestrial ecosystems is crucial for effective regional environmental management. However, current methods, primarily the Water Balance Model (WBM) using a lookup table (LTM) approach, face significant challenges. The LTM relies on generalized runoff coefficients derived from literature or small-scale experiments, which don't fully capture the spatial heterogeneity of ecosystems and their actual hydrological behavior. This leads to uncertainties and potential overestimations of water conservation, hindering accurate ecological protection and decision-making. The core problem is the lack of a robust method to quantify spatiotemporal runoff depth and, consequently, water conservation capacity at a regional scale using actual, localized hydrological data.
Methods
The study proposes a new method called the Runoff Observation Method (ROM). This approach integrates several key components:
- Ecosystem Type Data: Derived from Landsat satellite imagery, providing a high spatial resolution (30m) classification of ecosystems.
- Hydrological Soil Group (HSG) Data: Obtained from the Harmonized World Soil Database (HWSD).
- Runoff Coefficient (Rc) and Curve Number (CN) Data: Gathered through meta-analysis of relevant literature and the HWSD database, respectively.
- Runoff Index (RI) Calculation: A comprehensive runoff index ($R_i$) is calculated by combining the runoff index of ecosystem type ($R_{it}$) and the runoff index of soil ($R_{is}$) using an equal weighting approach. This step aims to capture the combined influence of ecosystem and soil characteristics on runoff.
- Integration of Hydrological Station Data: Crucially, ROM uses actual annual runoff data from inlet and outlet hydrological stations within the study basin. This data is used to spatially distribute runoff depth across the basin using the calculated comprehensive runoff index. This is a significant departure from LTM, which uses generalized coefficients.
- Water Balance Model (WBM): The WBM is then applied using the spatially distributed runoff depth, along with precipitation and evapotranspiration data, to calculate water conservation.
For comparison, the study also employs the traditional Lookup Table Method (LTM), which uses a pre-defined table of runoff coefficients based on ecosystem types and precipitation data to estimate runoff depth.
Datasets
- Landsat Satellite Imagery: Used for ecosystem classification at a 30m spatial resolution.
- Harmonized World Soil Database (HWSD) Version 1.2: Used to extract hydrological soil group (HSG) and curve number (CN) values.
- Meteorological Station Data: 251 stations were used for annual precipitation interpolation (500m resolution).
- MODIS Standard Products (MOD16A2): Used to generate annual evapotranspiration data (500m resolution).
- Annual Hydrological Report of P.R. China: Daily observation data from 2001-2020 at inlet and outlet hydrological stations of the Yellow River basin were used to calculate annual runoff.
Key Insights and Ideas
- The critical role of actual hydrological data: The study strongly emphasizes that integrating real-world runoff data from hydrological stations is paramount for accurate water conservation assessment, moving beyond theoretical coefficients.
- Spatial heterogeneity matters: ROM's ability to incorporate detailed ecosystem and soil data at a fine resolution (30m) reveals significant spatial variations in runoff and water conservation that are masked by the coarser LTM approach.
- LTM's tendency to overestimate: The research consistently shows that LTM overestimates water conservation capacity, sometimes by a factor of two to three, highlighting the limitations of generalized coefficients.
- Delayed impact of precipitation: ROM captures a more nuanced relationship between precipitation and runoff depth, demonstrating a delayed and cumulative effect, which LTM does not fully represent.
- Farmland's impact: The study identifies farmland as a significant factor in reducing regional water conservation capacity, a finding more clearly delineated by ROM.
Key Results
- Runoff Depth: ROM calculated an average runoff depth of 146.08 mm, significantly higher than LTM's 61.66 mm. ROM also showed more pronounced spatial differentiation and a higher rate of increase over the study period (91% increase vs. 24% for LTM).
- Water Conservation: ROM yielded an average water conservation of 68.96 mm, while LTM significantly overestimated it at 118.86 mm. Both methods showed a declining trend in water conservation over time, but ROM indicated a steeper decline (45% vs. 21% for LTM).
- Spatial Resolution: ROM achieved a finer spatial resolution (30m) for runoff depth compared to LTM (1000m), allowing for more detailed analysis.
- Sub-watershed Analysis: ROM detected significant increases in runoff depth across all sub-watersheds, whereas LTM only showed significant increases in specific regions.
Limitations and Future Work
- Concentrated Farmland and Urban Areas: The study acknowledges that the presence of concentrated farmland and small cities might lead to an underestimation of water consumption and, consequently, an overestimation of regional water conservation capacity.
- Terrain Slope: The impact of terrain slope was not explicitly considered in the study design, which could influence runoff production, especially at smaller watershed scales.
- Limited Hydrological Station Data: The evaluation relied on inlet and outlet hydrological stations, lacking data from internal sub-watershed nodes. This limits the accounting for scale effects on water conservation capacity.
- Future Work: The authors suggest that increasing the quantity of actual observation data is the most effective way to address these uncertainties and improve evaluation methods and models. They also highlight the need to consider terrain slope and scale effects in future research.
Summary updated