Water Resources Allocation for the Mainstream of the Tarim River: Principles, Methods, and Practice

1. Background: Why Ecological Rivers Need Dedicated Allocation Principles

Water resources allocation in arid inland river basins differs fundamentally from that in humid regions. Ecological rivers such as the mainstream of the Tarim River share three prominent characteristics: total water availability is limited and highly dependent on alpine snow and glacier melt, intra-annual runoff distribution is extremely uneven, and natural vegetation belts along the riverbanks depend on flood overbanking and groundwater recharge. This means allocation cannot merely answer “who gets water and how much” — it must also answer “how much water should be left in the channel itself, when, and with what flow process.”

Under the strictest water resources management system, the “three red lines” form rigid boundaries for allocation: the total water use control red line caps available supply, the water use efficiency control red line constrains water consumption per unit of output, and the water function zone pollution carrying capacity red line constrains water quality and discharge behavior. For ecological rivers, a further ecological constraint must be superimposed — safeguarding downstream ecological baseflow and specified discharge volumes at key cross-sections. The stacking of multiple constraints makes experience-based scheduling inadequate; a systematic analysis–decision–verification chain must be established.

2. Four Layers of Allocation Principles

2.1 Baseline Information Layer: Spatial-Temporal Characterization of Precipitation and Evapotranspiration

The rationality of an allocation scheme first depends on the accuracy of understanding the basin’s water cycle baseline. Precipitation analysis must address interannual variability, intra-annual distribution, spatial differences, and extreme event frequency; evapotranspiration analysis must distinguish potential from actual evapotranspiration and identify the main zones of vegetation water consumption and open-water evaporation. A common approach is to build gridded precipitation and evapotranspiration datasets from long-series meteorological observations and remote sensing products, then aggregate them by sub-basin or irrigation district.

The output of this layer is not a single number but a set of spatial-temporal distribution characteristics with uncertainty. For example, under the same total amount, whether precipitation concentrates in the flood season or spreads across spring and autumn has a completely different effect on the available water process for allocation. In engineering practice, mean values, quantiles, and scenario values for typical dry years should be provided simultaneously to offer interval-based references for subsequent decisions.

2.2 Scenario Analysis Layer: Wet-Dry Encounter and Virtual Water Balance

Wet-dry encounter analysis characterizes the joint probability of wet and dry combinations across different source regions. Mainstream inflow and major tributary inflow are not independent events; if upstream tributaries and the mainstream are simultaneously dry, available water will be significantly lower than if each were dry alone. After constructing a joint distribution, inflow combination scenarios at different reliability levels can be derived as input conditions for allocation schemes.

Virtual water balance examines the rationality of basin water use from a “water footprint” perspective. It incorporates the water embedded in agricultural products, industrial goods, and services into the accounting, revealing the degree of local consumption’s dependence on external water resources and the scale of local water resources exported through trade. For basins dominated by agricultural water use, virtual water balance analysis helps identify potential space for adjusting cropping structure and provides quantitative support for the efficiency red line.

2.3 Optimization Decision Layer: Water Allocation Under Red-Line Constraints

Once information and scenarios are sufficiently prepared, the allocation problem can be formalized as a multi-objective optimization model. Decision variables are the allocated water volumes to each water use unit (domestic, industrial, agricultural, ecological) in each period; objective functions typically include maximizing economic benefit, minimizing water shortage, and minimizing ecological deviation; constraints include:

• Total volume constraint: the sum of allocations to all units does not exceed available supply;
• Efficiency constraint: water consumption per unit area or per unit output value does not exceed the efficiency red line;
• Ecological constraint: discharge at key cross-sections is not below the ecological baseflow threshold, and recharge to important wetlands and vegetation areas is not below minimum requirements;
• Water balance constraint: flow routing between nodes satisfies the continuity equation;
• Water quality constraint: pollutant discharge does not exceed the water function zone’s carrying capacity.

For solution methods, linear programming or mixed-integer programming can handle rigid constraints, while multi-objective evolutionary algorithms generate Pareto fronts for decision-makers to trade off between efficiency and ecology. It must be emphasized that the value of optimization results lies not in producing a single optimal solution but in revealing the cost structure under different trade-offs.

Input: inflow scenario set, water demand set, constraint parameters
For each inflow scenario s:
    Build water balance and constraint matrix
    Solve multi-objective model to obtain Pareto solution set P_s
    Compute ecological deviation and water shortage indicators for solutions in P_s
Output: cross-scenario solution evaluation matrix

2.4 Implementation Layer: From Allocation Results to Scheduling Rules

The optimization model outputs idealized period-by-period allocation volumes; actual scheduling must be translated into executable rules. A common approach is to decompose the annual allocation scheme into monthly or ten-day control water levels or flow thresholds at key nodes, accompanied by a rolling correction mechanism that increases in wet years and decreases in dry years. A monitoring–evaluation–feedback loop should also be established: tracking scheme performance through cross-section flow, groundwater depth, vegetation indices, and other indicators, triggering scheme adjustment when deviations exceed thresholds.

3. Key Points in Engineering Practice

First, ecological water demand cannot be treated as a “residual item.” If domestic and industrial water uses are satisfied first and the remainder is allocated to ecology, ecological water in dry years will inevitably be squeezed out. The rational approach is to treat ecological baseflow as a rigid lower-bound constraint up front, with productive water uses allocated after it.

Second, inflow uncertainty must be handled explicitly. Scenario-based design is recommended rather than using a single mean inflow, and key schemes should undergo dry-year stress testing to avoid failure in below-normal years.

Third, implementing the efficiency red line depends on metering and accounting systems. Without reliable water withdrawal metering and consumption accounting, efficiency constraints cannot be translated into operable allocation parameters. Metering infrastructure and data quality control are prerequisites for allocation work.

Fourth, multi-objective trade-offs require an institutionalized consultation mechanism. Optimization models can provide technical trade-off curves, but the final allocation scheme involves inter-regional interest adjustments and requires a consultation platform to build consensus, with technical results serving as the factual basis for negotiation.

4. Summary

Water resources allocation for the mainstream of the Tarim River is essentially a systematic engineering effort to coordinate ecological protection and socio-economic water demands under extremely uneven water resource conditions. Its technical path can be summarized as: characterizing the baseline through precipitation-evapotranspiration and wet-dry encounter analysis, identifying water-saving potential through virtual water balance, generating solution sets under the “three red lines” and ecological constraints through multi-objective optimization, and achieving implementation through scheduling rules and monitoring feedback. The four layers are interlinked; missing information at any layer propagates downstream as decision risk. For governance practices of similar arid-zone ecological rivers, this “information–scenario–optimization–implementation” framework offers general reference value.