The initial emission rights allocation is the key measure to achieve the goal of total amount control and deepen the emission trading system. Although many studies have focused on the modeling of initial emission rights allocation, such as using game theory and multi-objective optimization methods, few studies have observed the hierarchical relationship of mutual interference and restriction between watershed management agency and local governments in each subarea during allocation. This relationship directly affects the rationality of the results of regional emission rights allocation. In this study, a leader-follower hierarchical decision model (LFHDM) for allocating initial emission rights in a basin is developed. Based on the bilevel programming approach, the model simulates the interactive decision-making process between the watershed management agency of the upper-level model (LFHDM-U) and the local government of the lower-level model (LFHDM-L) in the allocation under total amount control. A case study of China’s Yellow River Basin is conducted to demonstrate the feasibility and practicality of the model. Findings reveal that, compared with the single-level model, the developed LFHDM has higher satisfaction with the allocation scheme. Under different scenarios, the overall satisfaction of the configuration schemes of COD and NH₃-N in each province and autonomous region remains above 0.9. In addition, the allocation volumes of COD and NH₃-N in each province of the Yellow River Basin in planning year increase with the enhancement of allowable assimilative capacity of water bodies, but the interval gap of satisfaction with allocation schemes gradually narrows. It shows that when the allowable assimilation capacity of a water body is low, the decision-making of the allocation scheme needs to be more cautious. Moreover, for the Yellow River Basin, apart from Qinghai and Sichuan, the task of reducing water pollutants in other provinces in the next few years is very arduous. The average reduction of total COD and NH₃-N in the basin is about 48% and 46%, respectively.

Virtual water exerts an indispensable influence on water resources, yet the existing studies on the water rights allocation of transboundary rivers hardly consider virtual water transfer (VWT). Therefore, in this paper, we used Taihu Lake as an example with data collected in 2017 that described both physical and virtual water use. We used these data to evaluate water rights allocation schemes by coupling virtual and physical water use. In order to achieve this goal, we first determined the physical water rights allocated for the four regions connected to the Basin. Next, we employed the multi-regional input–output (MRIO) approach to calculate the VWT among the four regions; then, we converted the VWT to the riparian level via the water efficiency coefficient. Finally, with virtual water included in the physical water rights allocation, we formulated a final water rights allocation for Taihu Lake. The results showed the following findings: (1) The ranking of the amount of physical water rights allocation is: Jiangsu > Zhejiang > Shanghai > Anhui. (2) Anhui and Jiangsu produce a net export of virtual water (2.259 billion m³ and 1.78 billion m³, respectively), while Zhejiang and Shanghai have a net import of virtual water (2.344 billion m³ and 1.695 billion m³, respectively), indicating that Anhui houses more water-consuming industries and is in greater need of economic restructuring. (3) The integration of virtual water makes a difference: Jiangsu achieved 16.208 billion m³ in terms of the amount of water rights allocated, Zhejiang achieved 6.606 billion m³, Shanghai achieved 3.040 billion m³, and Anhui achieved 4.319 billion m³, with a ranking of Jiangsu > Zhejiang > Anhui > Shanghai. The results detailed above prove that virtual water exerts an indispensable influence, and integrating virtual water can make the physical water rights allocation of transboundary rivers more equal and reasonable.