基于迟滞排泄水箱模型模拟岩溶断流泉水文过程

传统的连续排泄水箱难以模拟岩溶泉断流现象，通过将表层岩溶带调蓄水箱对管道水箱的补给设置为迟滞排泄模式，模拟岩溶断流泉水文过程，并应用于丽江市黑龙潭岩溶泉域。研究结果表明，该模型较好地模拟了泉群流量动态及断流现象；当降雨量明显增大时，通过表层岩溶带调蓄水箱进入裂隙水箱的水量变幅不明显，而管道水箱的分配水量迅速增加，反映了集中补给的大气降水在岩溶系统中形成快速流的过程；黑龙潭泉群水量的绝大部分（82%~95%）来自于管道水箱，岩溶水流在时间和空间上分配的不均匀性易导致泉群断流。该研究提供了将水箱模型应用于岩溶断流泉模拟的参考，有助于理解该类型岩溶水系统的水文过程。

Modelling the hydrological process of the dried-up karst spring based on a reservoir model for hysteretic discharge

Karst is highly developed and widely distributed in Southwest China which is endowed with rich karst water resources but in different temporal and spatial distribution. As the main water source for the ecology and landscape of Lijiang Ancient City, Heilongtan springs have frequently dried up in recent years, which has seriously affected the production activities, daily life, and tourism quality of Lijiang Ancient City. The principle of reservoir model is to divide the karst aquifer system into different parts according to its structure or hydrological process. Based on the above division, each part has been generalized into a corresponding reservoir connected by a certain way in order to simulate the discharge of karst springs. The conventional reservoir model of continuous discharge has been widely used for reproducing the discharge of perennial springs. However, this continuous discharge model performs poorly in simulating the dry-up of karst springs. By setting the discharge law from epikarst reservoir to conduit reservoir as a hysteretic transfer function, this study reasonably reproduced the hydrological process of the dried-up springs. Several conclusions have been drawn as below. Firstly, Heilongtan spring area can be generalized as a reservoir model composed of epikarst regulation reservoir (E), matrix reservoir (M), and conduit reservoir (C). Discharge of Heilongtan springs has been successfully reproduced by the above model. Secondly, the rainy season and the normal season models can effectively simulate the spring dynamics based on the continuously discharge law. By setting the hysteretic discharge of the epikarst reservoir, the dry-up of the springs can be reproduced. To further characterize the multi-year regulation and storage characteristics of karst aquifers, additional reservoirs are needed. Thirdly, the simulation results show that the recharge from Reservoir E to Reservoir M is not sensitive to the rainfall, while the discharge from Reservoir E to Reservoir C is very sensitive to the rainfall, which indicates that the karst conduit in the Heilongtan spring area is developed and well connected, leading to the concentrated recharge and rapid increase of discharge at springs. Finally, the vast majority (82-95%) of Heilongtan spring water is recharged from Reservoir C. The threshold of rainfall required to trigger the rapid replenishment from Reservoir E to Reservoir C is relatively high. Moreover, the discharge from Reservoir C is characterized by sharp increase and decrease. The above characteristics cause uneven temporal and spatial distribution of karst groundwater, which can lead to the dry-up of springs. The research findings provide a reference for applying the hysteretic reservoir model to the simulation of karstic dried-up springs and help to understand the hydrological process of this type of karst system.