**5. System dynamics for groundwater management**

Possibly the first record of application of system dynamics to groundwater is found in Hong-gang [54]. Although the author does not develop formally a simulation model, provides an interesting analysis of the main feedback loops that characterize groundwater, which counteract each other. The first one is characterized by the lag that exists between water extraction and the population perception of the depletion, which is reinforced with government subsidies. The second is distinguished by the closure of water wells because of water quality deterioration. Both trigger more wells drilling to obtain more water in the short term, depleting aquifers' storage. Now, one of the first dynamic simulation models applied to aquifers is proposed in Assaf [55], which poses two modules: the physical and the economic. The former includes extractions, natural and artificial recharge, and geologic features of the aquifer, and the latter consists of productivity and the net present value of water resources. Therefore, although theoretical developments are traced back to 2001, it was until 2009 that simulation models started to thrive. Under this consideration, a systematic review was conducted.

The search was made on December 14th, 2021. The question to search was: What are the recent advances in the application of system dynamics to groundwater? One hundred and ninety seven records in Scopus were identified using the search string TITLE-ABS-KEY (system PRE/0 dynamics) AND TITLE-ABS-KEY (groundwater OR aquifer) and excluding the areas of biochemistry, business, physics, chemistry, materials science, and medicine. After reading the titles of the records, those who were not related to the question were excluded; this was the first exclusion criteria. The next exclusion criteria were the relevance of the abstract to the research question, after which 40 records were not retrieved. Next, 68 records were sought for retrieval during the entire timeframe 2001–2021. Twenty-two of these were published between 2017 and 2021 but three were not related to the research question. Hence, 19 studies were included in this systematic review, which are reported in **Table 2**.

It is interesting to note that 11 of these 19 studies were conducted in Iran. In addition, there are three studies in China, two in the Mediterranean region, one in the United States, one in Brazil, and one in Morocco. These regions are either arid zones or regions subjected to high water stress, therefore they rely entirely or partially on groundwater resources for their freshwater supply. In that sense, it is worth underscoring that groundwater is frequently modeled encompassed by the watershed or water system, which is aligned with the principles of IWRM. It is important to note that the most comprehensive study found is reported in Jiang et al. [67]: it is structured in four subsystems (socioeconomic, natural resources, water, and ecoenvironmental) and includes policies and impacts.

The classification purposed for the reviewed articles is strategic, tactical, and descriptive. They are distinguished by their time horizon and the objective they pursue. The first two look to project scenarios into the future, but with a different time horizon and objective: the former refers to strategic actions that look to increase groundwater sustainability, while the latter emphasizes short-term actions operation or management of infrastructure or irrigation. From the former kind, it is worth emphasizing that in Benabderrazik et al. [65] and Gómez Martín et al. [66], there was an involvement of stakeholders during the modeling process, which simultaneously validates the model and improves the understanding of the system's behavior. Regarding the latter case, in Barati et al. [70] the objective is to optimize cropping patterns, while in Silva and Teixeira [77] propose actions to improve institutional relationships in the context of water governance. Next, in Barati et al. [74], although


#### **Table 2.**

*Records from systematic review.*

not in a predictive fashion, the authors focus on the management of irrigation infrastructure. Finally, the third classification looks to describe the past behavior of the system and identify which policies modified the indicators of the system.

Another element worth comparing between the studies is the time resolution. The time resolution is the time-step considered during the performance of the simulation. In effect, eight of the studies used a yearly time-step, while six used a monthly time step, four did not state this information nor it is visible on the plots, and one uses a time step of 0.0625 months (1.875 days). The time-step is relevant, especially in those regions where there is intra-annual variability, to make more accurate decisions based on the simulations.

There is no doubt that the most pressing impact of climate change is felt through water resources, but the specific effects must be assessed in local studies. In effect, four studies considered climate change scenarios in their analysis. Interestingly, in Afruzi et al. [64] it was reported that, in the aquifer of Hamedan-Bahar watershed in Iran, current water management practices pose a more serious threat to groundwater resources sustainability than climate change. Also, in the Yangtze Economic Belt, climate change impacts in the economy, energy and food systems are negligible as long as the scenario is not RCP 8.5 [67]. Conversely, in the Hashtgerd Plain, Iran, climate change will intensify the water shortage [72] and, in Medina del Campo, Spain, it can turn Nature-Based Solutions insufficient in a worst-case climate change scenario. These effects were assessed through climate change effects in precipitation and temperature, but other effects must be also considered like soil degradation, plant pathogens, pests, increase of ground-level ozone [66], land yield, and extreme hydrometeorological disasters [67]. Considering these findings, studies of the impact of climate change in groundwater resources should be encouraged to include further effects but also to observe which are the effects in other regions of the world. For that matter, the studies presented in this review prove that system dynamics is a useful approach.

There is presence of coupling of system dynamics with other techniques like Compromise Programming, where it was used to rank strategies based on indicators [68]; Nash bargaining method, a quantitative model for conflict resolution in water allocation [69]; System of Environmental and Economic Accounts for Water (SEEA-Water) [75], to diagnose local water security indicators; CA-Marko model, to simulate the prediction of land use [71]; optimization [72, 73] and Agent-based model, to simulate the interaction between stakeholders and agencies [77]. In the context of groundwater management, there is a diversity of combination methods that have been used in the last 5 years, to complement system dynamics modeling, something that was suggested in Refs. [46, 54, 55].
