*4.2. The Ecosystem Aspect*

#### 4.2.1. Ecosystem Resilience

Disturbances such as urbanization and climate change affect the GSI facility as well as the entire ecosystem where GSI operates. As previously discussed, GSI is influenced by the interaction of ecological and hydrological processes in the ecosystem of the GSI planning framework; therefore, it is necessary to consider the ecosystem aspect, to which the concept of ecosystem resilience greatly contributes. The concept of resilience was introduced into the field of ecology in 1973 by Holling [108], who defined it as "a measure of the persistence of systems and of their ability to absorb change and disturbance and still maintain the same relationships between populations or state variables". Since then, many scholars have been devoted to clarifying the definition of resilience [109–111]. A clear formulation and application of ecological resilience can provide a basis for improving the ability of an ecosystem to cope with stressors and disturbances and help them tide over the reorganization period [37]. Systems with high resilience can absorb changes and maintain the same state in a series of disturbances and management actions [112]; these systems may possess favorable environmental conditions, strong multi-scale feedback, and a high level of diversity and redundancy [37]. As a comparison, systems with low resilience may react strongly to disturbances and move to another state [112], and these systems may contain poor environmental conditions, insufficiency of species or functional groups, and disturbances exceeding the range of historical changes [37]. Managers coping with the latter type of systems face the greatest challenge because they usually need to manage the systems actively. After management behavior improves the conditions, systems with high adaptability may be reorganized and restored to their original state [37]. Therefore, regarding the ecosystem aspect, if resilience is high enough, the ecosystem can absorb disturbances and return to the state before the disturbances, and the processes in the system can still operate normally; that is, ecological and hydrological processes can interact continuously and steadily. As a result, the management and promotion of ecosystem resilience is the guarantee for the stable operation of GSI. Therefore, the different purposes of facility and ecosystem aspects are obvious—the former is to harvest various ES from facilities directly, while the latter is to ensure that hydrological and ecological processes can maintain stable interaction in response to disturbances to indirectly support the continuous provision of ES.

Thus far, most studies have focused on theories, definitions, and conceptualizations to understand ecological resilience, focusing on the response of species diversity and functions to stress and disturbance on a small (i.e., local) scale [37]. In fact, integrating the concept of ecosystem resilience with landscape patterns provides an approach to understanding how ecosystem attributes and processes interact with landscape structure to affect ecosystem responses to disturbances and how the ecosystem supports resources, habitats, and species [37]. In the context of landscape, this integration provides a way of understanding the aforementioned processes within the ecosystem. Resilience-based management uses a spatially clear approach and contributes to selecting the type of management action that is most likely to succeed [37]. Ahern [113] explained the relationship of landscape composition, structure, and dynamics with resilience, and pointed out that a reasonable landscape pattern plays a significant role in buffering risks and helps the system to recover from disturbances. Therefore, landscape pattern optimization based on the interaction of spatial pattern and ecological processes can be an approach to the management and promotion of ecosystem resilience, as it provides a "spatial language" for concretely describing the interrelationship and the dynamics of spatiotemporal scales between landscape structure and function. Landscape patterns can be summarized as the shape, proportion, and spatial distribution of landscape elements. Patches, corridors, and matrices are the basic elements of landscape patterns, which are related to ecological processes in the landscape and affect the distribution and layout of resources and the physical environment. Landscape pattern optimization is essential to adjust the spatial structure of the landscape, with the goal of enhancing the integrity and connectivity of the

ecosystem and building a spatial pattern that maintains the regional ecological processes. The "patches–corridor–matrix" theory in landscape ecology is often used in landscape pattern optimization to identify and reorganize the key components of landscape patterns (i.e., patches, corridors, nodes, and matrices), which guide the protection and restoration of these components that are vital to the provision of expected ESs. The optimization process helps improve the integrity, connectivity, and diversity of landscape patterns, build ecological networks, enhance regional ecosystem resilience, protect or restore biodiversity, and sustainably provide multiple ESs [114]. Fu et al. [114] used InVEST software to evaluate two typical ecosystem services (water production and habitat quality) to identify ecological sources, and selected the minimum cumulative resistance model to identify ecological corridors; the landscape pattern was optimized by the improvement and reorganization of ecological sources to strengthen the material and energy flow between ecological sources and provide channels for species migration.

#### 4.2.2. Quantitative Assessment of Ecosystem Resilience

Obviously, a quantitative assessment of resilience is a way to visually express the results of resilience promotion management, but there is no single measurable variable that can represent ecosystem resilience. Relevant studies mostly evaluate urban resilience from the aspects of infrastructure, society, economy, and ecology [115]. Liu et al. [116] built an index system including diversity, connectivity, decentralization, and ecosystem service provision to assess the resilience of Shenyang, China, and established a link between resilience and landscape characteristics to guide the planning practice. Yi et al. [38] divided the existing quantitative assessment studies of resilience into three categories —forest resilience, soil microbial community resilience, and hydrological resilience; they found that many variables (e.g., tree-ring width, NDVI, microbiome mass, and catchment evapotranspiration index) can be used as indicators of system state variables, but it is hard to tell which one is better, as it depends on the objective and data availability in each study. They also pointed out that the measurement of resilience is not actually measuring itself, but its components, such as resistance, recovery, or combinations of them, i.e., elasticity. Dynamic system theory is a fundamental base of resilience research [38]. The uncertainty of resilience is based on the complexity of the nonlinear system, which contains many positive and negative feedback loops. Resistance, recovery, and resilience are the results of competition and cooperation between these feedback loops. Therefore, policymakers should understand the feedback structure of the nonlinear social–ecological system, and manage the related feedback loops to reduce disturbances or accelerate recovery, or to prevent the system from entering a new stable steady state [38]. In the future, it is necessary to grasp how to manage the dominant feedback to avoid catastrophic disasters.

As a matter of fact, the management of ecosystem resilience can bring many benefits, not just in terms of the ESs discussed in this review, but also the benefits of resources, habitats, and populations, etc. [37]. We focus on the effect of resilience promotion on the interaction between ecological and hydrological processes within the ecosystem. Future GSI planning should integrate facility and ecosystem aspects to explore the specific quantitative benefits of ecosystem resilience to the GSI system, and take more types of ESs into consideration. We recommend managing ecosystem resilience to ensure the stable operation of the GSI system through landscape pattern optimization. Exploring the interaction between the two aspects is also a point that needs to be considered in the future, for example, whether an ecosystem with high resilience can reduce GSI investment. We hold the assertion that the consideration of the two aspects in the GSI planning framework is equally crucial, and planners should strive to take into account both facility and ecosystem aspects when developing an overall understanding of the conditions in the study areas, so as to obtain ES in a comprehensive and sustainable manner. The quantitative assessment of ecosystem resilience also deserves more in-depth discussions in the future. A feasible solution is to select the indicators that can characterize the landscape pattern in view of the affirmation of landscape pattern optimization as an approach to enhancing ecological resilience, then

to determine the weights of selected indicators through appropriate methods, finally to form an evaluation system following the interaction mechanism of the eco-hydrological processes. In addition, the effective cooperation among managers, planners, scholars, and stakeholders helps to develop resilience-based management measures to strengthen and restore the ecosystem.

#### *4.3. Limitations*

This review included only 158 articles, although they help to identify the current research status of GSI planning regarding both facility and ecosystem aspects, and are the basis for us to predict future research directions, the papers that were not contained may contradict this review, which means the scope of follow-up studies needs to be expanded. We focused on water quality and water quantity regulation, but GSI practice has been extended to many areas of society, economy, and ecology. Therefore, it is inevitable for future research to explore more types of ESs in the social–ecological system, which is attached to greater complexity. In addition, the GSI planning framework we proposed may not be detailed enough, as some default planning steps were omitted. The objective formulation also includes steps such as collecting information on the conditions of the study areas; further, we did not describe the types of information required in detail, which may need to be discussed in forthcoming research.

In addition, most of the evaluation studies analyzing the performances of GSI were carried out in the laboratory or monitored conditions after the implementation of GSI planning. We do not deny the contribution of these evaluation articles to GSI and even agree with its positive effects. Although feedbacks can be given to GSI planning by evaluating its performances, the evaluation should be accomplished by the researcher through a long period of monitoring, and there are still many deficiencies in the pre-planning steps (i.e., objective formulation, type/scenario evaluation, quantity/scale determination, and site selection); therefore, the related articles regarding the performance evaluation of GSI were not considered as a planning step of the GSI planning framework in this review. It is worth mentioning that although we did not discuss the evaluation of GSI's performances in detail, we reviewed and quoted some relevant papers in the steps of objective formulation and type/scenario evaluation to support our arguments.

#### **5. Conclusions**

We developed a GSI planning framework that integrates the aspects of facility and ecosystems and made suggestions for future GSI planning to harvest stormwater management ESs through reviewing and synthesizing the literature. Regarding the facility aspect, quantitative and clear objectives are decisive for the entire GSI planning, since the social, economic, and environmental characteristics between study areas are discrepant; therefore, it is recommended that relevant authorities provide references to planning objectives that can vary with regional characteristics. It is foreseeable that these references should be scopes rather than a fixed value, thus allowing the actual planning to formulate clear objectives based on the trade-off of multiple anticipated ESs. Integrating indicators of multiple dimensions of social, economic, and environmental sustainability to evaluate GSI types/scenarios via MCDA and reinforce the sustainability of the GSI planning is the future research focus of type/scenario evaluation. Developing coupling models of hydrology and ecology to explore the quantitative relationship between the GSI type/scenario and the planning objective is the focus of future research; moreover, difficulty in difficulty in the determination of quantity/scale and finding the appropriate quantity/scale will receive much attention from stakeholders, which is helpful to the development of GSI. However, improving the completeness and accuracy of the coupling models will definitely increase the complexity in the meantime, and therefore, this trade-off needs to be considered in depth. A key factor in site selection is to evaluate the construction suitability of pixels in the study area based on the layer-cake theory, through which multiple considerations of social, economic, and environmental criteria should be covered.

In addition, the success of GSI planning is determined by the GSI facilities themselves as well as, the ecosystem, which also has a critical influence. Therefore, regarding the ecosystem aspect, in order to ensure that the ecosystem can withstand disturbances and still maintain the stable interaction of ecological and hydrological processes, and indirectly guarantee continuous ES production, we discussed the benefits of promoting ecosystem resilience. We suggested adopting landscape pattern optimization as an approach to resilience promotion, while it is necessary to consider more comprehensive and specific ways in the future. In addition, Future GSI planning should integrate facility and ecosystem aspects to explore the specific quantitative benefits of ecosystem resilience to the GSI system, and take more types of ESs into consideration. The quantitative assessment of ecosystem resilience also deserves more in-depth discussions. A feasible solution is to select the indicators that can characterize landscape patterns resulting from landscape pattern optimization as an approach to enhancing ecological resilience, then to determine the weights of selected indicators through appropriate methods, and finally, to form an evaluation system following the interaction mechanism of eco-hydrological processes. Exploring the interaction between these two aspects is also a point that needs to be considered in the future, for example. whether an ecosystem with high resilience can reduce GSI investment. We hold the assertion that the consideration of the two aspects in the GSI planning framework is equally crucial, and planners should strive to take into account both facility and ecosystem aspects when developing an overall understanding of the conditions in the studied areas. In addition, effective cooperation among managers, planners, scholars, and stakeholders helps to develop resilience-based management measures to strengthen and restore the ecosystem.

The most critical part that needs to be explored in detail urgently in this framework is the determination of quantity/scale. To advance the research in this area, we developed an indirect quantitative approach, where the relationship between GSI and ES is quantified precisely and operationally through a deep understanding, resulting from accurate simulations, of the interaction mechanism of ecological and hydrological processes. We encourage generating hydrological and ecological coupling simulations based on local social, economic, and environmental conditions in each planning scenario, then understanding the interaction mechanism between hydrological and ecological processes and identifying the interactions and changes in eco-hydrological processes caused by the ecological processes introduced by GSI. As a result, the number of ESs can be analyzed in accordance with these interactions and changes, as well as the quantitative relationship between GSI and ES that will instruct other steps of GSI planning. This approach is consistent with the spatiotemporal heterogeneity of the performance of GSI facilities. Furthermore, GSI planning using explicit data will be advantageous for its promotion, construction, and the reduction in the planners' and investors' concerns about selecting GSI as an alternative.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/su132413942/s1.

**Author Contributions:** Conceptualization, G.L.; methodology, G.L.; resources, G.L.; writing original draft preparation, G.L.; writing—review and editing, G.L. and L.W.; supervision, L.W.; project administration, L.W.; funding acquisition, L.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Key Research and Development Program of China (NO. 2018YFC0408000, 2018YFC0408004) and the Jinan Water Science and Technology Project (NO. JNSWKJ202103).

**Institutional Review Board Statement:** Not Applicable.

**Informed Consent Statement:** Not Applicable.

**Data Availability Statement:** Not Applicable.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
