1. Introduction
The ecosystem services concept covers ecological, economic, and social aspects by emphasizing the values of nature for humans. Thus, this approach provides a suitable framework to address complex problems related to the sustainable use of resources that societies face today [1,2]. Ecosystem services can also be regarded as benefits derived from ecosystem functions, for which there is a demand. Thus, in this context, ecosystem services also embrace the goods derived from ecosystems [3]. Rapid urban expansion has turned natural ecosystems into semi-natural or artificial ecosystems [4,5,6], resulting in dramatic changes in both ecosystem structure and functions [7,8,9]. Life in cities needs increasing amounts of ecosystem services (ESs) in the form of food, drinking water, clean air, and recreation [10]. Simultaneously, urbanization is an increasingly important driver of land-use change, biodiversity loss, and ecosystem service deterioration [11,12,13,14]. The ESs therefore became vulnerable with the rapid alteration of ecological landscapes caused by urbanization [15,16,17]. In a context of generalized decline of ESs [18], putting in place governance systems that secure long-term delivery and use of ESs is one of the biggest challenges for cities in the 21st century [19,20,21].
In this context, urban planning can be associated with ecosystem services supply and demand: in supply, by estimating the potential of these services in terms of conservation, management, and deployment of green infrastructure, and in demand, through the organization of land use and cover, so that the demand matches the available supply [22].
In addition to this assessment based on supply and demand, ecosystem services can be approached through conceptual models that analyze them from the structural characteristics in the environment that ensure the production of these benefits [22,23]. This is because the capacity to provide ecosystem services can be, and is, also affected by the pressures acting on the ecosystem. These pressures are most often associated with the spatial distribution of the population and anthropic factors arising, in particular, from urbanization [22].
As these pressures vary, so does the ability to generate ecosystem services and the demand for them. The urbanization of a given region, for example, results in a drop in the supply of ecosystem services because of soil sealing, while at the same time, the demand for ecosystem services increases as the number of urban inhabitants grows [22,24].
Correctly understanding the correlation between supply and demand of ecosystem services is a very important condition to promote a sustainable management of such services, contributing significantly to improving human welfare [25,26]. However, in practice, choosing appropriate tools to address ecosystem services in planning is still difficult because they differ in their complexity, transferability, and time and data requirements [1]. For example, urban forests benefit cities with different ecosystem services, such as groundwater recharge, retention of particulate emitted by motors, surface runoff of rainwater, thermal comfort and local climate regulation, and carbon uptake [27].
In natural landscapes, but more especially in urban landscapes, the relationship and impact caused by measurable variables such as size, shape, and structure of natural habitat fragments on their ecological function are still unclear [28]. But the identification of spatial and temporal changes in the supply and demand of ecosystem services faces challenges; it is still necessary to incorporate relevant indicators that relate the supply and demand of ecosystem services, considering a spatial and temporal analysis [25].
Most of the operational tools that address the provision of ecosystem services present their outputs as spatial information or in a form easily connected to GIS software (ArcGIS, version 10.8.1, ESRI). This is very interesting, as it allows the location of hotspots of provision of these services, the analysis of synergies and trade-offs between them, or the spatial comparison of supply and demand, also revealing areas under pressure [2].
In view of this, the present study aimed to demonstrate the potential of provision of regulating ecosystem services by forest remnants in Campinas/SP, Brazil, using analysis and weighting of geospatial indicators, taking the premises of supply and demand into account.
2. Theoretical Background
Given this need to approach ecosystem services associated with spatial planning, especially urban planning, multicriteria decision analysis (MCDA) approaches have proven to be quite suitable [29,30,31,32,33,34]. This is because they allow the integration of ecological and socioeconomic aspects in planning related to changes in land use and land cover, in addition to allowing dealing with subjectivity and the different demands of stakeholders involved in the decision-making process [1].
It is worth remembering, however, that these indicators of the state of the ecosystem provide an incomplete picture of the overall level of service provision, especially in an urban setting where many services result from a combination of human and ecosystem inputs [22]. Although progress has been made in decision support systems (DSSs), choosing the appropriate tools to assess ecosystem services, aiming at a specific decision process, is still complicated because there are no clear guidelines for the implementation of these tools [2].
Recent times have seen progress in research about the mapping and evaluation of ecosystem services, but these studies have mostly focused on aspects of ecological change and economic valuation; the discussion has not included the perspective of beneficiaries, i.e., stakeholders and their perceptions. And these groups have different interests and preferences regarding ecosystem services [35].
Thinking about this growing demand to incorporate management factors in the valuation of ecosystem services, a study proposed an analytical framework to identify similarities and differences in social preferences and values associated with ecosystem services and recognized by key stakeholders [35]. It can be highlighted that there is a gap between scientific perception and human preference, and therefore, studies aimed at broadening and including knowledge about these services are needed [36].
As for the approach to ecosystem services integrated into the planning process, it is worth noting that the different types of ecosystem services currently have a patchy participation in these processes. What we do find is that in public policy sectors with long traditions in natural resource management, such as forestry, agriculture, and water, the ecosystem services approach is already fairly well established. These are also the sectors that involve the most usual operational tools in terms of ecosystem services valuation [2].
Therefore, it is now necessary to find new methods that take into account the unique characteristics and scale demands of the urban environment. They should also consider the diversity of stakeholders involved, their interests and ideas as a support tool for sustainable ecosystem management practices; this is essential, especially for guiding land-use policies, lessening land-use conflict, and promoting the building of an ecological civilization [28,35].
However, few studies have considered human needs, i.e., demand, in approaches to assessing ecosystem services [37]. This is quite inconsistent, according to the authors, since the very definition of ecosystem services was born from the search for human well-being and the focus on sustainable development [26]. In this regard, some new approaches have been established to structure the assessment of ecosystem services considering three concepts: (1) the capacity or potential to provide services (PS); (2) the flow of provision of these services (AS); and (3) the demand for such ecosystem services [22,25,35].
Nevertheless, studies show that there is an incompatibility between the supply and demand of ecosystem services, especially in urban, peri-urban, and expanding areas. And generally, areas with high demand for these services are those with the lowest supply, hence the imbalance [38]. Therefore, only through strategies that combine both supply optimization and demand reduction can the supply of these services be enhanced [37].
It can thus be noted that changes in the type of land use are among the main reasons for the decrease in the supply–demand relationship, since they directly affect the potential of ecosystem services and the flow of services. But more than that, especially when it comes to regulating ecosystem services, which are intrinsically related to demand, urban planning decisions significantly affect the provision of such services and how they are related to the benefits provided in the city environment [25].
As example, Cortinovis & Geneletti [22] considered seven regulating services (air purification, global climate regulation, moderation of extreme events, noise reduction, runoff and flood control, urban temperature regulation, and sewage treatment), and correlated them to common indicators in urban planning. Such indicators include population density, census data, locations, and presence of infrastructure. With this association, the researchers sought to understand what the demands of ecosystem services are, multiplying the intensity of ecological pressure and the amount of urban population or physical assets exposed [22].
This is because it is also important to consider that inequalities between the supply and demand of ecosystem services change over time, influenced by biotic factors such as richness and diversity, abiotic factors, such as geospatial factors, and by anthropic factors such as population growth, for example [38,39]. These are concepts that are usually associated with the spatiality of land-use and land-cover changes and can be used as tools for rural and urban development planning at both local and regional levels [22,25,35,38].
The actual magnitude of changes in the supply of ecosystem services and the relevance of considering these services in decision-making also depend on factors such as landscape configuration and geopolitical and management issues. In other words, considering these other factors would ensure a more precise and accurate spatial assessment of the supply of ecosystem services, as would considering that the benefits promoted by ecosystems depend not only on the supply of these services but also on demand, cultural values, and interest [40].
Furthermore, by focusing on the assessment of ecosystem services, based on the peculiar characteristics of each small basin or sub-basin and integrated with socioeconomic development planning, it will be possible to develop a much more effective ecosystem services management policy, since it will be based on a specific management model that is more appropriate for each landscape [41].
3. Materials and Methods
3.1. Study Area
The present study was conducted in the municipality of Campinas/SP, in the interior of the state of São Paulo (Figure 1), and comprises six basins. It covers a total area of 794.571 km2 and has a population of 1,223,237 inhabitants [42]. According to the mapping of Brazilian biomes, provided by the Ministry of the Environment [43], the municipality of Campinas is located in a transitional region between the Atlantic Forest and Cerrado biomes. However, the Cerrado biome in the municipality of Campinas is residual, occupying the extreme northwest of the Anhumas and Atibaia river basins, i.e., 8.8% and 5.6% of the area of these basins, respectively.
Table 1 shows the main features of the basins under study. The Atibaia basin is the largest and has the most remnants, with a total of 1368 (12.8% of its total area).
The Capivari basin is the second largest in terms of area and the one with the highest population density. The Anhumas and Capivari basins have a very similar percentage of area occupied by remnants, respectively, 5.9% and 5.7%, while the Quilombo basin has the lowest population and area occupied by forest remnants (2.5%). In total, 2319 forest remnants were mapped [44,45,46,47].
3.2. Available Landscape Metrics
For each of the 2319 identified remnants, a set of metrics was determined (Table 2), such as size of forest remnant, central area index, circularity index, shape index, distance from nearest neighbor, proximity to watercourse, water production, proximity to road network, land use and occupancy at the border, degree of erodibility of soil, declivity, distance to sports and leisure facilities, distance to a conservation unit, and distance to woods or parks.
The data have been previously mapped [44], except for the distance to sports and leisure facilities, distance to a conservation unit, and distance to woods or parks metrics, which were added to this study and calculated in a GIS environment based on geospatial data made available by the City Hall of Campinas [48].
These metrics represent a significant set of landscape indicators commonly used in studies on forest fragmentation, especially in tropical forests [45,47,48,49,50,51,52,53,54].
Table 2Calculated landscape metrics.
| Metrics | Definition |
|---|---|
| Remnant Size (AREA) | The size of a forest remnant can be classed as very small (<0.50 ha), small (0.50–1.00 ha), medium (1.00–5.00 ha), good (5.00–20.00 ha), and proper (>20.00 ha) [49]. |
| Central Area Index (CAI) | Percentage of the central area (core) of a forest reserve disregarding its marginal strip, in this study taken to be 60 m, subject to edge effects [50,51]. |
| Circularity Index (CI) | Relationship between perimeter (L in m) and area (A in m²) through the |
| Shape Index (SI) | The shape index indicates the degree of cutout in the shape of a remnant and is related to its area and perimeter. It was calculated according to the following equation, [51,54]: Pij higher value implies more areas that are irregular and/or smaller, while values closer to 1 indicate fragments more simply and, therefore, beneficial for conservation. |
| Nearest Neighbor Distance (NND) | Euclidean distance in meters calculated from the edge of one remnant to the edge of the nearest remnant [55]. This metric refers to the connectivity of the landscape, as long as, after a certain degree of isolation, the biological populations of the fragments begin to show losses in terms of biological flow [50,54]. |
| Proximity to Watercourses (PWC) and Water Production (WP) | Euclidean distance in meters calculated from a forest reserve to the nearest water course and presence of springs in the remnants. Assessed from the survey of hydrography and springs updated by the Secretariat of Green, Environment, and Sustainable Development of the Municipality of Campinas at a scale of 1:5000 in 2014 [56]. |
| Proximity to the Road Network (PRN) | The proximity between forest fragments and the road network is an important indicator because the greater the proximity to the road network the greater the environmental disturbance of forest areas. Thus, the distance between the forest remnants and the nearest road was evaluated; for this, it was used the mapping of streets of the municipality of Campinas/SP-Brazil [49,52]. |
| Land Use and Occupancy in the Surroundings (LUOS) | Identification of the class of greatest modification present in the area surrounding each fragment, within a radius of 175 m. To this end, we used the reclassification of another study [45] for the mapping of land use and land cover of UGRHI 5, where: Class 0 (unmodified landscape); Class 1 (little modification); Class 2 (medium modification); Class 3 (high modification); Class 4 (very high modification). |
| Soil Erodibility Degree (EROD) | Evaluation of the predominant soil typology in each forest remnant from the semi-detailed pedological map of the Municipality of Campinas and correlation with the corresponding erodibility degree, where: gley soils Háplicos (very weak); red-yellow latosols and yellow latosols (weak); heterogeneous cambisols (strong/very strong); red-yellow argosols (very strong) [57,58]. |
| Slope (SLP) | Both the soil type and the slope of the terrain have a major influence on the degree of susceptibility to soil erosion, so the greater the slope, the greater the susceptibility [57]. |
| Distance to Sports and Leisure Facilities (DSLF) | Euclidean distance in meters calculated from a forest remnant to sports and leisure facilities. |
| Distance to the Conservation Unit (PCU) | Euclidean distance in meters calculated from the forest remnant to the conservation unit. |
| Distance to Woods or Parks (PWP) | Euclidean distance in meters calculated from a forest remnant to woods or parks. |
3.3. Application of Artificial Neural Networks
3.3.1. Overview
The potential to supply regulating ecosystem services was assessed by applying an artificial neural network. These networks consist of a cognitive simulation technique, consolidated from a data processing structure distributed by numerous small, connected units. These are the neurons, the processing elements, and the synapses, which comprise the weight of connections that establishes the degree of connectivity between neurons.
The goal is to understand and imitate the properties associated with the parallelism and connectivity of biological systems to solve complex problems [59]. The effectiveness of this methodology lies precisely in its learning capacity, which enables the networks to improve their performance through the training process. This is an iterative process of readjusting the interconnection pattern, that is, the weights of the aforementioned synapses [59,60,61,62,63,64,65,66]. Artificial neural networks (ANNs) include supervised and unsupervised approaches [61]. For this study, a supervised learning mechanism was used. In supervised learning, a response variable is specified a priori. The user first labels and groups the system input variables and supplies the algorithm with the target output variable. The algorithm then finds a function that links the inputs with the outputs, such that it can then predict what the output will be from a given set of input variables [61]. The ultimate goal in this case is to identify pre-established classes of regulation service provision potential (very high, high, medium, low, or very low). The software used to run the neural network was Jupyter Notebook from the Anaconda browser and the network used ExtraTreesClassifier.
The ExtraTreesClassifier creates a group of unpruned decision trees in accordance with the traditional top–down method [60]. It essentially involves randomizing both attribute and cut-point selection strongly while splitting a node of a tree. In the extreme situation, it creates fully randomized trees that have structures independent of the output values of the training sample. It mainly differs from other tree-based ensemble methods on two counts, which are that it splits nodes by picking cut points fully at random and that it uses the whole training sample (instead of bootstrap replica) to grow the trees. The predictions of all the trees are combined to establish the final prediction by majority vote. The idea behind the extra-trees classifier is that the full randomization of the cut point and attribute together with ensemble averaging will decrease variance better than a weaker randomization strategy used by other methods. The usage of all of the original training samples instead of bootstrap replicas is to decrease bias. Computational efficiency is a major strength of this algorithm [64]. Like the other algorithms, the extra-trees algorithm has also seen an extensive and diverse application in the literature. Some of the recent applications include classification of land cover using extremely randomized trees [65], and a multi-layer intrusion detection system with extra-trees feature selection, extreme learning machine ensemble, and softmax aggregation [66].
The basic structure of the supervised neural network methodology used in this study consists basically of the following steps: (1) supervision, which covered the random selection of a sample of 10% of the forest remnants for the application of a manual methodology to classify the potential supply of regulation services; (2) neural network training and verification of effectiveness, based on training and test data; (3) application of the network to the remaining 90% of remnants, classification, and evaluation of final results, as presented in Figure 2.
3.3.2. Correlation between Supply Potential and Physical Condition
There are three possibilities for describing the relationship between the intensity of ecological pressures and the capacity to supply ecosystem services, especially when applied to regulating services. In the first, the service provisioning capacity increases as ecological pressure increases—for example, air purification and climate regulation, as measured by pollution removal and carbon stocks. The second arises when the capacity to provide decreases as ecological pressure increases, as in the case of surface runoff mitigation and evapotranspiration. And finally, it can be a relationship for which the increase in ecological pressure does not generate a change in the provisioning capacity of the ecosystem service—for example, for moderating certain extreme events and noise reduction [22].
Other regulating services, such as protection against erosion and regulation of air quality, for example, are only provided by vegetation in areas with specific conditions—in the first case, in areas with medium to high erosion risk, mainly in regions with an important share of arable land on slopes and relatively high precipitation rates; in the second case, the service is deliberately restricted to large urban areas. These are important findings that can be and are highly relevant spatial indicators [3,24].
The regulation and support services are predominantly associated with forest areas, which in turn have less potential for provisioning services [25,35]. On the other hand, for example, agricultural lands have a high capacity for provisioning services, but a low potential for regulating, supporting, and cultural services [25,35].
Given these findings, the following correlations between landscape metrics and the potential to provide regulating ecosystem services have been mostly accepted, as presented in Table 3, and they served as the basis for the construction of a supervised neural network.
3.3.3. Supervision
In this stage of supervision, 10% of the remnants of each basin were selected for evaluation of the potential for provision of regulating services, according to the methodology described below, namely, Anhumas (18), Atibaia (137), Capivari (32), Capivari-Mirim (9), Jaguari (32), and Quilombo (5). Based on these correlations, the classification and weighting of the metrics (Table 4) were carried out for each of the selected remnants [44].
The potential to provide ecosystem services was then assessed by the sum of these parameters, normalized, and therefore lying between 0 and 1: . Finally, the service provisioning potential was classified as follows: very low (0.00–0.20); low (0.20–0.40); medium (0.40–0.60); high (0.60–0.80); very high (0.80–1.00).
As pointed out by some researchers, the use of broad and highly aggregated metrics has an implication, since they can oversimplify the relationships between urban ecosystem function and landscape structure. Moreover, such a tool can serve to help the establishment of other indicators capable of addressing the correlation between planning and ecosystem services [22].
3.3.4. Network and Application Training
With the data duly supervised, the ExtraTreesClassifier neural network training and testing processes were performed. In this step, all data from each forest patch were loaded, including those referring to metrics that were not weighted in the supervised step. This is because, despite not having identified clear correlations between all metrics and the potential for provisioning services, by analyzing all data, the neural network is able to identify other correlations.
The accuracy of the process was confirmed and provided by the algorithm itself. From the learning generated by the network, the data of the remaining forest remnants in each basin were loaded so that the neural network could predict the potential degree of provision of regulating ecosystem services.
4. Results and Discussion
4.1. Neural Network Supervision and Training
The result of applying the methodology in the supervision stage is presented in Table 5. None of the remnants studied in this stage exhibited a classification of the potential to provide regulating service as very low. In all basins, the biggest share was for the “medium potential” and “high potential” classes.
This is quite consistent when analyzing the general context and features of these hydrographic basins that have, above all, more urban characteristics. Especially in places that face problems with increasing urbanisation and disorganised land use and occupation, such as Latin American countries, the “forest remnants in urban areas are perhaps the last refuges for biodiversity protection and conservation” [67]. In fact, since the last five decades, the population living in the Municipality of Campinas has more than tripled (from around 375,000 inhabitants to 1,223,237 inhabitants), with around 99% currently living in urban areas [42]. This historic trend in population increase triggered significant land-use/cover changes, affecting native vegetation cover. The unsustainable use of natural resources and the loss of biodiversity caused by the conversion of native ecosystems into urban areas threatens essential processes for human well-being [68,69].
Joly et al. [70] highlight the nature and the history of agriculture in Brazil, which has been the main agent of change in land use, and the impact on ecosystem services and biodiversity. It is worth noting that this does not only occur in Latin America but in other regions of the globe, too, where a process of urbanization and increased overcrowding of people in cities is observed, which simply drives a growing demand for natural resources such as food, timber, and space for housing, provided by these natural areas [24].
These changes affect not only regulatory ecosystem services, such as flood control, water quality, and disease control, but also cultural services related to aesthetic, spiritual, religious, educational, ecotourism, and recreation values [70,71]. As an example, a growing number of studies have shown a wide range of health benefits for people as a result of contact with nature—not only physical health benefits, but also benefits for psychological, spiritual, social, and environmental well-being [71].
4.2. Application of the Neural Network
Applying the trained neural network to the remaining data of the forest remnants yielded the following results, as presented in Table 6. Figure 3 represents the correlation between the results obtained by the supervision method for sample remnants and the results arrived at by applying the neural network to the rest of the total remnants. The graph shows a great similarity between the results. In general, the largest portion of the remnants was included in the class with high potential for providing regulating services, followed by the classes “medium potential” and “very high potential”.
For the results obtained through the neural network, there were no remnants classed as having “low” or “very low” regulating potential for the provision of ecosystem services. Figure 4 presents the spatialization of the forest remnants in the hydrographic basins of Campinas and their classification regarding the potential to supply regulating services.
This result could be an interesting guide to urban environmental planning and management, especially when considering that, as already discussed above, for there to be a potential supply of each of these services there must be a demand that, in these cases, is manifested as an adverse condition which the remnant in its current state can mitigate, i.e., regulate. For example, providing a regulating service for erosion prevention only has high potential when there is a strong possibility of erosion, i.e., when there is demand for such a service [3,22,24]. Thus, the high potential for providing regulating services in all hydrographic basins indicates a significant need for regulating natural ecosystem conditions in these basins.
However, it is also worth remembering that there are numerous kinds of services within the category “regulation”, which were not included in the study, such as regulation of climate and local air quality, carbon sequestration and storage, moderation of extreme weather events, wastewater treatment, erosion prevention and maintenance of soil fertility, and pollination and biological control [20]. On the other hand, capacity, demand, and flow of regulating ecosystem services and related benefits are linked to the main variables controlled by the spatial distribution and vulnerability profile of population and physical assets, as well as urban planning, i.e., the location, typology, and size of urban green infrastructure [22,23].
In this regard, Shi et al. [37], based on a quantitative assessment, found the ecosystem services in Shanghai city were in short supply on the whole. This situation was especially serious in the central urban areas, and the surplus of ecosystem services in some suburbs was not enough to compensate for the shortage in the central urban area. The results also show that the central urban area was the main “cold spot” area of ecosystem service supply, in association with the distribution of high-density buildings, which eroded the ecological space, and that the high-density population and high-intensity human activities weakened the supply of ecosystem services.
The study carried out by González-García [38] found that the demand for the regulating service of carbon sequestration and storage was mainly concentrated in the central region of the study area, which corresponds to the essentially urban area of the city of Madrid and areas of high urban expansion, so there is clearly an urban–rural gradient for the demand for carbon sequestration and storage service. It is known that the central urban area was the main “cold spot” in ecosystem services supply because of its high-density characteristics.
As highlighted above, there is generally an incompatibility between supply and demand. This means that areas with the capacity to provide ecosystem services have an even higher potential when the spatial region in which they are located is considered to have a high demand. High potential can then be translated as supply capacity added to high demand. It is emphasized that rather than analyzing ecosystem services statically, it is important to consider that the demand–supply–provision interaction of these services is a dynamic process with a spatial starting point (area of supply of ecosystem services) and an end point (area receiving ecosystem services) [37]. Significant progress in research such as this is therefore essential, since it seeks to identify the relationships between the demand for services and the ecosystem and geospatial characteristics that enable the provision of these services. This is a crucial factor to drive the optimization and management of the provision of these services [39].
Analyzing ecosystem services at a local scale of basins and sub-basins is a practice that should be further encouraged, as it sheds light on important information about the interactions between ecosystem services and local characteristics, helping managers to devise policies targeted to that landscape and condition [41]. While it is a relatively straightforward and practical way to categorize the remaining natural spaces into classes, especially urban green areas, when it comes to establishing an ecological understanding of them, it is important to consider studying them individually as habitats, not just the matrix that contains them [28].
The spatial planning processes of cities need better assimilation of the economic, social, and ecological aspects, including the discussion about ecosystem services, in order to raise awareness about the limited resources of nature that otherwise would not be recognized by stakeholders [1]. It must be stressed that the importance of an ecosystem service also depends on its context, so that the priority ecosystem services demanded in a given area have to be identified, so as to lead to effective planning and decision-making consistent with the demands and goals [22].
It should also be remembered that tools for assessing ecosystem services need to be adapted to meet the conditions of limited data availability and technical capacity, particularly in developing countries. In addition, new methods also need to be developed to incorporate ecosystem services into decision-support tools, as these have been poorly explored. When these tools are designed to support decision-making at local and regional scales, ecosystem services end up being better handled [2].
It is an essential role of urban planning to determine which priority ecosystem services are in demand, the current availability of remaining green spaces, and how and in which spatial distribution they can contribute to the provision of such services in an efficient and affordable manner [72,73].
5. Conclusions
The present study allows us to conclude that the regulating ecosystem services have a peculiar characteristic: they are inseparable from the issues of demand. Therefore, to assess the potential provision of regulating services, there should be indicators of the conditions that demand regulation, i.e., indicators associated with erosion, adverse weather conditions, and the need for water infiltration into the soil, among others.
On the other hand, landscape metrics have been demonstrated to be very interesting tools with great potential to become key in assessing the potential for providing regulating services. This is because they allow the assessment of geospatial conditions intrinsically related to the main regulation demands in an ecosystem: water filtration, erosion protection, air purification, etc. Working with spatial analyses on a municipal scale generates an extensive amount of data; therefore, the use of artificial intelligence fundamentals based on artificial neural networks has proven to be quite effective, as it enables the combined analysis of several indicators, pattern analysis, and prediction of results from databases. Assessing the potential of ecosystem services at a municipal and hydrographic basin scale, especially from spatial analyses, seems to be an effective tool for municipal authorities, as it provides useful guidelines to inform urban planning and municipal environmental management. The application of neural networks to the evaluation of the potential supply of ecosystem services presented quite consistent results; however, further studies are still needed in this area to corroborate the effectiveness of this approach in other spatial contexts. Thus, future research needs to be performed considering other geographical contexts and ecosystems services provision and demand. Additionally, investigating the economic and social values associated with these services could add depth to this analysis, further aiding in urban planning and policy formulation.
Conceptualization, R.M.L., A.L.d.S., A.N.N. and A.I.R.; methodology, A.I.R., R.M.L. and A.L.d.S.; software, A.L.d.S.; validation, R.M.L., A.L.d.S., A.N.N. and A.I.R.; formal analysis, R.M.L., A.L.d.S., A.N.N. and A.I.R.; investigation, R.M.L., A.L.d.S., A.N.N. and A.I.R.; resources, R.M.L., A.L.d.S., A.N.N. and A.I.R.; data curation, A.L.d.S.; writing— R.M.L. and A.L.d.S.; writing—review and editing, D.d.M.C., F.C.S. and R.C.G.; visualization, D.d.M.C.; supervision, R.M.L., A.N.N. and A.I.R.; project administration, R.M.L., A.N.N. and A.I.R.; funding acquisition, R.M.L., A.N.N., D.d.M.C. and A.I.R. All authors have read and agreed to the published version of the manuscript.
Thanks to the Pontifical Catholic University of Campinas for providing the necessary infrastructure to carry out this study.
The authors declare no conflict of interest.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Enlarge this image.Figure 3. Percentage comparison between the results of the supervised classification methodology and the neural network.
Enlarge this image.Figure 4. Spatial distribution of the forest remnants in the hydrographic basins of Campinas and their potential to supply regulating services.
Main characteristics of the basins under study.
| Hydrographic Basin | Total Area (ha) | Population Density |
Altitude Range (m) | Number of |
Total Area Covered |
|
|---|---|---|---|---|---|---|
| Anhumas | 14507.97 | 2476 | 186.3 | 176 | 865.24 | 5.9 |
| Atibaia | 25782.70 | 1546 | 508.7 | 1368 | 3298.2 | 12.8 |
| Capivari | 21820.23 | 3776.8 | 209 | 323 | 1241.26 | 5.7 |
| Capivari-Mirim | 5544.46 | 1663.54 | 125.8 | 75 | 434.34 | 7.8 |
| Jaguari | 4554.03 | ** | 422.6 | 324 | 610.34 | 13.4 |
| Quilombo | 7325.3 | 2271 | 149.6 | 53 | 179.60 | 2.5 |
** Without population.
Likely correlation between landscape metrics and the potential to provide regulating ecosystem services.
| Correlation | Potentialized by | Equivalent to | |
|---|---|---|---|
| AREA | Positive | Large area | High potential |
| PWC | Negative | Shortest distance | High demand |
| PRN | Negative | Shortest distance | High demand |
| LUSE | Positive | Most anthropized areas | High demand |
| EROD | Positive | High erodibility | High demand |
| SLP | Positive | High slope | High demand |
| PCU | Negative | Shortest distance | High potential |
Where: AREA (total area); PWC (proximity to water course); PRN (proximity to road network); LUSE (land use and occupation of the land at the edge); EROD (degree of soil erodibility); SLP (slope); PCU (distance to the conservation unit).
Weighting of landscape metrics according to classification for potential to provide regulating services.
| Points | |||||||
|---|---|---|---|---|---|---|---|
| INDICATOR | 1 | 3 | 5 | 8 | 10 | ||
| AREA (ha) | <0.50 | 0.50–1.00 | 1.00–5.00 | 5.00–20.00 | >20.00 | ||
| PWC (m) | - | >200 m | 120–200 m | 60–120 m | <60 m | ||
| PRN (m) | - | >200 m | 120–200 m | 60–120 m | <60 m | ||
| LUSE | Class 0 | Class 1 | Class 2 | Class 3 | Class 4 | ||
| PCU (m) | ≥5000 m | 2000–5000 m | 700–2000 m | <700 m | |||
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | |
| EROD | Very weak | Weak/Very weak | Weak | Medium | Strong | Very strong/Strong | Very strong |
| SLP | 0–3% | 3–8% | 8–20% | 20–45% | 45–75% | >75% | |
Where: AREA (total area); PWC (proximity to water course); PRN (proximity to road network); LUSE (land use and occupation of the land at the edge); EROD (degree of soil erodibility); SLP (slope).
Potential to provide regulating services–supervised classification.
| Supervised |
Anhumas | Atibaia | Capivari | Capivari- |
Jaguari | Quilombo | |
|---|---|---|---|---|---|---|---|
| 0.00–0.20 | Very low | - | - | - | - | - | |
| 0.20–0.40 | Low | 2 (11.1%) | 1 (0.7%) | 1 (3.1%) | - | 2 (40.0%) | |
| 0.40–0.60 | Medium | 8 (44.4%) | 27 (19.7%) | 12 (37.5%) | 5 (55.6%) | 14 (43,8%) | 3 (60.0%) |
| 0.60–0.80 | High | 8 (44.4%) | 76 (55.5%) | 16 (50.0%) | 4 (44.4%) | 13 (40,6%) | - |
| 0.80–1.00 | Very high | - | 33 (24.1%) | 3 (9.4%) | - | 5 (15,6%) | - |
Neural network classification results for the basins under study.
| Neural Network Classification | Anhumas | Atibaia | Capivari | Capivari-Mirim | Jaguari | Quilombo |
|---|---|---|---|---|---|---|
| Very low | - | - | - | - | - | - |
| Low | - | - | 1 (0.3%) | - | - | - |
| Medium | 74 (46.8%) | 208 (16.9%) | 93 (32.0%) | 39 (59.1%) | 95 (32.5%) | 36 (75.0%) |
| High | 81 (51.3%) | 803 (65.2%) | 179 (61.5%) | 26 (39.4%) | 171 (58.6%) | 12 (25.0%) |
| Very high | 3 (1.9%) | 220 (17.9%) | 18 (6.2%) | 1 (1.5%) | 26 (8.9%) | - |
References
1. Grêt-Regamey, A.; Altwegg, J.; Sirén, E.A.; van Strien, M.J.; Weibel, B.; Grêt-Regamey, A.; Altwegg, J.; Sirén, E.A.; van Strien, M.J.; Weibel, B. Integrating ecosystem services into spatial planning—A spatial decision support tool. Landsc. Urban Plan.; 2017; 165, pp. 206-219. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2016.05.003]
2. Grêt-Regamey, A.; Sirén, E.; Brunner, S.H.; Weibel, B. Review of decision support tools to operationalize the ecosystem services concept. Ecosyst. Serv.; 2017; 26, pp. 306-315. [DOI: https://dx.doi.org/10.1016/j.ecoser.2016.10.012]
3. Maes, J.; Liquete, C.; Teller, A.; Erhard, M.; Paracchini, M.L.; Barredo, J.I.; Grizzetti, B.; Cardoso, A.; Somma, F.; Petersen, J.-E. et al. An indicator framework for assessing ecosystem services in support of the EU Biodiversity Strategy to 2020. Ecosyst. Serv.; 2016; 17, pp. 14-23. [DOI: https://dx.doi.org/10.1016/j.ecoser.2015.10.023]
4. Huang, G.; London, J. Mapping cumulative environmental efects, social vulnerability, and health in the San Joaquin Valley, California. Am. J. Public Health; 2012; 102, pp. 830-832. [DOI: https://dx.doi.org/10.2105/AJPH.2011.300466]
5. Qiu, B.; Li, H.; Zhou, M.; Zhang, L. Vulnerability of ecosystem services provisioning to urbanization: A case of China. Ecol. Indic.; 2015; 57, pp. 505-513. [DOI: https://dx.doi.org/10.1016/j.ecolind.2015.04.025]
6. Kehoe, L.; Romero-Muñoz, A.; Polaina, E.; Estes, L.; Kreft, H.; Kuemmerle, T. Biodiversity at risk under future cropland expansion and intensification. Nat. Ecol. Evol.; 2017; 1, pp. 1129-1135. [DOI: https://dx.doi.org/10.1038/s41559-017-0234-3]
7. Grimm, N.B.; Foster, D.; Groffman, P.; Grove, J.M.; Hopkinson, C.S.; Nadelhoffer, K.J.; Pataki, D.E.; Peters, D.P. The changing landscape: Ecosystem responses to urbanization and pollution across climatic and societal gradients. Front. Ecol. Environ.; 2008; 6, pp. 264-272. [DOI: https://dx.doi.org/10.1890/070147]
8. Eigenbrod, F.; Bell, V.A.; Davies, H.N.; Heinemeyer, A.; Armsworth, P.R.; Gaston, K.J. The impact of projected increases in urbanization on ecosystem services. Proc. R. Soc. B Biol. Sci.; 2011; 278, pp. 3201-3208. [DOI: https://dx.doi.org/10.1098/rspb.2010.2754]
9. Long, H.; Liu, Y.; Hou, X.; Li, T.; Li, Y. Effects of land use transitions due to rapid urbanization on ecosystem services: Implications for urban planning in the new developing area of China. Habitat Int.; 2014; 44, pp. 536-544. [DOI: https://dx.doi.org/10.1016/j.habitatint.2014.10.011]
10. Guo, Z.; Zhang, L.; Li, Y. Increased dependence of humans on ecosystem services and biodiversity. PLoS ONE; 2010; 5, e13113. [DOI: https://dx.doi.org/10.1371/journal.pone.0013113]
11. Mcdonald, R.; Marcotullio, P. Global effects of urbanization on Ecosystem Services. Urban Ecology: Patterns, Processes, and Applications; Niemelä, J. Oxford University Press: Oxford, UK, 2011; pp. 89-239.
12. Li, Y.; Li, Y.; Zhou, Y.; Shi, Y.; Zhu, X. Investigation of a coupling model of coordination between urbanization and the environment. J. Environ. Manag.; 2012; 98, pp. 127-133. [DOI: https://dx.doi.org/10.1016/j.jenvman.2011.12.025] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22265813]
13. Seto, K.C.; Güneralp, B.; Hutyra, L.R. Global forecasts of urban expansion to 2030 and direct impacts on biodiversity and carbon pools. Proc. Natl. Acad. Sci. USA; 2012; 109, pp. 16083-16088. [DOI: https://dx.doi.org/10.1073/pnas.1211658109] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22988086]
14. Dai, L.; Wang, Z.; Guo, T.; Hu, L.; Chen, Y.; Chen, C.; Yu, G.; Ma, L.Q.; Chen, J. Pollution characteristics and source analysis of microplastics in the Qiantang River in southeastern China. Chemosphere; 2022; 293, 133576. [DOI: https://dx.doi.org/10.1016/j.chemosphere.2022.133576] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35016956]
15. Su, C.-H.; Fu, B.-J.; He, C.-S.; Lü, Y.-H. Variation of ecosystem services and human activities: A case study in the Yanhe Watershed of China. Acta Oecologica; 2012; 44, pp. 46-57. [DOI: https://dx.doi.org/10.1016/j.actao.2011.11.006]
16. Estoque, R.C.; Murayama, Y. Landscape pattern and ecosystem service value changes: Implications for environmental sustainability planning for the rapidly urbanizing summer capital of the Philippines. Landsc. Urban Plan.; 2013; 116, pp. 60-72. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2013.04.008]
17. Das, M.; Das, A.; Pereira, P.; Mandal, A. Exploring the spatio-temporal dynamics of ecosystem health: A study on a rapidly urbanizing metropolitan area of Lower Gangetic Plain, India. Ecol. Indic.; 2021; 125, 107584. [DOI: https://dx.doi.org/10.1016/j.ecolind.2021.107584]
18. MEA, Millennium Ecosystem Assessment. Ecosystems and Human Well-Being: Synthesis; World Resources Institute: Washington, DC, USA, Island Press: Washington, DC, USA, 2005.
19. Elmqvist, T.; Fragkias, M.; Goodness, J.; Güneralp, B.; Marcotullio, P.J.; McDonald, R.I.; Parnell, S.; Schewenius, M.; Sendstad, M.; Seto, K.C. et al. Stewardship of the biosphere in the urban era. Urbanization, Biodiversity and Ecosystem Services; Elmqvist, T. 2013; pp. 1-38. [DOI: https://dx.doi.org/10.1007/978-94-007-7088-1_11]
20. TEEB. The Economics of Ecosystems and Biodiversity. Manual for Cities: Ecosystem Services in Urban Management. 2011; Available online: http//www.teebweb.org (accessed on 15 October 2018).
21. Ronchi, S. Ecosystem Services and Planning. Ecosystem Services for Spatial Planning: Innovative Approaches and Challenges for Practical Applications; Green Energy and Technology Springer: Cham, Switzerland, 2018; pp. 1-26. [DOI: https://dx.doi.org/10.1007/978-3-319-90185-5_1]
22. Cortinovis, C.; Geneletti, D. A framework to explore the effects of urban planning decisions on regulating ecosystem services in cities. Ecosyst. Serv.; 2019; 38, e100946. [DOI: https://dx.doi.org/10.1016/j.ecoser.2019.100946]
23. Haines-Young, R.; Potschin, M. The links between biodiversity, ecosystem services and human well-being. Ecosystem Ecology: A New Synthesis; Raffaelli, D.; Frid, C. BES Ecological Reviews Series; CUP: Cambridge, UK, 2010.
24. Maes, J.; Barbosa, A.; Baranzelli, C.; Zulian, G.; e Silva, F.B.; Vandecasteele, I.; Hiederer, R.; Liquete, C.; Paracchini, M.L.; Mubareka, S. et al. More green infrastructure is required to maintain ecosystem services under current trends in land-use change in Europe. Landsc. Ecol.; 2015; 30, pp. 517-534. [DOI: https://dx.doi.org/10.1007/s10980-014-0083-2]
25. Guan, Q.; Hao, J.; Ren, G.; Li, M.; Chen, A.; Duan, W.; Chen, H. Ecological indexes for the analysis of the spatial–temporal characteristics of ecosystem service supply and demand: A case study of the major grain-producing regions in Quzhou, China. Ecol. Indic.; 2020; 108, e105748. [DOI: https://dx.doi.org/10.1016/j.ecolind.2019.105748]
26. Owuor, M.A.; Icely, J.; Newton, A.; Nyunja, J.; Otieno, P.; Tuda, A.O.; Oduor, N. Mapping of ecosystem services flow in Mida Creek, Kenya. Ocean Coast. Manag.; 2017; 140, pp. 11-21. [DOI: https://dx.doi.org/10.1016/j.ocecoaman.2017.02.013]
27. Ferreira, M.L.; De Souza, L.C.; Conti, D.D.M.; Quaresma, C.C.; Tavares, A.R.; da Silva, K.G.; Kniess, C.T.; De Camargo, P.B. Soil Biodiversity in Urban forests as a consequence of litterfall management: Implications for São Paulo’s ecosystem services. Sustainability; 2018; 10, 684. [DOI: https://dx.doi.org/10.3390/su10030684]
28. Grafius, D.R.; Corstanje, R.; Harris, J.A. Linking ecosystem services, urban form and green space configuration using multivariate landscape metric analysis. Landsc. Ecol.; 2018; 33, pp. 557-573. [DOI: https://dx.doi.org/10.1007/s10980-018-0618-z] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31258244]
29. Costanza, R. Nature: Ecosystems without commodifying them. Nature; 2006; 443, 749. [DOI: https://dx.doi.org/10.1038/443749b] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17051185]
30. Dendoncker, N.; Boeraeve, F.; Crouzat, E.; Dufrêne, M.; König, A.; Barnaud, C. How can integrated valuation of ecosystem services help understanding and steering agroecological transitions?. Ecol. Soc.; 2018; 23, [DOI: https://dx.doi.org/10.5751/ES-09843-230112]
31. Gómez-Baggethun, E.; Martín-López, B. Ecological economics perspective in ecosystem services valuation. Handbook of Ecological Economics; Martínez-Alier, J.; Muradian, R. Edward Elgar: Cheltenham, UK, 2015; pp. 260-282.
32. Paetzold, A.; Warren, P.H.; Maltby, L.L. A framework for assessing ecological quality based on ecosystem services. Ecol. Complex.; 2010; 7, pp. 273-281. [DOI: https://dx.doi.org/10.1016/j.ecocom.2009.11.003]
33. Saarikoski, H.; Jax, K.; Harrison, P.A.; Primmer, E.; Barton, D.N.; Mononen, L.; Vihervaara, P.; Furman, E. Exploring operational ecosystem service definitions: The case of boreal forests. Ecosyst. Serv.; 2015; 14, pp. 144-157. [DOI: https://dx.doi.org/10.1016/j.ecoser.2015.03.006]
34. Sijtsma, F.J.; Van Der Bilt, W.G.; Van Hinsberg, A.; De Knegt, B.; Van Der Heide, M.; Leneman, H.; Verburg, R. Planning nature in urbanized countries. An analysis of monetary and non-monetary impacts of conservation policy scenarios in the Netherlands. Heliyon; 2017; 3, e00280. [DOI: https://dx.doi.org/10.1016/j.heliyon.2017.e00280]
35. Zhang, W.; Yu, Y.; Wu, X.; Pereira, P.; Borja, M.E.L. Integrating preferences and social values for ecosystem services in local ecological management: A framework applied in Xiaojiang Basin Yunnan province, China. Land Use Policy; 2020; 91, e104339. [DOI: https://dx.doi.org/10.1016/j.landusepol.2019.104339]
36. Riegels, N.; Lynggaard-Jensen, A.; Jensen, J.K.; Gerner, N.V.; Anzaldua, G.; Mark, O.; Butts, M.; Birk, S. Making the ecosystem services approach operational: A case study application to the Aarhus River, Denmark. Sci. Total. Environ.; 2019; 707, 135836. [DOI: https://dx.doi.org/10.1016/j.scitotenv.2019.135836]
37. Shi, Y.; Shi, D.; Zhou, L.; Fang, R. Identification of ecosystem services supply and demand areas and simulation of ecosystem service flows in Shanghai. Ecol. Indic.; 2020; 115, 106418. [DOI: https://dx.doi.org/10.1016/j.ecolind.2020.106418]
38. González-García, A.; Palomo, I.; González, J.A.; López, C.A.; Montes, C. Quantifying spatial supply-demand mismatches in ecosystem services provides insights for land-use planning. Land Use Policy; 2020; 94, e104493. [DOI: https://dx.doi.org/10.1016/j.landusepol.2020.104493]
39. Harrison, P.; Berry, P.; Simpson, G.; Haslett, J.; Blicharska, M.; Bucur, M.; Dunford, R.; Egoh, B.; Garcia-Llorente, M.; Geamănă, N. et al. Linkages between biodiversity attributes and ecosystem services: A systematic review. Ecosyst. Serv.; 2014; 9, pp. 191-203. [DOI: https://dx.doi.org/10.1016/j.ecoser.2014.05.006]
40. Paruelo, J.M.; Texeira, M.; Staiano, L.; Mastrángelo, M.; Amdan, L.; Gallego, F. An integrative index of Ecosystem Services provision based on remotely sensed data. Ecol. Indic.; 2016; 71, pp. 145-154. [DOI: https://dx.doi.org/10.1016/j.ecolind.2016.06.054]
41. Yao, J.; He, X.; Chen, W.; Ye, Y.; Guo, R.; Yu, L. A local-scale spatial analysis of ecosystem services and ecosystem service bundles in the upper Hun River catchment, China. Ecosyst. Serv.; 2016; 22, pp. 104-110. [DOI: https://dx.doi.org/10.1016/j.ecoser.2016.09.022]
42. BRASIL. IBGE CIDADES. Campinas. Available online: https://cidades.ibge.gov.br/brasil/sp/campinas/panorama (accessed on 12 October 2021).
43. Brasil. Ministério do Meio Ambiente. Download de Dados Geográficos. Available online: http://mapas.mma.gov.br/i3geo/datadownload.htm (accessed on 15 May 2018).
44. Silva, A.L. Análise da Qualidade Ambiental de Remanescentes Florestais por Meio de Métricas de Paisagem: Um Estudo no Município de Campinas/SP; Dissertação (Programa de Pós-Graduação em Sistemas de Infraestrutura Urbana)—Pontifícia Universidade Católica de Campinas: Campinas, Brazil, 2020.
45. da Silva, A.L.; Longo, R.M. Ecologia da paisagem e qualidade ambiental de remanescentes florestais na sub-bacia hidrográfica do Rio Atibaia dentro do município de Campinas-SP. Ciência Florest.; 2020; 30, pp. 1176-1191. [DOI: https://dx.doi.org/10.5902/1980509842640]
46. da Silva, A.L.; Ribeiro, A.; Longo, R.M. Environmental Quality of forest remnants in the Capivari River Sub-basin (Campinas–Sao Paulo–Brazil). Soc. Nat.; 2021; 33, pp. 1-10. [DOI: https://dx.doi.org/10.14393/sn-v33-2021-59129]
47. da Silva, A.L.; Ribeiro, A.; Longo, R.M. Tools for prioritizing Ecosystem Services provided by fragments forest in the context of cities. Ambient. Soc.; 2021; 24, pp. 2-23. [DOI: https://dx.doi.org/10.1590/1809-4422asoc20200012r1vu2021l2ao]
48. Campinas. Portal Geoambiental. Available online: https://geoambiental.campinas.sp.gov.br/ (accessed on 15 July 2021).
49. Freitas, E.P. Análise Integrada do Mapa de uso e Ocupação das Terras da Microbacia do Rio Jundiaí-Mirim Para Fins de Gestão Ambiental; Dissertação (Mestrado em Agricultura Tropical e Subtropical), Instituto Agronômico de Campinas: Campinas, Brazil, 2012.
50. Massoli, J.V.; Statella, T.; Santos, V.S. Estimativa da fragmentação florestal na microbacia Sepotubinha, Nova Marilândia—MT, entre os anos de 1990 a 2014. Caminhos Geogr.; 2016; 17, pp. 48-60. [DOI: https://dx.doi.org/10.14393/RCG176004]
51. Calegari, L.; Martins, S.V.; Gleriani, J.M.; Silva, E.; Buzato, L.C. Analysis of the dynamics of forest fragments in the city of Carandaí, MG, for forest restoration. Rev. Árvore; 2010; 34, pp. 871-880. [DOI: https://dx.doi.org/10.1590/S0100-67622010000500012]
52. Fengler, F.H.; de Moraes, J.F.; Ribeiro, A.I.; Peche Filho, A.; Storino, M.; de Medeiros, G.A. Environmental quality of forest fragments in Jundiaí-Mirim river basin between 1972 and 2013. Rev. Bras. Eng. Agrícola E Ambient.; 2015; 19, pp. 402-408. [DOI: https://dx.doi.org/10.1590/1807-1929/agriambi.v19n4p402-408]
53. Etto, T.L.; Longo, R.M.; Arruda, D.D.; Invenioni, R. Ecologia da paisagem de remanescentes florestais na bacia hidrográfica do Ribeirão das Pedras—Campinas-SP. Rev. Árvore Viçosa; 2013; 37, pp. 1063-1071. [DOI: https://dx.doi.org/10.1590/S0100-67622013000600008]
54. Lima, B.C.; Francisco, C.N.; Bohrer, C.B. Landslides and forest fragmentation in the mountain range region of Rio de Janeiro state. Ciência Florest.; 2017; 27, pp. 1283-1295. [DOI: https://dx.doi.org/10.5902/1980509830321]
55. Fernandes, M.; Ameida, A.; Gonzaga, M.I.; Gonçalves, F. Landscape Ecology of a Watershed in the Brazilian Coastal Flat Lands. Floresta E Ambiente; 2017; 24, e00025015. [DOI: https://dx.doi.org/10.1590/2179-8087.025015]
56. Campinas. Prefeitura Municipal de Campinas. Mapeamento de Áreas Verdes do Município de Campinas. Available online: http://ambientecampinas.wix.com/mapeamento#! (accessed on 26 May 2018).
57. Salomão, F.X.T. Controle e Prevenção dos Processos Erosivos. Erosão e Conservação dos Solos: Conceitos, Temas e Aplicações; Guerra, A.J.T.; Silva, A.; Botelho, R.G.M. Rio de Janeiro: Bertrand, Brazil, 1999; 340p.
58. Ross, J.L.S. Geomorfologia: Ambiente e Planejamento. Contexto, Coleção Repensando a Geografia; 8th ed. Editora Contexto,: São Paulo, Brazil, 2005; 285p.
59. Furtado, M.I.V. Redes Neurais Artificiais: Uma Abordagem Para Sala de Aula. Ponta Grasso (PR): Atena Editora. 2019; Available online: https://www.atenaeditora.com.br/wp-content/uploads/2019/05/e-book-Redes-Neurais-Artificiais-uma-Abordagem-para-Sala-de-Aula.pdf (accessed on 23 February 2021).
60. Valente, R.D.O.A. Definição de Áreas Prioritárias Para Conservação e Preservação Florestal por Meio da Abordagem Multicriterial em Ambiente SIG. Ph.D. Thesis; University of São Paulo: São Paulo, Brazil, 2005.
61. Dias, T.L.; Cataldi, M.; Ferreira, V.H. Aplicação de técnicas de redes neurais e modelagem atmosférica para elaboração de previsões de vazão na Bacia do Rio Grande (MG). Eng. Sanit. Ambient.; 2016; 22, pp. 169-178. [DOI: https://dx.doi.org/10.1590/s1413-41522016158015]
62. Scowen, M.; Athanasiadis, I.N.; Bullock, J.M.; Eigenbrod, F.; Willcock, S. The current and future uses of machine learning in ecosystem service research. Sci. Total. Environ.; 2021; 799, 149263. [DOI: https://dx.doi.org/10.1016/j.scitotenv.2021.149263] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/34426354]
63. Ampomah, E.K.; Qin, Z.; Nyame, G. Evaluation of Tree-Based Ensemble Machine Learning Models in Predicting Stock Price Direction of Movement. Information; 2020; 11, 332. [DOI: https://dx.doi.org/10.3390/info11060332]
64. Geurts, P.; Ernst, D.; Wehenkel, L. Extremely randomized trees. Mach. Learn.; 2006; 63, pp. 3-42. [DOI: https://dx.doi.org/10.1007/s10994-006-6226-1]
65. Zafari, A.; Zurita-Milla, R.; Izquierdo-Verdiguier, E. Land Cover Classification Using Extremely Randomized Trees: A Kernel Perspective. IEEE Geosci. Remote. Sens. Lett.; 2019; 17, pp. 1702-1706. [DOI: https://dx.doi.org/10.1109/LGRS.2019.2953778]
66. Sharma, J.; Giri, C.; Granmo, O.-C.; Goodwin, M. Multi-layer intrusion detection system with ExtraTrees feature selection, extreme learning machine ensemble, and softmax aggregation. EURASIP J. Inf. Secur.; 2019; 2019, 15. [DOI: https://dx.doi.org/10.1186/s13635-019-0098-y]
67. Muñoz, A.M.M.; Freitas, S.R. Importância dos serviços ecossistêmicos nas cidades: Revisão da publicação de 2003 a 2015. Rev. Gestão Ambient. E Sustentabilidade; 2015; 6, pp. 89-104. [DOI: https://dx.doi.org/10.5585/geas.v6i2.853]
68. Marselle, M.R.; Lindley, S.J.; Cook, P.A.; Bonn, A. Biodiversity and Health in the Urban Environment. Curr. Environ. Health Rep.; 2021; 8, pp. 146-156. [DOI: https://dx.doi.org/10.1007/s40572-021-00313-9]
69. Joly, C.A.; Scarano, F.R.; Bustamante, M.; Gadda, T.M.C.; Metzger, J.P.W.; Seixas, C.S.; Ometto, J.P.H.; Pires, A.P.F.; Boesing, A.L.; Sousa, F.D.R. et al. Brazilian assessment on biodiversity and ecosystem services: Summary for policy makers. Biota Neotrop.; 2019; 19, e20190865. [DOI: https://dx.doi.org/10.1590/1676-0611-bn-2019-0865]
70. Ballerini, A.P.; De Campos, L.A.; Galhardi, A.C. Ecosystem services and their importance in the sustainable management of a natural heritage: The case of Serra do Japi. Braz. J. Dev.; 2021; 7, pp. 119522-119537. [DOI: https://dx.doi.org/10.34117/bjdv7n12-627]
71. Romagosa, F. Physical health in green spaces: Visitors’ perceptions and activities in protected areas around Barcelona. J. Outdoor Recreat. Tour.; 2018; 23, pp. 26-32. [DOI: https://dx.doi.org/10.1016/j.jort.2018.07.002]
72. Elliott, M.; Burdon, D.; Hemingway, K.L.; Apitz, S.E. Estuarine, coastal and marine ecosystem restoration: Confusing management and science–A revision of concepts. Estuar. Coast. Shelf Sci.; 2007; 74, pp. 349-366. [DOI: https://dx.doi.org/10.1016/j.ecss.2007.05.034]
73. Almeida, M.; Loupa-Ramos, I.; Menezes, H.; Carvalho-Ribeiro, S.; Guiomar, N.; Pinto-Correia, T. Urban population looking for rural landscapes: Different appreciation patterns identified in Southern Europe. Land Use Policy; 2016; 53, pp. 44-55. [DOI: https://dx.doi.org/10.1016/j.landusepol.2015.09.025]
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; de Melo Conti, Diego 4
; Raissa Caroline Gomes 2
; Fabricio Camillo Sperandio 4 ; Ribeiro, Admilson Irio 5 1 Postgraduate Program in Urban Infrastructure Systems and Postgraduate Program in Sustainability, Pontifical Catholic University of Campinas (PUC Campinas), Campinas 13087-571, SP, Brazil;
2 Postgraduate Program in Urban Infrastructure Systems and Postgraduate Program in Sustainability, Pontifical Catholic University of Campinas (PUC Campinas), Campinas 13087-571, SP, Brazil;
3 Department of Geography and Tourism, Centre of Studies in Geography and Spatial Planning (CEGOT), University of Coimbra (UC), 3004-530 Coimbra, Portugal
4 Postgraduate Program in Sustainability, Pontifical Catholic University of Campinas (PUC Campinas), Campinas 13087-571, SP, Brazil;
5 Postgraduate Program in Environmental Sciences, São Paulo State University “Júlio de Mesquita Filho” (UNESP), Sorocaba 18087-180, SP, Brazil;