The Maule region is located within the Chilean Mediterranean (see Figure 1), and presents an area of 30296.1 km2, which represents 4.0% of the national surface, has a population of about 1 million inhabitants and a rich agrarian cultural diversity, where the rural population represents 33.6% of the regional total, about 330 thousand inhabitants (INE, 2017). It presents a warm and sub-humid climate of Mediterranean type, where there are four geomorphological zones: Andean Mountain range, intermediate depression, coastal mountain range and coastal plains. This allows the existence of native vegetation and the development of agricultural and forestry activities (ODEPA, 2018).

The area known as the intermediate depression, also known as the Central Valley, has a characteristic Mediterranean climate with cold, wet winters and hot, dry summers. It is considered a priority region for the conservation of world biodiversity (Myers et al., 2000) due to its high level of endemism and continuous habitat loss, as it is where most of the agricultural sector is located today. It is dominated by export crops, which occupy 90% of the land, out of a total of 811,480 ha available, and are represented in percentage terms by forestry plantations (60.8%), fruit trees and vineyards (12.4%), cereals (9.1%) and fodder crops (5.7%). With a much smaller area are vegetables, legumes, tubers and home gardens, which reach 22,236 ha planted, representing no more than 2.7% of the total regional area available (ODEPA, 2018).

Satellite images [Sentinel 2] available for March 2022 (see list of images in the Supplementary material) were used to characterize the landscape by photointerpretation, and only for the identification of patches of native vegetation and water bodies present in the landscape, vector information corresponding to the CONAF vegetation cadastre (CONAF, 2021) was used (Table 1). Patch extent metrics and distances between patches and agroecosystems were processed and analyzed with Qgis and RStudio software using the sf, terra and rgdal libraries (QGIS Development Team, 2021; R Core Team, 2023).

In order to characterize and analyze the relationship between the agroecosystems of the Maule region, specifically the area near the municipality of San Clemente and the surrounding landscape, the methodology proposed by León-Sicard et al. (2018). Main Agroecological Structure [EAM], expressed in equation [1], is used to characterize the design of agroecosystems and their relationship with the surrounding landscape over time in an integrated manner. The Kruskal-Wallis test was used to assess significant differences between agroecosystem types.

This methodological tool allows us to look at the ecological, social and cultural links that exist between an agroecosystem and its environment [buffer zone], with an emphasis on water bodies, semi-natural areas and other non-productive uses. The buffer zone of [500 m] was defined as an area within which, for example, insect movements (Raymond et al., 2015) of interest to participating farmers and researchers could be distinguished. It was calculated using the agroecosystem perimeter and corresponds to a measure to normalize differences in total area between the types of agroecosystems studied (livestock, orchard, and horticulture). The index focuses on the quantitative and qualitative measurement of agrobiodiversity, particularly in terms of structure. The indicators used to construct the index are described in the Table 2.

Qualitative and quantitative methods were combined to analyze the biophysical and agroecological conditions present in each agroecosystem. The following tools were used for data collection.

The workshops initially identified some of the problems that the group of participating farmers identified as priorities in their agroecosystem, see Table 3, including the lack of support for agroecological transition from the state and its agencies, technical difficulties such as pest, disease, and weed control, low yields, and low sales prices. The increase in external inputs and the general devaluation of traditional knowledge of the farmers were some of the most frequent observations made in the workshops held.

The results show that, in general, the Mediterranean agroecosystems studied cover an average area of 1.96 ± 0.1 ha, with a low presence of native vegetation patches and water bodies within the agroecosystems and in the surrounding landscape, reaching no more than 4.0 ± 0.2% of the total area studied covered by native forest and 0.3 ± 0.05% of the total area with water bodies. Most of the native vegetation types present are of the renoval type of sclerophyll forest, with formations dominated by Cryptocarya alba (Chilean peumo), Quillaja saponaria Mol. (Quillay) and [Lithrea caustica Mol. (Liter) species]. The connection between the agroecosystems and the few surrounding patches of native vegetation or water bodies present was low or zero, as can be seen from the average distance between the patches and the center of each agroecosystem [DFFCF DBWCF], where only for horticulture patches of water bodies and native vegetation were found.

In general, distances between native vegetation patches and water bodies were smaller in vegetable agroecosystems, but the presence of patches in the buffer zone was significantly higher in livestock systems. Mean distances between native vegetation patches were low for all three agroecosystem types, with a mean of 31.4 ± 1.8 m. The distance between water body patches was 17.8 ± 5.6 and the distance from the center of the agroecosystems to the water body patches was 37.3 ± 28.4 (see Table 4).

The MAS of the evaluated agroecosystems can be considered as slightly developed, with a calculated mean of 52.6 ± 0.27. Livestock agroecosystems received the highest and lowest scores for the main agroecological structure, i.e., agroecosystems with important proportions of native forest and water bodies in the buffer zone, which were also connected by vegetation edges, and others that did not have any of these types of patches were characterized. As there were no statistical differences between the types of agroecosystems assessed (Kruskal-Wallis, p = 0.14), the distribution of observations can be seen in Figure 2. The mean was calculated for livestock (53.4 ± 3.2), orchards (55.7 ± 2.9) and horticulture (49.5 ± 1.8) (More details in the Supplementary material).

The research process allowed us, through the application of the MAS methodology, to know in detail the management that each farmer carries out in his agroecosystem. A set of 11 practices was identified (see Table 5), recognized for their positive contribution to key ecological functions for Mediterranean agricultural systems, soil fertility, natural regulation of pest organisms and weed control. At least 50% of the agroecosystems studied use spatial and temporal diversification as a strategy to maintain soil fertility. These strategies include at least 3 agroecological practices, crop rotation [85.3%], crop diversification in the agroecosystem [73.5%] and integration of the animal component [67.6%], whether it is sheep, cattle or poultry production systems.

Natural pest regulation is another key element in agroecosystems with Mediterranean, in the group no strategies developed to optimize this ecological process were identified, however, at least 50% of farmers use, crop association [70.6%] and the inclusion of aromatic plants [55.8%]. It is important to highlight the use of chemical products in an important group of agroecosystems [76%] to replace the ecological processes of soil fertility and natural pest regulation.

In this study, the methodology proposed by León-Sicard et al. (2018) was used, as at the time of the fieldwork, the update of the methodology carried out in 2022 had not yet been officially published. The work of Quintero et al. (2022) promotes an equitable weighting, through the aggregation or balanced summation of each of the parameters involved in the construction of the MAS, which included the metrics of composition, configuration, heterogeneity, and landscape management practices used in each agroecosystem (Fahrig et al., 2011). Given the apparent link between management practices and ES provision (Palomo-Campesino et al., 2018, 2022), it is crucial to identify which agroecosystems have the potential to contribute to ES provision and which do not, as illustrated in Figure 3 and highlighted by Sirami et al. (2019). The MAS values recorded ranged from 35 to 79, indicating a gradient between the agroecosystems studied, with one group with a poorly developed agroecological structure and considerable cultural potential, and another group with an agroecological structure in a moderately developed state, with management differences observed between the agroecosystems studied and a high degree of isolation from the ecological structure of the surrounding landscape.

The use of maps and other GIS tools allowed for a participatory characterization of the landscape surrounding the agroecosystems studied, working together with farmers to identify strengths and weaknesses at the landscape scale. In this context, the MAS methodology allows aspects of landscape composition and configuration to be observed in an integrated manner, allowing for the standardization and local refinement of the landscape metrics used (Liere et al., 2017). Including the perspective of the farmers’ group (PA) on what they perceive as environmental degradation and biodiversity loss, which was little developed at the beginning of the workshops conducted. On the other hand, farmers also recognize an important individual and collective capacity for action (CA) that could significantly improve what is done in their production units.

Particularly in this region, the agriculture landscapes show a high homogeneity from an agricultural point of view, dominated in the last 20 years by the increase of forest plantations and agricultural export crops (Díaz-Hormazábal and González, 2016; Tapia and Morais, 2020). It is therefore not surprising that the valorization of the extension (EEC; EIC) and diversification (DEC and DIC) of the external and internal connectors of the agroecosystems are mostly low, since these agroecosystems do not have an established agroecological design, which is reflected in external connectors at the periphery that are abandoned and in some cases non-existent, internal connectors with low or no connectivity between the different areas within the agroecosystem, and in both cases, external and internal connectors, with a low diversity of tree and shrub plant species present. In this territory, agroecosystems play an important role in the conservation of organisms in an important global biodiversity hotspot (Henríquez-Piskulich et al., 2021).

A transversal characteristic of the agroecosystems studied is the integration of the animal component, which in practice is observed as different land uses (USC) in the agroecosystem, where these agrosilvopastoral subsystems are included, using an average area of less than 50% of each farm studied. Management of Weeds (MW) is mostly conventional, where mechanical control and the use of herbicides for weed control predominate. In general, the agroecosystems are characterized by conventional management, which includes at least 5 practices that, if properly applied, could serve as a basis for conversion to agroecological systems.

The use of landscape metrics for the construction of MAS, through methodological tools such as the creation of collective maps, makes it easier for farmers to understand the importance of the internal and external connectivity of the biodiversity of each agroecosystem with the agricultural matrix that surrounds it (Cattaneo et al., 2018). In addition, these tools allow researchers and other stakeholders to visualize some of the most common and necessary technical issues that need to be addressed in each context.

In order to promote the agroecological transition and expand the scale of agroecological experience (González de Molina et al., 2017), it is important to work with farmers through the use of practical, horizontal evaluation methods that take into account different scales, from the agroecosystem (Nicholls et al., 2020; Tamburini et al., 2020) to the landscape (León-Sicard et al., 2018; Vanbergen et al., 2020), which allow a better understanding of the impact of management on the agroecological landscape of all actors involved in this research, a complex and non-linear process that requires attention.

To achieve agroecological landscapes, it is crucial to understand the biodiversity patterns, biological interactions, and mechanisms of the natural ecosystems present in the territory (Brauman et al., 2020; Jeanneret et al., 2021), and to engage farmers in a bottom-up, context-specific approach to improve services at the landscape scale (Barrios et al., 2020; Brauman et al., 2020). In addition, it would be important to assess at the local scale the contribution that the portions of native vegetation cover [native forest or scrub] in the landscape adjacent to the agroecosystems included in this study could make to the provision of ecosystem services, such as natural regulation of pests and diseases at the landscape scale (Wratten et al., 2019). Considering that only a proportional 20% of the total area devoted to non-agricultural land could significantly improve the impact on local biodiversity and ES provision, reducing dependence on agricultural inputs by up to 50% (Garibaldi et al., 2021).

The MAS methodology is a useful collective planning tool in the process of socio-ecological transition, allowing the involvement of different actors of the territory (Quintero et al., 2022). Since participatory and quantitative methods are used in a combined way, accurate and relevant assessments of agroecological transitions can be made (Teixeira et al., 2018). This also allows a future work plan in terms of planning, seeking an integral connection of the environment of each agroecosystem, which can even be replicated and extended to peri-urban and urban production systems (Vaarst et al., 2018). In addition, successful cases of farmers were identified to become beacons that stimulate and guide the adoption of agroecological practices and principles in local communities of the area, where the recovery of traditional agricultural systems and their management, which have historically offered promising models of sustainability and resilience, can be observed (Nicholls and Altieri, 2018).