Impacts of Land Use Change on Landscape Structure and Ecosystem Services at Local Scale: A Case Study in Central Portugal

The study was conducted in Lousã municipality, in the central region of Portugal. Three selected areas are representative of the geodiversity of Lousã municipality (Fig. 1).

Fig. 1

The municipality of Lousã is relatively heterogeneous in terms of climate, geology and topography, so it was considered suitable for studying the effects of spatial and structural features on the patterns of landscape, biodiversity and forest patches. Thus, the selected areas of study include different types of landscape (LULC) because of their biophysical characteristics and human occupation. The first selected area, on the ‘western slope of Serra da Lousã’, includes a landscape of mountains marked by a strong altimetric gradient that varies between 348 m and 1134 m as well as quite steep slopes with an average of 20% (Table 1). The dominant bedrock type is schist, which generates nutrient poor soils. The second area of study is mostly drained by the Ceira River and occupies a lower altitude, with a mean of 200 m and an average slope of 8%. The third area is located in the Lousã Basin and it partly covers the urban perimeter of the municipality's parishes. The mean altitude is 82 m and the average slope is 3%. North and west aspects are dominant in the study area.

The use of time series of historical maps and aerial photography is a common practice in historical geography and has proven to be very useful (Ihse 1996, Skånes, Bunce 1997, Vuorela 2000). In this study, land use and cover change were based on maps produced by the Agrarian Survey and Management Service for 1974, and the LULC Map of Mainland Portugal for 2018 from the Directorate-General for Territorial Development (both at 1:25,000 scale). The main difficulty with the simultaneous use of these documentary sources has been the different classes of inventorying. Thus, we had to carry out a process of unification through classes which could effectively and immediately demonstrate the significant uses conferred on the territory and the changes carried out in the past few decades (Table 2).

For this purpose, the different classes of land occupation/cover were combined to facilitate the quantification of changes and the application of landscape metrics, which were implemented through the use of geographical information systems (GIS). All spatial analyses and procedures related to LULC change and landscape metrics were implemented through the ArcGIS software, using ArcMap version 10.5.1 and the extension Patch Analyst. The analysis of the landscape metrics used the variables of the Patch Analyst that were best suited to our objective (Table 3), related to the landscape composition, configuration and diversity.

The configuration metrics are mainly used to describe the spatial characteristics of individual patches or the spatial relationships between the multiple patches (Botequilha Leitão et al. 2002, Couto 2004). This group of metrics also integrates measurements of the landscape configuration, such as area and edge metrics.

In order to estimate the change in ES flow, the ES mean economic value, adapted from Marta-Pedroso et al. (2014), was used to describe both spatial and temporal changes of ES. The estimated economic values are systematised in Table 4, according to the reference values achieved for a pilot area, located in the Alentejo region, Portugal. The obtained results are included in the first The Economics of Ecosystems and Biodiversity (TEEB) study in Portugal, in which ES were identified based on the literature, expert judgement, stakeholders’ engagement and field work. The economic valuation of these ES was performed using available information and relied mainly on the use of avoided costs (carbon sequestration and soil protection), willingness to pay (biodiversity) and market prices (crop and fibre production) methods (Table 4). According to the authors, the economic value of the different services was based on estimates or proxies of values, namely data published in scientific journals, or information available in the public administration that allowed an estimate of the economic value. Thus, the economic coefficients used for this study should be interpreted as proxy values as the estimates of these values of ES can be biased for the study area.

The Ecosystem Services Total Value (ESV) at time T is estimated by multiplying the area of each land-cover category by the coefficient associated to each land use/land cover, using the following relationship: (1) SV=(Ak×VCk) {\rm{SV}} = \left( {{\rm{Ak}} \times {\rm{VCk}}} \right) where: –

Ak – the area in hectares of land-cover category ‘k’,

Changes in LULC in the period from 1974 to 2018 are presented in Figure 2 and summarised in Table 5. The main changes observed result from a substantial increase of areas occupied by EU (+878%), IS (+405%) and the social areas (SA) (+346%). Conversely, the areas devoted to agriculture and occupied by OGs saw a decrease in their areas, of 10 and 87%, respectively. Between 1974 and 2018, the pine forest remained the dominant land cover in the territory, suffering only a slight reduction.

Fig. 2

The conversion from pine forest, shrubland and CL into EU forest explains its relevant increment while the IS expanded mostly by soils previously occupied by pines, OGs, shrubs and CL. Social facilities have expanded into areas which had previously been cultivated or contained OGs. As for DFs, they increased in area chiefly due to the decline of pine forests (35%) and cultivated areas (26%). The abandonment of agriculture has led to part of the cultivated area now being occupied by DF (20.6%) and SA (16.6%).

Comparing the LULC dynamics in the selected areas (Fig. 3) we can conclude that agricultural land and OGs decreased in the three areas, while PW and shrubland also show a significant decrease in the River Ceira Basin and in the western slope of Serra da Lousã, respectively.

Fig. 3

At the same time, the EU enjoyed a general expansion, whilst DF and IS register an increase in the River Ceira Basin and on the western slope of Serra da Lousã. SA have increased significantly in the Lousã Basin, coinciding with the main town in the municipality.

Figure 4 systematises the metrics of the landscape, considering the two periods under analysis. The results show an almost triple increase in the NumP in all land uses/covers and a decrease in the average size, from 30 ha to 7 ha, meaning that the landscape has experienced a strong fragmentation process. Pine remains the species with the greatest spatial relevance, occupying around 40% of the total area, and has the largest NumP (71 No./P) and the highest average size, despite the significant decrease observed, that is, from 57 ha to 19 ha. The highest increase in the NumP is recorded by the EU and DF, although the respective mean patch area decreased from 13 ha to 7 ha and from 27 ha to 7 ha. The same trends are noted for shrubland and IS. The CL records a significant increase in the NumP between 1974 and 2018; however, the relevant size fell by about one-third. The social area is the only LULC to register an increase in the mean patch size. Consequently, the patch density (PD) has also expanded in all LULC classes, whilst total edge (TE) and edge density (ED) only decreased for the CL and OG.

Fig. 4

Both the Shannon Diversity Index (SDI) and the Simpson diversity index (SEI), which express the degree of landscape diversity given by the number of classes and proportion of the landscape area occupied by each class, indicate an increment from 1.2 to 1.6 and 0.6 to 0.7, respectively, from 1974 to 2018.

In the western slope of the Serra da Lousã (Table 6), the landscape matrix was dominated by shrubland in 1974, and it changed to PW by 2018. SA, CL and OGs almost disappeared in 2018, whereas IS and EU forest record the highest increases. With the lowest NumP among the selected areas, the western slope of the Serra da Lousã recorded an increase from 48 to 74 (+41%), a decrease in the median path size, from 36 ha to 13.6 ha, and in the PD, from 61 ha to 59/100 ha. Only the PW doubled their patch size, from 1974 to 2018. Both indexes of diversity – SDI and SEI – have recorded an increase from 1.59 to 1.78 and 0.72 to 0.81, respectively.

In the River Ceira Basin (Table 6), PW stands as the dominant landscape matrix in both periods, in spite of its significant decline in the total area recorded in recent decades. Conversely, EU and DFs have recorded a relevant increase. Overall, the NumP has more than doubled, and the respective average area has fallen from 30 ha to 7.5 ha. The SDI and SEI also recorded an increase, indicating that the landscape heterogeneity and evenness have slightly increased. Table 6 also shows that although CL has lost a significant part of its area in the Lousã Basin, in 2018 it is still the most important land use, with a value significantly above PW, which increased around 3%. SA and EU forest recorded the highest increment (around 600%). Conversely, OGs decreased by 70%. In general, the total NumP tripled, whereas the mean area per patch has fallen, from 37 ha to 5 ha. Considering all the LULC, the NumP has increased, whereas the mean area per patch has only grown in the SA (from 2.2 ha to 4.3 ha).

By using the value coefficients and areas of the main LULC categories changes (Tables 4 and 5, respectively), the ecosystem service value of land use category was calculated for both years (1974 vs 2018). The results showed that although the eucalyptus and deciduous forest increased the ecosystem service value, the overall total ecosystem service value fell around 10% between 1974 and 2018, as a consequence of the decline recorded in the olive groves (−74%), cultivated land (−44%) and shrubland (−54%) (Fig. 5).

Fig. 5

The estimated ecosystem service value for the three selected areas demonstrates different dynamics as we can see in Figure 6. On the western slope of Serra da Lousã an overall increase in the ecosystem service value mainly related to the growth observed in the pine woods area. Conversely, in the Ceira River and Lousã Basin the registered decrease in the overall ecosystem services value are mainly related to the significant decrease in the pine woods and agricultural area, respectively.

Fig. 6

Pine woods and olive groves are the LULC that contribute the most to the variability observed in the ecosystem services value, whilst the regulation services (sediment retention and timber production) are the ones with the highest value. Considering the various ES evaluated, the greatest losses occurred in the production of agricultural goods (−64% of the annual value) while the most evident gains occurred in biodiversity, where the value more than doubled (Fig. 7).

Fig. 7

If we take the three selected areas, a general trend in agricultural abandonment was observed that was more significant in the mountain area of Lousã municipality, which can be mainly attributed to rural exodus. In fact, villages such as Cerdeira and Candal, on the western slope of the Serra da Lousã, had suffered a truly significant reduction in the population (280 inhabitants in 1940 to 10 inhabitants by 2011, i.e., −96%) (INE 1947, 2012). River Ceira Basin also saw its population decrease between 1940 and 2011 (from 2400 to 1800 inhabitants, i.e., −25%) however, the rates of conversion were significantly lower. This agricultural abandonment has negatively impacted the supply of agricultural goods, which have been partially counterbalanced by an increase in fibre production and biodiversity, mainly associated with an expansion of native forest species. In fact, several authors (Lasanta et al. 2015, Nadal Romero et al. 2016, Gashaw et al. 2018, Perpiña Castillo et al. 2020) view agricultural abandonment as an opportunity to reverse the long-term decline of forests, provision of ES, and habitat enhancement that re-naturalisation of landscape provides. Other authors (Qin et al. 2013, Ramankutty et al. 2018, Zhang et al. 2018, Chaudhary et al. 2020, Crawford et al. 2022), however, regard agricultural abandonment as negative process with impacts on the local population and on society as a whole in terms of the production of goods (e.g. foodstuffs, livestock feed, fibre), biodiversity, as well as services provided by the multi-functionality (e.g. sociocultural practices, values, and norms) of the agricultural landscape.

Forests (evergreen plantation and deciduous woodland) occupy the major proportions of land use in the study area and also recorded the highest increase over the analysed period. Plantations of Pinus pinaster constitute the dominant element of the landscape matrix. The high percentage of PW is related to several afforestation campaigns promoted by the Portuguese government, mostly involving mountain common lands and the planting of Pinus pinaster. Since joining the European Union in 1986, new forestry policies have been adopted by Portugal. These prioritise the restoration of mixed woodlands and closer collaboration with private owners of woodlands and forests (Canaveira et al. 1998), explaining the increase observed in both forest plantation (mainly Eucalyptus species, due to given their fast growth) and deciduous. In the study area, as well as in other parts of the Mediterranean region, the increase of DF mainly results from agricultural abandonment that enhanced natural secondary succession and the spread of woodland (Nunes 2008, Almeida et al. 2009, Lasanta et al. 2009).

It is widely accepted that forest ecosystems provide diverse critical services and benefits to human society (Jenkins, Schaap 2018). Although the increase of EU monoculture plantations have a positive impact in the timber production, various authors consider that monoculture plantations could affect several regulating ES, related to soil, water and biodiversity. Tree harvesting by machines can promote soil erosion and compaction (Boltodano 2000, Bowyer 2006, Affek et al. 2017) which will adversely affect the growth of understory. Single-species plantations are also not efficient in trapping nutrients, because there are fewer roots near the surface, which may further lead to significant loss of nutrients from the harvest sites. In addition, some species, such as Eucalyptus, can acidify the soil and release specific substances that inhibit the growth of other plant species (Bowyer 2006), thus affecting the biodiversity. Some researchers (Morris et al. 2004, Bowyer 2006) have observed that Eucalyptus consumes more water than other species in natural forests, which may draw down the water table in some regions.

On the other hand, the impressive afforestation measures implemented on common and private land in the twentieth century went ahead without proper forest management, silviculture and fuel-hazard reduction, especially after the 1970s (Mateus, Fernandes 2014), have increased the wildfires hazard.

The municipality of Lousã has registered a very important increase in alien species, especially Acacia dealbata, mainly as a consequence of land abandonment and the recurrence of wildfires. The Acacia dealbata is considered as a problematic and widespread invasive plant; it has negative impacts on the ecosystem's structure and functioning, triggering ecological homogenisation and reducing biodiversity (Binggeli 1996, Williamson 1999, Chapin et al. 2000, Pysek, Richardson 2010, Correia 2012, Marchante et al. 2015, Gil 2017). Therefore, biological invasions by alien species have been recognised as one of the most important drivers of biodiversity loss and ES changes worldwide. Also, they have negative effects on socioeconomic, cultural and human health aspects by affecting all four categories of ES: supporting, provisioning, regulating and cultural services (Vilá, Weiner 2004, Marchante et al. 2015, Gil 2017). In this context, urgent preventative, eradication and control actions are required to impede their entry and establishment or minimise their long-term impacts (Robertson et al. 2020).

Urban expansion in the study area occurs mainly in the Lousã Basin, where the municipality's main town is located, as it responds to pressures on the territory to accommodate new residential, service or industrial areas. This process of transforming agricultural and forest land uses into urban land cause substantial losses of natural habitats and ES (e.g. food production, freshwater provision, and carbon storage) due to the loss of vegetation and increase in impervious surfaces (Zhang et al. 2017). Several authors (Schröter et al. 2005, Wade et al. 2009, Wu et al. 2014) also consider that it indirectly influences the delivery of ES, such as water retention, climate regulation and nutrient retention by altering the hydrologic cycling, atmospheric circulation and nutrient cycling processes.

Fragmentation was the most obvious characteristic of landscape change in all the areas studied, even though the obtained results might be partially overestimated by different details in the mapping between 1974 and 2018. The results indicate the threefold increase in the NumP and the decrease by half in the mean patch area.

Although both Diversity Indexes used denoted an increase in the study area, several authors consider that the habitat fragmentation commonly influence ES provided by forests, shrubs and grasslands across landscapes on both spatial and temporal scales (Kremen et al. 2006, Laurance et al. 2007, Midha, Mathur 2010, Berhane et al. 2013, Pinto-Ledezma, Rivero Mamani . 2014, Cuke, Srivastava 2016). As mentioned by many authors (Debuse et al. 2007, Çakir et al. 2008, Hartter, Herrera 2009, Zipperer et al. 2012, Qi et al. 2014), the high increase in the NumP, the shrinking mean patch size as well as increasing ED at the landscape level corresponded to the declining quality of ES such as biodiversity conservation, carbon sequestration and seed dispersal.