LAND COVER CHANGES IN THE NW SPANISH COASTAL ZONE: DRIVERS AND IMPACT ON ECOSYSTEM SERVICES

3.1. Artificial land cover changes and fluxes in the Galician coast

The land cover distribution observed in the Galician coast in 2014 (last year available) is shown in Figure 2. 65% of the coastal area was dedicated to forest and seminatural areas, occupying a total of 273534 hectares when the coastal municipalities were considered. Agricultural areas accounted for 21% of the coastal area whereas artificial surfaces, located mainly around the coastal cities, represented 12%. Wetlands and water bodies were much less represented. These results were very similar when the 10 km coastal zone or the municipalities’ coastal strip was considered, suggesting that both methodological approaches could be used indistinctly.

Figure 3A shows absolute and relative changes in land cover in the area delimited by the Galician coastal municipalities in the period 2005-2014. The largest relative change occurred in the artificial surfaces, which increased by ca. 7.4% in the study period. Agricultural areas decreased by 3% and the other land classes varied less than 1%. However, the most considerable absolute losses were found in the agricultural areas, showing a decrease of 2753 hectares, albeit only represented 3% in relative terms.

As expected, the same pattern was observed when the 10 km coastal zone was used, although differences were found in the absolute number of hectares, as the total extent of both areas differed (Fig. 1). Figure 3B shows daily rates of land cover changes in the municipalities of the coastal strip. In general, artificial areas increased and agricultural areas decreased. However, forest and seminatural areas decreased from 2005 to 2009, but increased in the 2009-2014 period. Artificial land cover rose from 47361 ha in 2005 to 50350 ha in 2009, which yields a daily growth rate of ca. 2 ha day^-1 (i.e. /day) in that period, while in the 2009-2011 and 2011-2014 periods these rates were 0.2 and 0.4 ha day^-1. The estimated rate for the 10 km coastal zone from 2005 to 2009 was 2.5 ha day^-1 due to the large size of some coastal municipalities, which penetrate into the continent where exposition to anthropogenic pressures is less intense.

When the results obtained in this study (using SIOSE), are compared with the data published by the Observatorio de la Sostenibilidad (2016b)Observatorio de la Sostenibilidad (2016b). Cambios de ocupación del suelo en la costa en España 1987-2011. Retrieved from https://www.observatoriosostenibilidad.com/informes/ , derived from CLC, substantial differences were found in both the artificial area and the annual land cover change rate. According to CLC, 36054 hectares were classified as artificial surfaces in the 10 km coastal zone in 2011, which represents 6.8% of the total coastal area, whereas the results obtained in this study show that 11.4% of the 10 km coastal zone was covered by artificial surfaces (60963 ha) in the same year. Previous studies have warned about the underestimation of the artificial area along the Spanish coast due to the scale of CLC data, an error enhanced in areas with a dispersed settlement, for example, the Galician coast (Observatorio de la Sostenibilidad, 2016bObservatorio de la Sostenibilidad (2016b). Cambios de ocupación del suelo en la costa en España 1987-2011. Retrieved from https://www.observatoriosostenibilidad.com/informes/ ). In addition, the CLC project discards linear data such as roads and rail networks as artificial surfaces, thus reinforcing this underestimation (Olazabal and Bellet, 2018Olazabal, E. & Bellet, C. (2018). Procesos de urbanización y artificialización del suelo en las aglomeraciones urbanas españolas (1987-2011). Cuadernos Geográficos, 57. https://doi.org/10.30827/cuadgeo.v57i2.5920 ). These authors also reported the potential effect of the minimum percentage (5%) considered by each SIOSE polygon when digitising large polygons representing disseminated urban settlements, which may result in abnormally high values of urban surfaces. The above-mentioned differences are also reflected in the between-periods variability of annual artificialization rates (Table 3).

The artificialization of the 10 km coastal zone in Spain between 1987 and 2011 showed a rate of 15.4 ha day^-1 according to CLC, increasing the artificial surface by 2.3% per year (Observatorio de la Sostenibilidad, 2016bObservatorio de la Sostenibilidad (2016b). Cambios de ocupación del suelo en la costa en España 1987-2011. Retrieved from https://www.observatoriosostenibilidad.com/informes/ ). Specifically, artificial surfaces in Spain increased by 1.48% per year in the period 2006-2012, the highest rate in the EU, where the average rate was 0.41% year^-1 (i.e. /year) (Dirección General de Medio Ambiente de la Comisión Europea, 2017Dirección General de Medio Ambiente de la Comisión Europea (2017). Revisión de la aplicación de la normativa medioambiental de la UE. Informe de España. Retrieved from https://ec.europa.eu/environment/eir/pdf/report_es_es.pdf ). The annual growth rate of artificial surface in the Galician coastal zone (10 km) according to CLC between 1987 and 2011, was 1.52%, which meant 1.1 ha day^-1 (Observatorio de la Sostenibilidad, 2016bObservatorio de la Sostenibilidad (2016b). Cambios de ocupación del suelo en la costa en España 1987-2011. Retrieved from https://www.observatoriosostenibilidad.com/informes/ ). However, the increase in artificial hectares in the Galician coastal zone (10 km) in the period 2006-2012 was 0.64 ha day^-1, representing an increase of 0.67% per year. When these data are compared with those obtained in this work using SIOSE, and choosing the closest available interval, 2005-2011, the artificial surface increased by 1.1% per year, with an annual rate of 1.7 ha day^-1. This is consistent with the underestimation of CLC and the overestimation of SIOSE in terms of artificial surfaces previously reported in the literature (Diaz-Pacheco and Gutiérrez, 2014Diaz-Pacheco, J. & Gutiérrez, J. (2014). Exploring the limitations of Corine Land Cover for monitoring urban land-use dynamics in metropolitan areas. Journal of Land Use Science, 9, 243-259. https://doi.org/10.1080/1747423X.2012.761736 ; Observatorio de la Sostenibilidad, 2016bObservatorio de la Sostenibilidad (2016b). Cambios de ocupación del suelo en la costa en España 1987-2011. Retrieved from https://www.observatoriosostenibilidad.com/informes/ , 2017Observatorio de la Sostenibilidad (2017). Sostenibilidad en España 2016 - SOS16. Retrieved from https://www.observatoriosostenibilidad.com/informes.; Olazabal and Bellet, 2017Olazabal, E. & Bellet, C. (2017). Análisis de las nuevas dinámicas de urbanización en España. Su estudio a través del uso de Corine Land Cover y SIOSE [Conference paper]. XXV Congreso de la Asociación de Geográfos Españoles, Madrid, Spain. https://libros.uam.es/uam/catalog/book/558., 2018Olazabal, E. & Bellet, C. (2018). Procesos de urbanización y artificialización del suelo en las aglomeraciones urbanas españolas (1987-2011). Cuadernos Geográficos, 57. https://doi.org/10.30827/cuadgeo.v57i2.5920 ; Ovejero-Campos et al., 2019Ovejero-Campos, A., Fernández, E., Ramos, L., Bento, R. & Méndez-Martínez, G. (2019). Methodological limitations of CLC to assess land cover changes in coastal environments. Journal of Coastal Conservation, 23(3), 657-673. https://doi.org/10.1007/s11852-019-00696-w ).

The relative contribution of the artificial land cover in the Galician coast in 2014 is represented in Figure 4, and shows the presence of areas intensely artificialized mainly linked to the main coastal cities. The municipalities with the highest percentage of artificialization were A Coruña (64%) and Vigo (51%).

The distribution of this variable (Fig. 4) agrees with the location of the urban aggregates defined in the regional littoral planning (Plan de Ordenación do Litoral de Galicia) which corresponds to areas of concentration of urban growth where favorable topographic conditions facilitate the development of multiple economic activities (Xunta de Galicia, 2011Xunta de Galicia (2011). Plan de ordenación do litoral de Galicia. Retrieved from http://www.xunta.es/litoral/ ). It also matches with the metropolitan areas of Vigo-Pontevedra and A Coruña-Ferrol.

Figure 5 shows the annual artificialization rate (%) of the Galician coastal municipalities calculated with respect to the municipal area, which represents the amount of natural surfaces (forest and seminatural areas, agricultural zones and land reclamation) has been “denatured”.

The highest artificialization rates with respect to the total area of the municipalities were found in the 2005-2009 period, in those areas contiguous to the larger coastal cities (Fig. 5). In the following period (2009-2011), clear differences were found between southern municipalities, showing relatively low land artificialization rates, and northern municipalities where this artificialization was practically non-existent. In the most recent period (2011-2014), a south to north artificialization gradient was observed, albeit rates were lower than in the first period (2005-2009).

To further investigate the patterns of land artificialization in this region, the Galician coast was divided into zones according to the different degrees of artificialization, derived from their physiographic characteristics, geographical location, climate, orientation, demographic burden, or even the economic and social characteristics. As a result, five groups of coastal municipalities were defined: Costa da Morte, Mariñas Coruñesas, Rías Altas, Rías Baixas and Sur de Galicia (Fig. 1). The maximum annual artificialization rates were generally observed in the period 2005-2009, albeit large differences among coastal groups were observed, with Mariñas Coruñesas almost doubling the artificialization rates of the Rías Baixas, Rías Altas and Costa da Morte, and the Sur de Galicia area showing an artificialization rate 5-fold lower than Mariñas Coruñesas. A similar pattern was observed in the 2011-2014 period. During the 2009-2011 period, land artificialization was only detected in the southern groups (Fig. 6).

Figure 7 shows fluxes between land cover classes in the municipalities’ coastal strip (2005-2014) resulting from the corresponding cross-matrix, as proposed by Pontius, Shusas & McEachern (2004)Pontius, R.G., Shusas, E. & McEachern, M. (2004). Detecting important categorical land changes while accounting for persistence. Agriculture, Ecosystem & Environment, 101, 2-3, 251-268. https://doi.org/10.1016/j.agee.2003.09.008 and Martínez-Fernández et al. (2015)Martínez-Fernández, J., Ruiz-Benito, P. & Zavala, M.A. (2015). Recent land cover changes in Spain across biogegraphical regions and protection levels: implications for conservation policies. Land Use Policy, 44, 62-75. https://doi.org/10.1016/j.landusepol.2014.11.021 . The total area transformed into the artificial class in this period was 4610 ha. Almost 3000 of them were previously classified as forest or seminatural areas, and around 1600 hectares as agricultural areas. 215 hectares of total land uptake by artificial surfaces correspond with land reclamation.

Therefore, 61.4% of the new artificial areas derive from forests or seminatural areas, 33.8% from agricultural areas, 4.7% from areas reclaimed from the sea and 0.1% from wetlands and water bodies.

We further explored fluxes involving land artificialization to analyse and characterise their dynamics, aiming at a better understanding of the processes involved in the generation of artificial land (Table 4). It is worth mentioning that artificialization fluxes in the coastal municipalities strip and in the 10 km coastal zone coincide.

The largest changes were concentrated again in the period 2005-2009, being the forest and seminatural classes those presenting the more intense annual changes. During that period, 0.13% of the municipalities’ coastal strip became artificialized at the expense of forests or seminatural areas. The most noticeable change from agricultural areas to artificial surfaces occurred in the same period (0.07% of the total area of the strip per year). When land cover annual changes (%) with respect to the total area of each class were considered, the agricultural class presented the higher rates in all the periods. A total of 1211 agricultural hectares were artificialized between 2005 and 2009, representing a rate 17-fold higher than in the 2009-2011 period and more than 3-fold higher than in the last one (2011-2014). In the forestry class, 2180 hectares classified as forest or seminatural areas in 2005 were transformed into artificial surfaces until 2009, 16-fold higher than in the period 2005-2009 and 2.5 higher than in the last period.

The high demand for human settlements and the development of economic activities in the coastal zone triggers a high rate of land consumption. In this study, we found the largest changes in the artificial class, which increased by more than 7% in the studied period, at the expenses of forest and seminatural areas. In most cases, the spatial pattern of this artificialization process was associated with disorganised growth detached from traditional urban settlements, resulting in a diffuse transformation of the Galician coast leading to significant territorial and landscape fragmentation. This fragmentation directly affects ecological connectivity, thereby amplifying several environmental problems such as soil sealing, deforestation, water, soil or air pollution, as well as biodiversity losses (Celliers and Ntombela, 2016Celliers, L. & Ntombela, C. (2016). Urbanization, coastal development and vulnerability, and catchments. In José Paula (Ed.), Regional State of the Coast Report (pp. 386-404). Nairobi, Kenya: Nairobi Convention Secretariat. https://doi.org/10.18356/dd8dca69-en ; Pasquali and Marucci, 2021Pasquali, D. & Marucci, A. (2021). The effects of urban and economic development on Coastal Zone Management. Sustainability, 13, 6071. https://doi.org/10.3390/su13116071 ).

3.2. Impact of artificialization on ecosystem services

Figure 8 shows the spatial distribution of land cover and the potential to provide ecosystem services (ES) of the three main categories (regulation, provisioning and cultural).

The lower potential to provide ecosystem services is generally associated with the anthropogenic land cover types (except some cultural ecosystem services), while agricultural and forest or seminatural areas show high potential in the regulation and provisioning ecosystem services. The two divisions characterised by the highest artificialized areas present the lowest values of ES (Mariñas Coruñesas and Rías Baixas).

Local Climate Regulation and Erosion Regulation showed the highest values within the regulation services whereas Wild Foods and Resources and Wood Fuel scored high in provisioning services. Recreation and Tourism, Landscape aesthetics and Inspiration and Knowledge Systems were the most important cultural services. Regulation and provisioning services were largely associated with forest and seminatural areas (the predominant land cover class in the study area, 65%). On the contrary, all the land cover classes contributed to cultural services.

The potential to deliver ecosystem services showed a temporal declining trend in the study period, showing annual decreasing rates of -0.13% (provisioning services), -0.05% (regulation services) and -0.003% (cultural services). The highest annual loss rates corresponded with two provisioning services, crops (-0.35%) and livestock (-0.32%), and it is related to the decreased by 3% of the agricultural areas.

Ecosystem services have been presented as a useful tool to assess de parallel evolution between natural and human processes (Burkhard et al., 2009Burkhard, B., Kroll, F., Müller, F. & Windhorst, W. (2009). Landscapes’ capacities to provide ecosystem services - A concept for land-cover based assessments. Landscape Online, 15, 1-22. https://doi.org/10.3097/LO.200915 ; Hasan et al., 2020Hasan, S., Shi, W. & Zhu, X. (2020). Impact of land use land cover changes on ecosystem service value - A case study of Guangdong, Hong Kong, and Macao in South China. PLoS ONE, 15, e0231259. https://doi.org/10.1371/journal.pone.0231259 ). This study revealed that land cover changes produced in the study area had a direct effect on ecosystem services, as previously detected by other authors (Burkhard, Kroll, Nedkov & Müller, 2012Burkhard, B., Kroll, F., Nedkov, S. & Müller, F. (2012). Mapping ecosystem service supply, demand and budgets. Ecological Indicators, 21, 17-29. https://doi.org/10.1016/j.ecolind.2011.06.019 ; Hasan et al., 2020Hasan, S., Shi, W. & Zhu, X. (2020). Impact of land use land cover changes on ecosystem service value - A case study of Guangdong, Hong Kong, and Macao in South China. PLoS ONE, 15, e0231259. https://doi.org/10.1371/journal.pone.0231259 ; Kandziora et al., 2013Kandziora, M., Burkhard, B. & Müller, F. (2013). Mapping provisioning ecosystem services at the local scale using data of varying spatial and temporal resolution. Ecosystem Services, 4, 47-59. https://doi.org/10.1016/j.ecoser.2013.04.001 ; Yirsaw, Wu, Temesgen & Bekele, 2016Yirsaw, E., Wu, W., Temesgen, H. & Bekele, B. (2016). Effect of temporal land use / land cover changes on ecosystem services value in coastal area of China: the case of Su-Xi-Chang region. Applied ecology and environmental research, 14, 409-422. http://dx.doi.org/10.15666/aeer/1403_409422 ). Therefore, it would be desirable to integrate the ecosystem services approach into land use planning as it demonstrated to be a useful tool for environmental sustainable management or (Baral, Keenan, Fox, Stork & Kasel, 2013Baral, H., Keenan, R.J., Fox, J.C., Stork, N.E. & Kasel, S. (2013). Spatial assessment of ecosystem goods and services in complex production landscapes: A case study from south-eastern Australia. Ecological Complexity, 13, 35-45. https://doi.org/10.1016/j.ecocom.2012.11.001 ; Bastian et al., 2012Bastian, O., Haase, D. & Grunewald, K. (2012). Ecosystem properties, potentials and services - The EPPS conceptual framework and an urban application example. Ecological Indicators, 21, 7-16. https://doi.org/10.1016/j.ecolind.2011.03.014 ; Burkhard et al., 2014Burkhard, B., Kandziora, M., Hou, Y. & Müller, F. (2014). Ecosystem service potentials, flows and demands-concepts for spatial localisation, indication and quantification. Landscape Online, 34, 1-32. https://doi.org/10.3097/LO.201434 ; Müller and Burkhard, 2007Müller, F. & Burkhard, B. (2007). An ecosystem based framework to link landscape structures, functions and services. In Ü. Mander, H. Wiggering, & K. Helming (Eds.), Multifunctional Land Use: Meeting Future Demands for Landscape Goods and Services (pp. 37-63). Heidelberg, Germany: Springer. https://doi.org/10.1007/978-3-540-36763-5_3 ; Vihervaara, Kumpula, Tanskanen & Burkhard, 2010Vihervaara, P., Kumpula, T., Tanskanen, A. & Burkhard, B. (2010). Ecosystem services-A tool for sustainable management of human-environment systems. Case study Finnish Forest Lapland. Ecological Complexity, 7, 410-420. https://doi.org/10.1016/j.ecocom.2009.12.002 ).

3.3. Relationship between artificialization rates and demographic and economic variables

In Figure 10, the temporal evolution of the population, artificial area and number of new buildings in the coastal municipalities of the Galician coast is shown.

The three variables show a sharp increase from 2005 to 2009, when the registration of new buildings was 6-fold higher than in the rest of the periods studied. Population increased until 2011, then decreased in the last period, although the number of new buildings was similar and, consequently, the artificialized area still increased with respect to the previous period. The number of hectares classified as artificial surfaces per capita raised continuously from 322 m² of artificial surface per inhabitant in 2005 to 343 m² in 2014.

Galicia showed a negative demographic growth (-0.06% year^-1) in the period 2005-2014. However, in the coastal municipalities, the temporal evolution of this variable was positive (+0.10% year^-1). Spanish population growth in the period 1991-2011 was positive (1.02% year^-1) and slightly lower (0.87% year^-1) when the 10 km coastal zone was considered (Observatorio de la Sostenibilidad, 2016bObservatorio de la Sostenibilidad (2016b). Cambios de ocupación del suelo en la costa en España 1987-2011. Retrieved from https://www.observatoriosostenibilidad.com/informes/ ).

The domestic gross disposable income was also used in this analysis as it measures the income of the residents of a given territory used for consumption and savings operations. This indicator, therefore, may reflect if the artificialization process carried out in the territory directly reverts to the families’ economy or, on the contrary, it does not represent an improvement in economic terms for the local populations.

A correlation analysis was performed among the different variables, including artificialization (artificialized hectares excluding those dedicated to transport networks or industrial development, which represent 55% of the total artificialized area between 2005 and 2014), the variation in the number of inhabitants, the number of new buildings and the changes in the domestic gross disposable income (by municipality in the period 2005-2014). Two municipalities were excluded from the analysis: Ferrol and Arteixo due to their demographic and economic singularities (supplementary material).

A bivariate Pearson correlation was calculated to assess the relationship between the selected variables in the coastal zone and in each of the geographical divisions. The four variables are significantly correlated when all the coastal municipalities were considered (Table 5). The significant correlation between population and land cover/land use changes found in this study is consistent with the results reported in previous investigations focused on the drivers of land use and land cover (Clement and York, 2017Clement, M.T. & York, R. (2017). The asymmetric environmental consequences of population change: an exploratory county-level study of land development in the USA, 2001-2011. Population and Environment, 39, 47-68. https://doi.org/10.1007/s11111-017-0274-2 ; Kamwi et al., 2015Kamwi, J.M., Chirwa, P.W.C., Manda, S.O.M., Graz, P.F. & Kätsch, C. (2015). Livelihoods, land use and land cover change in the Zambezi Region, Namibia. Population and Environment, 37, 207-230. https://doi.org/10.1007/s11111-015-0239-2 ; Oreinstein and Hamburg, 2010Orenstein, D.E. & Hamburg, S.P. (2010). Population and pavement: population growth and land development in Israel. Population and Environment, 31, 223-254. https://doi.org/10.1007/s11111-010-0102-4 ).

Neither of the tested variables showed statistically significant correlations in the Rías Altas division. This area is characterised by the absence of large urban centers concentrating the population and economic activity and, consequently, transforming the environment. Despite this division showed the second highest averaged rates of artificialization (3.33 ha year^-1), it is not related to parallel population increases (70% of the municipalities in this division show a demographic decrease between 2005 and 2014), nor to domestic gross income variability. Several processes could account for the observed pattern, such as second home developments, which accounted for a large fraction of the total artificialization process in some municipalities or the dispersed occupation of the territory, mainly with single-family buildings, surrounded by small yards, whose designation in the SIOSE classification represents an overestimation of artificial land cover (Olazabal and Bellet, 2017Olazabal, E. & Bellet, C. (2017). Análisis de las nuevas dinámicas de urbanización en España. Su estudio a través del uso de Corine Land Cover y SIOSE [Conference paper]. XXV Congreso de la Asociación de Geográfos Españoles, Madrid, Spain. https://libros.uam.es/uam/catalog/book/558.).

Almost half of the artificialization of the Galician coast (41.5%) occurring from 2005 to 2014, derived from non-artificial land covers corresponding to the categories: “Residential agricultural settlement” (areas with mainly residential use, characterised by a dispersed settlement of buildings, accompanied by arable and/or woody crops), or “Discontinuous urban fabric”, ie. areas occupied by buildings mainly for housing associated with vegetated areas and bare surfaces, including urban areas that may be consolidated or in the process of consolidation. In both cases they do not represent crowded areas of buildings, so it is unlikely that the number of buildings could account for the whole of the artificialized area. The existence of these scattered settlements is the reason why many polygons appeared as transformations from non-artificial land cover to discontinuous urban fabric land covers, thereby yielding inconsistent results.

The slopes of the linear regressions between changes in population or new buildings and artificialization rates did not significantly differ depending on the zone (ANCOVA p = 0.852 and p = 0.467, respectively). By contrast, significant differences were found in the slope of the regression lines derived from the relationship between artificialization rates and domestic gross income in the different divisions, although the p-value resulting from the correlation between artificialization and domestic gross income (0.08) is close to the reference value (0.05), it does not seem correct to rule out the existence of significant differences (Amrhein, Greenland & McShane, 2019Amrhein, V., Greenland, S. & McShane, B. (2019). Scientists rise up against statistical significance. Nature, 567, 305-307. https://doi.org/10.1038/d41586-019-00857-9 ) since, as shown in Figure 11, the regression line corresponding to Costa da Morte shows a different pattern from the other divisions.

On average, 700 m² became artificialized per each new building and 96 m² of the territory were transformed into artificial per each new inhabitant when the 2005-2014 period was considered. These values can be compared with the data provided by the European Environment Agency, although this comparison should be made with caution since they referred to the whole country and are divided into two periods (2000-2006 and 2006-2012). According to EEA, 115 m² was artificialized in Spain per each new inhabitant in the first period and 29 m² new inhabitant^-1 in the second one, while the values for France were 149 and 126 m² new inhabitant^-1 and 438 and 426 m² new inhabitant^-1 in the case of Portugal. It is worth noting that despite the higher number of artificialized hectares per new inhabitant estimated for Spain or France, the subtle demographic growth of Portugal in this period causes the higher amount of surface assigned to each new inhabitant.

Sanxenxo, located in the Rías Baixas division, was the municipality with the highest artificialization rates (12.79 ha), and the city of Vigo showed the largest number of new buildings built, with 104 new buildings year^-1. Unlike the other divisions, in which population decreased, more than 50% of the municipalities belonging to the Rías Baixas division increased their population (2005-2014). The highest averaged domestic gross income and population are concentrated in the Mariñas Coruñesas division, although this division did not show the highest average number of new buildings per year. This observation can be explained by the patterns found in urban areas of A Coruña and Ferrol, characterised by an increase in population and domestic gross income. These municipalities form a supramunicipal area, with one of the most important industrial conglomerates in Spain, where energy, construction, shipyards, clothing and metallurgy industries are concentrated. Both divisions (Rías Baixas and Mariñas Coruñesas) present very similar patterns, as shown in Figure 11, and as evidenced by the correlation analyses discussed above.

The relationships between artificialization and domestic gross income present different patterns in the divisions considered in this study. As noted earlier, Mariñas Coruñesas and Rías Baixas present similar results, with 1-2 m² of territory required to increase the domestic gross disposable income by one unit (thousand of €). On the contrary, the division Costa da Morte, characterised by decreasing population and domestic gross income values, required 12 m² to achieve the same goal. Two municipalities under the influence of the urban area of A Coruña, A Laracha and Carballo, stand out with respect to the others, concentrating the largest increases both in inhabitants and domestic gross income and presenting artificialization rates 2 to 3-fold higher than the average of the division, respectively. In the case of this division, the slopes of the regression lines between artificialization and population or income were positive. Most of the municipalities of this division showed increases in artificialization and new buildings that are not associated with parallel increases in population or income. This division contains 5 of the 6 most extensive municipalities of the Galician coastal zone, with a high percentage of non-coastal territory and with a markedly disseminated settlement, ie. the dominant land cover class is the “discontinuous urban fabric”, which as mentioned above, may involve an overestimation of the artificial surface resulting from the SIOSE classification. It should also be noted that, unlike the other two divisions, Costa da Morte has not any urban aggregate with the potential for development and growth, thus constraining population growth and economic development of this area.

The results obtained in this study show that land take occurring in the Galician coastal zone, mainly at the expense of forest and seminatural territories, affected the potential services provided by ecosystems but was not generally related to demographic or domestic economic drivers.