This paper aims to inform the scientific community of preliminary developments and of some research methodologies that have been elaborated as part of an ongoinggg scientific project financed by the Urban Planning and Housing Policies Department of the Tuscany Region. The objective of this research is to test innovative methodologies for elaborating of the geographical data required for territorial governance. This paper will reflect on techniques for monitoring and evaluating transformations in land use/cover and how this information can be used to verify changes in the provision of Ecosystem Services.

Ecosystem services (ESs) are generally defined as the benefits, direct or indirect, that people can obtain from ecosystems ( (Groot et al., 2002);(Costanza et al., 1997).

In the last twenty years, and especially since 2005, the year in which the second Millennium Assessment study was carried out, research into the classification of ESs and their spatial mapping has greatly increased. In Europe, the process of spatial identification and ESs evaluation was significantly boosted following the definition of the European Biodiversity Strategy to 2020: the objective of Target number 2 was to maintain and improve ecosystems and related services by creating green infrastructure and restoring at least 15% of degraded ecosystems (Maes et al., 2016). Furthermore, in the context of Action n. 5 of this Strategy, Member States, supported by the MAES working group, undertook to map and evaluate the ecosystem services in their countries by 2014.

An interesting initiative to map ecosystem is attributable to the formulation of the Ecosystem Map of Italy (Blasi et al., 2017) and the monetary values assigned to some ESs contained in the Third Report on the State of Natural Capital in Italy. The Ecosystem Map of Italy identified ecosystems on a scale of 1:100,000 using a system whereby data from the Corine Land Cover project was connected with an archive of potential vegetation. More specifically, the map was developed with a methodology that used data from the Corine Land Cover project up until level 4 for classes 31x and 32x the data of the potential vegetation classes. This methodology made it possible to identify the typologies of ecosystems, but it did not push towards specific elaborations to identify ESs.

The second project used a completely different methodology which simplified the ecosystem macro-categories and pushed towards a spatial assessment of the monetary value of some ecosystem services at the 1:100,000 scale: recreational, crop pollination, water supply, flood risk regulation (Notte et al., 2017).

A survey of the scientific literature reveals that in many studies the mapping of ESs has been carried out with proxy methods (Chan et al., 2006)(Egoh et al., 2008), Naidoo et al., 2008;(Eigenbrod et al., 2010), using data prepared in other contexts of research to compensate for the lack of primary data (Maes et al., 2012). While provisioning ESs can often be pg. 26 directly quantified with primary data, most regulating, supporting, and cultural services are less straightforward to be put on maps and researchers must rely on proxies for their quantification. Land Use and Land Cover (LULC) map are the most common proxy data used, but also environmental indicators such as the evaluation of nitrogen concentration for the measurement of the ES of water purification (Grizzetti et al., 2008)(Grizzetti et al., 2012), or the distribution of animal species for the evaluation of supporting ESs (Eigenbrod et al., 2010)) are used.

In Italy a similar methodology has been adopted for the first time to test the role of protected areas to prevent loss of ES with land transformation(Scolozzi & Geneletti, 2012),(Scolozzi et al., 2014).

The availability of LULC themes, readily obtainable from satellite image processing, together with the possibility of quantifying changes in land use over time, has encouraged many researchers to develop ESs mapping methodologies which, over time, have evolved at national, regional and local levels (Eigenbrod et al., 2010).

Many of these studies avoid a monetary evaluation of ESs, adopting a comparative approach. In this latter case, for each land-use class are assigned scores in accordance with a predefined scale of values measuring the supply capacity of each ES. Scores are assigned on the base of ratings provided by experts from related disciplines regarding the different ES categories (or groups); the scores are then used to evaluate the evolution over time of the performance of an ES within a given geographical area (Burkhard et al., 2009)(Burkhard et al., 2012).

This paper presents the result of preliminary ESs mapping trials that were conducted on the Tuscany territory, using the methodology proposed by Burkhard in his studies on the landscape of central-eastern Germany (Burkhard et al., 2009)(Burkhard et al., 2012) and used by many studies at European and international level (Campagne et al., 2020).

With the aim of evaluating the capacity of the different types of land cover to provide ecosystem services, Burkhard created a relation matrix with on the columns, the 44 types of land cover encoded by the Corine Land Classification Cover and on rows 29 ecosystem services: 7 ESs able to measure ecological integrity, 11 ESs aimed at measuring the provisioning services, 9 ESs for the regulating services, 2 ESs for cultural services.

In Burkard's work concerning ecological integrity service (corresponding to the support services as defined by MEA, 2005), reference was made to Muller, 2005, while, for the provisioning, regulating and cultural services reference to the works of de Groot, 2006; MEA, 2005; Costanza, 1997.

Therefore, each cell of the matrix reports a value (along a scale from 0 to 5 (0 = no relevant capacity of the land cover type to provide this particular ecosystem service, 1 = low relevant capacity, 2 = relevant capacity, 3 = medium relevant capacity, 4 = high relevant capacity and 5 = very high relevant capacity) which expresses the ability of each LULC class to provide a specific ecosystem service. In Burkhard's works, the method used for the attribution of scores is that of consulting with competent experts with respect to the different categories of ESs.

Once the capacity matrix has been built (Table 2), using GIS software, it is associated with the LULC database, for understanding the spatial distribution of service supply and for evaluating the dynamics on different time scales.

The twenty-nine indicators proposed by Burkhard have been grouped into four categories, according to TEEB, 2008: ecological integrity (supporting services), provisioning services, regulating services, cultural services. The score of the four main categories of ESs was obtained as the sum of the scores of each ES belonging to the specific category. Some values have been modified to adapt them to the LULC legend of Tuscany Region which have some difference from original CLC classification (Table 2). The total scores of the four ESs categories were subsequently normalized in the range from 0 to 1 so that the measurements of the four categories of ecosystem services would be comparable.

To calculate the ecosystem services capacity of the Tuscan territory, the research associated to each of the polygons in the database (for each of the four ESs categories) a score equal to the normalized value between 0 and 1 of the single LUCL class, multiplied by the area of the polygon.

The accuracy of the ESs supply is very dependent on the cartographic detail of the data used and its ability to recognize the presence of natural ecosystems that perform ecological functions. Small cartographic scales (eg. 1: 100.000) are more affected by the presence of polygons including too heterogeneous areas within them, failing to distinguish the presence of permeable areas and arboreal and shrubby vegetation. This capacity due to the greater detail are very important especially in the areas surrounding the small Tuscan cities, often characterized by heterogeneity of use and by the permanence of agricultural cultivations.

The limits of the cartographic scale are particularly important in proxy methods (Eigenbrod et al., 2010); the availability of a medium-scale LULC cartography for the Tuscan territory considerably improves the evaluation of the distribution of ESs, partly improving the capability of this data to function as a proxy (Figure 2).

The analysis was carried out by choosing the Landscape Systems as the territorial unit (Rossi et al., 1994) (Figure 3). The Landscape System Map is a classification of the landscape of Tuscany created using digital data on altitude, geology and geomorphology, climate and land use. The method is similar to that used in other existing landscape classification maps (Amadei et al., 2003), (Mücher et al., 2010). The Landscape System Map divides the region in 9 landscape system, within which a further 82 landscape sub-systems are articulated and is often used in environmental impact assessment, protection and management of natural resources and territorial planning policies.

Table 2 shows the capacity matrix used in our research. The assessment scale reaches from 0 / red = no relevant capacity of the land cover type to provide this particular ecosystem service, 1 / orange = low relevant capacity, 2 /yellow = relevant capacity, 3 / light green = medium relevant capacity, 4 / medium green = high relevant capacity and 5 / dark green = very high relevant capacity. Columns with gray headers give sums for the four ecosystem services groups.

The table differ from original one builted by Burkhard, 2009 in some codes of UCS: 1121 (land of relevance, scattered buildings), 1211 (urban wastewater), 1212 (photovoltaic system), 1221 (roads in forest areas), 1411 (cemetery), 210 (irrigated ed non irrigated arable land), 2101 (greenhouse), 2102 (plant nurseries), 2221 (agro-forestry areas), 3331 (firebreaks).

Classes 1121, 1211, 1212, 1221, 1411, 2221, 3331 were assigned the value of the class at the third level of classification; the UCS of the Tuscany Region does not contain, on the other hand, the items of UCS 211 (non-irrigated arable land) and 212 (permanently irrigated land): both these classes are grouped under an item called 210. Codes 2101 (greenhouses) have been attributed the value of class 121 while the codes 2102 (plant nurseries) have been given the same value used for the agro-forestry areas.

In this paragraph we analyse the results of the methodology described above which made it possible to spatialize and quantify the ecosystem services provided by the territory of Tuscany in 2016.

The maps in Figure 4 show the normalized values (scale 0 - 1) of the ESs supply of each UCS polygon and refer to the 4 main categories (support, provisioning, regulating, cultural). Table 2 shows the contribution (as a percentage of the total) of each type of landscape to the supply of the 4 categories of ESs. This allows us to evaluate the contribution that each sub-region offers in the provision of services on a regional scale.

The Apennines (AP), the Anti-Apennine Reliefs (RA) and the Pliocene Hills (CP) are the areas that give the greatest contribution to the supply of the four ESs categories while the Alluvial Plain (PA) and the Coastal Plain (PC) which are the most urbanized areas, contribute in a more limited way.

It should be noted that if, on the one hand, the territorial surface of the different Landscape Systems shows values similar to those of the ESs offer, on the other hand, the most anthropized Landscape Systems (PA, PC and CP) show lower values of the ESs offer than to the surface and more rural and mountainous Landscape Systems (AP, RA, AA). Looking at the contributions of the different Landscape Systems to the four categories of ESs, the Apennines (PA), the Anti-Apennine Reliefs (RA) and the Apuan Alps (AA) highlight the highest values of regulating and cultural services offer while Pliocene Hils (CP), Alluvional Plains (PA) and Coastal Plains (PC) show higher values for the supply of provisioning services and ecological integrity.

Many studies visualize and quantify changes over time and space of ESs supply ((Carreño et al., 2012); (Bürgi et al., 2015); (Ncube et al., 2018)). Also in our research, having available the LULC 2007-2016 data of the Tuscany Region, we have produced maps of the evolution of the different ESs for the entire territory of Tuscany using the capacity matrix developed by Burkhard (Table 2). The comparison was made taking into consideration the years 2010 and 2016, because the first cartographic survey (2007) is more affected by photointerpretation errors than the others. The capacity matrix developed by Burkard (Table 2) was linked to the 2010 and 2016 land use and land cover classes contained in the 2010 and 2016 LULC database: Table 4 shows the absolute and percentage change in the supply capacity of ESs in the Tuscan territory in the period 2010/2016 for the four categories of ESs (support service; provisioning services; regulating service; cultural service). The data (Table 4) reveal a general reduction of the four , with a higher value for the provisioning services and with lower values for the regulatory and cultural services. The quantities of the support services, on the other hand, remain substantially unchanged.

With the 2007/2016 UCS database linked to the capacity matrix, ESs losses or increases in the period 2010-2016 were determined, based on UCS transitions. Below are some summary tables related to changes in UCS at the first level of the Corine Land Cover classification. Table 5 shows the internal transitions at the same level (e.g. Code 141 in UCS 2010 which changes the LULC class to code 112 in UCS 2016), while Table 5, Table 6, Table 7, Table 8 show the variations between the different UCS levels ( e.g. code 311 in UCS 2010 which changes the LULC class to a code 223 in UCS 2016).The result of the query of the UCS 2007/2016 database, linked to the capacity matrix, allows to recognize the losses or increases of ESs in the period 2010-2016, according to the specific UCS transitions.

Considering the four categories of ESs, in the period 2010-2016, the most significant reduction of ESs is given by the passage of some portions of the Tuscan territory from level 2 (agricultural areas) to level 1 (artificial areas). Compared to the total, the contribution of this transition to the loss of supply of ESs was, respectively: -68.9% for support services (ecological integrity) (Table 6); -94.1% for provisioning services (Table 7); -26.6% for regulatory services (Table 8) and -37.8% for cultural services (Table 9). This loss is partially compensated by a reverse transition from level 1 to level 2 (from artificial to agricultural surfaces); these transitions are mainly due to the presence of areas of relevance for construction for public and infrastructural works in 2010 (code 133) which, in 2016, return to arable land or, in any case, returned to agricultural use. These compensations, while decreasing the loss of ESs, nevertheless lead to an overall balance of these transformations very negative values equal to: -44.3% for ecological integrity services (Table 6), -74.4% for provisioning services (Table 7), -21.1% for regulatory services (Table 8) and -30.8% for cultural services (Table 9).

Considering all the transitions towards codes 1 as phenomena of land consumption, it can be said that the reduction of support and provisioning services is largely due to the processes of artificialization of the soil. With regard to regulating services and cultural services, both the phenomena of land consumption and the transitions from natural surfaces to agricultural surfaces play an important role in the loss of ESs with values equal to -32.6% and -31.6% (Table 8, Table 9). The analysis of the data also shows the contribution provided by the internal transitions to class 2 (agricultural land) to the reduction of the loss of supply capacity of ESs in the Tuscan territory (Table 5). This is evident, in particular, for provisioning (+ 12.34%) and regulating (+ 9.56%) services, while cultural services recorded a reduction (-7.23%). A contribution determined by the transition from permanent crops (especially olive groves and orchards) to arable land and / or meadows and pastures which, in fact, determine an increase in the ESs of provisioning and regulating, but reduce the supply of cultural services (probably linked to the aesthetic-perceptive of the landscape).

Burkard's methodology, applied at LULC data of the Tuscany Region, was used to develop easy-to-read maps, to identify macro-phenomena of erosion of ESs supply and to identify possible conflicts and limits to managing environmental resources with particular reference to land consumption. Thanks to the availability of homogeneous data for the Tuscany Region, the methodology is easy to use, quickly provides results, ensures the replicability of the process and allows to verify the correlation between the supply of ecosystem services and the incidence of anthropic activity described in the LULC archive.

However there are also some critical issues: one is related to the difficulties in correlating the delivery level of some ecosystem services to single LULC classes, another to the lack of descriptive detail for some LULC classes, both in natural and urban areas, which makes it difficult to adequate evaluation of the ecosystem services provided by these areas.

With regard to the first issue, there is a pressing need for further research to be undertaken on the opportunities offered by combining the use of the LULC data with other geographical data of the Tuscany Region in order to increase the reliability of the evaluation of the ecosystem services offer that are not closely correlated with the LULC data base.

The possibility of using the Tuscany Region LULC archive to undertake multi-temporal measurements (2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016), of the variation in ESs supply is a very promising area of research that could lead to the introduction of payment mechanisms for ecosystem services (PES) (Pagiola & Platais, 2006) between different territorial areas, especially in light of the opportunities provided by Law 22/2015 for promoting the green economy and limiting the excessive use of resources: in fact Art. 70 of this law envisages the establishment of systems of remuneration for ecosystem and environmental services (PSEA). The adoption of expeditious standardized methods such as the one proposed in this paper would make it possible to rapidly assess the effectiveness of urban, territorial and sectorial policies and to introduce, compensation mechanisms where required.

The introduction of PSEA between territories requires identifying both the places where the services are generated and where they are used, so that the environmental economic balance is oriented towards sustainability principles. In this approach the maps of ESs supply capacities elaborated according to the illustrated methodology and the correlation of these with demographic data related to the population, can help to identify the spatial congruence or the disparities between supply, flow and demand of ESs ((Santolini & Morri, 2017); Morri et al. 2014).