- 1 See Agnoletti (2014) for various definitions.
1A cultural landscape can be considered a combination of anthropogenic and natural spatial structures, elements, and processes arranged by humans in a way that best suits their demands.1 For millennia, humans have been shaping different types of landscapes while seeking to survive on the available resources. Early activities such as hunting and gathering had only a small influence on the natural processes in landscapes. With the beginning of agriculture, the surroundings of settlements experienced profound changes on the landscape. Over time, organized agricultural landscapes were extended intensively in some regions of the earth – primarily at the expense of woodland.
- 2 Retrieved online from Eurostat statistics on 2017/05/17: http://ec.europa.eu/eurostat/statistics- (...)
- 3 Retrieved online from FAOstat on 2017/05/17: http://www.fao.org/faostat/en/#data/EL
2In densely populated Europe, cultural groups have had a strong relationship with agricultural and silvicultural activities for millennia. Today, 44% of Europe’s surface is agricultural land and 33% are covered by forests.2 In Portugal 75% of the total land area is considered cultural landscape (40% agricultural area and 35% forest area).3 Practices of farmers are the main drivers of landscape modifications. These include land management techniques, specialization, the degree of mechanization and modernization combined with market demands and regulations imposed by the EU’s Common Agricultural Policy. Farmers determine the visual aspect of the countryside and its amenity value by modifying the configuration and composition of cultural landscapes by addressing “the spatial patterns of the landscape; individual elements and the way in which they are organized” (Moreira 2006). Vos and Meekes (1999) argue that in bulk-production oriented agriculture, “any form of nature or scenery is an unintended by-product of agriculture”. Compared to recent structures, traditional agriculture has been characterized by considerably smaller sizes of farms and fields and high diversity regarding land use, crops, animal holding and landscape structures. This enormous variety, often combined with polyculture, resulted in fine-grained landscape mosaics with a high biodiversity. Altieri (1999) mentions that the “species richness of all biotic components of traditional agro-ecosystems is comparable with that of many natural ecosystems”.
- 4 See Ales et al. (1992), Vos & Meekes (1999), and Antrop (2005).
3Over the past decades, two processes have mainly threatened traditional agriculture in Europe: (a) the dramatic intensification and mostly insane regulation of agriculture in highly productive areas and (b) the abandonment of any land use in marginally productive areas.4 The mountainous landscapes of Mediterranean Europe in particular have drastically been affected by the ongoing process of land abandonment (MacDonald 2000; Agnoletti 2014). Several investigations of traditional cultural landscapes, mainly focusing on the shift of spatial patterns of human-landscape interactions and their effects, have recently been conducted in southern European countries. For instance, Tuscany (Italy), with its unique, widely terraced landscapes, has been studied intensely by Agnoletti (2007, 2015). Additionally, Lasanta et al. (2001) studied the degradation of abandoned terraced landscapes in Camero Viejo, La Rioja province (Spain). Ales et al. (1992) studied landscape dynamics in the Guadalquivir valley in southwest Spain. And finally, in Antikythera (Greece), Bevan et al. (2013) investigated the long-term ecology of terrace fields.
4In Portugal, various authors studied land use changes linked to the reduction of traditional agriculture either due to abandonment or modernization. In Alentejo (Southern Portugal), the effects of the abandonment of traditional wood pasture with pig husbandry in cork oak forests have been studied by Teresa Correia (1993). Cavaco & Marques (1966) studied the traditional agriculture in Loriga and Alvoco valleys in Serra da Estrela in detail, and Lourenço (Lourenço & Fernandes 2012; Lourenço et al. 2005-2006) studied the effects of agricultural terraces as prevention for natural disasters. In the mountainous, rural landscape of Montalegre (Northeast Portugal), Pôças et al. (2011) analyzed landscape changes and their driving forces. In 2003, the agricultural abandonment in Portugal was quantified, the driving forces evaluated, and presented in a technical report for the ministry of agriculture and the ministry of environment (Alves et al. 2003).
- 5 The Global Landscapes Forum is the world’s largest and only science-led multi-sectoral platform d (...)
- 6 Available at http://www.fao.org/giahs/background/goal-and-objectives/en/.
5In these studies, the driving forces behind the prevailing trends in abandonment of terraced mountain areas and pastoralism activities were analyzed. They proved that emigration and dislocation into economically stronger areas were of a similarly high importance in most southern European mountain areas. These population movements lead to an over aging of rural populations and put the century-long developed knowledge on sustainable farming in marginally productive and often environmentally sensitive areas at risk. Lately, the degradation and vanishing of traditional agricultural land use systems have been identified as a highly precarious development by many internationally recognized organizations such as Global Landscapes Forum, Council of Europe, and UNESCO.5 The need to maintain these valuable landscapes with all their functions and services has also become a central focus of FAO through the GIAHS programme,6 whose overall goal is:
To identify and safeguard Globally Important Agricultural Heritage Systems and their associated landscapes, agricultural biodiversity and knowledge systems through catalyzing and establishing a long-term programme to support such systems and enhance global, national and local benefits derived through their dynamic conservation, sustainable management and enhanced viability.
6To fully understand a traditional landscape with its structures, features, processes, functions and services, the history of the interactions between the local population and the landscape has to be studied. Over time, the intensity of the interactions between humans and their land shifted. The demands of societies have significantly shifted over centuries. Commodities produced on the soils have been altered. The way of working the soil (including the various tools) has been adapted to these changes and to the specific soil conditions of the agricultural land at each settlement over time. Farmers have introduced technologies to improve the given conditions: for example, extended irrigation systems.
7Studying these complex systems is challenging since every small region has its own way of shaping the land surface, often adapted to a certain product being favored and to varying physical conditions. These include factors such as regional climate, microclimate, soil type and quality, water availability, slope gradient, solar and wind exposure, and altitude above sea level, among others. As a result, farming communities created a strong identification with the individual landscape. Therefore a profound scientific analysis of the different mosaics of landscapes requires detailed studies of the individual pieces forming these mosaics to assure an in-depth understanding of the area in question.
8These detailed, historical studies are very important for our society since they are the only sound basis for political decisions concerning the possible future development of a landscape. Maintaining or even restoring landscape elements requires a certain amount of resources. To reasonably decide which value should be attributed to a specific landscape, a society needs to be informed about (1) the functions and services provided, (2) the environmental quality safeguarded by traditional practices in agriculture, forestry and pastoralism, and (3) the quality of the agricultural products. The historical formation of a landscape and the art of its transformation, for example terracing in mountainous areas, and the effects of landscape elements are key information to maintain the present elements and work toward a possible continuation of the valued landscape.
9In the case of the Portuguese Island of Madeira, little is known about construction, spatio-temporal development, use, and maintenance of agricultural terraces and soil quality of the created anthroposols. These key topics are investigated in this paper. In the first section the research area, and the research topics are explained; the second section focuses on the applied research methods, the results are presented in section 3, and then the results are discussed in section 4.
10During the fifteenth and sixteenth centuries, huge parts of the Portuguese Island of Madeira [32°23´N – 33°07´N, 16°15´W – 17°15´W] were transformed into a cultural landscape. The most eminent, visible feature of this change is countless agricultural dry-stone terraces which have been laboriously installed on almost every mountain slope from sea level up to an altitude of around 600 m a.s.l. The mid twentieth century Portuguese agronomist Joaquim Vieira Natividade described the terrace-building this way:
And the man, the dwarf, attacked the mountain. During centuries, he did not stop the hard work with pickax and lever, and on the price of lives, sweat and blood carved gigantic stairs into the mountain without that the steep escarpments, the profundity of the chasms or the vertigo of the abysms stopped the steps of the giant. This monument [is] unique in the world, because in no other part, with this big amplitude, so much human force was used to conquer [cultivate] the land. (Natividade 1954, 31)
- 7 These are relatively even accumulations of material along the coastline that resulted from coasta (...)
11In this quotation, Vieira Natividade clearly describes how adverse the physical conditions of Madeira’s mountain slopes used to be when farmers built new dry-stone terraces. This dangerous and labor-intensive way of creating flattish surfaces for crop cultivation had to be applied almost all over the island, with very few exceptions, such as the socalled fajãs.7 The latter have often been used to initialize agricultural production in the early phase of settlement and as preferred locations for living. Any agricultural activity extending beyond these small fajãs had to be practiced on steep mountain ridges characterized by an outstandingly rugged morphology. As a consequence, generations of farmers over several centuries created unique, interlinked systems of dry-stone terraces, locally called poios, to grow food for living and crops for sale.
- 8 See Natividade (1954) for some descriptions on terrace construction, and Martins (2017, 156ff.) f (...)
12The terraced landscape of Madeira has surprised visitors from the early days of scientific travel to present day tourism. There is hardly any guidebook on the island that does not mention these specific landscape elements and their remarkable beauty. Nevertheless, there is very little exact information available in scientific reports stating which terrace construction measures were applied and with which materials the terrace bodies and the supporting walls were built. The temporal development of the terrace systems has merely been described in general terms and no investigations on the exact temporal frame of the extension of terraces in a specific place have been conducted.8 Furthermore, the effects traditional agricultural practices have had on soil fertility are mostly unknown. Madeira et al. (2004, 29) have analyzed the C-org content in various types of Madeiran soils, but have not analysed anthroposols, e.g. terrace soils.
- 9 Locally the three hamlets of Primeira Lombada, Segunda Lombada and Terceira Lombada are often ref (...)
13Our study aims to identify (1) the spatio-temporal development of the terraced landscape and (2) the effects of traditional agricultural methods, applied over centuries, on soil quality parameters in an exemplary parish in the north of Madeira. The coastal parish of Ponta Delgada, located in the north-eastern part of the municipality of São Vicente, was selected for this study. Ponta Delgada consists of the parish itself, the place Lombinho, and the hamlets of Primeira Lombada, Segunda Lombada and Terceira Lombada. While the main settlement and Lombinho are located at altitudes ranging from sea level to approximately 120 m a.s.l., the Lombadas9 are situated on an elevated plateau ranging from 280 m to 350 m a.s.l. The settlements are circumscribed in the south by mountains rising steeply to elevations of 1500 m a.s.l.
- 10 The term sesmeiro signifies a person who was granted a certain amount of uncultivated land for im (...)
- 11 See Leça (2011), Nepomuceno (1994), and Veríssimo (2014).
14The main economic activities of Ponta Delgada have long been agriculture and forestry; even fishery played only a minor role (Vieira 1997). In 1466 or 1469, Manuel Afonso da Sanha received land as sesmeiro10 and thus started the process of clearing land and initializing cultivation in the area of Ponta Delgada. With its relatively even surface, the headland of the present village of Ponta Delgada was perfectly suited for agricultural use. In the first decades after colonization, the most important crops were cereals, particularly wheat. The production was sufficient to satisfy the local needs and an even larger part was available for export to mainland Portugal.11
- 12 See Pereira (1989), Vieira (2015), and Quintal (2013).
15With the introduction of sugar cane (Saccharum officinarum) in Madeira in the mid fifteenth century, the foundation for a highly profitable sugar cane industry was laid (Vieira 1993, 2004b). This profitable cash crop quickly occupied huge areas of agricultural land in areas at low altitudes, predominantly on the southern side of the island, up to a maximum altitude of 400 m a.s.l. (Vieira 2004a; 2004b; 1993, 12). Climatically, this region was most suitable for cane cultivation, providing subtropical conditions for optimal growth. To guarantee a constant supply of water for the cane fields, irrigation channels were built, leading water from the more humid zones higher in the mountains to the cropping areas. Locally these irrigation channels are called levadas. In addition to their agricultural function, they became an important part of the local culture (Santos & Correia 2017). The early constructions of levadas extended over short distances of only a few kilometers often collecting water from streams in higher mountain areas. Over time, the construction of levadas became more and more sophisticated. Most modern levadas collect water in the humid northern parts of the island and transport it to the south. Some of these levadas extend over several kilometers, like Levada do Norte (62 km), Levada Nova da Calheta (63 km) or Levada dos Tornos (nearly 105 km).12
Figure 1: Map of the research area
Source: Kiesow (2017).
- 13 On the history of wine production in Madeira see, inter alia, Hancock (2009), Pereira (1989), Rib (...)
16With the extension of sugar cane plantations, more and more of the limited area suitable for agricultural production was used for the cultivation of this cash crop. Food production successively moved towards higher altitudes, where the more profitable sugar cane would not be cost-effective. The extension of cash crop production implied further terracing of mountainous slopes and was a main driving force for the enhanced shaping of the mountain landscape. This process continued even after the decline of sugar cane production in the first quarter of the sixteenth century, when in many areas of the island cane fields were replaced by vineyards (Vieira 1993). Until present days these are still an important part of the agro economy of Madeira Island.13
- 14 Ponta Delgada used to be referred to as “Corte do Norte” (noble residence in the north) because m (...)
17In Ponta Delgada, the extension of terraces into the mountains started later than in the villages along the south coast of Madeira. It is reported that by the end of the fifteenth century and the beginning of the sixteenth century cultivation of sugar cane gained importance in the agro-economy of the northern villages. The spatial extension of sugar cane plantations was limited by less favorable climatic conditions, offering profitable growing conditions solely up to altitudes of approximately 200 m a.s.l. (Vieira 1993). Nevertheless, the territory of Ponta Delgada did not stay unaffected by the excessive monoculture of sugar cane cultivation taking place on the southern side of the island. The most productive agricultural land on the island was occupied by a merely export-oriented crop and the demand for food production had to be satisfied by production in other areas. The parish of Ponta Delgada with its vast, relatively even ridges of the Lombadas had the spatial possibilities to extend its agricultural area into the mountains and it was the preferred place of residence for several landlords from other parts of the island, who knew about the necessity of turning more natural landscape into a food-producing cultural landscape.14
18The process of clearing woodlands in higher altitudes by chopping trees and burning the smaller remains, like thin branches, on the spot of clearance was extremely labor intense, since it had to be done manually with the at that time typical tool: the axe. In 1466, Prince D. Fernando, donatary of the island, determined that the clearance of woodland by using fire was no longer permitted (Veríssimo 2014). To avoid soil erosion cleared steep slopes had to be terraced immediately. Therefore, only small areas were cleared at a time. Only when the terracing work was finished in a slope segment, a further extension of the cultivated area was possible. The created arable surface on one terrace depended much on the given relief conditions of the ridge where the terrace was constructed. In Ponta Delgada area it ranged between approximately 3 to 6 m of horizontal depth and approximately 8 to 40 m of length. In places where dry–stone walls were built, all remains of the trees that grew there had to be removed before wall building started. Taking out the lowest part of the stems and the thick roots was often impossible. So farmers tried to burn and char as much of these remains as possible, usually together with the trees’ thin branches.
19This generally applied method of terrace construction created a valuable geo-archive. The charcoals that resulted from the incineration of small and medium sized wood pieces were either left on the surface or worked into the terrace soil. Some charcoal was also buried underneath the foundation of the dry–stone wall. The space between the stone wall and the deepened slope segment above was then filled with the soil that had been removed before. Even the deeper soil filling contains charcoal pieces. These charcoal pieces can be dated and used to study the temporal dynamics of terrace extension into the mountains.
- 15 The literature on the historical development of agriculture and the economic cycles of cash crops (...)
20Studying the environmental history of a specific area in detail and especially the interactions between the inhabitants and the surrounding landscape requires a multidisciplinary approach to include all available sources of information. Methods from social and natural sciences have been combined in our multidisciplinary study. Our investigations are based on the conduction of qualitative interviews, the survey of historical literature,15 as well as geomorphological, pedological and geo-archaeological methods. They are described in this section.
21In order to understand the spatio-temporal dynamics of a landscape and to conceive the motivations that drove people to interact with it, it is eminently important to access the still existing local traditional knowledge. Some old people living in remote villages like Ponta Delgada and its hamlets still have a rich traditional knowledge. Over centuries the exchange with other settlements around was limited; respectively, traditions, specific tools, working techniques and the best practices have evolved with little external influence. As to the best practices applied in subsistence agriculture, farmers improved their ways to cultivate the land over generations and passed these practices orally to the next generation. The only way of accessing this knowledge, which has never been written down, is by conducting interviews with farmers still working traditionally in the respective area – a technique that has been applied in many studies addressing human-landscape interaction (e.g. Assefa & Bork 2014, 2017).
22In this study twelve interviews have been conducted in an informal way as qualitative in-depth-interviews, each lasting approximately one to four hours, with inhabitants of Ponta Delgada, Lombinho, Primeira Lombada, Segunda Lombada and Terceira Lombada. This technique of interviewing is “preeminently suited to the gathering of the kinds of qualitative and descriptive data that are difficult or time-consuming to unearth through structured data-gathering techniques such as questionnaire surveys” (Trembley 1949). The interview partners have been selectively chosen amongst farmers and retirees, who have long-term practical experience in crop cultivation on terraced land and have applied traditional best practices on the terraces they cultivated in the past or still work on presently. This selection of informants assures a good coverage of the topic in question and permits sampling of very specialized knowledge. For example, a farmer working on his own fields cannot give detailed comparative information on yields in neighboring areas, a land overseer responsible for a larger part of land can provide this information. In the interviews cross-validation of information given by interview partners was an important issue, questions on the same topic were directed to several people in order to reduce possible bias in the provided information. Due to limited article length, we selected only the most important interview results for this publication.
23In the Lombadas of Ponta Delgada today only the agricultural terraces in the vicinity of the houses remain under cultivation; the steeper slopes and the more distant locations are semi- or fully abandoned. For a detailed analysis of (a) the spatial extension of terraces over time and (b) the fertility of the terrace soils and (c) their hydrological characteristics regarding infiltration capacity, eight terrace sites have been selected. Usually people tended to first occupy accessible areas and later, when the demand for agricultural land could not be satisfied by the cultivated terraces remote steep slopes were utilized, too. The Lombadas are characterized by very rough and steep terrain with deeply carved valleys. It is very likely that the inhabitants extended the agricultural area along the ridgelines, but such factors as existing footpath infrastructure, the slopes’ exposure, proximity to waterlines for irrigation and property issues might have had an influence on the dynamics of spatial extension of the dry-stone terraces.
- 16 For the elaboration of the maps that illustrate the topographical situation in the area of the sa (...)
24In order to uncover the development of the terraced landscape and to determine the temporal frame of the first anthropogenic big scale landscape modification, a good spatial distribution of the sample sites was of main concern. A semicircular line of sampling sites around the main valley of the Lombadas, the Ribeira da Camisa and the smaller Ribeiro do Velho was established. All major ridges presenting wider areas suitable for agriculture are crossed by this line and on each non-interconnected ridge one sample site has been positioned at altitudes ranging from 380 m to 580 m above sea level. Furthermore, on some ridges with pronounced former agricultural activity and visible differences in wall structure or material an additional sampling site has been selected. Outside the main agricultural area, the terraces at the remote sampling site Albino Miséria (880 m a.s.l.) have been included in this study to illustrate extreme positions far from the settlement. The accessibility of the particular terraced areas range from 300 m of distance and 50 m of elevation to 4000 m of distance and 430 m of elevation from the nearest road. These differences result in a necessary time investment for a one-way trip to the terraces of approximately 15 minutes at the closest site and one hour to reach the southernmost one.16
- 17 The traditional hoe used on Madeira Island is locally known as enchada. The shape of this hoe var (...)
25At all sample sites only one terrace, taken as characteristic for the corresponding terrace system, was selected for analysis. It should be noted that a terrace system is in this study considered an interconnected area of terraces, which are generally constructed on one ridge or slope and limited by orographic conditions of the terrain, such as waterlines, deep gorges, steep or vertical rock walls or other limitations, like huge boulders of several meters of height. A terrace system often has one common levada, providing water for all terraces and thus interlinking them all. In the selected terrace, a soil profile was opened manually with the help of a spade and a hoe.17 The lengths of the soil profiles vary from 2.60 m to 8.60 m and have been adapted to the given conditions of the terrace system. The height of each soil profile was determined by the height of the dry-stone wall of the respective terrace under analysis and ranged between 1.60 m to 3.00 m.
Figure 2: Location of soil profiles opened
Source: Kiesow (2017).
26The opened profiles displaying the terrace wall and the terrace soil body were studied for two main purposes: (a) to understand and document the structure and arrangement of the dry-stone wall and (b) to perform a pedological analysis of the terrace soil for evaluation of important soil quality parameters. Furthermore, the stratigraphy of the terrace soil body has been studied. All soil profiles have been documented as drawings on scale paper, showing the stratigraphy of the different soil layers, the amount and distribution of stones of various sizes and roots with more than 2 cm of diameter in each layer. After finishing the field work the drawings were digitalized.
27For the pedological analysis two types of samples were taken: (a) undisturbed samples and (b) disturbed samples. The places where undisturbed samples were taken have been documented on the drawings; disturbed samples were taken in each stratigraphic layer. The undisturbed samples have been collected with a core steel sampling ring with a volume of 100 cm³. This method prevents compaction of the soil and thus distortion of the original volume used for analysis. In this study, the bulk density of the soil was the parameter in question since it contributes to soil quality and stability of terrace systems. The bulk density has been analyzed by oven-drying the samples for 24h at 105°C. Afterwards the dried samples were cooled down and their dry weight measured. The density was calculated with the following formula: Bulk density (g/cm³) = dry weight (g) / soil volume (cm³).
28The disturbed samples in this study have been used to perform an analysis of the soil organic matter content of the terrace soils. The organic carbon content has been analyzed with a Euro EA Elemental Analyzer 3000. For this process, the samples first had to be dried, then hand ground and finally ground in a planetary mill before being introduced into the EA analyzer.
- 18 For 14C analysis the age of the material is of importance, since it can disturb the dating result (...)
29Each soil profile was searched for charcoal pieces of more than 3 mm in diameter. Their preferred location was as low in the profile as possible and ideally underneath parts of the dry-stone wall. This would indicate that most probably the wood was burnt before the respective terrace was constructed or at approximately the same time. It is very unlikely that any material containing charcoals was introduced under or into an already existing terrace wall since it would require the removal of the wall. Not at all sampling sites charcoals that would fit the necessary dimensions for 14C analysis were found. Samples could only be taken at four profiles: Lombo das Coches, Lamaceiros/Poço Feiticeiro, Caminho do Bardo 2 and Poios da Ladeira. In each of the profiles the locations, where charcoal pieces have been found, have been documented on the drawing with size of the piece and its depth. After analyzing which of the charcoal pieces were suitable for 14C analysis,18 the selected samples were brought to the AMS-14C laboratory at Kiel University, currently named Leibniz Laboratory for Radiometric Dating and Stable Isotope Research, where the analysis took place. To visualize the temporal dynamics of the terrace area extension into the mountains the data have been included in GIS Arcmap 10.4.1.
- 19 The field measurements on water infiltration have been done by the authors of this paper and Clau (...)
30Infiltration was measured in the field at three exemplary locations, which represent different types of terraces. One site for infiltration measurement was on a terrace with merely an earth wall (Lombo das Coches), the second one on a terrace with a thin dry-stone wall of only one layer of stone (Lamaceiros / Poço Feiticeiro) and the third one on a terrace with a dry-stone wall of over one meter of width (Caminho do Bardo 3). Infiltration was measured with a double ring infiltrometer according to Parr & Bertrand (1960). This device consists of two different sized plastic rings. In this case one ring had a diameter of 30 cm and the second one a diameter of 60 cm. The big ring is placed on the soil and pressed into it until approximately 15 cm of the ring are in the soil. The remaining 15 cm stay above the soil surface. The smaller ring is now placed in the center of the big one and also pressed approximately 15 cm into the soil. Before measuring, water was poured into the rings to wet the surface. The measurements were then started by filling the outer ring with water, which was kept at a stable level by refilling it during the measurements. Afterwards one liter of water was poured into the inner ring and the time until it was fully infiltrated into the soil measured. This procedure was repeated five times or until the water infiltration took longer than 20 minutes.19
31In this section, the results from field and laboratory analysis are presented and the significance of the data is discussed. Information gained in the interviews is evaluated jointly with the results from the laboratory analysis. First, one exemplary soil profile is presented, then the results from the soil quality parameters investigated in this study are shown, followed by the parameters regarding stability of the terrace system under investigation, namely bulk density and water infiltration capacity. Finally, the data from the Radio Carbon Analysis are presented and the temporal extension of agricultural terraces shown.
32The investigated agricultural terraces differ significantly in structure, material and stratigraphy. A typical dry-stone wall consists of one or more layers of stones that support the soil behind the wall (FAO n.d.). This is the most common type of supporting walls in Madeiran bench terrace systems. We have identified three different types of dry-stone walls and an earth wall, constructed without the aid of any stone material. The profile Lamaceiros/Poço Feiticeiro (see figure 2) is representative for a dry-stone terrace in terms of layering and type of soil filled behind the dry-stone wall. The comparatively thin dry-stone wall with only one layer of stones as supporting wall is similar to the other profiles investigated in terms of rock type used.
33The most common rock types used for the construction of dry-stone walls in the Lombadas area are blackish basalt and greyish mugearites and trachytes. These materials were available at or close to the majority of sites that were terraced and required only short transport distances. Mugearite and trachyte are often exposed in volcanic dykes (Ribeiro & Ramalho 2007). The dry-stone wall in the Lamaceiros/Poço Feiticeiro profile is built with basalt and trachyte stones. It measures 220 cm in height and has a thickness of approximately 70 cm. Some of the wall stones were worked to match better and to provide stability. In between the large stones, smaller ones and stone wedges have been placed to further increase stability.
Figure 3: Profile of agricultural terrace at Lamaçeiros/Poço Fetiçeiro
- 20 Soil layers have been named according to the World Reference Base for Soil Resources often referr (...)
- 21 Colors have been identified according to Munsell soil color chart with humid soil conditions in t (...)
34Farmers started the construction of this terrace with a ditch at the foundation of the dry-stone wall. When removing the original soil from the proposed area of the terrace, farmers removed most of the stones in the soil and put them behind or next to the foundation of the future wall. The lower layers within the terrace, just next to the basis, show quite a few stone deposits, while the upper two layers of soil do not have any stones or even gravel mixed in. Altogether, five different layers of soil can be distinguished in this terrace body20: At the bottom, a first dark brown layer of fine soil (Y1) was identified. Above it, a mixed layer of stones and soil (Y2) and a thin layer (Y3) without stones were deposited. The two top layers Y4 and Y5, which are most important for agricultural activities and crop growth, consist of fine soil. They are each approximately 80 cm thick and have very good structure. According to the World Reference Base for Soil Resources (FAO 1998) the terrace soil is classified as a typical Anthrosol. The dark brown color already indicates a high content of organic matter.21 Since the soil was never ploughed, no zone of compaction exists within the profile. Thus, pore space is well developed and interlinked. The soil has an extremely low specific weight and is easy to work with the typical tool used in Madeiran traditional agriculture: the hoe.
35The laboratory investigations of the disturbed samples taken from the profile Lamaceiros/Poço Feiticeiro proved what the Munsell soil color chart already indicated: the two top layers (Y4 and Y5) have very high contents of organic matter. Y4 has an organic matter content of 15.9 % and Y5 a content of 14.2 %. These values are extraordinarily high and indicate a very sustainable and a highly effective long-term soil conservation. In all soil samples analyzed, the organic matter contents are high, in some profiles even in the lower layers. Table 1 shows the C-org contents of the investigated soil layers.
Figure 4: Digitalized drawing of soil profile Lamaceiros/Poço Feticeiro
Source: graphic design by Kramer, 2017.
Table 1. Organic matter contents of terrace soil layers
|
Layers
|
Lev_F_1
|
C_B_3
|
C_B_1
|
C_B_2
|
L_P_F_1
|
L_C_1
|
P_L_1
|
|
Top layer
|
3.14
|
8.993
|
8.71
|
8.003
|
13.904
|
19.942
|
5.984
|
|
Layer 1
|
3.239
|
10.734
|
5.263
|
8.05
|
4.68
|
17.04
|
6.014
|
|
Layer 2
|
|
2.499
|
3.622
|
7.338
|
|
5.192
|
5.54
|
|
Layer 3
|
|
|
2.288
|
|
|
12.044
|
3.626
|
|
Layer 4
|
|
|
0.34
|
|
|
13.156
|
|
36The varying contents of organic matter in the investigated terrace systems are the result of different land management measures taken by the farmers, and of changing land use intensities over time. The three profiles along the footpath Caminho do Bardo (C_B_1, C_B_2 and C_B_3) have similar high C-org contents in the upper layer as a result of the application and incorporation of manure and possibly the accumulation and incorporation of litter from surrounding forests. Mineral fertilizers like ashes from burned plant residues or in recent years from chemical fertilizers are also incorporated into this layer. The low variation, ranging from 8.0 % to 9.0 % C-org, indicates comparable soil conservation and management practices.
37The soils of the terrace systems Lamaceiros/Poço Feiticeiro (L_P_F_1), Poios José Mendes (P_J_M_1) and Lombo das Coches (L_C_1) indicate that they were treated similarly. The low value of 4.7 % C-org in the lowest soil layer of P_J_M_1 is the result of a very stone rich YX-soil horizon. The terrace system of Lombo das Coches is singular in most of its characteristics; it is the only one in the entire Lombadas da Ponta Delgada area that has no stone walls securing the terrace soils. The bench terraces were constructed with a thick soil cover on top of the ridge by digging the terraces into the soil and removing soil from above to slightly lower altitudes. In a former water storage basin, approximately 100m distant from the investigated terrace, the soil atop the solid rock has a thickness of 520cm.
38The terraces of Levada da Fonte (Lev_F_1) and Poios da Ladeira (P_L_1) have the lowest C-org contents and stone rich layers. For traditional agriculture, this implied more labor input and less favorable conditions for crop growth. Probably farmers did not have the same care for soil conservation on these less productive terraces. In the case of Poios da Ladeira with its pronounced slope inclination above 20 % it is possible that soil erosion of the upper layer, which is rich in organic matter, played a role over time.
- 22 On the island of Madeira small wooden huts, rarely exceeding 10m² and mostly with peaked roof are (...)
39The high C-org contents measured in the top layers of all investigated terraces can only be explained by intensive cultivation and successfully applied soil conservation measures. In the interviews most farmers explained three remarkable traditional practices in regard to maintaining a high soil organic matter content: (1) Since every farmer kept livestock (goats, sheep or cattle) in small stables, the so called palheiros,22 there was need for bedding. The material used was rarely straw, since it was a valuable product for thatching roofs. Farmers replaced straw with common bracken (Pteridium aquilinium), which was collected on deforested high mountain slopes and brought down into the valleys. This material became part of the manure and was brought onto the terrace fields. (2) Land scarcity made fodder acquisition complicated; it competed with valuable alimentation or market crops. This pressure caused a stronger use of forest resources: small branches with leaves from sweat chestnut trees (Castanea sativa) and from lily-of-the-valley-trees (Clethra arborea) were cut and taken down the mountain to serve as leaf fodder (Kiesow & Dierssen 2017). Some small branches were used afterwards as fuel wood and the eaten parts of the fodder became manure. (3) Farmers worked as much organic material into the terraces as possible. The annual pruning of vines resulted in huge amounts of tiny branches. They were collected and placed in small ditches with depths of approximately 15-20 cm into the terraces and covered with soil afterwards. Furthermore, weeds were collected on fields and public spaces along footpaths and taken onto the terrace fields for fertilization. In Terceira Lombada it was a common practice to collect leaves from the forest floor and bring them onto the terraces as additional fertilizer.
- 23 There are detailed studies on the severity of this event by Fragoso (2012) or Nguyen et al. (2013 (...)
40All investigated terraces have a common characteristic: a remarkable aggregate stability against erosive forces. In the terrace systems collapsed dry-stone walls or signs of wall reparations that have taken place earlier on are rare. Even frequent intensive precipitation events typical for Madeira, namely the very expressed one from 201023, did not cause damage or collapse of terrace walls within the research area. In the Konso terrace area in Ethiopia for example terraces have to be rebuild every few years (Watson 2009; Assefa & Bork 2014). The high stability of the terraces is partly a result of the perfect adaptation to the locally given conditions, namely high infiltration capacities and high aggregate stabilities of the redeposited andosols.
- 24 Original Portuguese designation in soil map: “Terreno acidentado”.
41Surprisingly, the soils of the research area are not represented in the soil map of Madeira Island (Direcção Regional da Agricultura da Madeira 1992), probably due to the difficult access. The terrain of the Lombadas is generally designated as steep terrain.24 In the neighboring valleys of Boaventura and São Vicente a few agricultural areas situated on an elevation above 500 m a.s.l. have been classified as umbric andosols. It is very likely that the original soils in the Lombadas area belonged to this soil group. There is further indication for the former, natural soil type: the climo-toposequencial layering of soils has been described for Madeira Island and the predominate soil type in elevations above 400 m a.s.l. is andosol. Since most terrains are situated above 400 m a.s.l., andosols are the dominating soil type on Madeira (Ricardo et al. 2004). According to FAO andosols are common on mountainous terrain and have high organic matter contents at the surface, ranging between 8 and 30% (FAO n.d.). Another important characteristic of andosols is a predominately low bulk density, resulting from a well-developed pore system. This characteristic is also present in the anthrosols of the terraces under investigation, as can be seen in table 2.
Table 2: Bulk densities of top soil layers of terraces
|
Terrace
|
L_P_F_1
|
L_C_1
|
C_B_1
|
C_B_2
|
C_B_3
|
|
g/cm³
|
0.67
|
0.51
|
0.96
|
0.89
|
0.66
|
42The infiltration measurements done in the field resulted in outstandingly high infiltration rates, which reflects the low bulk densities and well-established pore systems. The terraces used for infiltration measurements are located at Lamaceiros/Poço Feiticeiro (L_P_F_1), Lombo das Coches (L_C_1) and Caminho do Bardo (C_B_3). Table 3 shows the infiltration values recorded at the different terraces (Fárez Roman 2015). The extraordinarily high values recorded at C_B_3 are probably due to recent root channels of shrubs and trees and do not reflect the true infiltration capacity of this former agriculturally used terrace soil.
Table 3: Average infiltration capacities
|
Terrace
|
L_P_F_1.
|
L_C_1
|
C_B_3
|
|
mm/h
|
181.22
|
394.86
|
1146.93
|
43The spatio-temporal extension of the terrace systems from the Lombadas into the mountainous woodlands was assessed by radio carbon dating from charcoals gathered from terrace profiles in the three main investigation areas. In the soil profile at Lamaceiros/Poço Feiticeiro (L_P_F_1) charcoal was found at a depth of 120cm at the base of the terrace wall. Consequently, the piece of charcoal was most probably deposited there in the initial phase of the construction of the terrace system at L_P_F_1. According to the dating results of the sample (KIA50772) the charcoal is, with a certainty of 82.7 %, from the period between 1445 and 1515 AD. This early period of construction indicates that farmers did not only take the travel distance from the houses to the fields into consideration, but were more concerned with the terrain conditions. Flattish areas with thick layers of soil were preferred to areas with a likely better solar exposure, but less favorable terrain conditions. It is quite remarkable that the ridge at L_P_F_1 was terraced in such an early period, since it is quite far from the village and the terraces are exposed to the north. A factor that might have supported the decision of cultivating this area so early is the availability of water all year round. A tributary of the main stream Ribeira da Camisa provides sufficient quantities of water to irrigate the area.
44The terraces at Caminho do Bardo are younger: The soil profile that provided the best datable charcoal was C_B_2. Two samples were analyzed, KIA50391 from a depth of 120cm below the top of the dry-stone wall and KIA50388 from a depth of 80cm. With a probability of 52.4 % KIA50772 origins from the period between 1460 and 1525 AD, with a probability of 41.5 % from the period between 1570 and 1630 AD. The analysis of KIA50388 provided more precise information, with a probability of 69.3 % this sample grew in the period from 1440 to 1525 AD. The next highest probability, with 25.6 %, is that it origins from the period between 1570 and 1635 AD. It can be concluded that the Caminho do Bardo terraces have most probably been installed slightly later than the terraces at Lamaceiros/Poço Feiticeiro.
- 25 With a probability of 26% the extracted piece of charcoal dates from between 1719 and 1778 AD and (...)
- 26 Miséria is the Portuguese word for misery.
45The third area of terraces under investigation has provided two datable pieces of charcoal: one from a depth of 140cm (KIA50998) and the other (KIA50999) from a depth of 130cm. While KIA50998 did not provide reliable results,25 the second sample KIA50999 revealed that the charcoal was, with a probability of 55.0%, produced in the period from 1805 to 1894 AD. Combining the two samples, it is very likely that the terraces at the steep Ladeira site were constructed in the nineteenth century. Although this area is closest to the village and facing east, it was terraced last - probably due to a difficult access to irrigation water and the steepness of the terrain. The southernmost terrace located at approximately 880 m a.s.l. and over an hour of walking distance from the Lombadas has been dated by oral history. Senhor Ernesto and senhor Caldeira, two informants from Primeira Lombada, remembered this terrace was built in the early 1950s by a poor farmer with the nick name Albino Miséria.26 Figure 5 illustrates the most probable ages of the investigated terrace systems according to both 14C dating and the information obtained by the interviews.
46The combination of social investigations with methods of natural science leads to a profound and detailed understanding of the development of terrace systems in the north of Madeira. 14C dating revealed important information on the temporal scale of terrace extension. The introduction of these into a GIS added the spatial component, which was refined by visits in the area and interviews with local farmers. They allowed to uncover the farmers’ motivations and priorities, when they extended the agriculturally utilized area. Recent changes, namely the pronounced abandonment of terraces and of some of the most labor intensive agricultural practices, were observed in the landscape and described by the interview partners, too. In regard to accessing traditional knowledge on how the land was managed and which practices have been applied over centuries the qualitative interviews gave deep insights that could not have been accessed otherwise.
- 27 Senhor Caldeira original statement: “Foi uma bela freguesia para passar fome no caso duma colheita (...)
47Traditional agriculture practiced in Lombadas da Ponta Delgada is an example of a nearly perfectly adapted agricultural system that has managed to maintain the terrace soils in fertile condition over several centuries of intensive land use. The maintenance of the soil’s high fertility was a prerequisite for successful farming activity. Until the end of the twentieth century the survival of farming families directly depended on a good harvest. As senhor Caldeira ironically stated: “It was a nice parish to pass famine in case of a bad harvest”.27 This situation clearly turned maintenance of good soil quality into a priority for farming families. Farmers have developed the described techniques of bringing plant residues and manure onto the terraces to improve crop growth and nutrient availability over centuries.
Figure 5: Age of terraces according to 14C-dating results
Source: Kiesow (2017).
- 28 The topic of soil water budget has been discussed in detail by Scheffer & Schachtschabel (2002).
48The traditional system was regulated so that the amount of organic material that was removed from the terraces with harvested crops, was replaced by a sufficient amount of new organic material before the next crop was planted. The high content of organic matter in the soil proves this careful treatment. The organic matter in the soil fulfills a number of beneficial functions: It stabilizes the soil and its pore system, thus making it resistant against erosive forces and allowing plant roots to penetrate the soil more easily (Johnston 2011). Furthermore, the water storing capacity is positively influenced.28
49Until today not one of the interviewed farmers has reported diminishing yields due to a loss in soil quality. Nevertheless, in recent years some of the most labor-intensive practices, like the collection of leaf-fodder and bedding in the mountains, have already been abandoned. This is credited to the fact that animal husbandry has nearly disappeared. Consequently, the practice of adding organic material from outside the terraced area has disappeared, too. The ongoing abandonment of agricultural activities on the terraces, created and maintained by generations of farmers over centuries, does not only leave a fragile landscape with unmaintained terrace walls, but puts a huge amount of traditional knowledge on sustainable land use at risk. So far, the higher mountain slopes are affected most by abandonment, while the easier accessible terraces are still maintained and most traditional practices still kept alive.
50The terraced landscape fulfills a number of services to society which are not directly linked to agriculture or food production. The investigated terrace soils have such a high water-infiltration capacity that even heavy rainfall amounts infiltrate without causing surface runoff – thus preventing soil erosion. This function is very valuable for the inhabitants of the parish, especially since the region of Madeira is known for its frequent flash-flood events (Quintal 1999; Baioni 2011). It can be stated that well maintained terraces are actively reducing the hydrological risk for settlements in valleys, a function that has been observed and evaluated in other parts of the globe, too (Dorren & Rey 2000).
5114C analysis provided temporal information on specific interactions between the humans and the landscape. Written sources on landscape history of Ponta Delgada and the Lombadas solely offer a generalized vision of landscape dynamics. The lack of cadastral information prevents a detailed study of spatial dynamics concerning the extension of the cultural landscape. There are no sources available that clearly state, which area was turned into cultural land in which period.
52The combination of the determined ages of the terraces and their precise location in a GIS provides the spatial information on the extension of cultural land. The GIS allows uncovering the main factors that influenced farmer´s preferences on selection of arable land. If the early terraces are further away from the settlement, than younger ones, distance was not a main preoccupation. Since small plateaus higher up on tiny ridges were cultivated first, the terrain conditions for terracing were a more important preoccupation for farmers, maybe combined with other factors such as solar exposure of slopes, and proximity to water lines for the establishment of irrigation canals, the famous Madeiran levadas.
53Next to the social investigation this study showed some of the many services that this traditional cultural landscape offered and still offers to society. The century long, mutual adaptation between humans and landscape is evident. Not only in the constructed heritage of terraces and irrigation channels, but in the traditional practices of soil conservation, too. The traditional knowledge is a very valuable heritage and in terms of food security and prevention of natural disasters could regain importance in the future, particularly in times marked by climatic changes.
