The impacts of past human activities on ecosystems

Although there has been a growing emphasis on the interpretation of pollen analytic and other biologic records as proxies for past climate change (e.g. Birks 2003, this volume), the paleoecological evidence for major human impacts on past vegetation continues to grow, especially in Europe, where pollen taxonomy favors the identification of species and genera unambiguously indicative of human activities such as deforestation and the development of pastoralism and agriculture. Gaining some sense of the timing, sequence, and degree of alteration of previously forested landscapes is fundamental to addressing the subsequent themes.

In some environments, the millennia from the early Neolithic through to the early Medieval period were marked by successive alternations between episodes of deforestation and farming, and periods of forest recovery. This appears to have been the case in the Black Forest region of south-western Germany where Roesch (2000) records nine phases of forest clearance between the early Neolithic ca. 7600 years BP and the early Medieval period that saw the start of major, sustained deforestation. This record is strongly reminiscent of that from parts of north-west

England (e.g. Oldfield 1963). Such long-term records of recurrent recovery suggest that the forest ecosystems of these regions, hence the soils also, have been remarkably resilient (sensu Lal 1997) in the face of repeated phases of deforestation. This is despite the fact that several pollen diagrams from the north of England suggest that the area deforested during Late Iron Age and Romano-British times, between around 2500 and 1500 years BP, was as extensive as during the peak deforestation in the 17th century (Oldfield and Statham 1965; Dumayne-Peaty and Barber 1998; Wimble et al. 2000; Oldfield et al. 2003a; Figure 3.1). Strong evidence for the importance of deforestation during the period of Roman occupation also comes from many studies in France (e.g. Noel et al. 2001; Cyprien et al. 2004).

In some other parts of Europe, the evidence points to earlier and somewhat more progressive conversion of forest to less productive ecosystems as a result of human activities. In parts of Denmark, for example, there are indications of irreversible deforestation from ca. 500 bc onwards (Bradshaw et al. 2005a). Odgaard and Rasmussen (2000) reinforce this view by statistical comparisons between the mosaic of land-cover types documented in 1800 and those inferred from successive pollen records spanning the period 4500 bc to ad 1800 at 20 sites (Figure 3.2). In-depth, multi-disciplinary studies of ecological and cultural history at the regional scale are relatively rare, but one such is that of Berglund (1991) in the Ystad region of southern Sweden (Figure 3.3). As in the Danish study, he records the progressive deforestation of the landscape during prehistory. Intermediate between the studies showing progressive deforestation without significant recovery, and those marked by recurrent re-afforestation, are sites such as Lough Neagh in Northern Ireland where pulses of forest clearance, marked by peaks in weeds and bracken, are superimposed on evidence for a progressive increase in grassland (O'Sullivan et al. 1973).

Heathlands are distinctive habitats often of high conservation value in northern and western Europe. Understanding their origins and history is germane to any future management designed to preserve the special habitats and high biodiversity that they contain (Walker et al. 2003). They have often been seen as, in part at least, the result of forest clearance combined with subsequent soil degradation and/or sustained management practices favoring the persistence of ericaceous shrubs and acidophilous herbs. Fyfe et al. (2003) interpret their pollen records from Exmoor in south-west England as indicating the expansion of heathland from Neolithic times onwards, rather as did Godwin in his early study of the Breckland heaths of East Anglia (Godwin 1944), although they regard the use of fire as important for the maintenance of heathland. Savukynien et al. (2003) present evidence from Lithuania that points to the development of heathland from 1200 years BP onwards, as well as its maintenance largely through burning until recent times when its extent has been reduced. Following the earlier work of Kaland (1986), Prosch-Danielson and Simonsen (2000) show that the expansion of coastal heathland in south-west Norway took place from 4000 to 200 bc, but was mostly completed by the end of the Bronze Age. They suggest that the expansion can best be explained as a result of the interaction between land-use history, topography, and edaphic conditions, within a climatic regime that favored heathland development.



Hard IRM

0 20

Plantago Miscellaneous

Poaceae lanceolata Cultivated Calluna Herbs

Hard IRM

0 20

Figure 3.1 The impact of forest clearance during the late Iron Age and Romano-British period. Evidence from Gormire, North Yorkshire, UK. The chemical elements and "Hard-IRM" (a rock magnetic parameter) are indicative of the erosion of inorganic mineral soil. The pollen record links deforestation to evidence for human activities - both pastoralism and cultivation - as well as to the spread of moorland over the past 2000 years. The //-alkane record reflects the relative abundance in the sediments of abiomarker specific to open grassland. (Based on Oldfieldef «/. 2003a.)

LPZ Date 1770 ad

1200 ad 600 ad

200 bc

Figure 3.1 The impact of forest clearance during the late Iron Age and Romano-British period. Evidence from Gormire, North Yorkshire, UK. The chemical elements and "Hard-IRM" (a rock magnetic parameter) are indicative of the erosion of inorganic mineral soil. The pollen record links deforestation to evidence for human activities - both pastoralism and cultivation - as well as to the spread of moorland over the past 2000 years. The //-alkane record reflects the relative abundance in the sediments of abiomarker specific to open grassland. (Based on Oldfieldef «/. 2003a.)

Figure 3.2 Plot of the extent to which forest composition inferred from pollen diagrams from 10 sites in Denmark corresponds with forest composition at ad 1800. Note how the percentage of between-site variance explained increases steeply along with the significance of the relationship (*** = p < 0.001) from middle Neolithic to late Bronze Age times. The results show that both the differentiation between sites and the macroscale landcover patterns were established some 3000 years ago. LEB, late Erteb0lle; EN, early Neolithic; MNA, MNB, middle Neolithic A and B; LN, late Neolithic; EBA, early Bronze Age; BA, Bronze Age; PRIA, Pre-Roman Iron Age; RA, Roman; LIA, Late Iron Age; MED, Medieval. (From Odgaard B.V. & Rasmussen P. (2000). With permission from Blackwell Publishing Ltd.)


Time period


Water erosion

BIOTA Deforestation


Open land

Figure 3.3 Some ofthe human-induced changes inferred from the multi-disciplinary Ystad Project in southern Sweden. (Based on Berglund 1991; Dearing et al. 2006.)

Soil leaching

POPULATION 4000 BC - AD 2000

Fire impact

Arable land

4000 3000 2000 1000 0 1000 2000


Lake eutrophication


4000 3000 2000 1000 BC

0 1000 2000 AD

4000 3000 2000 1000 BC

0 1000 2000 AD

Bunting's (1996) evidence for the development of heathland in the Orkneys also points to a complex interaction between human, edaphic, and climatic factors. In less oceanic areas such as south-west Sweden, extensive heathlands were not formed until historical times (Malmer 1965; Bjorkman 2001), mainly through intensive grazing subsequent to deliberate deforestation.

It is entirely possible that anthropogenic transformation of upland ecosystems in Britain began before the Neolithic period. Evidence from sub-peat charcoal layers and from organic biomarkers (Chiverrell et al. in press) confirms that fire was prevalent in upland regions of northern England currently covered by moorland or blanket peat. Several authors (Jacobi et al. 1976; Simmons 1996; Innes and Simmons 2000; Innes and Blackford 2003) have suggested Mesolithic peoples used fire as part of their woodland management strategy in order to increase animal abundance. Drier climatic conditions during the Mesolithic period, however, may also have led to more widespread natural fires (Bradshaw et al. 1996; Brown 1997; Tipping and Milburn 2000).

Holocene pollen diagrams from east European sites that are outside the zone of Mediterranean climate record deforestation ranging in age from older than 7000 years BP in north-eastern Bulgaria (Bozilova and Tonkov 1998) to the 13th century ad in Russian Karelia (Vuorela et al. 2001). Kremenetski et al. (1999) find evidence for major human impacts from Bronze Age times onwards in southern Russia, whereas in Estonia, clear signs of human impact are delayed until around 1500 years BP (Niinemets and Saarse 2006).

Several studies in the Alps point to human activities, notably grazing and the use of fire, as the main factors responsible for shifts in the tree line and changes in forest composition from Neolithic times onwards (e.g. Gobet et al. 2003; Carcaillet and Muller 2005; Tinner et al. 2005). In the Lake Garda region of northern Italy, Valsecchi et al. (2005) estimate that over 60 percent of the pollen source area was deforested during the Bronze Age between 2000 and 1100 bc, whereas in the East Pyrenees, Guiter et al. (2005) point to significant anthropogenic reduction in forest cover from ca. 2500 years BP, accelerating during the past 2000 years.

The relative importance of the roles Holocene climate change and human activity played in the development of the present-day plant cover in those parts of Europe experiencing a Mediterranean-type climate remains a matter of debate. There is strong evidence from both marine and terrestrial archives for extensive desiccation beginning between 5000 and 4000 years BP and continuing at least until 3000 to 2000 years BP (e.g. Bar-Matthews and Ayalon 2004; Kallel et al. 2004; Roberts et al. 2004). Many pollen diagrams document the spread of xerophytic scrub and steppe biomes during this period and some authors stress climate change as the main cause (e.g. Jalut et al. 2000; Sadori and Narcisi 2001; Pantaleon-Cano et al. 2003). At the same time, many others, while acknowledging the importance of climate change, note widespread evidence for the degradation of Mediterranean ecosystems through human activities, beginning mainly during Bronze Age times towards the end of the third millennium bc, and often peaking during the Medieval period (van Andel et al. 1986; Jahns 1993, 2005; Brochier et al. 1998; Atherden and Hall 1999; Ramrath et al. 2000; Carrion et al. 2001; Oldfield

Figure 3.4 Evidence for human impact over the past 4000 years from an Adriatic core (RF 93-30) integrating evidence from a large area along the eastern flank of Italy. The pollen record shows deforestation from 3500 years BP onwards, with periods of peak human activity in the Bronze Age (ca. 3500-2000 years BP) and Medieval (post-800 years BP) periods. The magnetic properties (IRM 300 and Xfd, both also influenced by tephra layers) mainly record terrigenous input during these periods. The benthic foraminifers show distinctive faunal changes at the onset of both periods of erosion. The alkenone-inferred surface water temperature highlights the likelihood that some of the changes, especially between 3500 and 3000 years BP, may also reflect climatic forcing. A-D refer to the main horizons of change as identified in the original reference. (Based on Oldfield etal. 2003b.)

et al. 2003b, Sobrino et al. 2004; Butzer 2005; Figure 3.4). As noted in a later section, feedback from the creation of extensive garrigue and bare ground in place of woodland, through overgrazing and soil erosion, would have tended to reinforce any summer drought initially resulting from climate change.

Summarizing the evidence from Europe as a whole, the following generalizations are proposed.

• Paleoecological evidence distinguishes between landscapes characterized by relatively resilient woodland ecosystems within which repeated forest recovery occurred over most of the past 6000 years until the Middle Ages and open landscapes reflecting progressive deforestation from prehistoric times onwards.

• Among the latter are the heathlands of north-western Europe and Mediterranean garrigue. In both cases, human activities and climate change have acted synergisti-cally in edaphically sensitive environments to promote or hasten the transformation of ecosystems. The extent to which persistence of these has depended on their maintenance through sustained human impact and the use of fire has varied from place to place.

• Throughout Europe, from the western edge of Ireland (Molloy and O'Connell 1995; O'Connell et al. 2000) to the Urals, all but the most remote or northerly sites show episodes of human impact through deforestation from Neolithic or Bronze Age times onwards (cf. Berglund 2000). Even in the central Swedish forest region as far as 62° north, there is clear evidence of deforestation for sedentary farming from ca. ad 100 onwards (Emanuelsson 2001).

• Mesolithic cultures may also have had a significant impact on upland ecosystems in Britain at least, through the use of fire.

• At many sites in the Mediterranean region and western Europe there is evidence for extensive and, in some cases, permanent deforestation from an early stage in the past 2000 to 3000 years.

• Many studies confirm the importance of fire in the creation and maintenance of deforested ecosystems.

• In the Mediterranean region, it seems likely that climatically induced summer drought during the second half of the Holocene may have been reinforced by feedback from land-cover transformed by human impact, for example, from woodland to garrigue.

Even in Europe, where palynological evidence for agriculture and pastoralism is relatively unambiguous and there have been many studies of Holocene climate variability (Verschuren and Charman, this volume), the relations between human activities and climate change remain open to much debate (see e.g. Berglund 2003). There are cases where favorable climatic conditions appear to have encouraged the expansion of agriculture, as during Romano-British times in Britain. Equally, one may argue that challenging environmental conditions triggered critical advances in adaptive technology (e.g. Rosen 1995). These issues are explored further below, as is the challenge of developing more quantitative expressions of landscape openness from palynological data.

Looking briefly beyond Europe to other regions where a long history of human occupation and land management raises questions of early, long-term, and sustained human impact on ecosystems, the picture is less clear, partly because much less evidence is available and partly because pollen-analytic data are often less amenable to interpretation in terms of human impact. Hoelzmann et al. (2004), summarizing evidence from the arid and sub-arid parts of Africa, including the Sahara, point to climate change as the primary cause of mid-Holocene desiccation. All authorities now accept, however, that feedback from changing land-cover was critical in reinforcing the effects of externally driven changes in climate (Claussen et al. 1999; deMenocal et al. 2000). What remains in doubt, partly because of the absence of unambiguous pollen indicators of human activity (Waller and Salzmann 1999), is the degree to which human activities were responsible for some of the land-cover changes that helped to trigger the shift to aridification in parts of the region. Similar difficulties lie in the way of establishing the role of past human activities in the spread of savannah (Saltzmann 2000), although Sowunmi (2004) claims that human activities were responsible for a significant expansion of savannah from 3000 years BP onwards in coastal Nigeria. Indeed, Ballouche (2004) states that "most of the West African savannas are cultural landscapes which have been strongly shaped by humans". Evidence from East Africa also points to significant human impacts on land-cover in pre-colonial times (Taylor 1990; Mworia-Maitima 1997), though the evidence for strong hydrologic variability and recurrent drought is incontrovertible (Verschuren and Charman, this volume). In southern Africa (Scott and Lee-Thorp 2004), the main spread of agriculture and grazing associated with increasing population densities and deforestation appears to have taken place during the past 2000 years, peaking around 800 years BP. This is consistent with evidence from marine cores summarized by Shi et al. (1998).

Although the evidence from western, south-eastern, and eastern Asia for the early origins of different forms of agriculture is well documented (see summaries in Yasuda 2002), it is much more difficult to determine the timing and extent of human-induced land-cover change for most regions. Some of the strongest evidence is emerging from China, for early deforestation (Ren and Zhang 1998; Ren 2000), the long-term maintenance and expansion of open landscapes, and the likely impact of these processes on the climate (Fu 2003). Huang et al. (2002) find evidence for deforestation and cultivation on the southern Loess Plateau from 7500 years BP onwards, and Zhou et al. (2002) ascribe the south-eastern displacement of the desert-loess boundary to human activity over the past 3000 years, although evidence presented by Li et al. (2003) suggests that the vegetation of the east-central part of the Loess Plateau to the north of Xian was dominated by steppe and grassland throughout the Holocene. In the far north-east of China, Makohonienko et al. (2004) document deforestation from around 900 bc onwards and show that the spread of grassland over the Manchurian Plains was the result of human activities over the past 1000 years. Li et al. (2006) also show that in Inner Mongolia human impact on the landscape, already significant before 5400 years BP, intensified from 4700 years BP onwards. In the region of the Yellow River delta, successive episodes of deforestation and cultivation took place from 4000 years BP onwards (Yi et al. 2003). Ji et al. (2006) show that deforestation by human populations began as early as 7000 years ago around Erhai Lake in Yunnan Province, south-west China.

In the case of the Indian sub-continent, it is relatively easy to reconstruct in outline the main sequences and patterns of cultural changes over the past 6000 years, from early Neolithic times onwards (see e.g. Misra 2001). Many studies confirm that the establishment of settled agriculture was widespread from the fourth millennium bc onwards, not only in the Indus Valley where the Harappan Civilization flourished but also in southern India (Fuller et al. 2004). What is much more difficult to glean is any clear indication of the impact pre-colonial societies had on land-cover. Maloney (1980, 1981) finds evidence for early prehistoric deforestation in Indonesia.

Evidence from Oceania points to rapid deforestation on many islands spanning the past 3000 to 800 years, depending on location, with devastating human consequences in some cases (Bahn and Flenley 1992; Diamond 2005). Even in the New World, clear indications of deforestation and changed land-cover go back as far as 4000 years in central Mexico (Watts and Bradbury 1982; Bradbury 2000; Fisher 2005) and 2000 years in the Cuzco area of Peru (Chepstow-Lusty et al. 1998). Australian landscape evolution has been dominated by the traditional use of fire in land management for many thousands of years (see e.g. Bowman 1998), although there are a few indications of changing Holocene patterns and intensities of aboriginal fire use, which may well stretch back for over 40 000 years.

The above account, albeit partial and regionally selective, confirms the reality of extensive land-cover transformation through human activities long before ad 1700, the starting point for some recent evaluations of anthropogenic land-cover change (Ramankutty and Foley 1999; Goldewijk 2003). This is a key observation when we come to consider the likely impact of land-use and land-cover change on climate directly and on the carbon cycle. A further important conclusion to draw from this account is the extent to which many habitats of great value for amenity and biodiversity are themselves the result of human intervention, rather than a simple and direct response to "natural" climatic and edaphic conditions. This can be illustrated by studies both at regional scale (Segerstrom et al. 1996; Berglund et al. in press) and for individual sites (Oldfield 1969). Such studies, together with others (e.g. He et al. 2002; Latty et al. 2004) showing that legacies of disturbance events may have impacts on ecosystem processes on century time-scales and beyond, suggest that paleoresearch can contribute to developing plans for future conservation. Paleoresearch, embracing long-term perspectives, shows that it is impossible to generalize about the rates, trajectories or degree of recovery from disturbance using only a small number of isolated case studies. The degree of asymmetry or hysteresis between the sudden impact of fire or land clearance, for example, and the rates of recovery varies with a wide range of factors (see Figure 3.5). There is a need for more comparative studies taking the long-term view and oriented towards a stronger theoretical basis for understanding the ecosystem processes involved (Dearing et al. 2006). Such studies are needed in order to improve our understanding of vulnerability, provide realistic guidance for conservation and restoration, and add empirically derived, temporal dimensions to the evidence against which the performance of ecosystems models can be tested.

All the emphasis above has been on terrestrial ecosystems, but from the 1960s onwards it has been increasingly apparent that human activities have had a major impact on water quality, in both inland (e.g. Digerfeldt 1972; Battarbee 1978,

Figure 3.5 Projected time-scales of recovery from degradation resulting from land-use changes (Batchelor and Sundblad 1999). The figure is taken from Chhabra et al. (2006) and provides some indication of the strong hysteresis between the rapidity with which lake and catchment systems can be degraded and the long-term nature of several aspects of recovery.

River water quality Lake water quality Groundwater quality Re-establishment of vegetation Improved groundwater recharge Reduced flooding Re-established biodiversity Improved soil fertility Reduced soil erosion Re-establishment of eroded soils j_I_'''_I_■ ■ ■ ■_I_■ ■ ■ ■_I_■ ■ ■ ■_I_■ ■ ■ ■_I_■ ■ ■ ■

"T 10



10 000 100 000 1 000 000

1990) and near-shore marine (Andren et al. 1999, 2004) environments. Although the main impacts have taken place over the past 200 years, there is growing evidence for discernable impacts in Europe at least from Bronze Age times onwards (e.g. Gaillard et al. 1991; Renberg et al. 1993; Anderson et al. 1995; Oldfield et al. 2003b; Bradshaw et al. 2005b).

0 0


  • venla
    Is there any evidence of human activity in ecosystem?
    8 years ago

Post a comment