The Natural Environment

In addition to regular climatic cycles, the Mediterranean environment was also a world of sudden catastrophes throughout antiquity (Horden and Purcell 2000: 298-341), all the way from the Santorini volcanic eruption in the late seventeenth century BC to the complex of catastrophes at the end of antiquity in the middle of the sixth century ad - the extremely cold years following the dust-veil event of AD 536 (Arjava 2005), the plague of Justinian from ad 541 onwards (Sallares 2007), and the tsunami that struck Phoenicia in ad 551 (Elias et al. 2007). There were earthquakes, tidal waves, dust storms, disease epidemics, swarms of locusts, devastating floods, and volcanic eruptions. Mt Vesuvius played a prominent role in the Roman period, destroying Pompeii and Herculaneum in ad 79. The eruptions of Mt Etna, an important source of carbon dioxide and sulphur dioxide, in Sicily (44-42 BC) probably had a significant short-term impact on the climate (Stothers and Rampino 1983).

Tectonic activity caused numerous earthquakes, for example at Helike on the Corinthian Gulf in 373 bc, Pompeii in ad 62, and Olympia in ad 426 (Huntington 1910). Earthquakes caused substantial short-term damage to the urban economy of affected towns; whether they also affected agriculture in the countryside is unclear. Dust storms arrived in southern Europe from the Sahara, and also had local origins in soil erosion, as in the case of the dust storms which buried the city of Stobi in Macedonia in late antiquity. Swarms of locusts were one type of calamity which hit the countryside harder than the towns. Livy mentions a devastating locust onslaught on North Africa in 125 bc, for example. Early modern accounts of locust swarms suggest that they would have had a severe short-term impact on agriculture in antiquity (Sallares 2002: 183; Livy Per. 60).

Floods were the type of catastrophe which had the most substantial long-term impact by permanently altering the landscape. The Tiber river experienced a long series of severe floods in the past which deposited sediments in the valleys between the hills of Rome (Aldrete 2007). Strata in the Roman Forum dating to the time of the Roman empire are several meters above archaic levels and six to seven meters below the current ground level. The construction of the Cloaca Maxima shows that ancient Rome had drainage problems from the beginning of its history. High rainfall in mountains around the Mediterranean frequently resulted in deluges, causing erosion in the uplands and interior, while the eroded sediments were redeposited in deltas like those of the Po and Tiber in Italy, the Achelous in Greece, and the Ebro in Spain. The silt brought down by the Nile is responsible for the fertility of Egypt. In the case of the Tiber the existing delta has only developed since about ad 1500, but in antiquity large lagoons that existed on either side of the river near Ostia were being filled in (Bellotti et al. 1995). Such developments created new economic opportunities in the shape of extremely fertile agricultural land, but the expansion of easily flooded coastal plains which could turn into marshes facilitated the spread of malaria around the coastal regions of Italy during the Roman period. The recent remarkable finds of buried Roman ships at Pisa illustrate the importance of catastrophic floods in modifying many coastal regions of the Mediterranean in Roman times, a phenomenon also observed in the Rhone delta in southern France, for example (Benvenuti et al. 2006; Lippi et al. 2007).

The question of the causes of soil erosion leads us to the problem of deforestation, the most controversial issue in Mediterranean environmental history. Many historians have believed that extensive deforestation occurred in the Mediterranean in antiquity, leaving a denuded landscape. Undoubtedly by the late republic the city of Rome, for example, had huge requirements for building work, heating houses and baths, industrial activities and many other purposes, which could not be met locally (Meiggs 1982: 218-59). These requirements were mainly met by floating timber down the Tiber to Rome. Wood was also important for metal smelting, for example silver ores in Attica and Spain, copper in Cyprus, iron from Elba near Populonia in Tuscany, and for shipbuilding.

Despite the importance of the timber industry in antiquity, the “ruined landscape theory” of Mediterranean deforestation has attracted criticism (Grove and Rackham 2001 contra Hughes 1994 and 2005). There is a great diversity of opinion about this controversial issue among scientists who are specialists in Mediterranean ecology (e. g. Blondel and Aronson 1999: 201-6 contra Grove and Rackham 2001). Similarly, those historians who do believe in large-scale deforestation have differences of opinion regarding its chronology. One study concluded that there has indeed been extensive deforestation in five mountain zones of the Mediterranean world (McNeill 1992). However the conclusion was reached that it occurred principally in the early modern period, not in classical antiquity.

The view that little has changed is based on the observation that many regions of Mediterranean countries with low annual rainfall, a limestone-based geology that does not retain water, and a summer drought could never have supported significant forests. This theory maintains that savannah-style vegetation, with scattered trees in open country but no closed forest canopy (like the Spanish dehesas), is characteristic of many Mediterranean areas now as in the past, and little or nothing has changed over the last three thousand years; little deforestation has occurred and it is not responsible for soil erosion. Erosion is interpreted as predominantly gully erosion of badlands caused by deluges, for example in Basilicata in southern Italy, producing the alluvial deposits of Metapontum on the coast (Abbott and Valastro 1995). Where soil erosion from cultivated land has occurred it is attributed principally to plowing, not to deforestation.

The history of erosion is tied to the problem of the Younger Fill (originally described by Vita-Finzi 1969), a heterogeneous mix of depositional episodes that occurred at different times in different areas; some are definitely classical in date, for example those at Metapontum mentioned above, or on the coasts of western Anatolia (Kayan 1999); others date to late antiquity or the early medieval period, for instance the burial of Olympia in Greece (Huntington 1910), while yet others are only a few hundred years old (Grove and Rackham 2001: 291-4). Some of these episodes can be associated with human activity, while for others cyclical changes in the climate offer a more convincing explanation.

The critical point to emerge from the whole debate is that it is impossible to generalize about the Mediterranean as a whole. The human impact varied from area to area. Consequently the Mediterranean countries as a whole cannot be described either as an unchanged or as a ruined landscape. The theory that little has changed in the last few thousand years is reasonably convincing for some of the most arid parts of the Mediterranean, such as south-eastern Greece. However, even in southern Attica there is archaeological evidence for farming in the fourth century bc on limestone ground which has virtually no soil cover today (Lohmann 1994). The question of environmental degradation cannot be considered independently of the question of human population pressure on the landscape. Some of the details of the theory of no change seem to be self-defeating. For example, the importance of anthropogenic erosion is minimized, but it is then acknowledged that soil erosion can be caused by plowing arable land to grow rain-fed, autumn-sown cereals, leading to criticism of the cultivation in southern Italy of durum wheat, the type of wheat best suited to the local environment. What are large human populations supposed to have eaten in the past if they had not been allowed to grow cereals, to avoid soil erosion? (Grove and Rackham 2001: 89, 265 and 270).

By focusing on Mediterranean Europe, Grove and Rackham excluded North Africa from their consideration of the problem of “desertification,” but it is surely in the vicinity of the Sahara (in an area where megafauna such as lions and elephants existed until Roman times) that this problem was and is most acute. Literary sources suggest that some parts of North Africa had plenty of trees, while others did not (Contrast Caes. BC2.37 with BAf 20). Computer modeling of the climate about two thousand years ago to study the effects of the presence of substantial vegetation in such areas on the climate suggests that there was more rainfall in Armenia, Spain, North Africa, and Egypt than there is today (Reale and Dirmeyer 2000; Reale and Shukla 2000).

These conclusions help to make sense of Ptolemy’s weather diary, written in Alexandria during the second century ad, which describes a weather pattern with rain in every month except August and thunder throughout the summer (G. W. Murray 1935: 19-20). This helps to explain the agricultural prosperity of North Africa in antiquity, as well as the prosperity of southern Spain during the Roman period. The rise and fall of the kingdom of the Garamantes in North Africa has been linked to rainfall fluctuations.

Leaving aside marginal areas such as those bordering deserts, the greatest degree of human impact on the natural environment in antiquity is most likely to have occurred in the immediate vicinity of the largest human population centers, but such areas are actually rarely considered in detail in the debates between ecologists. It has been argued that the rate of soil erosion in Latium increased ten times in the second century bc (Judson 1968). This trend is surely associated with the increase in settlement numbers in south Etruria revealed by archaeological field surveys, linked to intensive agriculture and market gardening to feed the population of the city of Rome, which was increasing rapidly at the time (T. Potter 1979). Environmental degradation would then have spread away from large settlements along communication lines such as river valleys and roads. It has been suggested that the major Mediterranean river valleys were once generally forested, since perennial rivers compensated for the shortage of summer rainfall, but are now largely deforested, with a few exceptions, such as the River Strymon in northern Greece (Blondel and Aronson 1999: 122).

Livy described the Ciminian Forest north of Rome c.300 bc as if it was the Amazon jungle (Livy 9.36). His account is sometimes regarded as exaggerated, but in the opinion of Italian specialists in environmental studies little of the ancient beech forest in the area of the Monti Cimini remains today (Pratesi and Tassi 1977: 49 contra Grove and Rackham 2001: 172). A series of detailed local studies do support the idea of substantial human impact on the natural environment in many areas in classical antiquity. In the Biferno valley in the Molise region of eastern Italy a field survey revealed a massive expansion of rural settlements from the fourth century bc to the first century ad accompanying extensive erosion and sediment deposition (Barker and Hunt 1995). Similarly, in the hinterland of Metapontum in southern Italy ten meters of sediment was deposited during the period c.600 to c.300 bc, when this Greek colony was flourishing. The affected areas quickly became marshy, a development which accelerated the spread of malaria in this region and subsequently led to the decline of the human population (Henneberg et al. 1992).

A balanced interpretation of the problem of environmental degradation in antiquity is that it is difficult to generalize; there were different outcomes in different areas. In many areas the vegetation cover was probably much the same as it is today. For example palynology yields this conclusion in the vicinity of the Lago di Pergusa in central Sicily (Sadori and Narcisi 2001). In other areas there was a substantial degree of deforestation in the past. For instance the Roman boundary stones which enclose the famous cedar forest on Mt Lebanon include considerable areas where there are hardly any trees today, although the question arises of whether this already became the case in antiquity or whether deforestation only happened later (Mikesell 1969). In the lower Rhone valley in France the scarcity of tree pollen indicates that this area was largely deforested in the classical period (Andrieu-Ponel et al. 2000). Of course there are other areas where forests have spread, either naturally or through human planting, since antiquity. Pollen cores document the gradual spread of Aleppo pine (Pinus halepensis), which is now common in Attica for example. It has increased in frequency because it is a good colonizer of cleared terrain, since its seeds germinate readily after forest fires.

The effects of human pressure on the environment in classical antiquity took other forms besides deforestation. It is easy to think of pollution as a modern problem, but it has been demonstrated that there were considerable increases in the concentration of lead and copper in ice strata in the Greenland ice cap (Hong et al. 1994 and 1996), in peat bogs in Switzerland (Shotyk et al. 1998) and Spain (Martinez-Cortizas et al. 1999), and in lake sediments in Sweden from about 600 bc onwards (Renberg et al. 1994), by-products of the great increase in the scale of mining in classical antiquity. The Laurion silver mines of Attica, which paid for the Athenian navy constructed by Themistokles, and so for the Athenian empire, were also the first major source of anthropogenic pollution in world history. Mining for other metals, such as mercury, also made a contribution to atmospheric pollution. Thus cinnabar mining increased from the fifth century bc onwards at Almaden in Galicia in Spain, as shown by paleoenvironmental data (Martinez-Cortizas et al. 1999).

The Mediterranean Sea, since it is almost entirely landlocked and surrounded by large human populations, is one of the most heavily polluted seas on earth today. However, such problems are not unique to the modern epoch. Seneca clearly described the high level of atmospheric pollution in the city of Rome, which is not surprising in view of the extensive burning of wood for fuel that was mentioned earlier (Capasso 2000; Nutton 2000). Pollution is important not only for environmental history and for human health; it is also a key indicator of the scale of “industrial” activity in antiquity (see also Pyatt et al. 2002 for another approach to this problem). Although the metal deposition data do not tell us anything about productivity, they do demonstrate that the scale of mining activity in the period c.500 bc to c. ad 500 was substantially greater than anything seen before or for a long time afterwards. This supports the theory that substantial economic growth (at least in terms of total production levels) occurred during the time of the Roman empire.