The Mediterranean climate is characterized by a seasonal pattern of mild, wet winters and hot, dry summers. The summer drought favors annual rather than perennial vegetation. It also means that fire is an important agent of natural selection on vegetation communities. Bioclimatic indicators such as the distribution of the olive tree are sometimes used to provide alternative definitions of Mediterranean-climate regions (Blondel and Aronson 1999: 13-18). The distribution of the olive tree, which is killed by severe frost but requires temperatures to drop to a certain level to initiate flowering the following year, defines the Mediterranean winter. The total volume of winter rainfall can support evergreen trees, but is too low for deciduous trees (Sallares 1991: 307).
Dendrochronological evidence from the Parthenon in Athens shows patterns of climatic variability in the fifth century bc similar to the modern pattern. Evidence from ancient authors confirms that the climate of Greece was basically the same in the fourth century BC as it is today. Theophrastus wrote that date-palm trees planted in Greece could grow there but not set fruit, indicating that mean summer temperatures in the southern Aegean c.300 bc were very similar to recent values (Eginitis 1908 on Theophr. Hist. pl. 3.3.5). Rainfall exhibits extreme seasonal and interannual variability. This frequently created shortfalls in agricultural production in the past, with runs of several successive good or bad harvests (Garnsey 1988). In classical Greece there were periods of food shortages probably caused by drought c.360 and c.330 BC, but Theophrastus recorded heavy rainfall which raised Lake Kopais in Boiotia to an unusually high level in the years before the battle of Chaeronea in 338 BC (Sallares 1991: 390-5). However, documentary sources do not provide the statistical data required to investigate possible long-term climatic trends.
The pattern of hot dry summers and cool wet winters has only existed since about c.3000 BC. Evidence from pollen cores shows that annual Mediterranean rainfall was higher and more evenly distributed in the Neolithic period, permitting deciduous lime and oak trees to flourish in areas dominated by drought-resistant evergreen vegetation today (Vernet 1997). Many important Mediterranean plants (e. g. the vine) are dormant in winter and grow in the hot, dry summer, when it would be better to grow in winter instead, when more water is available. As a result irrigation in the summer is very important for Mediterranean agriculture today. The degree to which artificial irrigation was practiced in antiquity is an important question. The emphasis of the Roman agronomists on wheat and barley cultivation, crops of semi-arid drylands, and their lack of attention to rice, a much more productive way of exploiting coastal Mediterranean plains, suggests that irrigation was not important in the Mediterranean in antiquity (Sallares 1991: 22-4).
There were also periodic climatic cycles lasting for centuries or longer, caused by variations in annual-mean solar radiation. In the Levant, for instance, warm and dry periods alternated with cold and humid periods. The Iron Age was cold and humid, the Persian period dry, the Hellenistic period cold and humid, the Roman period warm, while the Byzantine period was yet again cold and humid (Issar and Zohar 2007). Cold and humid conditions signified more rainfall, which benefited cereal production and favored population growth. A decline in solar radiation leading to colder and more humid conditions c.850 bc may well have been the critical factor behind the simultaneous development of Iron Age cultures around the Mediterranean, which is otherwise hard to explain (Speranza et al. 2002).
Research on periodic advances and retreats of the Alpine glaciers during the Holocene indicates that most of the time of the Roman empire (c.100-400 ad) was a warm period (Rothlisberger 1986). Other types of climatological evidence yield similar conclusions. For example, analysis of atmospheric mercury deposition (a process influenced by temperature) in a peat bog in Galicia in north-western Spain suggests that the Roman Warm Period at its peak was about 2°C warmer than the present and was more prolonged than the Medieval Warm Period (Martinez-Cortizas et al. 1999). The reality of the Roman Warm Period can be corroborated by proxy data, for example the spread of viticulture into Roman Britain, as illustrated by recent archaeological finds in the Nene Valley in Northamptonshire (Brown et al. 2001). Olive presses and olive wood have been found near the Roman city of Sagalassos in south-western Anatolia in areas where it is too cold for the olive tree to survive today, suggesting that average temperatures might have been 2-3°C higher than today (Waelkens et al. 1999). The economic effects of climate change were complex. The warmer conditions of the Roman empire permitted the geographical extension of olive and vine cultivation and favored arable farming in northern Europe, but probably adversely affected it in some semi-arid areas such as the Near East. Whether climatic trends can be correlated with and played a causal role in human population fluctuations is a very important question for ancient history. Just to give one more example, favorable climatic conditions probably facilitated the spread of olive cultivation and human populations in Syria in late antiquity (Hirschfeld 2004).