Two steps forward, one step back…


*Correspondence author. E-mail:

The number of examples of ecological ‘fingerprints’ of recent climate change (Walther, Burga & Edwards 2001; Parmesan & Yohe 2003; Root et al. 2003) steadily increases, providing ample evidence that global warming is affecting a wide range of species in various habitats from the poles to the tropics (Walther et al. 2002; Parmesan 2006; Rosenzweig et al. 2007). Despite the general warming trend, however, climatic parameters vary from year to year and many follow periodic oscillations. Similarly, the ecological impacts, although following the general climatic trend, can be affected by the variability overlaying the general trend, and hence do not exhibit a uniformly directed pattern.

One example of such a varying response to the general trend is presented by Augspurger (2009). Based on phenological observations from 1993 to the present day of 20 canopy species in a forest northeast of Urbana, Illinois (USA), she assessed the climatic anomaly of the year 2007 and its impact on spring phenophases of the study species. Spring temperatures of 2007 were exceptional in two ways: (i) prolonged warmth in March brought multiple record-breaking temperatures and thus early and rapid phenological development of many studied species; (ii) this period of exceptional warmth was followed in April by a period of unusually cold temperatures including a record low temperature, well below the freezing point. Although the timing and severity of the cold spell were not that exceptional, the combination of enhanced March warming and subsequent cold phase in April resulted in the observed impacts on phenology and damage to actively growing plant organs.

Parallel observations of multiple tree species show that the impact of this unusual environmental variation varies considerably depending on plant phenological phase, and among organs, individuals, life stage, and species. These highly variable responses reveal a high degree of complexity in natural systems and caution against extrapolating from relatively simple single species studies to higher levels of complexity such as communities and ecosystems (cf. Walther 2007). This lack of uniformity of response direction, which includes short-term (partial) reversals, is also known from past (e.g. Iversen 1944; for an update see Walther, Berger & Sykes 2005) and present (e.g. Parmesan 2001) range shifts. Range expansions typically occur during favourable phases, with (partial) extinctions often occurring in unfavourable periods and re-colonisations occurring when conditions become favourable again. The crucial aspect is that environmental variation leads to continual alternation of expansion and retraction of geographic ranges, highlighting the difference between individual events vs. trends.

One example to illustrate this forward and (partial) backward process is the recent expansion of a palm species (Trachycarpus fortunei) beyond the limit of global palm distribution (Walther et al. 2007). Milder winter conditions of the past years and decades allowed the cultivation of palms in areas where this was not previously possible (Francko 2003; Stähler & Spanner 2007; Walther et al. 2007). North of the European Alps in central Europe, the chance of survival for palm species grown outdoors and unprotected during winters was very low until c. the 1980s (Klötzli 1995; cf. also Landolt 1993). However, since the 1990s the outdoor cultivation of palms has become less risky and thus more popular. As a consequence, palm cultivation increased in the milder western part of Germany and also extended to the more continental eastern Germany, especially in the last few years of mild winters (Fig. 1). This period reflecting the global warming trend was interrupted in 2008 with a relatively cold winter and temperatures down to c.−20°C (measured at stations in eastern part of Germany, data provided by Germany’s National Meteorological Service ( These low temperatures severely damaged unprotected outdoor palms. However, despite the unusually cold winter conditions of 2008, palms were not extirpated in all parts of Germany. Whereas gardeners in the eastern part reported severe damage and complete loss of unprotected palms, the damage to palms grown unprotected outdoors in the west was restricted primarily to leaf damage. Hence, from the two-step process of (assisted) colonization, with a first step in the west followed by a second step towards the east, only one step was reversed by an unusual climatic event, revealing a net one step forward for palms cultivated in central Europe.

Figure 1.

 A ‘two steps forward, one step back’ process with the introduction and outdoor cultivation of palms (here: Trachycarpus fortunei) in Germany (based on Black dots: outdoor planting >5 years ago (= first step forward); grey dots: outdoor planting <5 years ago (= second step forward); arrow: (partial) reversal due to cold winter of 2008 (= one step back) but only palms in the range of the leading edge were severely affected (see text) resulting a net one step forward for palms in Germany; dotted line: approximate limit of continuous range.

This process also shows the different consequences of climatic trends vs. events for plant distributions. Whereas colonization responds strongly and quickly to climatic events (here: the cold winter of 2008), the front of the continuous range reflects the pattern of the climatic trend (here: the warming trend since the late 1980s). Hence, the question is not whether events or trends are more important; they both act as important drivers for, e.g. range shifts, but are reflected in different parts of species’ ranges, with the individuals representing outposts or forerunners in the expansion process strongly exposed and responding to events, whereas the movement of the front of the continuous range follows more the direction and pattern of the trend. Similarly, but on the temporal scale, phenologically early and fast responding species – especially saplings of buckeye Aesculus glabra– were more affected by the unusual event reported by Augspurger (2009), but it remains to be seen whether this combination of unusual warm March temperatures followed by spells of sub-freezing conditions in April will return in this area, while the climate continues to warm.

  • image(2)

[  Variation in early spring leaf phenology in Trelease Woods results in interspecific variation in vulnerability to a late spring frost. Photo by John Cheeseman. ]