An analysis of northern ecosystems shows that the effects on plant growth of rising night-time temperatures are opposite to those of increasing daytime temperatures — a finding that has implications for carbon-cycle models.See Letter p.88
An under-appreciated aspect of climate change is the fact that Earth is warming at a higher rate at night than during the day: over the past 50 years, daily minimum temperatures have increased about 40% faster than daily maximum temperatures1. This asymmetric warming may have impor- tant biological consequences, particularly for fundamental ecosystem metabolic processes that are strongly sensitive to temperature vari- ations, such as photosynthesis and respiration. On page 88 of this issue, Peng et al.2 document regionally significant, and in many cases opposing, effects of year-to-year (interannual) variations in daytime and night-time temperatures on plant growth and carbon cycling in land regions of the Northern Hemisphere. Photosynthesis is driven by light and thus happens only during the day, whereas plant and microbial respiration occurs continu-
ously. Therefore, faster night-time warming presumably affects respiration more than it affects photo synthesis, and this could have far- reaching implications for how ecosystems react to expected increases in warming in coming decades. But remarkably little research has been done on how asymmetric warming influences ecological function, especially at large scales. To address this issue, Peng and colleagues have analysed satellite-derived data sets of plant greenness, which is a proxy for plant growth. The authors found that ecosystems in cool, wet temperate and boreal regions such as northwestern North America and Japan, and those in cold regions such as Siberia and the Tibetan plateau, seem to have benefited most from daytime temperature increases over the period considered (1982–2009). By contrast, ecosystems in dry temperate regions, such as central Eurasia and western China, showed the opposite effect: increasing daytime tempera- tures correlated with decreasing plant green- ness. These contrasting responses broadly agree with expectations for ecosystems in which plant growth is limited primarily by temperature (cool, wet climates) or moisture (warm, dry climates). More intriguingly, Peng and colleagues found that ecosystems in many of the boreal and wet temperate regions grew less well in response to increases in night-time minimum temperatures — the opposite effect to their response to increasing daytime maximum
temperatures (Fig. 1). Conversely, in many arid and semi-arid regions, such as the grass- lands of China and North America, increasing night-time minimum temperatures correlated positively with plant greenness. Peng et al. used a statistical approach to control for other contributing environmental vari- ables, such as solar radiation and precipitation. This allowed them to isolate the interannual greenness responses to daytime maximum and night-time minimum temperature variations. The authors confirmed the statistical validity of their findings using other techniques, and also analysed the sensitivity of the greenness response to alternative interpolated climate
data sets and at individual weather-station locations. Importantly, the different analyses all confirmed the same broad conclusions.
A strength of this study is that the research- ers explored ecosystem responses to asymmet- ric warming using a variety of other large-scale data sets, and found similar patterns. One data set was for the net exchange of carbon between land and the atmosphere — a quantity that integrates photosynthesis and respiration, and which was inferred from a multi-year analysis3. Peng and co-workers found that this quantity correlated positively with daytime temperature variations for cool and wet boreal ecosystems, but negatively with night-time temperatures for these ecosystems. They also observed that the amplitudes of the seasonal cycles of car- bon dioxide levels measured at Point Barrow, Alaska, and Mauna Loa, Hawaii, vary in the same way with daytime and night-time tem- perature variations in boreal regions, but not in temperate areas. Peng et al. focused only on boreal and temperate ecosystems. The response to asymmetric warming of tropical and subtropi- cal ecosystems, which account for most CO2 exchange between the land and the atmos- phere, is not clear and merits further investiga- tion. Previous work4 at a well-studied tropical
forest revealed a negative correlation between tree growth and annual mean daily minimum temperatures, a response broadly similar to Peng and colleagues’ findings for boreal forests.
Tropical forests are thought to be vulnerable to warming5, with some evidence6 suggesting that they are already near high-temperature thresh- olds above which growth could be restricted. Future research could help to fill major gaps in our understanding of thermal tolerance and acclimation in tropical and subtropical plant species, and thus their response to warming5,7. So what are the physiological mechanisms that drive large-scale correlations between temperature variations and ecosystem metabo- lism? The commonly discussed mechanisms involve biochemical responses to temperature, but with some interesting twists. For example, the positive correlation found between night- time minimum temperatures and greenness in semi-arid grasslands is puzzling, but might be related to greater night-time plant respiration that stimulates increased daytime photosyn- thesis8. Increases in night-time respiration have also been invoked in a pioneering study9 of nocturnal warming that documented differ- ent plant responses in grassland: the dominant grass species declined in response to increases in night-time temperature during spring, whereas other plant species that use a different photosynthetic pathway increased in number. A research agenda to investigate these mechanisms further should include manipu- lative field and mesocosm experiments (in which small parts of a natural ecosystem are enclosed and warmed). Experimental warm- ing studies are lacking for many ecosystems. Even fewer night-time warming experiments have been conducted so far, with most being in shrublands10 or grasslands and croplands8; warming experiments that truly impose asym- metry between day and night warming are rare11. There is a particularly urgent need for warming studies in forests, which dominate the global carbon cycle and climate feedbacks. However, there are substantial technological challenges to conducting such experiments in large-statured ecosystems. Forest mesocosm experiments would require exceedingly com- plex and expensive facilities. Despite these limitations, Peng and colleagues’ results argue strongly for an increased focus on the dif- fering ecological impacts of night-time and daytime temperatures, to improve our ability to understand and predict how warming will affect Earth’s ecosystems. ■