NEWS & VIEWS RESEARCH

non-ageing-related deaths in calorie-restricted Center at San Antonio, San Antonio, 6. Leutenegger, W. & Cheverud, Int.J. J. Prim atol. 3, young-onset animals than in controls. Texas 78245, USA. 387-402 (1982). 7. Turturro, A. eta/. Gerontol.J. A54 , B492-B501 What might account for the differences e-mail: [email protected] (1999). between the studies’ results? One possibil­ 8. Welndruch, P., Walford, P. L, Fllglel, S. & Guthrie, D. 1. Messaoudi, I.etal. Proc. Natl Acad. Sci. USA 103, ity is diet composition. Whereas the diets J.N utr. 116,641-654 (1986). 19448-19453 (2006). 9. Holloszy, J. 0. Appl.J. Physiol. 8 2399-403 , (1997). were broadly similar in their overall con­2. Colman, R. J., Beasley, T. M., Allison, D. B. & 10.Hellbronn, L. K.e ta l. J. Am . Med.Assoo. 295, tent of carbohydrates, proteins and fats, Welndruch, R.J. Gerontol. A63,556-559 (2008). 1539-1548(2006). they differed in the specific types of such3. Colman, R. J.eta l. Sc/ence 325, 201-204(2009). 11.Fontana, L.etal. Am. J. Physiol. Endoorlnol. Metab. 4. Mattlson, J. A. et al. Nature 4 8 9318-321 , (2012). 293,E197-E202 (2007). nutrients. For instance, sucrose made up5. Colman, R. J. & Kemnltz, J. W. InM ethods In A g ing 12.Eontana, L.. Meyer. I. E.. Klein, S. & Flolloszy, J. 0. 28.5% of the WNPRC diet, but only 3.9% Researoh (ed. Yu, B. P.) 249-267 (CRC Press, 1998). Proc. N atlA oad. Sol. USA 101,6659-6663 (2004). of the NIA diet. Possibly related to this dif­ ference, more than 40% of the WNPRC control animals and only 12.5% of the NIA BIOGEOCHEMISTRY controls developed diabetes — although this metabolic malfunction was completely absent in the WNPRC calorie-restricted animals, but not in the treated animals in Drought and tropical the NIA study. Another difference is that the NIA controls were given an apportioned amount of foodsoil to emissions prevent obesity, whereas the WNPRC con­ trols were fed ad libitum (that is, they could Past research implied that positive feedback might exist between climate change eat as much as they pleased). Consequently,and greenhouse-gas emissions from soil. A study finds that drought-induced NIA control animals weighed less and were declines in such emissions from tropical forests could counter climate change. considerably longer-lived than the WNPRC controls. One interpretation of this observa­ tion is that the NIA controls were partiallyCORY C. CLEVELAND forest (Fig. 1) to test how drought affects soil restricted, which would account for the lack & BENJAMIN W. SULLIVAN emissions of COj, methane and nitrous oxide. of a survival effect of the treatment in the NIA Such experimental manipulations remain the study. Nevertheless, all animals in both stud­ ropical forests have been aptly describedgold standard for directly assessing the poten­ ies — even in the calorie-restricted groups — as the planet’s lungs. Tropical vegetationtial effects of environmental change on eco­ weighed more than wUd-caught monkeys'^. inhales vast amounts of carbon dioxidesystem processes, and can provide critical input Taken together, the contrasting results raise Tand exhales oxygen to the atmosphere, andto the development of better Earth-system an intriguing question about the nature andin so doing, provides a vital climate benefitmodels. The authors simulated the effects of robustness of restricting calorie intake in pri­by absorbing some of the nearly 9 petagramsa prolonged dry season by preventing rainfall mates. Is calorie restriction anything more (1 petagram is 10'* g) of carbon that hum an from reaching the forest floor for three months, than the elimination of excess fat? That might activities release into the atmosphere each year'. and measured greenhouse-gas emissions from be concluded if one interprets the control ani­W hat’s more, tropical soUs influence climate bysoil before, during and after the experiment. mals in the NIA study as being restricted to releasing large amounts of COj as a by-product After the simulated drought. Wood and a healthy weight, such that further restriction of the decomposition of organic matter, andSilver documented profound declines in the had little additional effect on longevity, ffow-through several other soil microbial processesefflux of COj from soil and increases in soil ever, researchers who use rodent models havethat both emit and consume greenhouse gases,consumption of methane. In addition, the long assumed that calorie restriction resultsincluding nitrous oxide and methane*. drought elicited surprising reductions in in more than just leanness and that it extends Important questions remain about thedenitrification (the production of nitrous life beyond its norm al limits. It is know n, foreffect of climate change on greenhouse-gas oxide), a process that is carried out by a tiny instance, that calorie restriction improves sur­fluxes from soil, and hence to what extent thefraction of the soil’s microbial community. vival in obese and non-obese mice and rats^,beneficial climate services provided by tropi­Together, these effects lowered the global- even at very high restriction levels*, and thatcal forests will persist in the future. Writing warming potential ofthe soil emissions. When survival is affected by calorie restriction in ain Global Bio geo chemical Cycles, W ood and the heat-trapping ability of nitrous oxide is different way from exercise-driven leanness'’.Silver* report that an experimentally inducedtaken into account, a substantial reduction in The same question arises in ongoingdrought had reduced greenhouse-gas emis­total global-warming potential from all three studies comparingoverweightor lean'* sions from tropical soil to the atmosphere, angases was accounted for by the decline of people with normal-weight controls. If calorieeffect that represents a negative feedback to denitrification. Because tropical forest soils are restriction defines only the food intake neededclimate change. a major source of nitrous oxide emissions to to maintain ahealthybody weight, then phar­ Many ofthe scenarios usedto predict futurethe atmosphere*, this result implies that a drier macological mimicry of such an effect mightclimate focus on rising temperature, butclimate could cause substantial reductions in improve the health of only the fraction of thechanges in the water balance of tropical forests global emissions of nitrous oxide from soil. population that is overweight. And, in thatmight have more immediate and pronounced However, extrapolating results from experi­ case — somewhat disappointingly — no spec­effects. Although substantial uncertaintyments such as these is particularly challenging tacular increase in health or longevity shouldremains, m ost climate models predict signifi­ for tropical forests. In Wood and Silver’s be expected. ■ cant declines in rainfall across large areas of experiment, drought effects varied depending the tropics''"*. So how might such declines alteron landscape position: ridge and valley plots Steven N. Austadis at the Barshop the net balance of the gases that contribute to had larger responses to experimental drought Institute for Longevity and Aging Studies, climate change? than did those on slopes. This variation has Department of Cellular and Structural To address this question. Wood and Silverconsiderable implications for models used to Biology, University o f Texas Health Science manipulated rainfall in a Puerto Rican tropicalsimulate biogeochemical responses to climate

13 SEPTEMBER 2012 VOL 48 9 I NATURE I 211

© 2012 Macmillan Publishers Limited. All rights reserved RESEARCH NEWS & VIEWS

welfare, our understanding of fundamental eco­ system processes, and their potential response to climate change, remains woefuUy incomplete in complex and diverse tropical forests. ■

Cory C. Clevelandand Benjamin W. Sullivan are in the Department of Ecosystem and Conservation Sciences, University o f Montana, Missoula, Montana 59812, USA. e-mails: [email protected]; [email protected]

1. Le Quere, C. e ta l. N ature Geosci. 2, 8 3 1-836 (2009). 2. Montzka, S. A., Dlugokencl^, E. J. & Butler, J. H. N ature476 , 4 3 - 5 0 (2011). 3. Wood,I E. & Silver, W. L. Glob. Blogeochem. Cycles 26, GB3005 http://dx.dol. org/10.1029/2010GB004014 (2012). 4. Cox, P. M.etal. Theor.Appl. Cllmatol. 78, 137-156 (2004). Figure 1 1 Luquillo Experimental Forest. Wood and Silver have studied the effect of drought on 5. Neelln, J. D., Munnlch, M., Su, H., Meyerson, J. E. greenhouse gases emitted from soils in the Luquillo Experimental Forest, Puerto Rico. & Holloway, 0. E.Proc. NatlAoad. Sol. USA 103, 6110-6115(2006). 6. Matson, P. A. & Vltousek, P. M.BloSolenoe 40, change, none of which is able to capture thelarge-scale experiments that can more com­ 667-672 (1990). complex interplay of factors that can createpletely resolve the potential effects of climate7. Townsend, A.P., Asner, G. P. & Cleveland, C. C. substantial biogeochemical heterogeneity change on trace gas emissions in tropical forests. Trends Eool. Evol. 23, 424-431 (2008). 8. Vltousek, P. M.etal. Ecosystems6, 762-772 (2003). within and across tropical forests^. The The authors’ data are critical for improving and 9. Porder, S. & Hllley, G. E.Blogeoohemlstry 102, authors’ results are a prime example of thevalidating models that predict ecosystem and 153-166(2011). challenge facing scientists — many tropical climate responses over large spatial and tem­10.CIark, D. A., Piper, S. C, Keeling, C. D. & Clark, D. B. Proc. N atlA oad. Sol. USA 100, 5852-5857 (2003). regions have highly complex landscapes withporal scales. Considering the pivotal role of 11.Zhao, M. & Running, S. W. Sc/ence329 , 940-9 4 3 varying nutrient availability, which in turn cantropical ecosystems in basic human health and (2010). regulate biological processes that influence greenhouse-gas production*’’’. The possibility of positive feedback between SOCIAL SCIENCE climate change and greenhouse-gas emissions from soil has been recognized for decades. Wood and Silver’s results suggest the opposite: declines in greenhouse-gas emissions follow­Poked to vote ing drought would reduce climate forcing. But as the authors highlight, the handful of studiesA Facebook message sent out during the 2010 US congressional elections influenced in which rainfall has been excluded from tropi­the voting behaviour of millions of people. The experiment illustrates the power of cal forests have shown positive, negative or nodigital social networks to spread behavioural change.S e e Le t t e r r 2 9 5 net effects of drought on soil greenhouse-gas emissions. Moreover, although it is crucial to consider the effects of drought alone, real-SINAN ARAL public health. On page 295 of this issue. Bond world emissions will hinge on the combined etaU present some of the most convincing evi­ effects of changing precipitation and tempera­ 1 ocial networks are the pathways throughdence to date that peer influence and digital ture, along with chronic shifts in atmospheric .which information, advice, resources andsocial signals can affect political mobilization.

CO 2 levels and nutrient deposition — factorsOS' s' support flow between people. They are Political mobilization has been central to that were not manipulated in the authors’essential for many people’s decision-making,recent discourse about the transformative experiment. Finally, studies'”’” have showncooperation and complex interdependence.effects of social media — for example, the part that the growth and carbon uptake of trees Yet although humans have almost alwaysthat technologies such as Facebook or Twitter in the tropics are highly sensitive to climate,lived in networks, advances in computingplayed in the protests collectively known as the but the experimental plots used in Wood andpower and new social technologies have onlyArab Spring, or may play in the forthcoming Silver’s study (1.54 square metres) were notrecently facilitated the development of formsUS presidential election. The question is: what large enough to simulate the potentiallyof networked communication that are auto­role do peer influence and digital social signals negative effects of drought on carbon uptakemating and accelerating the social signals have in mobilizing political expression? Do our through tree growth. The overall effects ofthat pulse through the hum an netw ork on friends’a behaviours inspire us to be politically drought on the greenhouse-gas balance of thedaily basis. The rapid dissemination of social active, to protest or to vote? sites therefore rem ain unknown. signals in these digital networks — status These questions may seem relatively simple In recent decades, tropical forests have givenupdates, tweets, likes, posts, shares and so to answer, given the right data. But several sta­ us a discount on anthropogenic COj emis­on — raises serious scientific questions: how,tistical challenges make it difficult to quanti­ sions by absorbing more greenhouse gaseswhen and to what extent do these signals tatively estimate peer influence in networks. than human activity produces. Wood and Sil­influence decision-making and the spread ofFor example, networks are homophUous — we ver’s findings suggest that tropical soils maybehaviours in society? If social influence drivestend to make friends with people like ourselves continue to offset greenhouse-gas emis­behaviour, then digital social signals could beand thus have preferences that are highly cor­ sions during drought. Perhaps more notably,used to promote widespread behaviour changerelated with those of our friends”. If two friends their study highlights the need for additional and thus to transform commerce, politics andadopt a behaviour, one immediately after the

212 I NATURE I VOL 489 13 SEPTEMBER 2012

© 2012 Macmillan Publishers Limited. All rights reserved