• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Gourdji etal prsb_2012
 

Gourdji etal prsb_2012

on

  • 850 views

 

Statistics

Views

Total Views
850
Views on SlideShare
422
Embed Views
428

Actions

Likes
0
Downloads
1
Comments
0

5 Embeds 428

http://dapa.ciat.cgiar.org 422
http://ciatcorporate.staging.wpengine.com 2
http://dapablog.local 2
http://translate.googleusercontent.com 1
http://www.networkedblogs.com 1

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Gourdji etal prsb_2012 Gourdji etal prsb_2012 Document Transcript

    • doi: 10.1098/rspb.2012.2190,2802013Proc. R. Soc. BSharon M. Gourdji, Ky L. Mathews, Matthew Reynolds, José Crossa and David B. Lobellin hot environmentsAn assessment of wheat yield sensitivity and breeding gainsSupplementary datatmlhttp://rspb.royalsocietypublishing.org/content/suppl/2012/11/30/rspb.2012.2190.DC1.h"Data Supplement"Referenceshttp://rspb.royalsocietypublishing.org/content/280/1752/20122190.full.html#ref-list-1This article cites 32 articles, 3 of which can be accessed freeSubject collections(18 articles)plant science(181 articles)environmental scienceArticles on similar topics can be found in the following collectionsEmail alerting service hereright-hand corner of the article or clickReceive free email alerts when new articles cite this article - sign up in the box at the tophttp://rspb.royalsocietypublishing.org/subscriptionsgo to:Proc. R. Soc. BTo subscribe toon December 5, 2012rspb.royalsocietypublishing.orgDownloaded from
    • rspb.royalsocietypublishing.orgResearchCite this article: Gourdji SM, Mathews KL,Reynolds M, Crossa J, Lobell DB. 2012 Anassessment of wheat yield sensitivity andbreeding gains in hot environments. Proc R SocB 280: 20122190.http://dx.doi.org/10.1098/rspb.2012.2190Received: 14 September 2012Accepted: 9 November 2012Subject Areas:environmental science, plant scienceKeywords:climate change, wheat, heat tolerance,breedingAuthor for correspondence:Sharon M. Gourdjie-mail: sgourdji@stanford.eduElectronic supplementary material is availableat http://dx.doi.org/10.1098/rspb.2012.2190 orvia http://rspb.royalsocietypublishing.org.An assessment of wheat yieldsensitivity and breeding gains inhot environmentsSharon M. Gourdji1,2, Ky L. Mathews3, Matthew Reynolds3, Jose´ Crossa3and David B. Lobell1,21Department of Environmental Earth System Science, Stanford University, Stanford, CA 94305, USA2Center on Food Security and the Environment, Stanford University, Stanford, CA 94305, USA3International Maize and Wheat Improvement Center (CIMMYT), Apdo. Postal 6-641,06600 Mexico D.F., MexicoGenetic improvements in heat tolerance of wheat provide a potential adap-tation response to long-term warming trends, and may also boost yields inwheat-growing areas already subject to heat stress. Yet there have been fewassessments of recent progress in breeding wheat for hot environments.Here, data from 25 years of wheat trials in 76 countries from the Inter-national Maize and Wheat Improvement Center (CIMMYT) are used toempirically model the response of wheat to environmental variation andassess the genetic gains over time in different environments and for differ-ent breeding strategies. Wheat yields exhibited the most sensitivity towarming during the grain-filling stage, typically the hottest part of theseason. Sites with high vapour pressure deficit (VPD) exhibited a less nega-tive response to temperatures during this period, probably associated withincreased transpirational cooling. Genetic improvements were assessed byusing the empirical model to correct observed yield growth for changesin environmental conditions and management over time. These ‘climate-corrected’ yield trends showed that most of the genetic gains in thehigh-yield-potential Elite Spring Wheat Yield Trial (ESWYT) were madeat cooler temperatures, close to the physiological optimum, with noevidence for genetic gains at the hottest temperatures. In contrast, theSemi-Arid Wheat Yield Trial (SAWYT), a lower-yielding nursery targetedat maintaining yields under stressed conditions, showed the strongest gen-etic gains at the hottest temperatures. These results imply that targetedbreeding efforts help us to ensure progress in building heat tolerance,and that intensified (and possibly new) approaches are needed to improvethe yield potential of wheat in hot environments in order to maintain globalfood security in a warmer climate.1. IntroductionWheat is the most widely grown crop in the world in terms of total harvestedarea [1], and currently provides an average of about 20 per cent of human cal-orie consumption [2]. Improvements in yield are essential to keep pace withpopulation growth and increased demand, yet long-term climate trends threa-ten to reduce wheat yields, or at least slow yield growth, in many regions.Spring wheat is already grown in many tropical and sub-tropical environmentsnear or past the optimal temperatures for wheat [3], particularly during thelater grain-filling portion of the season [4–6]. Modelling studies have shownthat even with adaptive agronomic changes to planting date and cultivar, pro-jected warming will still have a negative impact on wheat yields around theglobe [7]. Although elevated CO2 levels associated with warming may impartbenefits, which in many regions could outweigh the negative impacts of warm-ing for the next few decades [8], a warming climate still represents an importantadaptation challenge to the maintenance of past productivity gains.& 2012 The Author(s) Published by the Royal Society. All rights reserved.on December 5, 2012rspb.royalsocietypublishing.orgDownloaded from
    • Beyond relatively straightforward agronomic changes, anoften cited adaptation strategy is to breed new wheat var-ieties that combine improved heat tolerance with otherdesirable traits, such as disease resistance and high yieldpotential. The International Maize and Wheat ImprovementCenter (CIMMYT), based in Mexico, has been a leader inbreeding and disseminating improved varieties of wheat indeveloping countries since its inception in 1943, funded bythe Rockefeller Foundation and the government of Mexico.In the 1990s, it was estimated that 90 per cent of breadwheat releases in developing countries contained ancestryfrom one or more CIMMYT varieties [9], and today, morethan 75 per cent of the area planted to modern wheatvarieties in developing countries uses varieties developedby CIMMYT or its national-level partners (http://www.cimmyt.org/en/about-us/who-we-are). Recent studiesassessing long-term genetic gains of wheat lines released bythe CIMMYT Global Wheat Program show a continuousyield increase of approximately 0.7 per cent per year inboth low-yielding areas, and well-irrigated and high-rainfallareas [10,11].Given the major role of CIMMYT in international wheatimprovement, and the evidence of widespread warmingin major wheat growing regions in the past few decades[6,12], a relevant issue is the relative performance ofCIMMYT lines under different temperature conditions.A related question is whether nurseries that focus on breed-ing for targeted drought or heat stress show evidence ofmore rapid yield gains at high temperatures than the morestandard approach of breeding for high yield potentialunder optimal management, since this knowledge couldhelp us to guide future efforts.The current study addresses these questions using histori-cal datasets from three different spring bread wheat nurseriesat CIMMYT with different breeding goals: the Elite SpringWheat Yield Trial (ESWYT), which contains the highest-yield-ing varieties under ideal environmental and managementconditions; the Semi-Arid Wheat Yield Trial (SAWYT),where wheat is specifically bred to maintain yields underdry conditions that are frequently accompanied by heatstress; and the High Temperature Wheat Yield Trial(HTWYT), where wheat is bred for high temperature, irri-gated environments. Data from these nurseries are firstused in regression analysis to define the sensitivity of wheatyields to temperature and other environmental parameters.These regressions are then used to adjust observed yieldsfor changes in the locations and environmental conditionsof trials over time, a step necessary in order to assesstrue genetic gains under theoretically constant conditions.Inferred genetic gains are then compared across a range ofcool to hot temperatures in the grain-filling stage, in orderto determine the relative rate of gains in hot environments.2. Methods(a) DatasetsFor each year and breeding nursery, a new set of varieties issent annually by CIMMYT to a network of international collab-orators (called the International Wheat Improvement Network,IWIN), who grow this common germplasm under a range ofenvironmental conditions. For example, seasonal average temp-eratures vary in the database between 78 and 278C (see figure S1in the electronic supplementary material). The IWIN serves pri-marily to distribute improved germplasm globally, but the datareturned from these trials provide a valuable resource to assessgenotype  environment (G  E) interactions and long-termtrends in breeding. IWIN trial datasets have been used instudies to help us understand the impact of breeding nurseriessuch as the ESWYT [10,13], SAWYT [11,14] and HTWYT [15],among others.All trials in this dataset were generally requested to be wellmanaged in terms of water and fertilizer application, withtrials affected by lodging or disease filtered out. One exceptionis in the SAWYT nursery, where collaborators were encouragedto apply only enough irrigation to achieve germination, withfinal yields being largely dependent on in-season rainfall and/or stored soil moisture. As might be expected in this 76-country,25-year dataset, the implementation of management instructionsmost probably varied across trials within the database, as evi-denced by the wide range of yields in the database (fromapprox. 1 to approx. 10 tonnes ha21grown with the samevarieties in any given year; figure 1). However, among inter-national agricultural datasets, this one contains a relatively0–2wheat yields (tonnes ha–1)2–44–66–8>8Figure 1. Map of 349 trial locations included in analysis, colour-coded by average yields. Also shown (in grey) are the 31 877 stations used for weatherinterpolation.rspb.royalsocietypublishing.orgProcRSocB280:201221902on December 5, 2012rspb.royalsocietypublishing.orgDownloaded from
    • minimal amount of confounding factors, along with a widerange of environmental conditions, thereby enabling us toempirically assess relationships between wheat yield andenvironmental parameters throughout the crop life cycle. Suchan empirical analysis can both confirm current understandingand elucidate mechanisms for future prediction of crop yieldsin a changing climate.Yield data in this study represent means across genotypesand replications for a given trial location, sowing date and nur-sery. The trial means were calculated using only the subset ofgenotypes within a nursery each year that had similar phenology(i.e. +3.5 days of the mean trial heading date) in high-yielding(i.e. more than 5 tonnes ha21) environments, in order to excludethe confounding effects of large maturity ranges in stressenvironments. However, mean yields for the selected genotypeswere calculated for all trials, and included in the empirical modelregardless of yield level.Yield data were paired with reconstructed daily weather data,as described in the detailed methods section in the electronicsupplementary material. In short, daily temperature data wereobtained by combining high-spatial-resolution climatologieswith interpolation of anomalies from nearby station data, whiledaily relative humidity and radiation were obtained from satel-lite-based datasets. While water was assumed not to be alimiting factor for the irrigated trials, unfortunately little infor-mation was available regarding timing and amount of irrigationwater, nor were suitable soil moisture datasets available.(b) Empirical modelMean yields from a total of 1353 trials, pooled across nurseriesand planted from 1980 to 2009 in 349 unique locations(figure 1), were paired with weather data in a panel regression:yield ¼ cj þ an þ (gn  year) þ (b  W) þ 1;where cj are country fixed effects, an are nursery fixed effects, gnare yield trends by nursery, W is a set of environmental variablesdefined by growth stage and b are the coefficients on thesevariables.The environmental variables in W include: air temperature(both linear and squared terms), diurnal temperature range(DTR), shortwave radiation, day length, vapour pressure deficit(VPD), and interaction terms between VPD and temperature(linear and quadratic). Vapour pressure deficit was calculatedas the difference between saturation and actual vapour pressures,which were derived from daily minimum and maximum temp-eratures and relative humidity data. Each environmentalvariable included in the regression was averaged for threestages throughout the growing season [16]: vegetative (fromsowing to 300 growing degree-days, GDD, before heading),reproductive (from 300 GDD before heading to 100 GDD after)and grain-filling (from 100 GDD after heading to harvest).The linear and quadratic temperature terms allowed themodel to choose an ‘optimal’ temperature per growth stage,while DTR allowed for a differential response to day-versusnight-time temperatures. Radiation and day length affect photo-synthesis and development rates, respectively, and whileradiation tends to covary with temperature (especially in thevegetative stage), day length, along with temperature, is alsoan important determinant of phenology. VPD interacts with airtemperature through its impact on transpiration and, hence,canopy temperatures. A number of alternative models werealso tested (e.g. excluding day length and/or DTR, excludingthe temperature quadratic terms, or additionally includingstage length terms and their interaction with temperature). Theresults using these alternative models confirmed that the mainconclusions of the study were not sensitive to model formulation.Country fixed effects in the model accounted for averagedifferences in management or soil type by country, afteraccounting for variability explained by the weather-based predic-tors in the regression. We assumed that any remaining variationsin management or soil type within countries were not correlatedwith weather, and therefore did not bias our regression estimates.Trials were pooled across nurseries into a single model inorder to increase statistical power, and because of large differ-ences in the number of trials per nursery (959 from ESWYTversus 259 from SAWYT and 135 from HTWYT). However, struc-tural differences exist in germplasm, environment andmanagement between the nurseries (e.g. irrigation in ESWYTand HTWYT, but none in SAWYT). Nursery fixed effects andnursery-specific year trends, corresponding to varying levels ofgenetic yield growth, help us to account for these differences.As a sensitivity test, we also ran three separate nursery-specificregression models.(c) Assessment of genetic gains by nurseryand temperature binsGenetic gains were assessed by using the regression model tocorrect observed yield trends for changing environmental andmanagement conditions over time. (Here, ‘genetic gains’ refersto the relative performance of the changing germplasm in thetrial means over the lifetime of the nurseries.) Specifically, theregression model was used to predict yield changes in the datasetcaused by changes in environmental variables and countryeffects over time, and these partial fitted values are then sub-tracted from the observed yields. Linear time trends are finallyfitted to the residuals to assess ‘climate-corrected’ yield trends,or inferred genetic gains. Time trends in residuals can also beassessed for subsets of the data (e.g. by nursery and/or tempera-ture ranges). For this analysis, four temperature bins weredefined based on average temperature quartiles during thegrain-filling period, typically the hottest portion of the season.Genetic gains were not analysed for HTWYT, given the short life-span (1993–2004) and lack of significant observed yield trends inthis nursery. (Regardless, the HTWYT trials were retained in theempirical model in order to help increase statistical power.)Trends in environmental variables in the database primarilyreflect the changing mix of sites over time, rather than the cli-matic trends at the sites themselves. For example, there was astrong warming trend across trials in the grain-filling stage,which rose from an average daily temperature of 198C in 1983to 248C in 2009. However, the annual average global warmingtrend in the station database compiled for this study was onlyapproximately 0.88C over this same period. There was also asignificantly positive trend in radiation (by about 7%) in thegrain-filling stage over the period. These strong trends in temp-erature and radiation in the trial dataset were probably becauseof the growing proportion of trials in India, which rose from 5per cent in the earliest decade (1983–1992) to 30 per cent in thelast decade (2000–2009). India has some of the highest averagetemperatures in the grain-filling stage (26.08C versus a mean of21.98C), which may have depressed overall observed yieldtrends in recent years, although the higher radiation wouldhave had an opposing effect. Our method for assessing geneticgains should correct for any trends in environmental variablesand country makeup in the database, regardless of their source.3. Results and discussion(a) Results from empirical modelThe regression results exhibited a clear influence of tempera-ture on trial mean yields, with significant interactionsbetween temperature and VPD. Nearly half of all yieldvariability was captured by the regression model (adjustedrspb.royalsocietypublishing.orgProcRSocB280:201221903on December 5, 2012rspb.royalsocietypublishing.orgDownloaded from
    • r2¼ 0.44), with weather providing a substantial fraction of theexplanatory power (i.e. the adjusted r2¼ 0.23 for a model withonly weather and no country fixed effects). The country fixedeffects showed a substantial and significant variation acrosscountries, with countries such as Zimbabwe and Canadahaving a strong yield benefit (approx. 3 tonnes ha21) relativeto what is predicted by weather alone, and Nepal and Algeriahaving a significant yield penalty (approx. 21 tonnes ha21).Figure 2 shows the inferred yield response to temperaturefrom the regression model for each growth stage under bothhigh and low VPD conditions. These curves represent thepartial fitted values from the model, by growth stage, oftemperature (linear and quadratic terms), VPD, and the inter-action terms between temperature and VPD. Response curvesare shown for both high and low VPD in order to illustratethe importance of temperature  VPD interactions. (In thesecurves, the VPD values at each temperature were specifiedas the 10th and 90th percentiles of VPD at that temperature,as predicted from a quantile regression.)Yieldsdeclinesignificantlymorerapidlyathightemperaturesunder low VPD, or humid conditions, than under high VPD,especially in the grain-filling stage (figure 2). This relationshipprobably reflects the fact that, assuming sufficient soil moisture,more plant transpiration occurs under high VPD conditions, inwhich canopy temperatures are cooled below air temperature.This result agrees with past observations that yields in hot,low-humidity environments are strongly positively correlatedwith stomatal conductance and canopy temperature depression[17–19]. While VPD is positively correlated with radiation(correlation is roughly 0.35 for each of the three growth stages),the regression model includes a separate term for radiation,and thus the VPD  temperature interaction is unlikely to bean artefact of higher radiation in high-VPD environments.In the reproductive stage, interactions between tempera-ture and VPD were less significant than in the vegetative orgrain-filling stages. In this stage, it is probable that higherVPD for a given temperature has opposing effects on yield(i.e. higher transpiration cooling, but also more sensitivityto water stress, and hence closed stomata). For example, inmodels fitted to just the SAWYT and HTWYT trials(see electronic supplementary material, figure S2a–c), therewas a more negative response to warming for the high rela-tive to low VPD trials in the reproductive stage, whereasthe converse is true for an ESWYT-only model, which wasfitted to trials with presumably more sufficient soil moisture.Especially for the non-irrigated trials in SAWYT, it is probablethat a higher exposure to water stress during this periodplayed a role in negating the benefits of high VPD.Overall, warming was beneficial during the vegetativestage up to approximately 208C. In the reproductive stage,optimal temperatures were approximately 128C for both highand low VPD trials, with significantly negative impacts fromwarming at temperatures more than 168C. In the grain-fillingstage, warming had a negative impact on yields across thefull range of temperatures in the database. Given the higheraverage air temperatures during grain-filling relative to thoseearlier in the season (see the electornic supplementarymaterial, figure S3a–c), it may be that canopy temperaturesin this growth stage (especially in humid conditions) oftenreached physiological limits in terms of plant metabolism [20].The regression results also allowed us to infer relation-ships between the ancillary variables and yield, in additionto temperature (results not shown). For example, we saw avery positive and significant relationship between radiationand yield during the grain-filling stage with an inferred coef-ficient of 0.1 tonnes ha21(MJ (m2 day)21)21. Coefficientson radiation were negative, but insignificant, during thevegetative and reproductive stage, most probably becauseof their correlation with other variables and/or processescounteracting what would otherwise be a positive associ-ation. Day length has a significantly negative coefficient inthe vegetative stage, most probably because of faster develop-ment and lower potential grain number associated withlonger photoperiod [21]. Although day- and night-time temp-eratures have been shown to have differential impacts ongrain yield in previous studies [22], results here did notshow a significant relationship between diurnal temperaturerange (DTR) and yield in any of the three growth stages.(b) Inferred response to þ28C warmingAs an overall summary of the regression results, figure 3 dis-plays the estimated yield loss (or gain) in tonnes ha21from28C warming throughout the growing season for trials inthe 1990s and 2000s. The projected yield changes owing towarming were calculated by comparing the actual fittedvalues from the regression with recomputed fitted valuesthat reflect historical temperatures þ28C across stages,along with associated changes in VPD and DTR. Radiationand relative humidity values were assumed to stay constant.Radiation trends are primarily affected by trends in air pol-lution and aerosol-cloud feedback effects [23,24], whereasrelative humidity is projected to stay constant on a globalbasis with greenhouse-gas-induced warming [25,26].The model predicted that 95 per cent of trials would havea lower mean yield from a þ28C warming, with a meanloss of approximately 0.3 tonnes ha21, and a range of0.3 tonnes ha21gain to 1.4 tonnes ha21loss. This translatedinto an average loss of approximately 11 per cent of currentyields across the globe. In general, the regions that weremost subject to warming-related losses already had high sea-sonal average temperatures (see the electronic supplementary−4−3−2−101veg − high VPDveg − low VPDrep − high VPDrep − low VPDGF − high VPDGF − low VPD0 5 10 15 20 25 30average temperatures by growth stage (°C)yieldresponse(tonnesha–1)Figure 2. Inferred yield response to temperature from regression model forthree growth stages (veg ¼ vegetative; rep ¼ reproductive; GF ¼ grainfilling), with the response curves fitted separately for high and low VPD trials.The curves have been normalized to equal 0 at 128C. The line thicknesscorresponds to the significance of the slope (i.e. thin: NS, medium: p 0.1,thick: p 0.05, where the p-values are from a two-sided t-test.)rspb.royalsocietypublishing.orgProcRSocB280:201221904on December 5, 2012rspb.royalsocietypublishing.orgDownloaded from
    • material, figure S4), such as in Sudan, Myanmar and Paraguay,where projected losses average approximately 61, 58 and 35 percent, respectively, of current yields. Humidity also played arole, with regions such as the Nile basin in Egypt, Iran andnorthwest Mexico showing only modest projected losses forrelatively high current seasonal temperatures, because oftheir dry, high VPD conditions. Overall, the Mediterraneanbasin showed the least amount of losses from warming, andin some cases slight gains, owing to low humidity and lowertemperatures associated with winter planting in the region.Nursery-specific models fitted to only ESWYT or SAWYTtrials showed that SAWYT germplasm is more resilient thanESWYT to warming up to approximately 218C, when bothmodels began to converge in terms of their negative responseto future warming (see the electronic supplementarymaterial, figure S4b). Finally, we note that higher atmosphericCO2 should offset some of the temperature-related declines inyield shown here for wheat, a C3 crop sensitive to CO2fertilization. However, the magnitude of CO2 fertilization infield conditions, with associated interactions between nutri-ent, water and temperature limitations, is still subject todebate [27–29].(c) Estimated genetic gains by nursery and temperaturebinsBoth the observed and climate-corrected yield trends inESWYT were positive in all of the grain-filling temperaturebins since 1983 (figure 4a), although trends were only signifi-cant in the two coolest bins, closer to the optimal temperaturefor wheat yields [30–32]. Inferred genetic gains in thewarmer bins were insignificant after accounting for trendsin environment and location (i.e. country effects). Theenvironmental trends since 1983 had small net effects inESWYT on average, with negative impacts of a warmingtrend counteracted by other positive environmental effects(mainly in radiation; figure 4c).In contrast to ESWYT, the largest and only significant pro-gress for climate-corrected yields for SAWYT was observed inthe hottest temperature bin (figure 4b). Also in contrast toESWYT, the overall trends in environmental effects werenegative for SAWYT because of rising temperature trendsacross growth stages, which were not counteracted by radi-ation increases (figure 4d). For both ESWYT and SAWYT,country effects have been negative, because of the increasingproportion of trials in South Asia (figure 4c,d).Two robustness checks help one to support the findingthat ESWYT gains were concentrated at cooler sites, whileSAWYT gains came mainly from hotter sites. First, trendswere computed for varying start years from the beginningof the nursery to 10 years before the end (i.e. 2000). Weshow a version of figure 4a based on a start date of 1993,which is the year in which SAWYT started (see the electronicsupplementary material, figure S5). Over this shorter period,the ESWYT climate-corrected yield trends, or genetic gains,were even more skewed towards the coolest grain-fillingtemperature bin (less than 19.58C), with a ratio of 21 timesgrowth in the coolest relative to the warmest bin.Second, the analysis was repeated using nursery-specificmodels to correct for environment and country effects.Results were very similar (see the electronic supplementarymaterial, figure S6), with the exception that SAWYT trialsshowed stronger genetic gains across all temperature binsas compared with results from the pooled model. This dif-ference reflects the fact that the SAWYT-specific modelexhibited stronger responses to warming, and therefore pro-duced a stronger correction for the observed warming trends.Overall, these results demonstrate that genetic gains candiverge strongly from observed yield trends, especially soin the SAWYT nursery, where relatively flat yield trendsmask much stronger genetic gains evident in this breedingprogramme. Moreover, the significant genetic gains at hightemperatures in SAWYT, but not ESWYT, indicate that a tar-geted breeding programme helps one to ensure success inbreeding for heat tolerance. It should be noted that theseresults can also be explained by the environments inMexico, in which new varieties were sown and selected forthe two nurseries. For example, the median seasonal temp-erature across ESWYT trials in Mexico is 17.98C, whereasthat for SAWYT is 20.18C, with most probable even highercanopy temperatures owing to drought conditions and alack of evaporative cooling.<–0.8–0.8 to – 0.6–0.6 to – 0.4–0.2 to 0> 0–0.4 to – 0.2losses from +2°Cwarming (tonnes ha–1)Figure 3. Map of trial locations since 1990 with estimated loss/gain from þ28C warming; multiple years and sites clustered within a 100 km distance are averagedfor illustration purposes.rspb.royalsocietypublishing.orgProcRSocB280:201221905on December 5, 2012rspb.royalsocietypublishing.orgDownloaded from
    • CO2 fertilization has also probably played a role in theinferred ‘genetic’ gains shown here for both nurseries, givena 45 ppm rise in atmospheric CO2 from 1983 to 2009 (asmeasured at Mauna Loa, HI). Rising atmospheric CO2 mayhave especially promoted yield gains in SAWYT, because ofdecreased stomatal conductance and increased water savingsat higher CO2 under drought conditions [33]. However, giventhe covariance between variety improvement and increasingatmospheric CO2 in recent years, it is difficult to statisticallyidentify the CO2 effect in this study.Understanding the underlying mechanisms behind thedifferential yield progress for ESWYT and SAWYT at hottemperatures is beyond the scope of this paper. However,we offer a few observations. First, one strategy to withstandhotter temperatures while maintaining similar growth dur-ations would be to lengthen the accumulated temperature(or GDD) requirements for development. In the CIMMYTdatabase, both ESWYT and SAWYT showed positive trendsfor GDD in the vegetative stage. Consistent with the greaterinferred genetic gains for SAWYT versus ESWYT at hot temp-eratures in the grain-filling stage, the positive trend invegetative GDD requirements was more than two timeshigher for SAWYT than for ESWYT over a common time-frame (1993–2009, 13 versus 5 degree-days per year). GDDrequirements for the vegetative stage increased by 23 percent in SAWYT over the lifetime of the nursery, which wasenough to maintain a constant duration of this period,despite significant warming because of a combination of cli-mate trends and a changing mix of sites. Since the potentialgrain number is positively associated with both vegetativeduration [21,34] and yields, the increased GDD requirementsin this period have probably played a role in maintainingyield performance in hot conditions.A second potential mechanism relates to grain-fillingrates. The grain-filling period became significantly shorterin SAWYT over time owing to rising temperatures in thisgrowth stage (i.e. 12 days shorter for a 4.58C average risefrom 1993 to 2009, with no evidence of increasing thermalrequirements in this final growth stage). Yet grain weightdata, available for only approximately 20 per cent of therecords in the database, show a 4 per cent increase forSAWYT. A higher grain weight, together with a shortergrain-filling period, implies an increased grain-filling rateper day (perhaps to a small extent owing to temperature[35], but probably owing to variety improvement). The datasupport a significant increase in grain-filling rates for bothESWYT and SAWYT of 0.004 and 0.008 mg (kernel Âday)21year21, respectively.Thus, increased thermal time to flowering and highergrain-filling rates appear to be two sources of yield0.15(a) (b)(c) (d)0.100.05trend(tonnesha–1year–1)trend(tonnesha–1year–1)–0.05start year of trend start year of trend1985 1990 1995 2000<19.5°C 19.5–21.8°C 21.8–24.8°C ≥24.8°C <19.5°C 19.5–21.8°C 21.8–24.8°C ≥24.8°C00.10observedclimate-corrected1985 1990 1995 2000observedenvironmentcountrygenetic0.05–0.050Figure 4. (a,b) Observed and climate-corrected yield trends from earliest start year of nursery, binned by average temperatures in the grain-filling period for(a) ESWYT and (b) SAWYT. The error bars are at a p ¼ 0.05 significance level from a one-sided t-test. (c,d) Trends in environmental and country effects, andobserved and ‘climate-corrected’ yields plotted as a function of the start year of the trend for (c) ESWYT and (d) SAWYT nurseries.rspb.royalsocietypublishing.orgProcRSocB280:201221906on December 5, 2012rspb.royalsocietypublishing.orgDownloaded from
    • maintenance and growth at high temperatures in the SAWYTdatabase. Increased water savings at higher atmospheric CO2may also play a role, given increased evaporative demand athigher temperatures. However, it should be noted that increas-ing GDD requirements, faster grain-filling rates and reducedstomatal conductance at higher atmospheric CO2 will notprevent irreversible damage from extreme heat episodes [36],particularly during the reproductive period. Therefore, otherbreeding strategies, such as speeding development to forceflowering earlier in the season, may be beneficial in someenvironments and for risk-averse farmers [20].4. ConclusionsDecades of wheat breeding efforts at CIMMYT have resulted inan extensive trial database of wheat yields under varyingenvironmental conditions. This database provides a valuablemeans of empiricallyassessing the response of wheat to environ-mental variation, and also the genetic gains over time and indifferent environments that are associated with different breed-ing strategies. Consistent with previous studies, our empiricalmodel showed the most negative response to high temperaturesin the grain-filling phase under low VPD, or humid conditions.Assuming sufficient water supply, higher VPD for a given temp-erature leads to more transpiration cooling, lower canopytemperatures and a less negative response to warming. A nega-tive response to warming was also seen during the reproductivephase at average temperatures above 138C, but a higher sensi-tivity to water stress during this phase reduced the relativeadvantage of high VPD trials.With current breeding strategies, projected future climatechange will probably put a drag on growth in global springwheat yields, and may even depress them, especially inlocations where wheat is already grown in hot conditions(particularly in south Asia, and also parts of sub-SaharanAfrica, the Middle East and Latin America). Agronomicchanges (e.g. shifts in planting dates or locations, andimproved access to inputs) in combination with CO2 fertiliza-tion, can potentially help to mitigate these losses. However,new varieties of wheat with high yield potential in hotenvironments are required in order to adequately preparefor projected temperature rises of approximately 28C by 2050.ESWYT and SAWYT epitomize two different breedingstrategies, to develop wheat varieties that are (i) high-yieldingin irrigated and high rainfall conditions, but potentially sen-sitive to abiotic stresses, and (ii) tolerant to drought and heatstress under rain-fed conditions, but with lower yield poten-tial. This study finds that most progress in ESWYT to date hasbeen achieved at the cooler temperatures in the grain-fillingphase, closest to the optimal temperatures for wheat pro-duction. In contrast, progress has been made in SAWYTacross temperature bins, but most significantly in the hottestbin, thereby building greater amounts of heat tolerance intothe germplasm. Two potential mechanisms for the relativelyhigher genetic gains in SAWYT at high temperatures relate tolonger vegetative GDD requirements and faster grain-fillingrates. It will also be imperative to build resilience to extremeheat into future germplasm, especially during the reproduc-tive phase, in order to avoid the risk of complete cropfailure with more frequent heat waves.The lack of yield increase to date for the highest-yieldingvarieties under hot conditions, as shown in this study, indicatesthe need for new and intensified efforts to achieve these gains.This will require a combined effort, using genetic diversity withphysiological and molecular breeding, and bioinformatic tech-nologies, along with the adoption of improved agronomicpractices by farmers. Although many trade-offs exist betweenhigh yield potential and stress adaptation, in our view itshould be feasible to achieve both goals as long as hot environ-ments are systematically included in the selection process for abreeding strategy like ESWYT. Given that disease resistance,pest resistance and maintaining grain quality will also continueto be priorities, additional resources may be needed to simul-taneously achieve all of these targets.We gratefully acknowledge Mateo Vargas, who prepared much of thedata for analysis in this study, and Thomas Payne, who contributedvaluable interpretation of the results. This work was supported by agrant from the Rockefeller Foundation. The data associated with thisstudy are deposited in the Dryad Repository: http://dx.doi.org/10.5061/dryad.525vm.References1. Leff B, Ramankutty N, Foley JA. 2004 Geographicdistribution of major crops across the world. Glob.Biogeochem. Cycle 18, GB1009. (doi:10.1029/2003GB002108)2. Food and Agriculture Organization 2012 FAOSTAT:Food Balance Sheets. See faostat.fao.org.3. Kosina P, Reynolds M, Dixon J, Joshi A. 2007Stakeholder perception of wheat productionconstraints, capacity building needs, and researchpartnerships in developing countries. Euphytica157, 475–483. (doi:10.1007/s10681-007-9529-9)4. Wardlaw I, Wrigley C. 1994 Heat tolerance intemperate cereals—an overview. Aust. J. PlantPhysiol. 21, 695–703. (doi:10.1071/PP9940695)5. Ortiz R et al. 2008 Climate change: Can wheat beatthe heat? Agric. Ecosyst. Environ. 126, 46–58.(doi:10.1016/j.agee.2008.01.019)6. Asseng S, Foster I, Turner NC. 2011 The impact oftemperature variability on wheat yields. Glob.Change Biol. 17, 997–1012. (doi:10.1111/j.1365-2486.2010.02262.x)7. Deryng D, Sacks WJ, Barford CC, Ramankutty N.2011 Simulating the effects of climate andagricultural management practices on global cropyield. Glob. Biogeochem. Cycle 25, GB2006. (doi:10.1029/2009GB003765)8. Easterling W et al. In press. Food, fibre, and forestproducts. In Climate change 2007: impacts, adaptationand vulnerability (eds M Parry, O Canziani, J Palutikof,P van der Linden, C Hanson), pp. 273–313. Cambridge,UK: Cambridge University Press.9. Heisey PW, Lantican MA, Dubin HJ. 2002 Impacts ofInternational Wheat Breeding Research in DevelopingCountries, 1966–97. Mexico DF, Mexico: CIMMYT.10. Sharma RC, Crossa J, Velu G, Huerta-Espino J,Vargas M, Payne TS, Singh RP. 2012 Geneticgains for grain yield in CIMMYT spring breadwheat across international environments. CropSci. 52, 1522–1533. (doi:10.2135/cropsci2011.12.0634)11. Manes Y, Gomez HF, Puhl L, Reynolds MP, Braun HJ,Trethowan RM. 2012 Genetic yield gains of theCIMMYT international semi-arid wheat yield trialsfrom 1994 to 2010. Crop Sci. 52, 1543–1552.(doi:10.2135/cropsci2011.10.0574)12. Lobell DB, Schlenker W, Costa-Roberts J. 2011Climate trends and global crop production since1980. Science 333, 616–620. (doi:10.1126/science.1204531)13. Trethowan RM, van Ginkel M, Rajaram S. 2002Progress in breeding wheat for yield and adaptationrspb.royalsocietypublishing.orgProcRSocB280:201221907on December 5, 2012rspb.royalsocietypublishing.orgDownloaded from
    • in global drought affected environments. Crop Sci.42, 1441–1446. (doi:10.2135/cropsci2002.1441)14. Trethowan RM, Crossa J, van Ginkel M, Rajaram S.2001 Relationships among bread wheatinternational yield testing locations in dry areas.Crop Sci. 41, 1461–1469. (doi:10.2135/cropsci2001.4151461x)15. Lillemo M, van Ginkel M, Trethowan RM, HernandezE, Crossa J. 2005 Differential adaptation of CIMMYTbread wheat to global high temperatureenvironments. Crop Sci. 45, 2443–2453.(doi:10.2135/cropsci2004.0663)16. Dreccer MF, Chapman SC, Ogbonnaya FC,Borgognone MG, Trethowan RM. 2008 Crop andenvironmental attributes underpinning genotype byenvironment interaction in synthetic-derived breadwheat evaluated in Mexico and Australia.Aust. J. Agric. Res. 59, 447–460. (doi:10.1071/AR07220)17. Fischer RA, Rees D, Sayre KD, Lu ZM, Condon AG,Saavedra AL. 1998 Wheat yield progress associatedwith higher stomatal conductance andphotosynthetic rate, and cooler canopies. Crop Sci.38, 1467–1475. (doi:10.2135/cropsci1998.0011183X003800060011x)18. Suyker AE, Verma SB. 2010 Coupling of carbondioxide and water vapor exchanges of irrigated andrainfed maize–soybean cropping systems andwater productivity. Agric. For. Meteorol. 150, 553–563. (doi:10.1016/j.agrformet.2010.01.020)19. Reynolds MP, Balota M, Delgado M, Amani I, FischerRA. 1994 Physiological and morphological traitsassociated with spring wheat yield under hot,irrigated conditions. Aust. J. Plant Physiol. 21,717–730 (doi:10.1071/PP9940717).20. Reynolds MP, Ortiz-Monasterio JI, McNab A (eds)2001 Application of physiology in wheat breeding.Mexico DF, Mexico: CIMMYT.21. Gonzalez FG, Slafer GA, Miralles DJ. 2005 Pre-anthesis development and number of fertile floretsin wheat as affected by photoperiod sensitivitygenes Ppd-D1 and Ppd-B1. Euphytica 146,253–269. (doi:10.1007/s10681-005-9021-3)22. Lobell DB, Ortiz-Monasterio JI. 2007 Impacts of dayversus night temperatures on spring wheat yields: acomparison of empirical and CERES modelpredictions in three locations. Agron. J. 99,469–477. (doi:10.2134/agronj2006.0209)23. Stanhill G, Cohen S. 2001 Global dimming: areview of the evidence for a widespread andsignificant reduction in global radiation withdiscussion of its probable causes and possibleagricultural consequences. Agric. Forest Meteorol.107, 255–278. (doi:10.1016/S0168-1923(00)00241-0)24. Wild M. 2012 Enlightening global dimming andbrightening. Bull. Amer. Meteorol. Soc. 93, 27–37.(doi:10.1175/BAMS-D-11-00074.1)25. Dai A. 2006 Recent climatology, variability, andtrends in global surface humidity. J. Clim. 19,3589–3606. (doi:10.1175/JCLI3816.1)26. Willett KM, Gillett NP, Jones PD, Thorne PW. 2007Attribution of observed surface humidity changes tohuman influence. Nature 449, 710–712. (doi:10.1038/nature06207)27. Lobell DB, Field CB. 2008 Estimation of the carbondioxide (CO2) fertilization effect using growth rateanomalies of CO2 and crop yields since 1961. Glob.Change Biol. 14, 39–45. (doi:10.1111/j.1365-2486.2007.01476.x)28. Li W, Han X, Zhang Y, Li Z. 2007 Effects of elevatedCO2 concentration, irrigation and nitrogenous fertilizerapplication on the growth and yield of spring wheatin semi-arid areas. Agric. Water Manage. 87,106–114. (doi:10.1016/j.agwat.2006.06.027)29. Long SP, Ainsworth EA, Leakey ADB, Nosberger J,Ort DR. 2006 Food for thought: lower-than-expected crop yield stimulation with rising CO2concentrations. Science 312, 1918–1921.(doi:10.1126/science.1114722)30. Farooq M, Bramley H, Palta JA, Siddique KHM. 2011Heat stress in wheat during reproductive and grain-filling phases. Crit. Rev. Plant Sci. 30, 491–507.(doi:10.1080/07352689.2011.615687)31. Hatfield JL, Boote KJ, Kimball BA, Ziska LH,Izaurralde RC, Ort D, Thomson AM, Wolfe D. 2011Climate impacts on agriculture: implications for cropproduction. Agron. J. 103, 351–370. (doi:10.2134/agronj2010.0303)32. Luo Q. 2011 Temperature thresholds and cropproduction: a review. Clim. Change 109, 583–598.(doi:10.1007/s10584-011-0028-6)33. Wall GW et al. 2006 Interactive effects of elevatedcarbon dioxide and drought on wheat. Agron. J. 98,354–381. (doi:10.2134/agronj2004.0089)34. Fischer RA. 2011 Wheat physiology: a review ofrecent developments. Crop Pasture Sci. 62, 95–114.(doi:10.1071/CP10344)35. Tashiro T, Wardlaw I. 1989 A comparison of theeffect of high-temperature on grain development inwheat and rice. Ann. Bot. 64, 59–65.36. Lobell DB, Sibley A, Ortiz-Monasterio JI. 2012Extreme heat effects on wheat senescence in India.Nat. Clim. Chang. 2, 186–189. (doi:10.1038/NCLIMATE1356)rspb.royalsocietypublishing.orgProcRSocB280:201221908on December 5, 2012rspb.royalsocietypublishing.orgDownloaded from