climate shifts within major agricultural seasons for +1.5 ... · 96 scenarios on agriculture and...
TRANSCRIPT
1
Climate Shifts within Major Agricultural Seasons for +1.5 and +2.0 °C Worlds: 1
HAPPI Projections and AgMIP Modeling Scenarios 2 3
Alex C. Ruane1 4
Meridel M. Phillips2,1 5
Cynthia Rosenzweig1 6
7 1NASA Goddard Institute for Space Studies, New York, NY, USA 8
2Columbia University Center for Climate Systems Research, New York, NY, USA 9
10
Original submission: October 31st, 2017 11
Revised submission: April 22nd, 2018 12
Accepted at Agricultural and Forest Meteorology: May 12th, 2018 13
14
Corresponding Author: 15
Alex Ruane 16
NASA Goddard Institute for Space Studies 17
2880 Broadway 18
New York, NY 10025 19
21
22
https://ntrs.nasa.gov/search.jsp?R=20180007597 2020-05-31T10:08:45+00:00Z
2
Abstract: This study compares climate changes in major agricultural regions and current 23
agricultural seasons associated with global warming of +1.5 or +2.0 ºC above pre-industrial 24
conditions. It describes the generation of climate scenarios for agricultural modeling 25
applications conducted as part of the Agricultural Model Intercomparison and 26
Improvement Project (AgMIP) Coordinated Global and Regional Assessments. Climate 27
scenarios from the Half a degree Additional warming, Projections, Prognosis and Impacts 28
project (HAPPI) are largely consistent with transient scenarios extracted from RCP4.5 29
simulations of the Coupled Model Intercomparison Project phase 5 (CMIP5). Focusing on 30
food and agricultural systems and top-producing breadbaskets in particular, we distinguish 31
maize, rice, wheat, and soy season changes from global annual mean climate changes. 32
Many agricultural regions warm at a rate that is faster than the global mean surface 33
temperature (including oceans) but slower than the mean land surface temperature, leading 34
to regional warming that exceeds 0.5 ºC between the +1.5 and +2.0 ºC Worlds. Agricultural 35
growing seasons warm at a pace slightly behind the annual temperature trends in most 36
regions, while precipitation increases slightly ahead of the annual rate. Rice cultivation 37
regions show reduced warming as they are concentrated where monsoon rainfall is 38
projected to intensify, although projections are influenced by Asian aerosol loading in 39
climate mitigation scenarios. Compared to CMIP5, HAPPI slightly underestimates the 40
CO2 concentration that corresponds to the +1.5 ºC World but overestimates the CO2 41
concentration for the +2.0 ºC World, which means that HAPPI scenarios may also lead to 42
an overestimate in the beneficial effects of CO2 on crops in the +2.0 ºC World. HAPPI 43
enables detailed analysis of the shifting distribution of extreme growing season 44
temperatures and precipitation, highlighting widespread increases in extreme heat seasons 45
3
and heightened skewness toward hot seasons in the tropics. Shifts in the probability of 46
extreme drought seasons generally tracked median precipitation changes; however, some 47
regions skewed toward drought conditions even where median precipitation changes were 48
small. Together, these findings highlight unique seasonal and agricultural region changes 49
in the +1.5°C and +2.0°C worlds for adaptation planning in these climate stabilization 50
targets. 51
52
53
4
1. Introduction 54
Changes in the Earth’s climate present an array of challenging consequences for 55
humankind in the coming decades (IPCC, 2013). Food systems, in particular, are at risk 56
from rising temperatures, droughts, floods, and extreme events (Beddington et al., 2012; 57
Porter et al., 2014; Brown et al., 2015; FAO, 2016). In light of observed and projected 58
impacts, representatives of 196 countries signed the United Nations Framework 59
Convention on Climate Change (UNFCCC) Paris Agreement (UNFCCC, 2016) in 60
December 2015, agreeing to pursue efforts that would limit the global mean surface 61
temperature increase to 2°C above pre-industrial temperature levels, with ambition to limit 62
the rise below +1.5°C (both would require substantial deviation from the current business-63
as-usual greenhouse gas emission pathway). These commitments do not eliminate 64
anthropogenic climate change or the need to sustainably adapt natural and human systems, 65
but seek to diminish the risk of increasingly dangerous impacts associated with more 66
extreme climate changes (IPCC, 2014). 67
68
The Paris Agreement is largely focused on major policy and technological issues related to 69
how climate change may be mitigated, but has also given rise to the scientific question: 70
What type of world would be achieved should efforts to mitigate greenhouse gas emissions 71
and land use change be successful at keeping the world at either +1.5 ºC or +2.0 ºC above 72
pre-industrial temperature levels? Answers to this question are aided by the creation and 73
analysis of hypothetical climate states stabilized at +1.5 ºC and +2.0 ºC above pre-industrial 74
conditions, hereafter referred to as the ‘+1.5 ºC World’ and ‘+2.0 ºC World’. These 75
projections allow us to compare future and current conditions and evaluate potentially 76
5
beneficial or detrimental climate impacts across regions, sectors, and populations. These 77
impacts, the resulting need for adaptation, and pressure on current land use patterns must 78
be considered along with the costs of mitigation in determining future societal pathways 79
and climate targets. 80
81
Here we describe likely challenges for the agricultural sector under +1.5 and +2.0 ºC 82
Worlds, identifying key characteristics of projected changes for current growing seasons 83
in major breadbaskets. Agro-climatic indicators may contribute to more efficient 84
adaptation interventions in the agricultural sector, potentially including shifts in planting 85
date, seed varieties, irrigation applications, crop insurance, farm systems, food value 86
chains, and the management of pests, diseases, and weeds (Howden et al., 2007; 87
Rosenzweig and Tubiello, 2007; Yadav et al., 2011; Rickards and Howden, 2012). 88
89
The agro-climatic projections serve as the driving conditions for agricultural model 90
analyses utilizing protocols developed within the Agricultural Model Intercomparison and 91
Improvement Project (AgMIP; Rosenzweig et al., 2013; 2015; Ruane et al., 2017). 92
AgMIP’s Coordinated Global and Regional Assessments (CGRA) utilize a simulation 93
framework linking biophysical and economic impacts on global and local scales to 94
consistently evaluate the impacts of climate, socioeconomic development, and policy 95
scenarios on agriculture and food security (Rosenzweig et al., 2016, 2018). CGRA analyses 96
of the +1.5 and +2.0 ºC Worlds are thus grounded in climate scenarios with consistent 97
methodologies and assumptions, allowing CGRA findings to focus on emergent cross-98
disciplinary and cross-scale findings. 99
6
100
101
2. Methodology 102
The targets of +1.5 and +2.0 ºC global warming, here defined as global increase in average 103
near-surface air temperature above pre-industrial conditions, are dependent on the selection 104
of a pre-industrial period that was not explicitly delineated by the UNFCCC. A number of 105
pre-industrial period definitions have been utilized in prior Intergovernmental Panel on 106
Climate Change (IPCC) assessments (IPCC, 2013, 2014; Clarke et al., 2014), some 107
pegging it to the early industrial period starting in 1760 or more often connecting it to the 108
1850-1900 period (which is also the beginning of historical period climate simulations; 109
Taylor et al., 2012). Historical station-based temperature datasets are available beginning 110
in 1850 (Hadley Center Climate Research Unit Temperature, HadCRUT4, Morice et al., 111
2012; Berkeley Earth Surface Temperature, BEST, Rohde et al., 2013), with additional 112
datasets available by the 1880-1900 period (Goddard Institute for Space Studies Surface 113
Temperature, GISTEMP Team, 2017; Hansen et al., 2010; the National Oceanic and 114
Atmospheric Administration, NOAA, Smith et al., 2008). The climate between 1760 and 115
1900 has relatively small long-term trends, characterized by internal variability and 116
punctuated by brief cooling events caused by aerosol loading from several major volcanic 117
eruptions (Hartmann et al., 2013). We define a pre-industrial period in the next sub-118
section, which allows us to determine the relative global changes between current 119
conditions and the +1.5 and +2.0 ºC Worlds. 120
121
7
From a climate perspective it is theoretically not necessary to specify the exact year for the 122
hypothetical +1.5 or +2.0 ºC Worlds as we define this period to be in a steady state. 123
However, the selection of a stabilization year is affected by long-term trends in land use, 124
ozone hole recovery, and aerosol concentrations, which alter the carbon and radiation 125
balances needed to stabilize climate conditions (Mitchell et al., 2017). The stabilization 126
target year also has large implications beyond climate conditions themselves, owing to 127
transient (time-evolving) pathways for socioeconomic development (O’Neill et al., 2014), 128
greenhouse gas concentration emissions (Moss et al., 2010), climate policy (Kriegler et al., 129
2014), and technological and policy developments affecting a sector like agriculture (Antle 130
et al., 2015; Ruane et al., 2018). Here we concentrate on stabilized climates rather than the 131
temporal progressions (e.g., peak and overshoot, linear, or asymptotic approaches) that 132
could achieve this state. 133
134
Representing the stabilized end climate states for the +1.5 and +2.0 ºC Worlds is a 135
challenge given that climate model simulations are typically driven by sets of greenhouse 136
gas emissions and socioeconomic development pathways that result in evolving trajectories 137
of climate changes extending from the present into the future (SRES, 2000; Moss et al., 138
2010; Taylor et al., 2012; Eyring et al., 2016). Given that the Coupled Model 139
Intercomparison Project (CMIP) is currently between its 5th and 6th phases (Taylor et al., 140
2012; Eyring et al., 2016), investigations of +1.5 and +2.0 ºC Worlds require either new 141
simulations performed outside of the CMIP protocols or the extraction of a representative 142
warming signal from existing CMIP5 simulations. 143
144
8
2.1. HAPPI stabilization scenarios of +1.5 and +2.0 ºC global warming 145
The primary source of climate information examined in this study are projections from the 146
Half a degree Additional warming, Projections, Prognosis and Impacts project (HAPPI; 147
Mitchell et al., 2017). HAPPI is designed to compare current climate (defined as the 2006-148
2015 period) with stabilization scenarios for the +1.5 and +2.0 ºC Worlds (nominally run 149
for the 2106-2115 period). HAPPI’s rapid mobilization of the climate modeling 150
community enables an ensemble of simulations focused on the +1.5 and +2.0 ºC Worlds, 151
providing updated models, specified scenarios, and enhanced focus on extreme events in 152
these projections of relatively low global temperature changes. Key attributes of the 153
HAPPI Tier 1 experiments are presented in Table 1, with further information in Table 1 of 154
the supplementary material and enhanced detail provided by Mitchell et al. (2017). 155
156
Table 2 shows the GCMs that provided HAPPI experiment output analyzed in this study 157
according to two analysis approaches. First, aggregate statistics were calculated to 158
represent the full GCM ensemble’s changes in monthly mean maximum and minimum 159
temperatures, total precipitation, the number of rainy days, and the standard deviation of 160
daily maximum and minimum temperatures for each grid cell. These were then re-gridded 161
to a common ½º x ½º grid. Climate changes based on the aggregate statistics were then 162
imposed on observational data using a quantile-mapping approach to produce bias-163
corrected future driving conditions (Ruane et al., 2015a,b), which we applied to point-164
based simulations and gridded analyses (Ruane et al., 2018; Liu et al., 2018). Second, daily 165
outputs were bias-corrected to the same grid using a quantile-mapping method to match 166
observational statistics in order to capture shifts in inter-annual and intra-seasonal extremes 167
9
while maintaining long-term trends (Hempel et al., 2013; Schleussner et al., 2018). Daily 168
scenarios have been applied to West African analyses (Webber et al., 2018) and a global 169
examination of agricultural extreme events (Schleussner et al., 2018). Both scenario 170
generation approaches capture mean climate changes and represent shifts in the distribution 171
of extreme events. Further dynamical and statistical downscaling approaches, which could 172
have provided additional detail on shifts in the characteristics of extreme events, were not 173
available for this rapid assessment. 174
175
2.2. Application of the 1980-2009 recent observed climatology 176
Many existing AgMIP model analyses and projection frameworks participating in the 177
AgMIP CGRA use the 1980-2009 period for crop model calibration (‘1995 climate’; as 178
provided by the Agricultural Modeling version of the Modern-Era Retrospective Analysis 179
for Research and Applications, AgMERRA; Ruane et al., 2015a). To enable the use of 180
these resources in agricultural assessments of the +1.5 and +2.0 ºC Worlds, we extend 181
HAPPI simulations to estimate differences between the 1995 climate and HAPPI 182
simulations of its current climate (2006-2015; ‘2010 climate’) and the +1.5 and +2.0 ºC 183
Worlds. HAPPI simulations directly provide the difference between the 2010 climate and 184
the +1.5 and +2.0 ºC Worlds. However, representing the remaining difference between the 185
1995 and 2010 climates is a challenge as we cannot observe a similar 30-year climatology 186
centered on 2010 and HAPPI simulations do not include 1995. 187
188
We employ a simple pattern-scaling approach to estimate how a climatological variable Q 189
(e.g., maximum daily temperatures, precipitation, the number of rainy days, the standard 190
10
deviation of minimum temperatures) changes between 2010 and 1995 at any given 191
longitude (i) and latitude (j) from the ½ º x ½ º global grid: 192
193
∆𝑄𝑖,𝑗,[2010−1995] = ∆𝑄𝑖,𝑗,[1.5−2010]
∆𝑇𝐺𝑙𝑜𝑏𝑎𝑙,[1.5−2010]× ∆𝑇𝐺𝑙𝑜𝑏𝑎𝑙,[2010−1995] (Eqn 1) 194
195
The ratio in Eqn 1 represents the rate of local change in Q in response to a change in global 196
temperature, a pattern that may then be multiplied by any similar global temperature 197
change. This is generated by calculating the local change in Q between the HAPPI 198
ensemble average +1.5 ºC World and current period simulation (2010 climate) and then 199
dividing by the additional global temperature changes expected in the +1.5 ºC World 200
simulation (0.58 ºC). This scaling factor is then multiplied by the observed warming 201
between 1995 and 2010 (ΔTGlobal,[2010-1995] = 0.25 ºC) drawn from the ensemble of historical 202
temperature products (BEST, GISS, HadCRUT4, and NOAA; note that these observations 203
indicate that the 2010 climate is 0.92 ºC warmer than pre-industrial conditions). This 204
results in an estimate of the local offset for the 1995-2010 period, with the process repeated 205
for each month. The same offset is used to calculate climate shifts associated with both 206
the +1.5 and +2.0 ºC Worlds so that their differences are determined solely by HAPPI 207
simulations. More complex pattern scaling approaches were not feasible given temporal 208
constraints related to stakeholder interest in the +1.5 ºC and +2.0 ºC Worlds (Tebaldi and 209
Arblaster, 2014). While this approach likely contributed to local biases (Lopez et al., 210
2014), patterns for HAPPI models were largely consistent when compared against patterns 211
derived from the +2.0 ºC World for all but the MIROC5 GCM, which showed polar 212
11
regions accounting for an increasing proportion of global temperature rise at the expense 213
of other land areas (not shown). 214
215
2.3. CMIP5 transient scenarios of +1.5 and +2.0 ºC global warming 216
Alternative representations of the +1.5 and +2.0 ºC Worlds are derived from CMIP5 217
transient simulations whereby climate evolves from anthropogenic and natural influences 218
over the historical period (1850-2005) and future decades (RCPs covering 2006-2100; 219
Moss et al., 2010). As our interest is in understanding the challenges today’s agricultural 220
sector will face from future climate change, we utilize the recent observable 1980-2009 221
climatology (1995 climate) used in many AgMIP analyses for each GCM (drawing from 222
RCP4.5 for 2006-2009). Note that this period had already warmed 0.67 ºC above the 1861-223
1880 pre-industrial period according to the observational products ensemble. For each 224
GCM, +1.5 and +2.0 ºC Worlds are thus defined by their transient thresholds, or the period 225
in which global mean surface temperatures in the transient simulation reach 0.83 ºC and 226
1.33 ºC, respectively, above this 1995 climate. 227
228
We calculate corresponding threshold years for each GCM and RCP, then construct a 229
centered 30-year average climatology (to elevate the mean climate changes above internal 230
variability; WMO, 1989) for comparison to the current period. Threshold years differ by 231
GCM (depending on transient climate sensitivity) and among RCPs, with earlier thresholds 232
reflecting higher climate sensitivity in some GCMs or the accelerated accumulation of 233
greenhouse gases in RCP8.5 compared to RCP2.6. As many of the RCP2.6 simulations do 234
not reach the +2.0 ºC threshold, we focus on the ensemble response of 31 RCP4.5 GCMs 235
12
in Table 2 given that GCM differences in regional patterns of temperature and precipitation 236
change tend to be robust across RCP and time horizon in the 21st century (Ruane and 237
McDermid, 2017). 238
239
Differences in climate change projections between the transient threshold and the same 240
GCM at stabilization merit further study, with differences relating to aerosol and land-use 241
patterns within the transient RCPs as well as a lag between radiative imbalances and 242
equilibration within the climate system. For this study we assume that this effect is small 243
compared to differences between the GCMs captured in the CMIP5 ensemble. Model 244
versions (among models that contributed to both CMIP5 and HAPPI) likely also contribute 245
to differences between transient and stabilization scenarios, as HAPPI simulations (which 246
were conducted largely in 2016 and 2017) used updated versions compared to CMIP5 247
simulations completed before 2013. 248
249
Rising CO2 concentrations are the largest radiative forcing factor contributing to 250
anthropogenic climate change (IPCC, 2013). The direct impact of CO2 on photosynthesis 251
merits additional focus for agricultural impacts. CO2 concentrations in the CMIP5 252
transients were above HAPPI’s recommended 423ppm at the +1.5 ºC crossing point, but 253
below HAPPI’s 487ppm at the +2.0 ºC crossing point (Figure S1). These results suggest 254
that the change in CO2 concentration and its impacts on agricultural systems are likely 255
overestimated in HAPPI-driven +2.0 ºC analyses (Boote et al., 2010; O’Leary et al., 2015; 256
Kimball, 2016; Deryng et al., 2016). Radiative disequilibrium at transient crossing points 257
also likely overestimates stabilization CO2 concentrations. For example, van Vuuren et al. 258
13
(2015) indicated that CO2 concentrations must be kept below 450ppm to have a 67% 259
probability of keeping global temperature rise below +2.0 ºC. 260
261
2.4. Regional growing seasons 262
HAPPI and CMIP5 transient projections were distributed onto a ½º x ½º grid using a 263
nearest-neighbor approach in order to facilitate ensemble statistics and agro-climatic 264
analyses. Growing seasons (Julian day ranges) for each grid cell were drawn from the 265
AgMIP Global Gridded Crop Model Intercomparison and were assumed to remain constant 266
in current and future worlds (Elliott et al., 2015). Each grid cell’s growing season aims to 267
represent the practice of the majority of farming systems even though many regions include 268
a wide diversity in sowing dates. Static (~2005) cropped areas from the Spatial Production 269
Allocation Model database (SPAM) provide the basis for weighted spatial averages of 270
agro-climatic changes associated with the +1.5 and +2.0 ºC Worlds (You et al., 2014). They 271
also form the basis of agricultural area masks to focus figures on grid cells with >10ha 272
rainfed crop area. 273
274
Here we highlight rainfed seasons as they are vulnerable to both temperature and 275
precipitation change impacts. Irrigated systems will face similar temperature challenges 276
but are largely unaffected by precipitation changes other than through supply of managed 277
water resources. Wheat growing seasons include both spring and winter wheat systems – 278
in regions where both are grown we examine only the system with higher mean production 279
in the AgMIP Global Gridded Crop Model Intercomparison (GGCMI) historical period 280
simulations (Müller et al., 2017). For winter wheat we examine climate changes during 281
14
the final 90 days before average maturity to avoid the dormant vernalization period that 282
follows early planting. Wheat tends to be grown earlier in the season than maize, rice, and 283
soy, which are most often grown in the local spring and summer (Figure S2). 284
285
Comparison of climate shifts during the local growing season (ΔTGS and ΔPGS) against the 286
annual climate changes (ΔTAnn and ΔPAnn) helps us identify regions where agricultural 287
challenges are projected to be different than the annual signal suggests. We first look for 288
large magnitudes of difference across multiple models, classifying local growing season 289
changes as ‘substantially warmer than annual changes’ or ‘substantially cooler than annual 290
changes’ if ΔTGS - ΔTAnn > 0.2 ºC or ΔTGS - ΔTAnn < -0.2 ºC (respectively) for 3 of the 5 291
HAPPI models. We next examine areas that do not meet those criteria for consistency in 292
the direction of the difference, classifying local growing season changes as ‘consistently 293
warmer than annual changes’ if ΔTGS - ΔTAnn > 0 ºC for all HAPPI models. We do the 294
same in the opposite direction for ‘cooler’ classifications, and then all remaining grid cells 295
are designated as ‘small or inconsistent changes’. A similar procedure classified ‘wetter’ 296
and ‘drier’ local growing seasons using a threshold of +/-5% difference (compared to 297
current period climate) for substantially different ΔPGS. Note that areas meeting both the 298
‘substantially’ and ‘consistently’ different criteria are classified as ‘substantially’ different. 299
300
3. Results of climate change projections for agricultural regions 301
The +1.5 and +2.0 ºC Worlds are defined by their global mean temperature increases, yet 302
projections reveal important variations across regions, seasons, and the characteristics of 303
climate extremes. Uncertainty is heightened for phenomena simulated at finer temporal 304
15
and spatial scales where there is greater influence from physical processes beyond the limit 305
of climate model resolution and a lack of aggregation. Assessments of agricultural risks 306
are reliant on this fine-scale information given the potential for non-linear climatic 307
responses and the application of crop models on a hectare scale in AgMIP (Asseng et al., 308
2013, 2015; Bassu et al., 2014; Ruane et al., 2014; Li et al., 2015; Pirttioja et al., 2015; 309
Fleisher et al., 2017). Simulations for the CGRA assessment of +1.5 and +2.0 ºC Worlds 310
likewise need information with more detail than the global annual mean temperature 311
analysis, particularly given that the ½º x ½º resolution grids utilized for global crop model 312
simulations are finer than the horizontal resolution of most global climate models (Ruane 313
et al., 2018; Rosenzweig et al., 2018). 314
315
3.1. Mean changes for +1.5 and +2.0 ºC Worlds 316
Mean annual climate changes in HAPPI projections of the +1.5 and +2.0 ºC Worlds 317
compared to the 2010 current climate (Figures 1a-b and 2a-b, respectively) generally 318
reflect the major patterns described in recent IPCC assessments, with the signal of change 319
becoming more clear as global mean temperature increases (Collins et al., 2013). These 320
simulations highlight uncertainties in HAPPI models related to the distribution of 321
additional energy trapped in the system by greenhouse gases and other anthropogenic 322
influences, differentiating climate challenges for agricultural regions. Rosenzweig et al. 323
(2018) provide a comprehensive example of HAPPI uncertainty in displaying each GCM 324
projection for rainfed maize climate changes. 325
326
16
Warming over land generally exceeds the global mean temperature rise since 2010 (~0.58 327
ºC and ~1.08 ºC for the +1.5 and +2.0 ºC Worlds, respectively; recall that 2010 was 0.92 328
ºC above pre-industrial) owing to the ocean’s larger aggregate heat capacity. The poles 329
also warm faster than the tropics given snow-ice albedo feedbacks and the prominence of 330
high-latitude longwave radiation emissions in the global energy balance. Arid region 331
temperatures rise more rapidly as a lack of available soil moisture limits evaporation and 332
causes a larger portion of excess energy to go into sensible heat. Farmlands are not evenly 333
distributed around the globe, however, as agricultural regions are generally rare in arid 334
regions and at higher latitudes where sub-freezing temperatures shorten the viable growing 335
season. Agricultural area-weighted temperature changes are therefore closer to the global 336
average but depend on a given GCM’s distribution of the warming signal owing to ocean-337
atmosphere heat exchanges, radiative feedback processes, and atmospheric circulation 338
patterns. Projections of the +1.5 ºC World show stronger agreement in temperature 339
changes at low and high latitudes as overall temperature changes are comparable to the 340
range between GCM projections; however, many of the mid-latitude agricultural regions 341
have substantial GCM spread. In the +2.0 ºC World, projected temperature changes are 342
at least double the GCM range for most areas and seasons. Uncertainty is larger over 343
prominent wheat-growing regions of Eastern Europe and central North America in addition 344
to regions where agriculture is less prominent, such as over the Tibetan Plateau, the inner 345
Amazon, and north of the Caspian Sea. 346
347
In many regions and agricultural seasons there is not strong agreement in the direction of 348
precipitation changes (Figure 1b). Mean annual precipitation generally increases over land 349
17
given an overall increase in evaporation and transpiration as warmer conditions lead to 350
higher vapor pressure deficits even when relative humidities are maintained. Regional 351
differences reflect a more vigorous water cycle that exacerbates given atmospheric 352
moisture convergence patterns wherein convergence (wet) areas generally become wetter 353
and divergence (dry) areas are more prone to drying (Trenberth et al., 2011; Collins et al., 354
2013). Notable exceptions include drying over the Amazon (due in part to land cover 355
changes) and an increase in precipitation over Eastern Africa and Western Asia (likely 356
connected to large-scale circulation patterns associated with the South Asian monsoon; 357
Christensen et al., 2013). The South Asian monsoon benefits from more water vapor 358
transport as higher atmospheric specific humidities overwhelm slight reductions in the 359
strength of the monsoon circulation (Christensen et al., 2013). A reduction in aerosols 360
associated with RCP2.6 end-of-century forcings utilized in HAPPI simulations also 361
reduces the inhibition of precipitation in South and Southeast Asia (Carl-Friedrich 362
Schleussner, personal communication). 363
364
Changing precipitation patterns and shifts in the hydrologic cycle will drive shifts in soil 365
moisture and evapotranspiration demand, which HAPPI and CMIP earth system models 366
simulate using dynamic vegetation models of varying complexity (most have a very simple 367
representation of agricultural systems; McDermid et al., 2017). Projected changes in soil 368
moisture and evapotranspiration are beyond the present study’s focus on climate scenarios 369
for agricultural modeling applications given that AgMIP’s agricultural models simulate 370
these quantities internally. 371
372
18
3.2. Unique characteristics of changes for growing seasons 373
Agricultural systems are most vulnerable to climate changes affecting major production 374
regions and core growing seasons. HAPPI projections for the +1.5 and +2.0 ºC Worlds 375
feature substantial seasonal variation reflecting the annual cycle of atmospheric circulation 376
patterns, local radiation balances, and shifts in land surface characteristics. 377
378
Figures 1c-j and 2c-j show the growing season temperature and precipitation change 379
patterns over regions with substantial maize, wheat, rice, and soy production. Many of the 380
highest temperature and precipitation change regions are not agriculturally prominent (e.g., 381
high latitudes, high elevations, and arid regions), and thus the overall comparison reveals 382
more muted climate changes for agricultural regions and seasons than for the annual mean 383
global picture. Regional temperature patterns from the annual change map are generally 384
maintained from season to season, indicating that these are larger than monthly differences 385
in temperature change at any given location. Growing season and annual precipitation 386
changes are similar because rainfed cultivation tends to occur in the local rainy season, 387
which has a disproportionate influence on total annual rainfall totals. Process-based 388
models and empirical studies show that global cereal production responds negatively to 389
increasing temperatures (Rosenzweig et al., 2013; Asseng et al., 2013,4; Bassu et al., 2014; 390
Challinor et al., 2014; Li et al., 2015; Zhao et al., 2017), and the food system implications 391
of these projections are explored further within the AgMIP CGRA (Ruane et al., 2018; 392
Rosenzweig et al., 2018). 393
394
19
Zonal patterns are readily apparent in growing season climate change characteristics using 395
the classifications of ‘substantially’ and ‘consistently’ different than annual climate 396
changes presented in Figure 3 for each growing season in the +1.5 ºC World (Figure S3 397
displays the corresponding +2.0 ºC World maps). Growing seasons are generally projected 398
to warm at a slower rate than the annual mean in the humid tropics (e.g., in Brazil, Central 399
and Western Africa, Southern India, and Southeast Asia). In these regions cultivation is 400
aligned with a strongly seasonal rain band or wet monsoon phase that reduces local 401
warming compared to the dry seasons. Growing seasons in the semi-arid tropics (e.g., 402
Mexico, East Africa, and northern India) tend to warm consistently faster than the annual 403
signal, and in Mediterranean climates (including Eastern Australia) ΔTGS is accelerated in 404
association with drying trends. ΔTGS for high-latitude agricultural regions is lower than 405
the annual average given the pronounced snow-ice feedbacks that dominate the non-406
agricultural winter season. The region around Uruguay is an exception to this zonal 407
pattern, with heightened ΔTGS potentially linked to a precipitation pattern that also 408
enhances seasonal rainfall. 409
410
ΔPGS is substantially and consistently wetter than ΔPAnn over the South Asian and Southeast 411
Asian monsoon, with exceptions where the local growing season does not correspond with 412
the main monsoon rains (e.g., Indonesian maize, Indian wheat; see Figure S3). 413
Mediterranean drying affects agricultural seasons more strongly than the annual average 414
as it is most acute during the agricultural spring and summer seasons (see also Collins et 415
al., 2013). A dipole over North America indicates enhanced growing season drying in the 416
continental interior and wetter conditions in the Southern and Eastern United States, 417
20
exacerbating the annual trend. ΔPGS is drier than ΔPAnn for most Southern Hemisphere 418
wheat growing seasons, suggesting the potential for shifts in planting and harvest dates as 419
seasonal patterns change amidst a generally drying trend in Australia, Southern Africa, and 420
portions of South America. 421
422
Regional patterns in temperature and precipitation changes are largely consistent between 423
the +1.5 and +2.0 ºC Worlds. The magnitude of these regional changes is broadly 424
consistent with an approximate doubling in deviation from today’s conditions between the 425
1.5 and 2.0 ºC Worlds (tracking the global mean temperature rises of 0.58 and 1.08 ºC, 426
respectively). 427
428
3.3. Changes in agro-climatological extremes 429
The stability of agricultural systems under a +1.5 or +2.0 ºC World will depend not only 430
on mean climate conditions but also on interannual variations that exacerbate the risk of 431
poor harvests that stakeholders must manage. Figure 4 demonstrates the ways in which 432
the distributions of extreme heat and drought years shift in comparison to current climate, 433
utilizing the subset of HAPPI daily bias-corrected and downscaled data (each with 200 434
seasons from 20 ensemble-members’ 10-year simulations). In most tropical and extra-435
tropical regions, rainfed maize season temperatures in the +1.5 ºC World exceed the current 436
period ‘extreme heat’ year (defined as the 90th percentile growing season mean 437
temperature) in more than half of all years (Fig. 4a). This is true even in Southeast Asia 438
where mean temperatures rise more slowly, as the extreme temperature threshold is low 439
owing to limited interannual variability in the current climate. A dramatic increase in 440
21
extreme years is also projected in mid-latitude regions including portions of Northeast 441
Asia, Southern Europe, and the Eastern United States. Thus, even relatively low levels of 442
warming can lead to substantial anomalies, and these are particularly noticeable where 443
current interannual variation is low even as mean climate changes are lower than those at 444
higher latitudes (Figure 1c). 445
446
Diffenbaugh et al. (2018) likewise noted dramatic increases in extreme conditions under 447
+1.5 and +2.0 ºC warming scenarios. Changes in mean conditions alone can increase the 448
probability of crop failures as growing season climate moves closer to dangerous 449
biophysical thresholds (Hatfield et al., 2011; Lobell et al., 2011, 2013; Zhao et al., 2017). 450
The frequency of ‘extreme drought’ years in the +1.5 ºC World (defined as the 10th 451
percentile growing season total precipitation) generally become less common (occurring 452
less than 10% of the time) in areas where rainfall is increasing Fig. 1d), but the frequency 453
of drought conditions increases in many places where median rainfall changes were small 454
or negative (e.g., Indonesia, Peru, Mexico, the Northern US Plains, Europe) (Fig. 4b). 455
456
The additional half-degree warming between the +1.5 and +2.0 ºC Worlds is projected to 457
have substantial impacts on extreme conditions. Many mid-latitude regions see 20-40% 458
more years becoming extreme heat years in the +2.0 ºC World (Fig 4c), with smaller 459
increases in the tropics largely because many of those regions already had a high percentage 460
of extreme heat years in the +1.5 ºC World (Fig 4a). Changes between the frequency of 461
extreme drought between the +1.5 and +2.0 ºC Worlds (Fig. 4d) show the general 462
exacerbation of changes in the +1.5 ºC World patterns (Fig 4b). However, many areas show 463
22
a reduction or even slight reversal of these drought frequency trends as the extreme drought 464
years had already been reduced by mean precipitation increases (e.g., in the monsoon 465
regions of South and Eastern Asia). 466
467
It is possible that this increase in climate extremes comes simply from changes in mean 468
conditions, so we additionally examine changes in the temperature and precipitation 469
distributions by comparing changes in the extreme percentiles against median changes for 470
each bias-corrected GCM (Figs. 4d,e). If the 90th percentile of the temperature distribution 471
is warming more rapidly than the median this indicates a skew toward more extreme heat 472
years. This is apparent for many tropical and Asian monsoon regions. The opposite 473
skewness toward less extreme heat years is projected for more northerly latitudes. The 474
Eastern United States inverts this meridional pattern, with more extreme heat skewness in 475
the Great Lakes region and less in the Southern US. Skewness toward more or less extreme 476
drought years (whereby the 10th percentile precipitation dries in comparison to the median 477
change) shows much larger spatial variation and no clear large-scale patterns. Schleussner 478
et al. (2018) examine the effects that changing distributions of climate extremes have on 479
yield reliability within the CGRA crop model simulations. 480
481
482
4. Agro-climatic changes for major breadbaskets 483
Climate change impacts on top cereal-producing countries merit closer investigation as 484
they have a disproportionate influence on global food markets because they drive livestock 485
prices in addition to cereal prices, given the importance of grain-based animal feed. 486
23
Climate shocks in these countries may also threaten global food security through 487
disruptions in food exports. As agricultural production is concentrated in breadbaskets 488
within these countries, Figures 5 and 6 show the mean growing season temperature and 489
precipitation changes (weighted by area growing the rainfed crop within each country) 490
from each HAPPI GCM for the top 10 maize, wheat, rice, and soy production countries for 491
the +1.5 and +2.0 ºC Worlds, respectively (You et al., 2014). 492
493
Overall temperatures in these high-producing countries show mean warming that is larger 494
than the inter-GCM spread, providing higher confidence in projections even as warming is 495
lower in the tropics than at higher latitudes. Precipitation changes are mixed, with large 496
variation in sign and magnitude across GCMs and breadbaskets. Wheat growing seasons 497
demonstrate the largest precipitation change variations, while rice seasons are consistent 498
in projecting wetter conditions given the prominence of the Asian monsoons in global rice 499
production (results in Indonesia, Brazil, and Nigeria are less consistent). 500
501
Projected changes between the current period and the +1.5 ºC World are generally extended 502
in the +2.0 ºC World, with similar variation among breadbaskets and GCMs. Although all 503
GCMs have nearly identical global mean surface temperature increases for a given climate 504
stabilization, the patterns of warming (between oceans, land, specific land areas, and 505
different seasons) depend on internal model parameters and structures. MIROC5 stands 506
out as exceptionally warm for many breadbasket regions in the +1.5 ºC World, but 507
HadAM3P is most often the warmest over main agricultural countries for the +2.0 ºC 508
World. Warming between the +1.5 and +2.0 ºC Worlds exceeds the global average 0.5 ºC 509
24
in several breadbaskets (as discussed in Section 3.1), notably including US and Romanian 510
maize, US and Canadian wheat, and US soy. Many leading rice-producing countries are 511
projected to lag the global warming pace given their tropical location and prevailing wetter 512
conditions. 513
514
515
5. Comparing stabilization and transient projections for agricultural model 516
applications 517
Figure 7 compares aggregate temperature and precipitation changes over land for HAPPI 518
and CMIP5. Despite being a 5-GCM subset representing a much larger GCM ensemble, 519
the HAPPI simulations display an overall spread that is comparable to that of the 31-520
member CMIP5 ensemble. 521
522
5.1 Annual changes 523
Annual changes over land in HAPPI and CMIP5 are larger than the global signal, with the 524
land-sea contrast increasing as median land temperatures rise by more than 0.5 ºC between 525
the +1.5 and +2.0 ºC Worlds (0.59 ºC in HAPPI, 0.64 ºC in CMIP5). An earlier arrival of 526
+1.5 and +2.0 ºC conditions and a higher stabilization ceiling was also noted in projections 527
over the US (Karmalker and Bradley, 2017). MIROC5 produces the largest HAPPI land 528
temperature increases for the +1.5 ºC World, but HadAM3P is more extreme for the +2.0 529
ºC World. Precipitation over land increases between the +1.5 ºC and +2.0 ºC Worlds on 530
average, with increases (0.8% in HAPPI and 0.98% in CMIP5) slightly above the 1% K-1 531
25
noted for global precipitation changes in the CMIP5 ensemble due to an overall 532
amplification of the water cycle and moisture transport (Pendergrass and Hartmann, 2014). 533
534
5.2 Growing season changes 535
ΔTGS for each season (weighted by harvested rainfed area across all land) is more uncertain 536
than the all-land and all-season ΔTAnn across both the stabilization and transient scenarios, 537
as would be expected for more specific geographic regions and seasons. ΔTGS is also lower 538
than ΔTAnn as agricultural lands are disproportionately located away from arid and polar 539
regions that warm most rapidly, in addition to cultivation occurring in the wetter and frost-540
free months of the year. Much of the accelerated ΔTAnn over land comes from high-latitude 541
regions, so agricultural lands fall increasingly behind the pace of land ΔTAnn in the +2.0 ºC 542
World. The largest ΔTGS and ΔPGS uncertainties are in the wheat season, potentially owing 543
to the location of wheat systems predominantly in mid-latitude regions affected by variable 544
storm tracks (recall Figure 1e). ΔPGS in both HAPPI and CMIP5 ensembles lag behind the 545
Δ PAnn for wheat (owing largely to drier conditions in Europe and central North America), 546
while the rice season has the most dramatic lag in warming and an accelerated increase in 547
precipitation (likely due to the importance of monsoon precipitation changes in the global 548
land signal). 549
550
The projected rice-growing season features the largest differences between the HAPPI 551
stabilization and the CMIP5 transient ensembles. Global rice production is dominated by 552
intense rice cultivation in the monsoon regions of Asia that currently experience some of 553
the highest aerosol concentrations in the world. The HAPPI simulations include a dramatic 554
26
reduction in aerosols according to end-of-century RCP2.6 conditions, while these levels 555
are not achieved in CMIP5 RCP4.5 transients (nor would the same levels be seen in RCP2.6 556
transients that first reach +1.5 or +2.0 ºC warming well before the end of the century). 557
With reduced aerosols the HAPPI simulations warm rice production regions more rapidly, 558
raising global aggregated median rice-season temperatures by 0.19 ºC compared to CMIP5 559
for both stabilization worlds. The signal of aerosols is less clear in rice season 560
precipitation, as there remains substantial uncertainty in GCM simulation of aerosol effects 561
on evaporation, heat pumping, and precipitation inhibition (Ramanathan et al., 2005; Lau 562
and Kim, 2006; Rind et al., 2009; Boucher et al., 2013). 563
564
565
6. Discussion 566
Mean growing season climate changes and shifts in the distribution of interannual events 567
characterize agricultural system challenges for the +1.5 and +2.0 °C Worlds, although 568
climate changes associated with the +1.5 and +2.0 ºC Worlds are small in comparison to 569
end-of-century shifts in transient RCP8.5 simulations (~4 to 6 ºC above pre-industrial 570
conditions from ensemble of 39 GCMs in Collins et al., 2013). However, projected 571
conditions in major agricultural regions are likely to require substantial adaptation even 572
under these highly-mitigated climate stabilizations. While cereal-producing regions 573
generally experience a slower pace of warming in growing seasons than other land areas, 574
projected precipitation changes demonstrate high uncertainty and spatial variability that 575
will lead to regional wetting and drying trends even as global precipitation increases 576
slightly. Precipitation patterns over Norther America indicate that the dipole of a drying 577
27
continental interior and wetter East and Southeast will be more pronounced in the 578
summertime growing seasons than in the annual average. 579
580
Even with these relatively low levels of mean warming farmers are projected to experience 581
a dramatic increase in the probability of what are currently considered extremely hot 582
growing seasons, and many regions experiencing drier average conditions would also 583
experience an increase in extreme drought seasons. These challenges are apparent even in 584
the +1.5 ºC World, and are exacerbated in the +2.0 ºC World, particularly as regional 585
warming patterns cause many agricultural regions to warm by more than 0.5 ºC between 586
Worlds. HAPPI simulations provide useful insight into these targeted end states, but 587
uncertainties remain owing partially to difficulties in representing desired climate 588
stabilizations rather than simulating transient scenarios. Major production regions for 589
maize, wheat, and soy face larger average temperature increases than rice production 590
regions, which also are more likely to see increases in precipitation although both rice 591
effects may be influenced by HAPPI assumptions about reduced aerosol loading and 592
aerosol effects on Asian monsoon regions. For agricultural applications CO2 levels in the 593
+1.5 and +2.0 ºC Worlds (and their impact on agricultural systems) remains a major 594
uncertainty that can reverse the overall direction of production changes (Ruane et al., 2018; 595
Rosenzweig et al., 2018). 596
597
Agricultural sector impacts for the +1.5 and +2.0 ºC Worlds may be characterized by 598
linking the climate conditions described here with biophysical models (to determine 599
climate impacts on plant processes and production), and economic models (to understand 600
28
profit, trade, and resource allocation). Interdependence of global and regional 601
stakeholders can be explored through markets, policies, global climate, and resource 602
competition. AgMIP Coordinated Global and Regional Assessments utilize the climate 603
scenarios described in Section 2 to represent +1.5 and +2.0 ºC World climates for a 604
prototype integrated, multi-disciplinary, multi-scale, and multi-model assessment that 605
explores direct biophysical impacts and their implications for regional and global markets 606
(Ruane et al., 2018; Rosenzweig et al., 2018). 607
608
Several important climate-related mechanisms are not directly evaluated here and are 609
needed to further inform the development of effective adaptation strategies for agriculture. 610
Priority areas for scenario generation and future evaluations include: 611
Damages from floods and water-logging in farming regions where precipitation is 612
projected to increase (Ruane et al., 2013); 613
CO2 fertilization effects on crops and their interactions with drought (Boote et al., 614
2010; O’Leary et al., 2015; Kimball, 2016; Deryng et al., 2016); 615
Effects of sea-level rise and shifting hurricane intensities on agricultural production 616
in coastal floodplains (Church et al., 2013); 617
Acute damages from short-duration heat waves affecting critical crop development 618
stages (Asseng et al., 2015; Maiorano et al., 2017); 619
Changes in sowing and harvest dates in response to changing climatic conditions 620
(Tadross et al., 2009; Crespo et al., 2011; Glotter and Elliott, 2016); 621
Shifts in major modes of climate variability (e.g., the El Niño/Southern Oscillation 622
or North Atlantic Oscillation; Iizumi et al., 2014); 623
29
Competition over managed water resources (Elliott et al., 2014); 624
Shifts in cropping areas associated with climate and socioeconomic development 625
(Wiebe et al., 2015); and 626
Crop losses from damaging pests, diseases, and weeds (Savary et al., 2012; 627
Donatelli et al., 2017). 628
A primary limitation remains the collection and dissemination of high-quality experimental 629
and observational data along the climate-crop-economic-human outcome pathway. 630
Through AgMIP, the agricultural research community is developing improved models and 631
systematic protocols to address many of these gaps. 632
633
The combination of HAPPI and CMIP5 projections analyzed here indicate climate 634
challenges for agricultural regions even under high mitigation scenarios. Results and 635
agricultural model impact assessments, particularly those that compare against lower 636
mitigation scenarios with more dramatic climate changes, can inform cost-benefit analysis 637
that weigh the relative burden placed on adaptation, mitigation, and disruption to the 638
agricultural system. By projecting the challenges facing future farmers, we also aim to 639
provide climate condition targets for agricultural technologies that will increase resilience 640
for vulnerable farm systems around the world. 641
642
Acknowledgements 643
The authors thank the HAPPI team (Dann Mitchell, Myles Allen, Carl-Friedrich 644
Schleussner, Peter Uhe, Mamunur Rashid) for calculating aggregate statistics and 645
providing daily bias-corrected and downscaled outputs. We acknowledge the World 646
30
Climate Research Programme's Working Group on Coupled Modelling, which is 647
responsible for CMIP, and we thank the climate modeling groups (noted in Table 2) for 648
producing and making available their model output. For CMIP the U. S. Department of 649
Energy's Program for Climate Model Diagnosis and Intercomparison provides 650
coordinating support and led development of software infrastructure in partnership with 651
the Global Organization for Earth System Science Portals. This work was supported by 652
the National Aeronautics and Space Agency Science Mission Directorate (WBS 653
281945.02.03.06.79) and the US Department of Agriculture (USDA OCE grant 58-0111-654
16-010). We also thank Greg Reppucci for assistance in preparing figure graphics. Results 655
reflect the findings of the authors and do not necessarily represent the views of the 656
sponsoring agencies. 657
31
References 658
Antle, J.M., Valdivia, R.O., Boote, K.J., Janssen, S., Jones, J.W., Porter, C.H., Rosenzweig, 659
C., Ruane, A.C., Thorburn, P.J., 2015. AgMIP’s Trans-disciplinary Agricultural 660
Systems Approach to Regional Integrated Assessment of Climate Impact, 661
Vulnerability and Adaptation. In: The Handbook of Climate Change and 662
Agroecosystems: The Agricultural Model Intercomparison and Improvement 663
Project Integrated Crop and Economic Assessments, Part 1 [Rosenzweig, C., Hillel, 664
D. (eds.)]. Imperial College Press, London. 27-44. 665
doi:10.1142/9781783265640_0002. 666
Antle, J.M., Cho, S., Tabatatie, H., Valdivia, R., 2018. Economic and Environmental 667
Performance of the U.S. Pacific Northwest Wheat System in a Low Greenhouse 668
Gas Emissions World. Journal of Environmental Management, submitted. 669
Asseng, S., Ewert, F., Rosenzweig, C., Jones, J.W., Hatfield, J.L., Ruane, A.C., Boote, 670
K.J., Thorburn, P., Rötter, R.P., Cammarano, D., Brisson, N., Basso, B., Martre, P., 671
Aggarwal, P.K., Angulo, C., Bertuzzi, P., Biernath, C., Challinor, A.J., Doltra, J., 672
Gayler, S., Goldberg, R., Grant, R., Heng, L., Hooker, J., Hunt, L.A., Ingwersen, 673
J., Izaurralde, R.C., Kersebaum, K.C., Müller, C., Naresh Kumar, S., Nendel, C., 674
O’Leary, G., Olesen, J.E., Osborne, T.M., Palosuo, T., Priesack, E., Ripoche, D., 675
Semenov, M.A., Shcherbak, I., Steduto, P., Stöckle, C., Stratonovitch, P., Streck, 676
T., Supit, I., Tao, F., Travasso, M., Waha, K., Wallach, D., White, J.W., Williams, 677
J.R., Wolf, J., 2013. Uncertainty in simulating wheat yields under climate change. 678
Nature Clim. Change 3, 827-832. doi: 10.1038/NCLIMATE1916. 679
Asseng, S., and co-authors, 2015: Rising temperatures reduce global wheat production. 680
Nature Clim. Change, 5, no. 2, 143-147, doi:10.1038/nclimate2470. 681
Bassu, S., Brisson, N., Durand, J.-L., Boote, K., Lizaso, J., Jones, J.W., Rosenzweig, C., 682
Ruane, A.C., Adam, M., Baron, C., Basso, B., Biernath, C., Boogaard, H., Conijn, 683
S., Corbeels, M., Deryng, D., De Sanctis, G., Gayler, S., Grassini, P., Hatfield, J., 684
Hoek, S., Izaurralde, C., Jongschaap, R., Kemanian, A.R., Kersebaum, K.C., Kim, 685
S.H., Kumar, N.S., Makowski, D., Müller, C., Nendel, C., Priesack, E., Pravia, 686
M.V., Sau, F., Shcherbak, I., Tao, F., Teixeira, E., Timlin, D., Waha, K., 2014. How 687
do various maize crop models vary in their responses to climate change factors? 688
Glob. Change Biol., 20, no. 7, 2301-2320, doi:10.1111/gcb.12520. 689
Beddington, J., Asaduzzaman, M., Clark, M., Fernández, A., Guillou, M., Jahn, M., Erda, 690
L., Mamo, T., Van Bo, N., Nobre, C.A., Scholes, R., Sharma, R., Wakhungu, J., 691
2012. Achieving food security in the face of climate change: Final report from the 692
Commission on Sustainable Agriculture and Climate Change. CGIAR Research 693
Program on Climate Change, Agriculture and Food Security (CCAFS). 694
Copenhagen, Denmark. Available online at: www.ccafs.cgiar.org/commission. 695
Boote, K.J., Allen Jr., L.H., Prasad, P.V.V., Jones, J.W., 2010. Testing Effects of Climate 696
Change in Crop Models. In: The Handbook of Climate Change and 697
Agroecosystems [Rosenzweig, C., Hillel, D. (eds.)]. Imperial College Press, 698
Singapore. 109-129. 699
Boucher, O., Randall, D., Artaxo, P., Bretherton, C., Feingold, G., Forster, P., Kerminen, 700
V.-M., Kondo, Y., Liao, H., Lohmann, U., Rasch, P., Satheesh, S.K., Sherwood, S., 701
Stevens, B., Zhang, X.Y., 2013. Clouds and Aerosols. In: Climate Change 2013: 702
32
The Physical Science Basis. Contribution of Working Group I to the Fifth 703
Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, 704
T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., 705
Xia, Y., Bex, V., Midgley, P.M. (eds.)]. Cambridge University Press, Cambridge, 706
United Kingdom and New York, NY, USA. 707
Brown, M.E., Antle, J.M., Backlund, P., Carr, E.R., Easterling, W.E., Walsh, M.K., 708
Ammann, C., Attavanich, W., Barrett, C.B., Bellemare, M.F., Dancheck, V., Funk, 709
C., Grace, K., Ingram, J.S.I., Jiang, H., Maletta, H., Mata, T., Murray, A., Ngugi, 710
M., Ojima, D., O’Neill, B., Tebaldi, C., 2015. Climate Change, Global Food 711
Security, and the U.S. Food System. 146 pages. Available online at 712
http://www.usda.gov/oce/climate_change/FoodSecurity2015Assessment/FullAsse713
ssment.pdf. doi: 10.7930/J0862DC7. 714
Clarke, L., Jiang, K., Akimoto, K., Babiker, M., Blanford, G., Fisher-Vanden, K., 715
Hourcade, J.-C., Krey, V., Kriegler, E., Löschel, A., McCollum, D., Paltsev, S., 716
Rose, S., Shukla, P.R., Tavoni, M., van der Zwaan, B.C.C., van Vuuren, D.P., 2014. 717
Assessing Transformation Pathways. In: Climate Change 2014: Mitigation of 718
Climate Change. Contribution of Working Group III to the Fifth Assessment Report 719
of the Intergovernmental Panel on Climate Change [Edenhofer, O., Pichs-Madruga, 720
R., Sokona, Y., Farahani, E., Kadner, S., Seyboth, K., Adler, A., Baum, I., Brunner, 721
S., Eickemeier, P., Kriemann, B., Savolainen, J., Schlömer, S., von Stechow, C., 722
Zwickel, T., Minx, J.C. (eds.)]. Cambridge University Press, Cambridge, United 723
Kingdom and New York, NY, USA. 724
Challinor, A.J., and co-authors, 2014: A meta-analysis of crop yield under climate change 725
and adaptation. Nature Climate Change 4(4), 287-291. 726
doi:10.1038/NCLIMATE2153 727
Christensen, J.H., Krishna Kumar, K., Aldrian, E., An, S.-I., Cavalcanti, I.F.A., de Castro, 728
M., Dong, W., Goswami, P., Hall, A., Kanyanga, J.K., Kitoh, A., Kossin, J., Lau, 729
N.-C., Renwick, J., Stephenson, D.B., Xie, S.-P., Zhou, T., 2013. Climate 730
Phenomena and their Relevance for Future Regional Climate Change. In: Climate 731
Change 2013: The Physical Science Basis. Contribution of Working Group I to the 732
Fifth Assessment Report of the Intergovernmental Panel on Climate Change 733
[Stocker, T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., Boschung, J., 734
Nauels, A., Xia, Y., Bex, V., Midgley, P.M. (eds.)]. Cambridge University Press, 735
Cambridge, United Kingdom and New York, NY, USA. 736
Church, J.A., Clark, P.U., Cazenave, A., Gregory, J.M., Jevrejeva, S., Levermann, A., 737
Merrifield, M.A., Milne, G.A., Nerem, R.S., Nunn, P.D., Payne, A.J., Pfeffer, W.T., 738
Stammer, D., Unnikrishnan, A.S., 2013. Sea Level Change. In: Climate Change 739
2013: The Physical Science Basis. Contribution of Working Group I to the Fifth 740
Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, 741
T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., 742
Xia, Y., Bex, V., Midgley, P.M. (eds.)]. Cambridge University Press, Cambridge, 743
United Kingdom and New York, NY, USA. 744
Collins, M., Knutti, R., Arblaster, J., Dufresne, J.-L., Fichefet, T., Friedlingstein, P., Gao, 745
X., Gutowski, W.J., Johns, T., Krinner, G., Shongwe, M., Tebaldi, C., Weaver, A.J., 746
Wehner, M., 2013. Long-term climate change: projections, commitments and 747
irreversibility. In: Climate Change 2013: The Physical Science Basis. Contribution 748
33
of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel 749
on Climate Change [Stocker, T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., 750
Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M. (eds.)]. Cambridge 751
University Press, Cambridge, United Kingdom and New York, NY, USA. 752
Crespo, O., Hachigonta, S., Tadross, M., 2011. Sensitivity of southern African maize yields 753
to the definition of sowing dekad in a changing climate. Climatic Change 106:267-754
283. 755
Deryng, D., Elliott, J., Ruane, A.C., Folberth, C., Müller, C., Pugh, T.A.M., Schmid, E., 756
Boote, K.J., Conway, D., Gerten, D., Jones, J.W., Khabarov, N., Olin, S., 757
Schaphoff, S., Yang, H., Rosenzweig, C., 2016. Regional disparities in the 758
beneficial effects of rising CO2 concentrations on crop water productivity. Nature 759
Clim. Change, 6(8), 786-790, doi:10.1038/nclimate2995. 760
Diffenbaugh, N.S., Singh, D., Mankin, J.S., 2018. Unprecedented climate events: 761
Historical changes, aspirational targets, and national commitments. Sci. Adv. 4, 762
eaao3354. 763
Donatelli, M., Magarey, R.D., Bregaglio, S., Willocquet, L., Whish, J.P.M., Savary, S., 764
2017. Modelling the impacts of pests and diseases on agricultural systems. 765
Agricultural Systems 155, 213-224. doi:10.1016/j.agsy.2017.01.019. 766
Elliott, J., Deryng, D., Müller, C., Frieler, K., Konzmann, M., Gerten, D., Glotter, M., 767
Flörke, M., Wada, Y., Best, N., Eisner, S., Fekete, B.M., Folberth, C., Foster, I., 768
Gosling, S.N., Haddeland, I., Khabarov, N., Ludwig, F., Masaki, Y., Olin, S., 769
Rosenzweig, C., Ruane, A.C., Satoh, Y., Schmid, E., Stacke, T., Tang, Q., Wisser, 770
D., 2014. Constraints and potentials of future irrigation water availability on 771
agricultural production under climate change. Proc. Natl. Acad. Sci., 111, no. 9, 772
3239-3244, doi:10.1073/pnas.1222474110. 773
Elliott, J., Müller, C., Deryng, D., Chryssanthacopoulos, J., Boote, K.J., Büchner, M., 774
Foster, I., Glotter, M., Heinke, J., Iizumi, T., Izaurralde, R.C., Mueller, N.D., Ray, 775
D.K., Rosenzweig, C., Ruane, A.C., Sheffield, J., 2015. The Global Gridded Crop 776
Model Intercomparison: Data and modeling protocols for Phase 1 (v1.0). Geosci. 777
Model Dev., 8, 261-277, doi:10.5194/gmd-8-261-2015. 778
Eyring, V., Bony, S., Meehl, G.A., Senior, C.A., Stevens, B., Stouffer, R.J., Taylor, K.E., 779
2016a. Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) 780
experimental design and organization, Geosci. Model Dev., 9, 1937-1958, 781
doi:10.5194/gmd-9-1937-2016. 782
FAO, 2016. The State of Food and Agriculture: Climate Change, Agriculture and Food 783
Security. 194 pages. Available online at: http://www.fao.org/3/a-i6030e.pdf. 784
Faye, B., et al., 2018: What does 1.5 versus 2.0°C of global warming mean for cereal yields 785
in the West African Sudan Savannah? Environmental Research Letters 13, 034014. 786
doi:10.1088/1748-9326/aaab40. 787
Flato, G., Marotzke, J., Abiodun, B., Braconnot, P., Chou, S.C., Collins, W., Cox, P., 788
Driouech, F., Emori, S., Eyring, V., Forest, C., Gleckler, P., Guilyardi, E., Jakob, 789
C., Kattsov, V., Reason, C., Rummukainen, M., 2013. Evaluation of climate 790
models. In: Climate Change 2013: The Physical Science Basis. Contribution of 791
Working Group I to the Fifth Assessment Report of the Intergovernmental Panel 792
on Climate Change [Stocker, T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., 793
Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M. (eds.)]. Cambridge 794
34
University Press, Cambridge, United Kingdom and New York, NY, USA. 741–795
882, doi:10.1017/CBO9781107415324.020. 796
Fleisher, D.H., Condori, B., Quiroz, R., Alva, A., Asseng, S., Barreda, C., Bindi, M., Boote, 797
K.J., Ferrise, R., Franke, A.C., Govindakrishnan, P.M., Harahagazwe, D., 798
Hoogenboom, G., Naresh Kumar, S., Merante, P., Nendel, C., Olesen, J.E., Parker, 799
P.S., Raes, D., Raymundo, R., Ruane, A.C., Stockle, S., Supit, I., Vanuytrecht, E., 800
Wolf, J., Woli, P., 2017. A potato model inter-comparison across varying climates 801
and productivity levels. Glob. Change Biol., 23, no. 3, 1258-1281, 802
doi:10.1111/gcb.13411. 803
Glotter, M., Elliott, J., 2016: Simulating US agriculture in a modern Dust Bowl drought. 804
Nature Plants, 3, 16193, doi:10.1038/nplants.2016.193. 805
Hansen, J., Sato, M., Hearty, P., Ruedy, R., Kelley, M., Masson-Delmotte, V., Russell, G., 806
Tselioudis, G., Cao, J., Rignot, E., Velicogna, I., Tormey, B., Donovan, B., 807
Kandiano, E., von Schuckmann, K., Kharecha, P., LeGrande, A.N., Bauer, M., Lo, 808
K.-M., 2016. Ice melt, sea level rise and superstorms: Evidence from paleoclimate 809
data, climate modeling, and modern observations that 2°C global warming could 810
be dangerous. Atmos. Chem. Phys., 16, 3761-3812. 811
Hartmann, D.L., Klein Tank, A.M.G., Rusticucci, M., Alexander, L.V., Brönnimann, S., 812
Charabi, Y., Dentener, F.J., Dlugokencky, E.J., Easterling, D.R., Kaplan, A., 813
Soden, B.J., Thorne, P.W., Wild, M., Zhai, P.M., 2013. Observations: Atmosphere 814
and Surface. In: Climate Change 2013: The Physical Science Basis. Contribution 815
of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel 816
on Climate Change [Stocker, T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., 817
Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M. (eds.)]. Cambridge 818
University Press, Cambridge, United Kingdom and New York, NY, USA. 819
Hatfield, J.L., Boote, K.J., Kimball, B.A., Ziska, L.H., Izaurralde, R.C., Ort, D., Thomson, 820
A.M., Wolfe, D., 2011: Climate Impacts on Agriculture: Implications for Crop 821
Production. Agronomy, 103(2), doi:10.2134/agronj2010.0303. 822
Hempel, S., Frieler, K., Warszawski, L., Shewe, J., Piontek, F., 2013. A trend-preserving 823
bias correction – the ISI-MIP approach. Earth Syst. Dynam., 4, 219–236. 824
Howden, S.M., Soussana, J.F., Tubiello, F.N., Chhetri, N., Dunlop, M., Meinke, H., 2007. 825
Adapting agriculture to climate change. Proceedings of the National Academy of 826
Sciences of the United States of America, 104, 19691–19696. 827
doi:10.1073/pnas.0701890104. 828
IPCC, 2013. Summary for Policymakers. In: Climate Change 2013: The Physical Science 829
Basis. Contribution of Working Group I to the Fifth Assessment Report of the 830
Intergovernmental Panel on Climate Change [Stocker, T.F., Qin, D., Plattner, G.-831
K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, 832
P.M. (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New 833
York, NY, USA. 834
IPCC, 2014: Summary for policymakers. In: Climate Change 2014: Impacts, Adaptation, 835
and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working 836
Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate 837
Change [Field, C.B., Barros, V.R., Dokken, D.J., Mach, K.J., Mastrandrea, M.D., 838
Bilir, T.E., Chatterjee, M., Ebi, K.L., Estrada, Y.O., Genova, R.C., Girma, B., 839
Kissel, E.S., Levy, A.N., MacCracken, S., Mastrandrea, P.R., White, L.L. (eds.)]. 840
35
Cambridge University Press, Cambridge, United Kingdom and New York, NY, 841
USA. 1-32. 842
Iizumi, T., Luo, J.-J., Challinor, A.J., Sakurai, G., Yokozawa, M., Sakuma, H., Brown, 843
M.E., Yamagata, T. 2014. Impacts of El Niño Southern Oscillation on the global 844
yields of major crops. Nature Communications 5 (3712), 845
doi:10.1038/ncomms4712. 846
Karmalker, A.V., Bradley, R.S., 2017. Consequences of Global Warming of 1.5 ºC and 2 847
ºC for Regional Temperature and Precipitation Changes in the Contiguous United 848
States. PLOS One, 12. doi:10.1371/journal.pone.0168697. 849
Kimball, B.A., 2016. Crop responses to elevated CO2 and interactions with H2O, N, and 850
temperature. Current Opinion in Plant Biology, 31, 36-43. 851
Kriegler, E., Edmonds, J., Hallegatte, S., Ebi, K.L., Kram, T., Riahi, K., Winkler, H., van 852
Vuuren, D.P., 2014. A new scenario framework for climate change research: the 853
concept of shared climate policy assumptions. Climatic Change, 122, 401–414, 854
doi:10.1007/s10584-013-0971-5. 855
Lau, K.-M., Kim, K.-M., 2006. Observational relationships between aerosol and Asian 856
monsoon rainfall, and circulation, Geophys. Res. Lett., 33, L21810, 857
doi:10.1029/2006GL027546. 858
Li, T., Hasegawa, T., Yin, X., Zhu, Y., Boote, K., Adam, M., Bregaglio, S., Buis, S., 859
Confalonieri, R., Fumoto, T., Gaydon, D., Marcaida III, M., Nakagawa, H., Oriol, 860
P., Ruane, A.C., Ruget, F., Singh, B., Singh, U., Tang, L., Tao, F., Wilkens, P., 861
Yoshida, H., Zhang, Z., Bouman, B., 2015. Uncertainties in predicting rice yield by 862
current crop models under a wide range of climatic conditions. Glob. Change Biol., 863
21, no. 3, 1328-1341, doi:10.1111/gcb.12758. 864
Lobell, D.B., Hammer, G.L., McLean, G., Messina, C., Roberts, M.J., Schlenker, W., 2013. 865
The critical role of extreme heat for maize production in the United States. Nature 866
Climate Change 3:497. 867
Lobell, D.B., Bänziger, M., Magorokosho, C., Vivek, B., 2011. Nonlinear heat effects on 868
African maize as evidenced by historical yield trials. Nature Climate Change 1:42-869
45. 870
Lopez, A., Suckling, E.B., Smith, L.A., 2014. Robustness of pattern scaled climate change 871
scenarios for adaptation decision support. Climatic Change, 122, 555–566, 872
doi:10.1007/s10584-013-1022-y. 873
Maiorano, A., Martre, P., Asseng, S., Ewert, F., Müller, C., Rötter, R., Ruane, A.C., 874
Semenov, M., Wallach, D., Wang, E., Alderman, P., Kassie, B., Biernath, C., 875
Basso, B., Cammarano, D., Challinor, A.J., Doltra, J., Dumont, B., Eyshi Rezaei, 876
E., Gayler, S., Kersebaum, K.C., Kimball, B.A., Koehler, A.K., Liu, B., O'Leary, 877
G.J., Olesen, J.E., Ottman, M.J., Priesack, E., Reynolds, M.P., Stratonovitch, P., 878
Streck, T., Thorburn, P.J., Waha, K., Wall, G.W., White, J.W., Zhao, Z., Zhu, Y., 879
2017. Crop model improvement reduces the uncertainty of the response to 880
temperature of multi-model ensembles. Field Crops Res., 202, 5-20, 881
doi:10.1016/j.fcr.2016.05.001. 882
McDermid, S.P., Mearns, L.O., Ruane, A.C., 2017. Representing agriculture in Earth 883
System Models: Approaches and priorities for development. J. Adv. Model. Earth 884
Syst., 9, no. 5, 2230-2265, doi:10.1002/2016MS000749. 885
36
Mitchell, D., AchutaRao, K., Allen, M., Bethke, I., Beyerle, U., Ciavarella, A., Forster, 886
P.M., Fuglestvedt, J., Gillett, N., Haustein, K., Ingram, W., Iversen, T., Kharin, V., 887
Klingaman, N., Massey, N., Fischer, E., Schleussner, C.-F., Scinocca, J., Seland, 888
Ø., Shiogama, H., Shuckburgh, E., Sparrow, S., Stone, D., Uhe, P., Wallom, D., 889
Wehner, M., Zaaboul, R., 2017. Half a degree additional warming, prognosis and 890
projected impacts (HAPPI): background and experimental design. Geosci. Model 891
Dev., 10, 571–583. 892
Morice, C.P., Kennedy, J.J., Rayner, N.A., Jones, P.D., 2012. Quantifying uncertainties in 893
global and regional temperature change using an ensemble of observational 894
estimates: The HadCRUT4 dataset. J. Geophys. Res., 117, D08101, doi: 895
10.1029/2011JD017187. 896
Moss, R.H., Edmonds, J.A., Hibbard, K.A., Manning, M.R., Rose, S.K., van Vuuren, D.P., 897
Carter, T.R., Emori, S., Kainuma, M., Kram, T., Meehl, G.A., Mitchell, J.F.B., 898
Nakicenovic, N., Riahi, K., Smith, S.J., Stouffer, R.J., Thomson, A.M., Weyant, 899
J.P., Wilbanks, T.J., 2010. The next generation of scenarios for climate change 900
research and assessment. Nature, 463(7282), 747–756, doi:10.1038/nature08823. 901
Müller, C., Elliott, J., Chryssanthacopoulos, J., Arneth, A., Balkovič, J., Ciais, P., Deryng, 902
D., Folberth, C., Glotter, M., Hoek, S., Iizumi, T., Izaurralde, R.C., Jones, C., 903
Khabarov, N., Lawrence, P., Liu, W., Olin, S., Pugh, T.A.M., Ray, D., Reddy, A., 904
Rosenzweig, C., Ruane, A.C., Sakurai, G., Schmid, E., Skalsky, R., Song, C.X., 905
Wang, X., de Wit, A., Yang, H., 2017. Global gridded crop model evaluation: 906
Benchmarking, skills, deficiencies and implications. Geosci. Model Dev., 10, 1403-907
1422, doi:10.5194/gmd-10-1403-2017. 908
O’Leary, G.J., Christy, B., Nuttall, J., Huth, N., Cammarano, D., Stöckle, C., Basso, B., 909
Shcherbak, I., Fitzgerald, G., Luo, Q., Farre-Codina, I., Palta, J., Asseng, S., 2015. 910
Response of wheat growth, grain yield and water use to elevated CO2 under a Free-911
Air CO2 Enrichment (FACE) experiment and modelling in a semi-arid 912
environment. Glob. Change Biol., 21, 2670-2686. 913
O’Neill, B.C., Kriegler, E., Riahi, K., Ebi, K.L., Hallegatte, S., Cater, T.R., Mathur, R., 914
van Vuuren, D.P., 2014. A new scenario framework for climate change research: 915
the concept of shared socioeconomic pathways. Climatic Change, 122, 387–400, 916
doi:10.1007/s10584-013-0905-2. 917
Pendergrass, A.G., Hartmann, D.L., 2014. Changes in the distribution of rain Frequency 918
and intensity in response to global warming. J. Climate 27, 8372-8383, 919
doi:10.1175/JCLI-D-14-00183.1. 920
Pirttioja, N., Carter, T., Fronzek, S., Bindi, M., Hoffmann, H., Palosuo, T., Ruiz-Ramos, 921
M., Tao, F., Acutis, M., Asseng, S., Baranowski, P., Basso, B., Bodin, P., Buis, S., 922
Cammarano, D., Deligios, P., Destain, M.-F., Dumont, B., Ewert, R., Ferrise, R., 923
François, L., Gaiser, T., Hlavinka, P., Jacquemin, I., Kersebaum, K.C., Kollas, C., 924
Krzyszczak, J., Lorite, I.J., Minet, J., Minguez, M.I., Montesino, M., Moriondo, M., 925
Müller, C., Nendel, C., Öztürk, I., Perego, A., Rodríguez, A., Ruane, A.C., Ruget, 926
F., Sanna, M., Semenov, M.A., Slawinski, C., Stratonovitch, P., Supit, I., Waha, 927
K., Wang, E., Wu, L., Zhao, Z., Rötter, R.P., 2015. A crop model ensemble analysis 928
of temperature and precipitation effects on wheat yield across a European transect 929
using impact response surfaces. Clim. Res., 65, 87-105, doi:10.3354/cr01322. 930
37
Porter, J.R., Xie, L., Challinor, A.J., Cochrane, K., Howden, S.M., Iqbal, M.M., Lobell, 931
D.B., Travasso, M.I., 2014. Food security and food production systems. In: Climate 932
Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral 933
Aspects. Contribution of Working Group II to the Fifth Assessment Report of the 934
Intergovernmental Panel on Climate Change [Field, C.B., Barros, V.R., Dokken, 935
D.J., Mach, K.J., Mastrandrea, M.D., Bilir, T.E., Chatterjee, M., Ebi, K.L., Estrada, 936
Y.O., Genova, R.C., Girma, B., Kissel, E.S., Levy, A.N., MacCracken, S., 937
Mastrandrea, P.R., White, L.L. (eds.)]. Cambridge University Press, Cambridge, 938
United Kingdom and New York, NY, USA. 485-533. 939
Ramanathan, V., Chung, C., Kim, D., Bettge, T., Buja, L., Kiehl, J.T., Washington, W.M., 940
Fu, Q., Sikka, D.R., Wild, M., 2005. Atmospheric Brown Clouds: Impacts on South 941
Asian Climate and Hydrological Cycle. PNAS, Vol. 102, No. 15, 5326-5333. 942
Rickards, L., Howden, S.M., 2012. Transformational adaptation: agriculture and climate 943
change. Crop & Pasture Science, 2012, 63, 240–250, doi:10.1071/CP11172. 944
Rind, D., Chin, M., Feingold, G., Streets, D., Kahn, R.A., Schwartz, S.E., Yu, H., 2009. 945
Modeling the effects of aerosols on climate, in atmospheric aerosol properties and 946
impacts on climate. In: Aerosol Properties and Their Impacts on Climate, U.S. 947
Climate Change Science Program Synthesis and Assessment Product 2.3 [Chin, M., 948
Kahn, R.A., Schwartz, S.E. (eds.)]. National Aeronautics and Space 949
Administration. 64-97. 950
Rohde, R., Muller, R., Jacobsen, R., Perlmutter, S., Rosenfeld, A., Wurtele, J., Curry, J., 951
Wickham, C., Mosher, S., 2013. Berkeley Earth Temperature Averaging Process. 952
Geoinfor. Geostat: An Overview, 1:2, doi:10.4172/2327-4581.1000103. 953
Rosenzweig, C., Tubiello, F.N., 2007. Adaptation and mitigation strategies in agriculture: 954
an analysis of potential synergies. Mitig. Adapt. Strat. Glob. Change, 12, 855–873, 955
doi:10.1007/s11027-007-9103-8. 956
Rosenzweig, C., Jones, J.W., Hatfield, J.L., Ruane, A.C., Boote, K.J., Thorburn, P., Antle, 957
J.M., Nelson, G.C., Porter, C., Janssen, S., Asseng, S., Basso, B., Ewert, F., 958
Wallach, D., Baigorria, G., Winter, J.M., 2013. The Agricultural Model 959
Intercomparison and Improvement Project (AgMIP): Protocols and pilot studies. 960
Agric. Forest Meteorol., 170, 166-182, doi:10.1016/j.agrformet.2012.09.011. 961
Rosenzweig, C., Elliott, J., Deryng, D., Ruane, A.C., Müller, C., Arneth, A., Boote, K.J., 962
Folberth, C., Glotter, M., Khabarov, N., Neumann, K., Piontek, F., Pugh, T.A.M., 963
Schmid, E., Stehfest, E., Yang, H., Jones, J.W., 2014. Assessing agricultural risks 964
of climate change in the 21st century in a global gridded crop model 965
intercomparison. Proc. Natl. Acad. Sci., 111, no. 9, 3268-3273, 966
doi:10.1073/pnas.1222463110. 967
Rosenzweig, C., Jones, J.W., Hatfield, J.L., Antle, J.M., Ruane, A.C., Mutter, C.Z., 2015. 968
The Agricultural Model Intercomparison and Improvement Project: Phase I 969
activities by a global community of science. In: The Handbook of Climate Change 970
and Agroecosystems: The Agricultural Model Intercomparison and Improvement 971
Project (AgMIP) [Rosenzweig, C., Hillel, D. (eds.)]. ICP Series on Climate Change 972
Impacts, Adaptation, and Mitigation, Vol. 3. Imperial College Press, 3-24, 973
doi:10.1142/9781783265640_0001. 974
Rosenzweig, C., Antle, J., Elliott, J., 2016. Assessing Impacts of Climate Change on Food 975
Security Worldwide. Eos, 97, doi:10.1029/2016EO047387. 976
38
Rosenzweig et al., 2018: Coordinating AgMIP data and models across global and regional 977
scales for 1.5°C and 2.0°C assessments. Phil. Trans. R. Soc. A 20160455. 978
doi:10.1098/rsta.2016.0455. 979
Ruane, A.C., McDermid, S.P., 2017. Selection of a representative subset of global climate 980
models that captures the profile of regional changes for integrated climate impacts 981
assessment. Earth Perspect., 4, 1, doi:10.1186/s40322-017-0036-4. 982
Ruane, A.C., Major, D.C., Yu, W.H., Alam, M., Hussain, S.G., Khan, A.S., Hassan, A., Al 983
Hossain, B.M.T., Goldberg, R., Horton, R.M., Rosenzweig, C., 2013. Multi-factor 984
impact analysis of agricultural production in Bangladesh with climate change. 985
Glob. Environ. Change A, 23, 336-350, doi:10.1016/j.gloenvcha.2012.09.001. 986
Ruane, A.C., McDermid, S., Rosenzweig, C., Baigorria, G.A., Jones, J.W., Romero, C.C., 987
Cecil, L.D., 2014. Carbon–Temperature–Water change analysis for peanut 988
production under climate change: a prototype for the AgMIP Coordinated Climate-989
Crop Modeling Project (C3MP). Glob. Change Biol., 20, 394–407, doi: 990
10.1111/gcb.12412. 991
Ruane, A.C., R. Goldberg, and J. Chryssanthacopoulos, 2015a: Climate forcing datasets 992
for agricultural modeling: Merged products for gap-filling and historical climate 993
series estimation. Agric. Forest Meteorol., 200, 233-248, 994
doi:10.1016/j.agrformet.2014.09.016. 995
Ruane, A.C., Winter, J.M., McDermid, S.P., Hudson, N.I., 2015b. AgMIP climate datasets 996
and scenarios for integrated assessment. In: The Handbook of Climate Change and 997
Agroecosystems: The Agricultural Model Intercomparison and Improvement 998
Project Integrated Crop and Economic Assessments, Part 1 [Rosenzweig, C., Hillel, 999
D. (eds.)]. ICP Series on Climate Change Impacts, Adaptation, and Mitigation, Vol. 1000
3. Imperial College Press, 45-78, doi:10.1142/9781783265640_0003. 1001
Ruane, A.C., Rosenzweig, C., Asseng, S., Boote, K.J., Elliott, J., Ewert, F., Jones, J.W., 1002
Martre, P., McDermid, S., Müller, C., Snyder, A., Thorburn, P.J., 2017. An AgMIP 1003
framework for improved agricultural representation in IAMs. Environ. Res. Lett., 1004
in press, doi:10.1088/1748-9326/aa8da6. 1005
Ruane, A.C., Antle, J., Elliott, J., Folberth, C., Hoogenboom, G., Mason-D’Croz, D., 1006
Müller, C., Porter, C.H., Phillips, M., Raymundo, R., Sands, R., Valdivia, R., 1007
White, J., Wiebe, K., Rosenzweig, C., 2018. Global and Regional Agricultural 1008
Implications of +1.5 and +2.0 °C Warming. Glob. Environ. Change, under review. 1009
Savary, S., Ficke, A., Aubertot, J.-N., Hollier, C., 2012. Crop losses due to diseases and 1010
their implications for global food production losses and food security. Food Sec. 1011
4(2), doi:10.1007/s12571-012-0200-5. 1012
Schleussner et al., 2018: Crop productivity changes at 1.5°C and 2°C under climate 1013
response uncertainty, submitted. 1014
Smith. T.M., Reynolds, R.W., 2008. Improvements to NOAA’s historical merged land-1015
ocean surface temperature analysis (1880-2006). J. Climate, 28, 911-930. 1016
SRES, 2000. Special Report on Emissions Scenarios, A Special Report of Working Group 1017
III of the Intergovernmental Panel on Climate Change [Nakicenovic, N., Alcamo, 1018
J., Grubler, A., Riahi, K., Roehrl, R.A., Rogner, H.-H., Victor, N.]. 599 pages. 1019
Cambridge University Press, Cambridge, UK. 1020
Tadross, M., Suarez, P., Lotsch, A., Hachigonta, S., Mdoka, M., Unganai, L., Lucio, F., 1021
Kamdonyo, D., Muchinda, M., 2009. Growing-season rainfall and scenarios of 1022
39
future change in southeast Africa: implications for cultivating maize. Climate 1023
Research 40:147-169. 1024
Taylor, K.E., Stouffer, R.J., Meehl, G.A., 2012. An overview of CMIP5 and the experiment 1025
design. Bull. Am. Meteorol. Soc., 93(4), 485–498, doi:10.1175/BAMSD-11-1026
00094.1. 1027
Tebaldi, C., Arblaster, J.M., 2014. Pattern scaling: Its strengths and limitations, and an 1028
update on the latest model simulations. Climatic Change, 122, 459, 1029
doi:10.1007/s10584-013-1032-9. 1030
Trenberth, K., 2011. Changes in precipitation with climate change. Clim. Res., 47, 123–1031
138. doi:10.3354/cr00953. 1032
UNFCCC, 2015. Adoption of the Paris Agreement. United Nations Framework 1033
Convention on Climate Change. 32 pages. Available online at: 1034
http://unfccc.int/resource/docs/2015/cop21/eng/l09r01.pdf. 1035
van Vuuren, D., van Sluisveld, M., Hof, A.F., 2015. Implications of long-term scenarios 1036
for medium-term targets (2050). PBL Netherlands Environmental Assessment 1037
Agency publication number 1871. 30 pages. Available online at: 1038
http://www.pbl.nl/sites/default/files/cms/publicaties/pbl-2015-implications-for-1039
long-term-scenarios-for-medium-term-targets-2050_01871.pdf. 1040
Wiebe, K., Lotze-Campen, H., Sands, R., Tabeau, A., van der Mensbrugghe, D., Biewald, 1041
A., Bodirsky, B., Islam, S., Kavallari, A., Mason-D’Croz, D., Müller, C., Popp, A., 1042
Robertson, R., Robinson, S., van Meiji, H., Willenbockel, D., 2015. Climate change 1043
impacts on agriculture in 2050 under a range of plausible socioeconomic and 1044
emissions scenarios, Environ. Res. Lett., 10, 085010, doi: 10.1088/1748-1045
9326/10/8/085010. 1046
WMO, 1989. Calculation of Monthly and Annual 30-Year Standard Normals (WCDP928 1047
No.10, WMO-TD/No.341), World Meteorological Organization, Geneva. 1048
Yadav, S.S., Redden, R., Hatfield, J.L., Lotze-Campen, H., Hall, A.J., 2011. Crop 1049
adaptation to climate change. 632 pages. John Wiley & Sons, Oxford, United 1050
Kingdom, doi:10.1002/9780470960929. 1051
You, L., Wood-Sichra, U., Fritz, S., Guo, Z., See, L. Koo, J., 2014. Spatial Production 1052
Allocation Model (SPAM) 2005 v2.0. Available online at: http://mapspam.info. 1053
Zhao, C., Liu, B., Piao, S., Wang, X., Lobell, D.B., Huang, Y., Huang, M. Yao, Y., Bassu, 1054
S., Ciais, P., Durand, J.-L., Elliott, J., Ewert, F., Janssens, I.A., Li, T., Lin, E., Liu, 1055
Q., Martre, P., Müller, C., Peng, S., Peñuelas, J., Ruane, A.C., Wallach, D., Wang, 1056
T., Wu, D., Liu, Z., Zhu, Y., Zhu, Z., Asseng, S., 2017. Temperature increase 1057
reduces global yields of major crops in four independent estimates. Proc. Natl. 1058
Acad. Sci., 114, no. 35, 9326-9331, doi:10.1073/pnas.1701762114. 1059
Zhang, H.-M., Huang, B., Lawrimore, J., Menne, M., Smith, T.M., 2017. NOAA Global 1060
Surface Temperature Dataset (NOAAGlobalTemp), Version 4.0 Global Anomalies 1061
and Index Data. NOAA National Centers for Environmental Information. 1062
doi:10.7289/V5FN144H [accessed March 7, 2017]. 1063
1064
40
Tables and Figures 1065 Table 1: HAPPI reference and simulation periods. 1066
Climate Period Years [CO2]* Notes
Pre-Industrial** 1861-1880 288 ppm Pre-industrial period with relatively few volcanic
eruptions
Current 2006-2015 390 ppm Based on CMIP historical simulation (coupled models
initiated 100+ years prior; not expected to exactly
match observed interannual variability)
+1.5 ºC World 2106-2115 423 ppm Based on CO2 and sea-surface temperature patterns
from end-of-century RCP2.6 simulations.
+2.0 ºC World 2106-2115 487 ppm Based on CO2 and sea-surface temperature patterns
from end-of-century RCP2.6 and RCP4.5 simulations.
*Current and pre-industrial [CO2] determined from RCP scenarios described by Moss et 1067
al., 2012. 1068
**The Pre-Industrial period was used to determine the +1.5 and +2.0 ºC levels but was 1069
not directly simulated. 1070
1071
41
Table 2: Global climate model (GCM) ensemble members included in the AgMIP CGRA 1072
study, including initial condition ensembles for multiple models in the a) HAPPI 1073
simulations and b) a multi-model ensemble from CMIP5. For further information on 1074
HAPPI GCMs see Mitchell et al. (2017), and on CMIP5 GCMs see Ruane et al. (2017) and 1075
Flato et al. (2013). Aggregate HAPPI outputs are statistical summaries of broader 1076
ensemble used for mean impacts, while daily outputs included enabled distributional 1077
analysis across all ensemble members. Access date reflects output versions. 1078
a) 1079 Global Climate Model
HAPPI
Current
HAPPI
+1.5 ºC
HAPPI
+2.0 ºC
Horizontal
Resolution (Latitude x
Longitude)
Access
Date
CanAM4 (aggregate) 100 100 100 2.81˚ x 2.81˚ 06/2017
CAM4-2degree (aggregate) 500 500 500 1.88˚ x 2.5˚ 06/2017
CAM4-2degree (daily) 20 20 20 0.5˚ x 0.5˚ 09/2017
HadAM3P (aggregate) 94 83 86 1.25˚ x 1.88 06/2017
MIROC5* (aggregate) 160 100 100 1.41˚ x 1.41˚ 06/2017
MIROC5* (daily) 20 20 20 0.5˚ x 0.5˚ 08/2017
NorESM1* (aggregate) 125 125 125 0.94˚ x 1.25˚ 06/2017
NorESM1* (daily) 20 20 20 0.5˚ x 0.5˚ 09/2017
1080
b) CMIP5 GCMs from which Historical and RCP4.5 outputs were analyzed (one 1081
simulation each). 1082 Global Climate
Model
Horizontal
Resolution (Latitude x
Longitude)
Access
Date
Global Climate
Model
Horizontal
Resolution (Latitude x
Longitude)
Access
Date
ACCESS1-0 1.25˚ x 1.88˚ 02/2012 GISS-E2-H 2˚ x 2.5˚ 03/2013
BCC-CSM1-1 2.81˚ x 2.81˚ 07/2011 GISS-E2-R 2˚ x 2.5˚ 05/2012
BCC-CSM1-1-m 1.13˚ x 1.13˚ 04/2016 HadGEM2-AO 1.25˚ x 1.88˚ 05/2015
BNU-ESM 2.81˚ x 2.81˚ 04/2012 HadGEM2-CC 1.25˚ x 1.88˚ 01/2012
CanESM2 2.81˚ x 2.81˚ 04/2011 HadGEM2-ES 1.25˚ x 1.88˚ 01/2011
CCSM4 0.94˚ x 1.25˚ 11/2011 INMCM4.0 1.5˚ x 2˚ 06/2011
CESM1-BGC 0.94˚ x 1.25˚ 03/2012 IPSL-CM5A-LR 1.88˚ x 3.75˚ 06/2011
CMCC-CM 0.75˚ x 0.75˚ 05/2015 IPSL-CM5A-MR 1.27˚ x 2.5˚ 10/2011
CMCC-CMS 1.88˚ x 1.88˚ 05/2015 IPSL-CM5B-LR 1.88˚ x 3.75˚ 05/2015
CNRM-CM5 1.41˚ x 1.41˚ 05/2015 MIROC5 1.41˚ x 1.41˚ 07/2011
CSIRO-Mk3-6-0 1.88˚ x 1.88˚ 04/2012 MIROC-ESM 2.81˚ x 2.81˚ 10/2011
EC-Earth 1.13˚ x 1.13˚ 04/2016 MPI-ESM-LR 1.88˚ x 1.88˚ 08/2011
FGOALS-g2 2.81˚ x 2.81˚ 03/2012 MPI-ESM-MR 1.88˚ x 1.88˚ 10/2011
GFDL-CM3 2˚ x 2.5˚ 07/2013 MRI-CGCM3 1.13˚ x 1.13˚ 08/2011
GFDL-ESM2G 2˚ x 2.5˚ 11/2011 NorESM1-M 1.88˚ x 2.5˚ 06/2011
GFDL-ESM2M 2˚ x 2.5˚ 11/2011
Note: * HAPPI versions of MIROC5 and NorESM1 are updated versions of similar 1083
models used for CMIP5. 1084
42
1085 Figure 1: +1.5 ºC World (a,c,e,g,i) temperature and (b,d,f,h,j) precipitation changes for 1086
(a,b) all land areas and seasons; (c,d) rainfed maize; (e,f) rainfed wheat; (g,h) rainfed rice; 1087
(i,j) rainfed soy. These maps are the basis for estimating local climate change patterns 1088
between the 1995 Climate used by AgMIP protocols and the 2010 climate used by HAPPI. 1089
Hatch marks for temperature indicate that median changes are greater than twice the range 1090
among GCMs, and hatch marks for precipitation indicate agreement on the direction of 1091
change by 4 of the 5 HAPPI models. Grid cells with <10ha rainfed crop area in the SPAM 1092
dataset (You et al., 2014) are omitted to focus on regions with substantial production. 1093
43
1094 Figure 2: +2.0 ºC World (a,c,e,g,i) temperature and (b,d,f,h,j) precipitation changes for 1095
(a,b) all land areas and seasons; (c,d) rainfed maize; (e,f) rainfed wheat; (g,h) rainfed rice; 1096
(i,j) rainfed soy. Hatch marks and area mask as in Figure 1. 1097
1098
44
1099 Figure 3: Classes of +1.5 ºC World local (a,c,e,g) temperature and (b,d,f,h) precipitation 1100
change deviations for growing seasons compared to annual changes for (a,b) rainfed maize; 1101
(c,d) rainfed wheat; (e,f) rainfed rice; (g,h) rainfed soy. ‘Substantially’ different indicates 1102
that growing season change differences in at least 3 out of the 5 HAPPI models changed at 1103
a pace that was ‘faster’ or ‘slower’ than 0.2 ºC for temperatures or 5% for precipitation 1104
compared to the annual mean, while ‘Consistently’ different indicates that all 5 HAPPI 1105
GCMs agreed on the sign of difference. Areas that meet both the ‘substantially’ and 1106
‘consistently’ criteria are designated as ‘substantially’ different. Area mask as in Figure 1. 1107
Note that a Figure S3 presents corresponding results for the +2.0 ºC World. 1108
1109
1110
45
1111 Figure 4: Shifts in extreme conditions for rainfed maize season in +1.5 ºC World. (a) 1112
Change in the frequency of extreme heat years (above 90th percentile in current climate); 1113
(b) change in the frequency of extreme drought years (below 10th percentile in current 1114
climate); (c) additional extreme heat years in the +2.0 C World compared to the +1.5 C 1115
World; (d) additional extreme drought years in the +2.0 C World compared to the 1.5 C 1116
World; (e) change in the skewness of seasonal temperature distribution; (f) change in the 1117
skewness of seasonal precipitation distribution. Hatch marks in (a) and (b) indicate 1118
reduction in frequency of more than ¼ or increases of more than ½. Regions marked as 1119
inconsistent in (e,f) unless 2 of the 3 HAPPI daily GCMs demonstrate substantial shifts in 1120
skewness (percentile change differences greater than 0.1 ºC for temperature and 2% for 1121
precipitation). Area mask as in Figure 1c-d. 1122
1123
46
1124 Figure 5: +1.5 ºC World HAPPI projected (a,c,e,g) temperature and (b,d,f,h) precipitation 1125
changes, weighted by cropped area within the top 10 producing countries (presented in 1126
descending order from left to right) for (a,b) maize; (c,d) wheat; (e,f) rice; (g,h) soy. 1127
47
1128 Figure 6: +2.0 ºC World HAPPI projected (a,c,e,g) temperature and (b,d,f,h) precipitation 1129
changes, weighted by cropped area within the top 10 producing countries (presented in 1130
descending order from left to right) for (a,b) maize; (c,d) wheat; (e,f) rice; (g,h) soy. 1131
48
1132 Figure 7: HAPPI and CMIP5 RCP4.5 (a) temperature changes and (b) precipitation 1133
changes for annual conditions across all land areas as well as maize, wheat, rice, and soy 1134
seasons weighted according to cropped area. Symbols represent HAPPI simulations, while 1135
box-and-whisker diagrams indicate CMIP5 median projections and their interquartile 1136
range, while whiskers extend to the last point within an additional 1.5 times the 1137
interquartile range. 1138
1139
1140 1141
49
Climate Shifts for Major Agricultural Seasons in +1.5 and +2.0 °C Worlds: HAPPI 1142
Projections and AgMIP Modeling Scenarios 1143 1144
Alex C. Ruane1 1145
Meridel M. Phillips2,1 1146
Cynthia Rosenzweig1 1147
1148 1NASA Goddard Institute for Space Studies, New York, NY, USA 1149
2Columbia University Center for Climate Systems Research, New York, NY, USA 1150
1151
Corresponding Author: 1152
Alex Ruane 1153
NASA Goddard Institute for Space Studies 1154
2880 Broadway 1155
New York, NY 10025 1156
[email protected] 1157
1158
50
S1. Brief summary of the HAPPI Tier 1 simulation approach (see further details in 1159
Mitchell et al., 2017): 1160
HAPPI defines the pre-industrial period as 1861-1880, a 20 year climatology with 1161
little long-term global trend that is also largely devoid of major volcanic eruptions. 1162
Simulations of the 2006-2015 current period climate utilized SSTs provided by 1163
RCP8.5 considering its closest match with observations in this period. Global mean 1164
temperature for this current period is 0.92 ºC above pre-industrial conditions, and 1165
thus we would expect the +1.5 and +2.0 ºC Worlds to be only 0.58 ºC and 1.08 ºC 1166
warmer than current conditions, respectively. 1167
All driving conditions for the +1.5 ºC World were drawn from the lowest emissions 1168
representative concentration pathway RCP2.6 (Moss et al., 2010) end-of-century 1169
(2091-2100) period as the corresponding CMIP5 ensemble averaged 1.55 ºC over 1170
pre-industrial conditions. Note that this does not guarantee that each GCM 1171
produces exactly +1.5C global warming, rather the experiment is designed so that 1172
it is probable that the multi-model ensemble represents the +1.5 ºC World. RCP2.6 1173
also contains substantial reductions in tropospheric aerosols under an assumption 1174
of increased regulation toward the end of the 21st century, with skies particularly 1175
clearer over South and East Asia. 1176
The +2.0 ºC World was created from a weighted average of driving conditions from 1177
the RCP2.6 and RCP4.5 end-of-century scenarios, as the corresponding CMIP5 1178
ensemble of both RCPs averaged ~2.0 ºC global warming. For this scenario HAPPI 1179
derived only new core driving conditions (CO2, sea surface temperatures (SSTs), 1180
51
sea ice, and natural forcings); all others (including aerosol concentrations) were 1181
held to be the same as the +1.5 ºC World to isolate the effects of core forcings. 1182
SSTs in both future worlds were modified to impose mean changes from end-of-1183
century simulations on top of the SSTs used in the 2006-2015 simulation. 1184
Large (83-500 member) initial condition ensembles were conducted using 1185
atmosphere-only model versions for current and future worlds to more broadly 1186
capture changes in internal variability and extreme events. Given that these 10-1187
year simulations are on a similar time scale to some major modes of variability 1188
(such as the Pacific Decadal Oscillation), it is possible that initial conditions or the 1189
imposition of mean SST changes from the end of the century could over-represent 1190
particular phases of natural variability. 1191
HAPPI recommends CO2 concentrations of 423 ppm (1.5 ºC World) and 487 ppm 1192
(2.0 ºC World). The exact CO2 level at which a global climate model (GCM) will 1193
settle into a given stabilization temperature is related to its equilibrium climate 1194
sensitivity (which varies considerably across models; Flato et al., 2013); however, 1195
HAPPI aims to focus applications on the emergent patterns of regional climate 1196
change and shifts in the distribution of extreme events. As a key component of 1197
photosynthesis and stomatal gas transfer in plants, the selection of CO2 1198
concentration introduces further uncertainties (Rosenzweig et al., 2014; Deryng et 1199
al., 2016; Ruane et al., 2018). 1200
1201
52
1202 Figure S1: CO2 concentrations for current period and +1.5 and +2.0 ºC Worlds. HAPPI 1203
CO2 concentrations presented as gray dashed line for ‘current’ period (year 2010 = 390 1204
ppm) and black boxes for the +1.5 and +2.0C Worlds (423 ppm and 487 ppm, respectively). 1205
Crossing point CO2 concentrations in the CMIP5 RCP4.5 transient ensemble simulations 1206
are presented as box-and-whisker diagrams (indicating median, interquartile range, and 1207
whiskers extending to the last point within an additional 1.5 times the interquartile range). 1208
1209
53
1210 Figure S2: Middle day of growing season (between planting and harvesting dates in the 1211
AgMIP Global Gridded Crop Model Intercomparison) for (a) rainfed maize; (b) rainfed 1212
wheat; (c) rainfed rice; (d) rainfed soy. 1213
1214
54
1215 Figure S3: Classes of +2.0 ºC World local (a,c,e,g) temperature and (b,d,f,h) precipitation 1216
change deviations for growing seasons compared to annual changes for (a,b) rainfed maize; 1217
(c,d) rainfed wheat; (e,f) rainfed rice; (g,h) rainfed soy. ‘Substantially’ different indicates 1218
that growing season differences in at least 3 out of the 5 HAPPI models exceeded 0.2 ºC 1219
for temperatures or 5% for precipitation compared to the annual mean, while ‘Consistently’ 1220
different indicates that all 5 HAPPI GCMs agreed on the sign of difference. Area mask as 1221
in Figure 1. Note that the corresponding +1.5 C World analysis is presented in Figure 3. 1222
1223