CGIAR modeling approaches for resource‐constrained scenarios: I. Accelerating crop breeding for a changing climate

The quote only partly states this claim.

The meta-analysis of 19 modelling studies found positive mean simulated benefits from genotypic adaptation for wheat, sorghum, pearl millet, groundnut, chickpea, rice, maize, and barley.

“Average simulated future potential benefits from genotypic adaptation (including ideotype design) as derived from 19 modelling studies for wheat (n = 15 simulations), sorghum (n = 4), pearl millet (n = 48), groundnut (n = 12), chickpea (n = 48), rice (n = 159), maize (n = 19), and barley (n = 48).”

• AssumptionRests on an assumption: Publication bias in the systematic review does not overturn the conclusion that simulated gains from genotypic adaptation are positive for several crops.

Probably not a substantial claim: may not pass S3 (load-bearing).

The 12 CGIAR model-based ideotype design studies listed in Table 1, published over 9 years, indicate CGIAR Centers are very active in this area.

“The 12 studies listed in Table 1, published in a span of 9 years, indicate that CGIAR Centers are very active in this area of work.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

CGIAR involvement and leadership in trait-trait relationship modeling is very limited.

“CGIAR involvement and leadership in this area is very limited.”

May combine two claims.

CGIAR is undergoing major transformations in breeding programs, as shown by the Excellence in Breeding Platform, Crops to End Hunger Initiative, and existing genomic selection research.

“The CGIAR is no exception to these transformations, as shown by the establishment of the Excellence in Breeding Platform and the Crops to End Hunger Initiative 1 , and the existing research on genomic selection (Sect.2.4.1).”

May combine two claims.

Many CGIAR modeling studies, especially ideotype design studies, fail to capture the range of traits relevant in crop improvement, useful model outputs and scales, and parts of the breeding process that need informing.

“Many CGIAR modeling studies, especially those focused on ideotype design, fail to capture the range of traits relevant in crop improvement, the range of model outputs and spatial and temporal scales that would be useful to them, and the parts of the breeding process that need to be informed (see Sect. 2.3).”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Leadership from the CGIAR and coordination among its Centers vary in how they integrate modeling into breeding programs.

“In addition, varying degrees of leadership by the CGIAR and coordination between CGIAR Centers also exists with respect to the integration of modeling into breeding programs.”

May combine two claims.

Crop modeling work is seen to have, and actually has, minimal influence on breeding decisions and the breeding process.

“As a result, crop modeling activities have little perceived and actual impact on breeding decisions and the breeding process itself.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

CGIAR direct engagement and leadership in gene-based crop simulation modeling is very limited, likely due to center-specific focus, funding sources, and limited uptake/applicability in CGIAR breeding programs.

“The current level of direct engagement and leadership by the CGIAR in this line of work is very limited, likely due to a combination of CGIAR center-specific focus, funding sources for modelers, and limited uptake and applicability of these models in CGIAR breeding programs.”

May combine two claims.

The earliest crop model developed and used in CGIAR was the cassava model by CIAT, upon which GUMCAS, CROPSIM-Cassava, Gabriel et al. model, and MANIHOT were later developed.

“Perhaps the earliest crop model developed and used in the CGIAR was the cassava model developed by the International Center for Tropical Agriculture (CIAT) (Cock et al., 1979), upon which various components of the GUMCAS model (Matthews & Hunt, 1994), the CROPSIM-Cassava model (Jones et al., 2003), the model of Gabriel et al. (2014), and the MANIHOT model (CIAT, unpublished), were later developed.”

• cites Cock et al., 1979Not checked yet

  Not checked in the free run.

• cites Matthews & Hunt, 1994Not checked yet

  Not checked in the free run.

• cites Jones et al., 2003Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The release of semi-dwarf rice variety IR8 by IRRI in the 1950s and 1960s increased yield potential by 70%.

“One notable example is the 70% yield potential increase from the release of the semi-dwarf rice variety IR8 by the International Rice Research Institute (IRRI) in the 1950s and 1960s (Peng et al., 2008).”

• cites Peng et al., 2008Not checked yet

  Not checked in the free run.

May combine two claims.

These studies provide initial evidence that genotypic adaptation can deliver greater yields in stress-prone environments and likely contribute to future adaptation outcomes.

“These studies provide initial evidence that genotypic adaptation can indeed deliver greater yields in stress-prone environments, thus likely contributing to future adaptation outcomes.”

May combine two claims.

Modern varieties developed through crop improvement have contributed to large increases in crop production over the last 60 years and will continue to do so.

“Modern varieties developed through years of crop improvement have contributed to large increases in crop production in the last 60 years, and they will continue to do so.”

• cites Evenson & Gollin, 2003Not checked yet

  Not checked in the free run.

May combine two claims.

Accurately predicting genotype × environment responses allows identifying well-adapted genotypes for specific sites or stress situations and testing virtual genotypes to inform breeding programs.

“Accurately predicting G × E responses allows identifying well-adapted genotypes for specific sites or stress situations (Banterng et al., 2006;Hammer & Broad, 2003), or testing 'virtual' genotypes to inform breeding programs (Bogard et al., 2020;Cock et al., 1979;Hammer et al., 2020;Suriharn et al., 2011).”

• cites Banterng et al., 2006Not checked yet

  Not checked in the free run.

• cites Hammer & Broad, 2003Not checked yet

  Not checked in the free run.

• cites Bogard et al., 2020 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

On-farm maize trials in Africa showed drought-tolerant maize can yield 20-25% more than current commercial varieties, with no yield penalty in good years.

“On-farm maize trials in Africa have shown that drought-tolerant maize can yield between 20-25% more than current commercial varieties, with no yield penalty in 'good' years (Cairns & Prasanna, 2018;Setimela et al., 2017).”

• cites Cairns & Prasanna, 2018Not checked yet

  Not checked in the free run.

• cites Setimela et al., 2017Not checked yet

  Not checked in the free run.

May combine two claims.

During the early 1980s, progress in process understanding by CGIAR scientists led to development of three crop models that captured G×E×M responses and contributed to other models.

“During the early 1980s, progress in process understanding by CGIAR scientists led to the development of three crop models that ably captured G × E × M responses, while also contributing data and knowledge to many other models.”

May combine two claims.

Anticipatory and predictive tools using crop-climate models can enable preemptive breeding and help enhance and accelerate breeding gains, ensuring crop improvement contributes effectively to climate change challenges.

“Anticipatory and predictive tools using crop-climate models (reviewed in Sect. 2 and 3) can enable preemptive breeding and can help enhance and accelerate breeding gains, ultimately ensuring crop improvement contributes effectively to addressing major challenges for agriculture within the context of climate change.”

May combine two claims.

Accurate prediction of genotype performance across environments and management options depends on a model representing necessary physiological processes at appropriate complexity, a well-constrained parameter set, and high-quality environmental data.

“Accurate prediction of genotype performance across environments and management options is contingent on various elements, including (i) the development of a model with the necessary physiological processes represented at an appropriate level of complexity (Boote et al., 2013;Challinor et al., 2009;Hammer et al., 2019); (ii) the development of a well-constrained parameter set (Angulo et al., 2013;I”

• cites Boote et al., 2013Not checked yet

  Not checked in the free run.

• cites Challinor et al., 2009Not checked yet

  Not checked in the free run.

• cites Hammer et al., 2019 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

CGIAR has contributed to soil and crop geography datasets for crop modeling in recent years.

“Contributions to soil (Jones & Thornton, 2015;Piikki et al., 2017;Vågen et al., 2016) and crop geography (IFPRI, 2019;You et al., 2009You et al., , 2017) ) for crop modeling have also been made in recent years.”

• cites Jones & Thornton, 2015Not checked yet

  Not checked in the free run.

• cites Piikki et al., 2017Not checked yet

  Not checked in the free run.

• cites Vågen et al., 2016Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

ORYZA1 rice model developed at IRRI incorporated many years of eco-physiological research and evolved into ORYZA2000 and ORYZAv3.

“The ORYZA1 Crop Science rice model (Kropff et al., 1993a(Kropff et al., , 1994)), developed at the International Rice Research Institute (IRRI), incorporated many years of eco-physiological research from IRRI and elsewhere.ORYZA1 quickly evolved into ORYZA2000 (Bouman et al., 2001), and later into ORYZAv3 (Li et al., 2017).”

• cites Kropff et al., 1993aNot checked yet

  Not checked in the free run.

• cites Bouman et al., 2001Not checked yet

  Not checked in the free run.

• cites Li et al., 2017Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

CIAT led early development of BEANGRO, now part of CROPGRO module within DSSAT.

“Lastly, CIAT also led the early development of the BEANGRO model, which is currently part of the 'CROPGRO' module within DSSAT (Decision Support System for Agrotechnology Transfer) (Hoogenboom et al., 1993;White & Izquierdo, 1991;White et al., 1995).”

• cites Hoogenboom et al., 1993 sourceNot checked yet

  Not checked in the free run.

• cites White & Izquierdo, 1991Not checked yet

  Not checked in the free run.

• cites White et al., 1995Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Virtually all CGIAR centers have done modeling for their mandate crops.

“Virtually all centers have done modeling for their mandate crops.”

May combine two claims.

ICARDA and ICRISAT contributed to development and improvement of SSM, CROPGRO for peanut and chickpea, and CERES for sorghum and pearl millet in DSSAT.

“The International Center for Agricultural Research in the Dry Areas (ICARDA) and the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) have extensively contributed to the development and improvement of the Simple Simulation Model (SSM) (Sinclair et al., 2020;Soltani & Sinclair, 2012), the CROP-GRO for peanut and chickpea, and the CERES model for sorghum and pearl millet,”

• cites Sinclair et al., 2020 sourceNot checked yet

  Not checked in the free run.

• cites Soltani & Sinclair, 2012Not checked yet

  Not checked in the free run.

May combine two claims.

Attempts to estimate crop model parameters from genetic information date to 1990s CIAT work on common bean, where gene-based parameter estimates were generally as accurate as manually estimated parameters.

“Attempts to estimate model parameters from genetic information date to work in the 1990s at CIAT on common bean, where simulations from gene-based estimates of model parameters were generally as accurate as manually estimated parameters (Hoogenboom et al., 1997;White & Hoogenboom, 1996).”

• cites Hoogenboom et al., 1997Not checked yet

  Not checked in the free run.

• cites White & Hoogenboom, 1996Not checked yet

  Not checked in the free run.

May combine two claims.

CGIAR contributed to development of key spatially explicit climate datasets used as crop model inputs, including WorldClim, CCAFS-Climate database, and methods to generate daily weather data.

“The CGIAR has also contributed to the development of key spatially explicit climate datasets that are used as inputs into crop models.These include WorldClim (Hijmans et al., 2005) and the Climate Change, Agriculture and Food Security (CCAFS)-Climate database (Navarro-Racines et al., 2020), as well as methods to generate daily weather data for crop model simulations (Jones & Thornton, 2000, 2013).”

• cites Hijmans et al., 2005Not checked yet

  Not checked in the free run.

• cites Navarro-Racines et al., 2020Not checked yet

  Not checked in the free run.

• cites Jones & Thornton, 2000, 2013Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

CGIAR modelers have studied G×E×M interactions extensively in close coordination with breeding programs.

“Using available data and models, CGIAR modelers have studied G × E × M interactions extensively in close coordination with breeding programs.”

May combine two claims.

Li et al. (2015) used IRRI trial data from Los Baños and Ludhiana to calibrate and evaluate 13 rice crop models, and Fleisher et al. (2017) used a CIP Bolivia experiment for nine potato crop models.

“Li et al. (2015)used IRRI trial data from Los Baños (Philippines) and Ludhiana (India) to calibrate and evaluate 13 different rice crop models, and Fleisher et al. (2017) used an experiment from Bolivia conducted by the International Potato Center (CIP) as part of the calibration and evaluation dataset for nine potato crop models.”

• cites Li et al. (2015)Not checked yet

  Not checked in the free run.

• cites Fleisher et al. (2017)Not checked yet

  Not checked in the free run.

May combine two claims.

A recent on-farm wheat trial study used a univariate model to assess combined effects of wheat lines and their interactions with farmer-irrigation-year combinations for several traits, identifying stable wheat lines and environments causing most G×E interaction for grain yield.

“A recent study with onfarm wheat trials (Vargas-Hernández et al., 2018) used a univariate model to assess the combined effects of the wheat lines and their interactions with the farmer-irrigation-year combinations for several traits.For single trait grain yield, the study identified stable wheat lines across all environments, as well as the environments that caused most of the G × E interaction.”

• cites Vargas-Hernández et al., 2018Not checked yet

  Not checked in the free run.

May combine two claims.

Multivariate models have been implemented using Bayesian analysis and deep machine learning regression.

“Multivariate models have been implemented using Bayesian analysis (Montesinos-López et al., 2016b) as well as deep machine learning regression (Montesinos-López et al., 2018a,2018c).”

• cites Montesinos-López et al., 2016bNot checked yet

  Not checked in the free run.

• cites Montesinos-López et al., 2018a,2018cNot checked yet

  Not checked in the free run.

May combine two claims.

TPE allows prioritizing stress types by intensity and frequency across geographies and identifying sites where selection for given stresses is likely more successful.

“This, in turn, allows prioritizing stress types by their intensity and frequency across geographies, as well as identifying sites where selection for given stresses is likely to be more successful.”

May combine two claims.

The TPE approach has been used successfully by wheat breeding programs in Australia and more recently applied to maize in Europe.

“The TPE approach has been used with success by wheat breeding programs in Australia (Chenu et al., 2011(Chenu et al., , 2013;;Lobell et al., 2015), and more recently has been applied to maize in Europe (Harrison et al., 2014).”

• cites Chenu et al., 2011Not checked yet

  Not checked in the free run.

• cites Lobell et al., 2015Not checked yet

  Not checked in the free run.

• cites Harrison et al., 2014Not checked yet

  Not checked in the free run.

May combine two claims.

At CIMMYT, grid-based global-scale simulations are used to assess the value of certain traits, developed in a consortium of UF, CIMMYT, and IFPRI incorporating CERES-wheat, CROPSIM, and NWheat.

“At CIMMYT, grid-based global-scale simulations are used to assess the value of certain traits.This modeling capacity was developed in a consortium of UF, CIMMYT, and the International Food Policy Research Institute (IFPRI) that incorporated three crop models, including CERES-wheat, CROPSIM, and NWheat (Gbegbelegbe et al., 2017;Hernandez-Ochoa et al., 2018).”

• cites Gbegbelegbe et al., 2017Not checked yet

  Not checked in the free run.

• cites Hernandez-Ochoa et al., 2018Not checked yet

  Not checked in the free run.

May combine two claims.

At IRRI, simulations have been used to simulate potential yield across environments, identify ideotypes for increasing rice yield potential, and inform national certification processes for variety release.

“At IRRI, simulations have also been used to simulate potential yield across environments (Kropff et al., 1993b), identify ideotypes for increasing rice yield potential (Aggarwal et al., 1997;Dingkuhn et al., 2015Dingkuhn et al., , 2016;;Kropff et al., 1995), and to inform national certification processes for the release of crop varieties (Li et al., 2016).”

• cites Kropff et al., 1993bNot checked yet

  Not checked in the free run.

• cites Aggarwal et al., 1997Not checked yet

  Not checked in the free run.

• cites Li et al., 2016Not checked yet

  Not checked in the free run.

May combine two claims.

CGIAR's work on TPEs for breeding programs covers fewer crops and published studies than the applications described in Sect. 2.1.

“Compared with the applications described in Sect.2.1, CGIAR's work on TPEs for breeding programs is less in terms of number of crops covered and published studies (Figure 2a, b).”

May combine two claims.

At CIAT, crop models have been used to understand drought responses across genotype and management for rice and beans and to assess the value of drought tolerance traits.

“At CIAT, crop models have been used to understand drought responses across G and M for rice and beans (Heinemann et al., 2016;Ramirez-Villegas et al., 2018), as well as to assess the value of drought tolerance traits (Heinemann et al., 2019).”

• cites Heinemann et al., 2016Not checked yet

  Not checked in the free run.

• cites Ramirez-Villegas et al., 2018Not checked yet

  Not checked in the free run.

• cites Heinemann et al., 2019Not checked yet

  Not checked in the free run.

May combine two claims.

CIP used SOLANUM and LINTUL models to compare native and hybrid potato genotypes under extreme climatic conditions and climate change, demonstrating that appropriate germplasm and management choices could secure and increase potato production under future climate.

“CIP has used the SOLANUM (Condori et al., 2010) and LINTUL (Spitters, 1988(Spitters, , 1990) ) models to compare the performance of native and hybrid potato genotypes under extreme climatic conditions (Condori et al., 2010(Condori et al., , 2014) ) and climate change (Quiroz et al., 2018), demonstrating that appropriate choice of germplasm and crop management practices could significantly secure”

• cites Condori et al., 2010Not checked yet

  Not checked in the free run.

• cites Quiroz et al., 2018 sourceNot checked yet

  Not checked in the free run.

• cites Spitters, 1988 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

At ICRISAT, crop models are used to investigate whether changes in genotype and management produce positive yield change across environments, to prioritize breeding and agronomic decisions including sowing density, irrigation, genotype-irrigation combinations, and traits.

“At ICRISAT, crop models are used to investigate whether and how changes in G and M result in positive change in yield across different environments, as a way to prioritize breeding and agronomic intervention decisions, including sowing density (Vadez et al., 2017), irrigation (Vadez et al., 2012), the combination of better-adapted genotypes and irrigation (Soltani et al., 2016), and different”

• cites Vadez et al., 2017Not checked yet

  Not checked in the free run.

• cites Vadez et al., 2012Not checked yet

  Not checked in the free run.

• cites Soltani et al., 2016 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Such similarity could also suggest that models may fail to capture cropping system dynamics realistically when subjected to these parameter modifications.

“Such similarity could also suggest that the models may fail to capture cropping system dynamics realistically when subjected to these parameter modifications.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Similar TPE analysis has been done for chickpea in India.

“Similar TPE analysis has also been done for chickpea in India (Hajjarpoor et al., 2018).”

• cites Hajjarpoor et al., 2018Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Although less mainstream in CGIAR, TPE and Mega-Environment work has potential to help better-target breeding programs across scales.

“Though less mainstream in the CGIAR in terms of centers and traits, TPE and Mega-Environment work has the potential to help bettertarget breeding programs across scales (see Figure 2c).”

May combine two claims.

Significant breeding progress resulted from these studies, including improvements in drought phenotyping in a drought-prone environment that controls timing, intensity, and duration of drought, reducing uncertainty from climate variability trials in the main season.

“Significant breeding progress has resulted from these studies, including improvements in drought phenotyping in a drought-prone environment which allows controlling the timing, intensity, and duration of drought, reducing the uncertainty associated with climate variability trials in the main season (Martinez et al., 2014).”

• cites Martinez et al., 2014Not checked yet

  Not checked in the free run.

May combine two claims.

All studies in Table 1 use systematic parameter modifications to create ideotypes, suggesting opportunities to explore optimization methods and more direct gene-to-phenotype modeling.

“Foremost, we note that all studies use systematic parameter modifications to create ideotypes, which suggests opportunities to explore optimization methods as well as more direct gene-tophenotype modeling (e.g.van Eeuwijk et al., 2019).”

• cites van Eeuwijk et al., 2019Not checked yet

  Not checked in the free run.

May combine two claims.

The accuracy of forward predictions for grain yield using a previous nursery/year to predict the next is low and highly influenced by environment.

“The accuracy of forward predictions for grain yield (using a previous nursery/year to predict the next nursery/year) is, however, low and highly influenced by the environment (Juliana et al., 2018), thereby highlighting the importance of incorporating environmental data in genomic prediction models for grain yield (Bhandari et al., 2019;van Eeuwijk et al., 2019).”

• cites Juliana et al., 2018Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

This process requires delivering information on available genetic diversity, heritability, and high-throughput phenotyping methods for the trait.

“Most notably, it requires delivering information on (i) the available genetic diversity, (ii) heritability, and (iii) high-throughput phenotyping methods for the trait in question.”

May combine two claims.

Ideotypes can be developed for current and future climates via iterative testing of model parameter changes, optimization to maximize mean yield and minimize variability, or gene/trait-specific components in crop models.

“Ideotypes can be developed for current as well as for future climates via a variety of methods ranging from iterative testing changes in model parameters (Dingkuhn et al., 2015;Suriharn et al., 2011), optimization to maximize mean yield and minimize yield variability (Hammer et al., 2020;Semenov & Stratonovitch, 2013), or by developing gene-or trait × gene-specific components into the crop models”

• cites Dingkuhn et al., 2015Not checked yet

  Not checked in the free run.

• cites Suriharn et al., 2011Not checked yet

  Not checked in the free run.

• cites Hammer et al., 2020 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Gene-based crop simulation models can be achieved through a highly dynamic approach at the expense of increased uncertainty, or by predicting crop state variables at coarser temporal scales at the expense of mechanistic detail.

“This can be achieved through a highly dynamic approach (Hwang et al., 2017), at the expense of increasing uncertainty, or through prediction of crop state variables at coarser time scales, at the expense of mechanistic detail.”

• cites Hwang et al., 2017Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The Table 1 list covers five crops across five countries and offers valuable insights into methods and potential value of these analyses for breeding programs.

“The list, while not extensive in terms of crops or countries (i.e.covers five crops across five countries), offers valuable insights as to the methods used and the potential value of these analyses for breeding programs.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

Taking into account the TPE definition as part of breeding product definition will allow modelers to impact breeders and allow breeders to discuss model results from the start of the breeding process.

“Yet, taking into account the TPE definition as part of the definition of the breeding products will allow modelers to impact breeders, while allowing breeders to discuss model results from the start of the breeding process.”

May combine two claims.

Despite early findings and GS revolutionizing animal breeding by increasing selection accuracy and reducing cycle time and cost, implementation in CGIAR crop breeding programs is still limited, focusing primarily on major cereals, partly due to costs and relatively low prediction accuracy due to G×E.

“However, in spite of these early findings and the fact that GS has revolutionized animal breeding by increasing the accuracy of selections and reducing cycle time and cost (Hayes et al., 2013;Hickey et al., 2017), its implementation in CGIAR crop breeding programs is still limited (focusing primarily on the major cereals), in part due to costs associated with routine evaluation and relatively low”

• cites Hayes et al., 2013Not checked yet

  Not checked in the free run.

• cites Hickey et al., 2017Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

Moving from prescribed changes in model parameter values to a range of phenotypic screens feasible in breeding trials is not trivial.

“Finally, we note that moving from a set of prescribed changes in model parameter values (as reported in the studies listed in Table 1) to a range of phenotypic screens that can be feasibly measured and selected for in breeding trials is not a trivial process.”

Probably not a substantial claim: may not pass S2 (checkable).

Similarity in proposed ideotypes across different studies suggests need for refinement in traits assessed through discussion with crop improvement teams or use of more detailed eco-physiological models.

“Additionally, the similarity in the ideotypes proposed for different studies (e.g.chickpea, sorghum and groundnut) suggests the need for refinement in the traits assessed through discussion with crop improvement teams, or through the use of more detailed eco-physiological models (Dingkuhn et al., 2016;Rebolledo et al., 2015).”

• cites Dingkuhn et al., 2016Not checked yet

  Not checked in the free run.

• cites Rebolledo et al., 2015Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Mega-Environments have been used by CIMMYT to target breeding since the 1990s.

“A related larger-scale method, referred to as Mega-Environments, has been used by CIMMYT to target breeding since the 1990s (Cairns et al., 2013;Rajaram et al., 1994;van Ginkel et al., 2002).”

• cites Cairns et al., 2013Not checked yet

  Not checked in the free run.

• cites Rajaram et al., 1994Not checked yet

  Not checked in the free run.

• cites van Ginkel et al., 2002Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Studies comparing predictive abilities of some of these approaches have been conducted.

“Notably, studies comparing the predictive abilities of some of these approaches have also been conducted (Juliana et al., 2017b;Montesinos-López et al., 2018a;Pérez-Rodríguez et al., 2012).”

• cites Juliana et al., 2017bNot checked yet

  Not checked in the free run.

• cites Montesinos-López et al., 2018aNot checked yet

  Not checked in the free run.

• cites Pérez-Rodríguez et al., 2012Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

For rainfed beans, EMBRAPA initiated a drought tolerance breeding program following results of Heinemann et al. (2016, 2017), though implementation was halted due to the Brazilian economic crisis.

“For rainfed beans, EMBRAPA (Empresa Brasileira de Pesquisa Agropecuaria) initiated a drought tolerance breeding program following the results of Heinemann et al. (2016Heinemann et al. ( , 2017)), though its implementation was halted due to the Brazilian economic crisis.”

• cites Heinemann et al., 2016Not checked yet

  Not checked in the free run.

• cites Heinemann et al., 2017Not checked yet

  Not checked in the free run.

May combine two claims.

Process-based crop models have substantial shortcomings in the processes included, the accuracy and precision of their simulation, and the actual meaning of their parameters.

“Furthermore, as summarized by Rötter et al. (2011) and Challinor et al. (2018), major limitations exist in process-based crop models regarding the processes they consider, the accuracy and precision with which they do so, and the true significance of their parameters.”

• cites Rötter et al. (2011)Not checked yet

  Not checked in the free run.

• cites Challinor et al. (2018)Not checked yet

  Not checked in the free run.

May combine two claims.

QuLine can simulate selection of inbred lines, covering most major food cereals and leguminous crops.

“For instance, QuLine, can simulate the selection of inbred lines, which means most major food cereals in the world, plus basically all leguminous crops (Wang & Pfeiffer, 2007;Wang et al., 2003Wang et al., , 2005)).”

• cites Wang & Pfeiffer, 2007Not checked yet

  Not checked in the free run.

• cites Wang et al., 2003Not checked yet

  Not checked in the free run.

• cites Wang et al., 2005Not checked yet

  Not checked in the free run.

May combine two claims.

Over the past ten years, model enhancements have generally been slow, constrained by data, incremental, and concentrated on few crops.

“In the last decade, most model improvements have been relatively slow (compared to the rate of knowledge generation), limited by data availability, typically incremental (i.e.without thinking out of the box), and focused on a small range of crops (Challinor et al., 2014;Maiorano et al., 2017).”

• cites Challinor et al., 2014Not checked yet

  Not checked in the free run.

• cites Maiorano et al., 2017Not checked yet

  Not checked in the free run.

May combine two claims.

Approaches that dynamically connect genetic data to crop simulation models have succeeded for phenological development variables and for more complex traits.

“Dynamic approaches that link genetic information with crop simulation models have proved successful for crop development variables (i.e.phenology) (White & Hoogenboom, 1996;White et al., 2008;Yin et al., 2004) as well as for more complex traits (Bertin et al., 2010;Chenu et al., 2009).”

• cites White & Hoogenboom, 1996Not checked yet

  Not checked in the free run.

• cites White et al., 2008Not checked yet

  Not checked in the free run.

• cites Yin et al., 2004Not checked yet

  Not checked in the free run.

May combine two claims.

QuLine has served to compare and improve conventional selection strategies, to forecast cross performance from known gene information, and to optimize marker-assisted selection for pyramiding multiple genes.

“QuLine has been used to compare and optimize conventional selection strategies (Li et al., 2013b;Wang et al., 2003Wang et al., , 2009)), to predict cross performance using known gene information (Wang et al., 2005), and optimize marker assisted selection to pyramid multiple genes (Wang et al., 2007).”

• cites Li et al., 2013bNot checked yet

  Not checked in the free run.

• cites Wang et al., 2003Not checked yet

  Not checked in the free run.

• cites Wang et al., 2009Not checked yet

  Not checked in the free run.

May combine two claims.

Crop modeling can aid breeding in numerous ways, such as evaluating genotype adaptability and stability, defining target breeding environments, finding traits that boost yield, and forecasting genotype breeding values.

“Crop modeling can support breeding efforts in many ways, including assessing genotypic adaptability and stability, characterizing and identifying target breeding environments, identifying traits and/or eco-physiological characteristics that maximize yield for such environments, and making predictions about the breeding value of the genotypes.”

Probably not a substantial claim: may not pass S2 (checkable).

Substantial advances are still needed for crop modeling to meaningfully support the faster breeding rates necessary for climate change adaptation.

“However, much progress remains to be made if crop modeling is to effectively contribute to the accelerated breeding rates required to adapt to climate change (see Sect. 1.2).”