US Corn Research Articles

Structural complexity biases vegetation greenness measures.

Vegetation 'greenness' characterized by spectral vegetation indices (VIs) is an integrative measure of vegetation leaf abundance, biochemical properties and pigment composition. Surprisingly, satellite observations reveal that several major VIs over the US Corn Belt are higher than those over the Amazon rainforest, despite the forests having a greater leaf area. This contradicting pattern underscores the pressing need to understand the underlying drivers and their impacts to prevent misinterpretations. Here we show that macroscale shadows cast by complex forest structures result in lower greenness measures compared with those cast by structurally simple and homogeneous crops. The shadow-induced contradictory pattern of VIs is inevitable because most Earth-observing satellites do not view the Earth in the solar direction and thus view shadows due to the sun-sensor geometry. The shadow impacts have important implications for the interpretation of VIs and solar-induced chlorophyll fluorescence as measures of global vegetation changes. For instance, a land-conversion process from forests to crops over the Amazon shows notable increases in VIs despite a decrease in leaf area. Our findings highlight the importance of considering shadow impacts to accurately interpret remotely sensed VIs and solar-induced chlorophyll fluorescence for assessing global vegetation and its changes.

Harvest index has increased over the last 50 years of maize breeding

Quantifying historical changes in maize harvest index (HI), the fraction of above-ground biomass allocated to grain yield, can enhance our ability to explain grain yield trends and estimate stover carbon inputs for sustainability assessments. However, the HI genetic gain has not been the primary focus of previous era studies. The aim of this study is to enhance our knowledge of maize HI genetic gain. Our first objective is to quantify HI genetic gain in Bayer Crop Science Legacy hybrids and investigate the contribution of breeding and agronomic management. Our second objective is to develop a general-use model to describe the temporal evolution of maize HI. We studied 54 commercial hybrids (103-day and 111-day relative maturities) released from 1983 to 2020 across 13 environments, including plant density (current and historical increasing rate) and N-fertilizer (low and sufficient N rates) treatments. The HI was estimated at physiological maturity by destructively sampling plants. Then we synthesize new experimental data with literature findings (n = 16) to provide a robust HI genetic gain estimate. Results showed that HI has increased over the years from 0.516 to 0.571 in 103-day hybrids and from 0.537 to 0.584 in 111-day hybrids. The genetic gains were similar across environments and management treatments within the studied range, indicating that this increase is attributed to maize breeding. The N-fertilizer treatments affected the magnitude of the HI, but plant density did not. Our results, combined with 16 literature datasets, revealed a 0.26% year−1 relative increase in HI since 1964. We estimated that the increase in HI accounts for ca. 15% of the historical maize yield increase in the US Corn Belt over the past 50 years. The maize HI has increased over the last 50 years, and this increase was attributed to breeding, not to management. Our findings enhance our knowledge of maize HI, will support robust estimations of carbon inputs in sustainability studies, and inform crop models to better capture historical yield increases.

Heat stress to jeopardize crop production in the US Corn Belt based on downscaled CMIP5 projections

CONTEXTGlobal food security faces increasing challenges from the changing climate. Changes of agricultural output from some of the most productive regions such as the US Corn Belt can largely affect the world's food market. Developing predictive understanding of the agricultural risk of climate change and potential mitigation strategies is critical for the global food security. OBJECTIVEThe objective of this study is to assess the responses of maize and soybean yield to projected climate changes in the Corn Belt, identify the shifting environment stressors on crop yield, and tackle potential climate adaptation strategies. METHODSWe drive a process-based model, the Decision Support System for Agrotechnology Transfer, with high-resolution statistically downscaled and bias-corrected historical and future climates from ten CMIP5 models in the MACA-2 database. RESULTS AND CONCLUSIONSThe multi-model ensemble mean suggests a 12% decrease of maize yield by mid-century and 40% by late century, with a high degree of model consensus in the direction of changes; for individual models, the projected decrease of maize yield by late century ranges from <5% to over 80%, with the worst crop outcome corresponding to the most sensitive climate models. Soybean yield is projected to increase by midcentury with a high degree of model consensus, but such consensus is lost by late century as some projections shift to significant decreases. Crop yield in the Corn Belt is currently limited by water stress, but is projected to be increasingly limited by heat stress as well after the midcentury. The mounting heat stress will drive the most productive zone for maize to shift from central to northern part of the Corn Belt, but the projected increase in the northern states cannot fully compensate for the decrease in the south, causing the total production to decrease if agricultural practice stays the same. Earlier planting can alleviate only a small fraction of the heat-induced crop loss in a warmer climate. Climate change will (at least partially) offset the yield boost caused by agricultural technology and intensification. SIGNIFICANCEThis study advances our predictive understanding of crop yield responses to climate change, and suggests that a multitude of strategies will be needed to address the climate change challenges for the U.S. agriculture.

Improving generalisability and transferability of machine-learning-based maize yield prediction model through domain adaptation

Modern problems in agricultural management require non-traditional solutions, one of which is by utilizing domain adaptive machine learning models for crop yield prediction which are able to perform reliably in different temporal or spatial domains. However, most studies have focused on the application of domain adaptation to classification tasks such as crop type identification, while the application to regression tasks such as crop yield prediction have been limited. In this study, we explore the generalisability and transferability of ordinary Deep Neural Network (DNN) and domain adaptive neural network models created using three domain adaptation algorithms, namely Discriminative Adversarial Neural Network (DANN), Kullback-Leibler Importance Estimation Procedure (KLIEP), and Regular Transfer Neural Network (RTNN). These three algorithms represent feature-based, instance-based, and parameter-based domain adaptations, respectively. Maize yield records, weather variables, and remotely sensed features from 11 states in the US corn belt acquired in 2006–2020 were compiled and segregated into classes according to temporal (year) and spatial characteristics (annual growing degree days [GDD], vapor pressure deficit [VPD], soil organic content [SOC], and green chlorophyll vegetation index/GCI). We found that models trained using datasets from temperate regions with medium-high GDD and moderate VPD perform well whereas SOC does not significantly affect the generalisability. It is not advisable to train models with datasets constrained by GCI as this feature correlates significantly with the maize yield, and adaptation between two domains that rarely intercept will not work well. We also demonstrate that Kullback-Leibler divergence computed using features from source and target domains can be used to justify the feasibility of domain adaptation. Based on the divergence, a model trained in the US (or another region with sufficient data) is expected to work reliably in other regions through domain adaptation, especially feature-based adaptation.

Economic and risk analysis of sustainable practice adoption among U.S. corn growers

CONTEXTAgricultural sustainability has three main pillars: a healthy environment, economic profitability, and social equity. Most programs, however, focus on environmental benefits when establishing protocols for farms, de-emphasizing or ignoring the economic and social aspects. This focus on the environmental aspects of agricultural sustainability misses the importance of economic factors that are commonly found to be key determinants of farmer adoption of best management practices. More research on the economic effects of sustainability practices on farm income and risk would improve understanding and facilitate communication with farmers about the tradeoffs and risks when using the protocols. OBJECTIVESOur study has two primary objectives. First, we quantify the effects of sustainable practice adoption on the mean, variance, and skewness of yield for U.S. corn farmers. Second, based on the estimated yield risk model, we quantify the effects of sustainable practice adoption on farmer returns and the cost of risk to better understand the impacts of sustainability on farmer welfare. METHODSUsing the 2010 Agricultural Resource Management Survey, we construct a composite indicator proxying the intensity of sustainable practice adoption for 73 practices mostly intended to reduce soil erosion and improve nutrient and pest management. We use this indicator in a flexible moment-based approach to analyze field-level corn yield data. Estimation results are used to quantify expected farmer returns and the cost of risk (the risk premium). The cost of risk is decomposed into costs from variance (symmetric variation around the mean) and from skewness (downside risk from unusually low yields). RESULTS AND CONCLUSIONSResults indicate that substantial opportunities exist for US corn growers to increase sustainable practice adoption. Increased sustainable practice adoption significantly increased mean yield and decreased yield variance and skewness. In addition, increased adoption of sustainable practices increased expected farmer returns and, in most cases, also increased the cost of risk mostly from increased costs from lower skewness (greater downside risk). The positive effect of increased expected income always dominated the negative effect of increased costs of risk, so that on average farmer welfare increased with higher levels of sustainable practice adoption. SIGNIFICANCEOur study helps fill a literature gap by evaluating the economic aspects of sustainability, specifically farm income and the cost of risk. The results provide critical information for policy making and program establishing.

Cover crops and soil health in rainfed and irrigated corn: What did we learn after 8 years?

AbstractDuration of cover crop (CC) management, CC biomass production, and other factors could impact how CC affects soil health. We studied the 8‐year cumulative impacts of winter rye (Secale cereale L.) CC on soil physical, chemical, and biological properties in rainfed and irrigated no‐till corn (Zea mays L.)‐based systems in the western US Corn Belt. Average annual CC biomass production was 0.56 ± 0.51 Mg ha−1 at the rainfed site and 0.98 ± 0.95 Mg ha−1 at the irrigated site. After 8 years, CC improved particulate organic matter (POM) and mean weight diameter of water‐stable aggregates (MWD) compared with no CC in the 0–5 cm soil depth at both sites. Cover crop increased total POM concentration by 2.8 mg g−1 at the rainfed site and by 13.4 mg g−1 at the irrigated site, while it increased MWD by 0.39 mm at the rainfed site and by 0.79 mm at the irrigated site. Also, CC increased soil C at a rate of 0.125 Mg ha−1 year−1 in the 0–5 cm depth but only at the rainfed site. Cover crop affected neither water infiltration nor available water but improved microbial biomass. Changes in other properties were site‐dependent. Cover crop improved many soil properties after 8 years even though measurement taken after 4 years showed no significant effect of CC, which indicates CC slowly impacts properties in this environment. Low CC biomass production and high biomass input from corn‐based systems may explain the slow soil response. In general, winter rye CC enhances near‐surface soil properties in the long term.

Accelerated leaf appearance and flowering in maize after four decades of commercial breeding

AbstractMaize breeding programs have indirectly altered many plant traits; however, our knowledge of some important phenological traits remains unexplored. One such trait is leaf appearance rate, which is crucial for predicting maize development. We studied 40 short‐season (103‐day) and 38 long‐season (111‐day) hybrids released from 1980 to 2020 by Bayer Crop Science. Measurements included weekly counting of collared leaves across 13 experiments in the US Corn Belt. The progression of leaf number was expressed as a function of thermal time and described with a trilinear model. Results indicated that new 111‐day hybrids produce leaves faster than old hybrids throughout the vegetative phase (7.4% and 3.1% faster before and after the ninth leaf stage, respectively), whereas new 103‐day hybrids produce leaves faster only after the ninth leaf stage (9.4%). Thermal time to silking and anthesis decreased by about 1 and 0.56°C day year−1, respectively. Our data revealed that silking and anthesis can precede the final collared leaf by 96°C day (3.3 days under optimal conditions), which indicates an overlap between vegetative and reproductive phases. We concluded that maize breeding has indirectly altered the rate of vegetative development of maize hybrids without affecting the final leaf number. Present results expand our knowledge base on the genotypic variability in maize development traits, which can improve empirical and process‐based models used for crop stage and yield prediction.

Characterization and Comparison of Wet Milling Fractions of Export Commodity Corn Originating from Different International Geographical Locations

AbstractSoft endosperm corn has evident characteristics suitable for wet milling but is susceptible to breakage and fracture during transportation. In this study the millability of commodity corn samples with different endosperm hardness originating from different international locations is compared, and its economic impact on corn importers for wet milling is discussed. The broken corn and foreign material (BCFM) for the soft endosperm US commodity corn ranges from 1.13% to 5.57% which is higher than other commodity corn from other international locations exported to the same country. US corn exported to different international markets shows higher starch yield in a range of 5–8% compared to the corn from different geographical locations exported to the same market. The excess starch from US corn directly translates to an additional revenue of 8.13–13 million USD per year for a 2540 MT day−1 wet mill plant. Therefore, it can be concluded that the soft endosperm US commodity corn, despite higher breakage, has superior millability and gives higher starch yields compared to the hard endosperm corn from other international locations which has comparatively lower BCFM. There are technologies that can be used to process broken corn separately after the initial cleaning process.

Field scale SWAT+ modeling of corn and soybean yields for the contiguous United States: National Agroecosystem Model Development

CONTEXTDespite a steady increase in staple crop yields over the past ten years, current agricultural production must escalate even more to keep pace with the expected world population growth, which in turn will require improved agricultural methods that are adapted to many environmental pressures. Comprehensive models that can simulate crop production systems and the impact of management and conservation practices on natural resources and the environment, including water quality at large scale present important contributions to this challenge. OBJECTIVETo this end we developed the National Agroecosystem Model (NAM): a comprehensive model that uses the updated Soil and Water Assessment Tool (SWAT+) to accurately simulate staple crop yields across the contiguous United States (CONUS), with an initial focus on Corn (Zea mays L.) and Soybean (Glycine max L. Merr.) yields. METHODSAvailable open-access data was used to setup this high-resolution modeling system, where every 8-digit hydrologic unit (HUC8) is represented as an individual SWAT+ simulation. A total of 2201 HUC8 simulations across the CONUS were interconnected from upstream to downstream to make the NAM. Field boundary data was used to setup the NAM in such a way that every identified cultivated field is modeled as a unique Hydrologic Response Unit (HRU). Simulated corn and soybean yield from over 2.5 million field-type HRUs were compared to reported average annual corn and soybean yields for the respective area for the 2015–2020 period. RESULTS AND CONCLUSIONSResults show a good agreement between simulated and reported yields (R2 = 0.90 for corn and R2 = 0.70 for soybeans), with a very good model performance in the high corn and soybean production region of the US Corn Belt (Relative Error < ±5%). SIGNIFICANCEApart from assessing the capability of the updated SWAT+ model, we also demonstrate the new crop yield calibration module embedded in SWAT+, highlight changes to the plant growth module, and model parameterization. Results of an analysis of possible crop production differences for corn and soybeans in irrigated, tiled, and non-irrigated-non-tiled fields are also discussed. The versatility of the NAM provides the possibility to analyze information on impacts of changing conservation practices and enables identification of conservation gains and remaining conservation needs at the national scale.

It doesn't bother me: An intersectional analysis of discrimination among white women farmers in the US Corn Belt

While women in United States agriculture are increasingly asserting control over land and assuming identities as primary producers, they continue to face significant challenges in being “read” as legitimate producers and in accessing the material resources (land, labor, capital) to do the work of farming. Although scholarship documents how women are generating new strategies to gain legitimacy as farmers and how programs have emerged to provide agricultural outreach to women, much of this work has been “colorblind” in its lack of a critical analysis of race and its intersections with gender. In this paper, we analyze the complex intersections between white supremacy and patriarchy that may benefit white women farmers as they negotiate and normalize gender-based discrimination. Viewing race and gender as materially lived and negotiated, we analyze in-depth qualitative data from 43 conventional white women producers in the US Corn Belt to identify the nuanced ways they socially navigate the white male-dominated world of production agriculture. We find while respondents universally feel that women claiming the role of farmer are subject to both dismissals and outright aggressions in agricultural encounters, many of the respondents claim this discrimination “doesn't bother” them. We argue, that in the semi-dynastic world of Corn Belt production, these white women socially position themselves by drawing on both their embodied histories of labor along with their relationalities to land and people (particularly men) in ways that facilitate a sense of collective belonging contingent on others' exclusion. We conclude by reflecting on how agricultural research on gender could incorporate more critical analyses of race to uncover continuing forms of discrimination in agricultural spaces.

Can upscaling ground nadir SIF to eddy covariance footprint improve the relationship between SIF and GPP in croplands?

Ground solar-induced chlorophyll fluorescence (SIF) is important for the mechanistic understanding of the dynamics of vegetation gross primary production (GPP) at fine spatiotemporal scales. However, eddy covariance (EC) observations generally cover larger footprint areas than ground SIF observations (a bare fiber with nadir), and this footprint mismatch between nadir SIF and GPP could complicate the canopy SIF-GPP relationships. Here, we upscaled nadir SIF observations to EC footprint and investigated the change in SIF-GPP relationships after the upscaling in cropland. We included 13 site-years data in our study, with seven site-years corn, four site-years soybeans, and two site-years miscanthus, all located in the US Corn Belt. All sites’ crop nadir SIF observations collected from the automated FluoSpec2 system (a hemispheric-nadir system) were upscaled to the GPP footprint-based SIF using vegetation indices (VIs) calculated from high spatiotemporal satellite reflectance data. We found that SIF-GPP relationships were not substantially changed after upscaling nadir SIF to GPP footprint at our crop sites planted with corn, soybean, and miscanthus, with R2 change after the upscaling ranging from -0.007 to 0.051 and root mean square error (RMSE) difference from -0.658 to 0.095 umol m − 2s − 1 relative to original nadir SIF-GPP relationships across all the site-years. The variation of the SIF-GPP relationship within each species across different site-years was similar between the original nadir SIF and upscaled SIF. Different VIs, EC footprint models, and satellite data led to marginal differences in the SIF-GPP relationships when upscaling nadir SIF to EC footprint. Our study provided a methodological framework to correct this spatial mismatch between ground nadir SIF and GPP observations for croplands and potentially for other ecosystems. Our results also demonstrated that the spatial mismatch between ground nadir SIF and GPP might not significantly affect the SIF-GPP relationship in cropland that are largely homogeneous.

Combining corn N recommendation tools for an improved economical optimal nitrogen rate estimation

AbstractImproving corn (Zea mays L.) nitrogen (N) rate fertilizer recommendation tools can improve farmers’ profits and mitigate N pollution. Numerous approaches have been tested to improve these tools, but to date improvements for predicting economically optimum N rate (EONR) have been modest. This work's objective was to use ensemble learning to improve our estimation of EONR (for a single at‐planting and split N application timing) by combining multiple corn N recommendation tools. The evaluation was conducted using 49 corn N response trials from eight states in the US Corn Belt and three growing seasons (2014–2016). Elastic net and decision tree approaches regressed EONR against three unique tools for each N application timing. Tools used in various combinations included a yield goal method, two soil nitrate tests (pre‐plant and late season), a computer simulation crop model (Maize‐N), and canopy reflectance sensing. Any combination of two or three N recommendation tools improved or maintained performance metrics (R2, root‐mean square error , and number of sites close to EONR). The best results for a single at‐planting recommendation occurred when combining the three at‐planting N recommendation tools (including interactions) with an elastic net regression model. This combined recommendation tool had a significant linear relationship with EONR (R2 = 0.46), an increase of 0.27 over the best tool evaluated alone. Combining multiple tools increased the implementation cost, but it did not reduce profitability and, sometimes, improved profitability. These results show tools can be combined to better match EONR, and thus could aid farmers in improving N management.

A data-driven crop model for maize yield prediction

Accurate estimation of crop yield predictions is of great importance for food security under the impact of climate change. We propose a data-driven crop model that combines the knowledge advantage of process-based modeling and the computational advantage of data-driven modeling. The proposed model tracks the daily biomass accumulation process during the maize growing season and uses daily produced biomass to estimate the final grain yield. Computational studies using crop yield, field location, genotype and corresponding environmental data were conducted in the US Corn Belt region from 1981 to 2020. The results suggest that the proposed model can achieve an accurate prediction performance with a 7.16% relative root-mean-square-error of average yield in 2020 and provide scientifically explainable results. The model also demonstrates its ability to detect and separate interactions between genotypic parameters and environmental variables. Additionally, this study demonstrates the potential value of the proposed model in helping farmers achieve higher yields by optimizing seed selection.

Poorly drained depressions can be hotspots of nutrient leaching from agricultural soils.

Much of the US Corn Belt has been drained with subsurface tile to improve crop production, yet poorly drained depressions often still flood intermittently, suppressing crop growth. Impacts of depressions on field-scale nutrient leaching are unclear. Poor drainage might promote denitrification and physicochemical retention of phosphorus (P), but ample availability of water and nutrients might exacerbate nutrient leaching from cropped depressions. We monitored nitrate, ammonium, and reactive P leaching across multiple depression-to-upland transects in north-central Iowa, using resin lysimeters buried and retrieved on an annual basis. Crops included conventional corn/soybean (Zea mays/Glycine max) rotations measured at fields with and without a winter rye (Secale cereale) cover crop, as well as juvenile miscanthus (Miscanthus × giganteus), a perennial grass. Leaching of nitrogen (N) and P was greater in depressions than in uplands for most transects and years. The median difference in nutrient leaching between paired depressions and uplands was 56kg N ha-1 year-1 for nitrate (p=0.0008), 0.6kg N ha-1 year-1 for ammonium (p=0.03), and 2.4kg P ha-1 year-1 for reactive P (p=0.006). Transects managed with a cover crop or miscanthus tended to have a smaller median difference in nitrate (but not ammonium or P) leaching between depressions and uplands. Cropped depressions may be disproportionate sources of N and P to downstream waters despite their generally poor drainage characteristics, and targeted management with cover crops or perennials might partially mitigate these impacts for N, but not necessarily for P.

Field-scale dynamics of planting dates in the US Corn Belt from 2000 to 2020

Crop planting dates are a dynamic feature of agricultural systems that respond to short- and long-term climate signals, crop and cultivar selection, and technology changes. Planting date records are essential for yield gap analyses, accurate crop modeling, and tracking farmer adaptations to weather and climate change. Although planting dates have high variation at local scales due to heterogeneity in farm resources and decision-making, available long-term data on planting dates are largely restricted to aggregated regional statistics or, at best, satellite-derived datasets with limited spatiotemporal extent and at resolutions unable to distinguish individual fields. Here, we generated retrospective annual field-scale (30 m) planting date maps for both maize and soybeans spanning 2000–2020 across a 12-state region in the United States Corn Belt based on Landsat satellite data and a large ground sample of over 28,000 maize and soybean fields. Using training data from 2015 to 2020 for model selection, we found that planting date predictions improved with harmonic regression of Landsat data and additional annual weather covariates. The preferred random forest model approximately doubled performance compared to a null model based on state median planting dates, capturing 47% of field-level variation for maize (mean absolute error, MAE = 7.4 days) and 44% for soybeans (MAE = 7.5 days) against held-out ground truth test data for 2008–2014. We also evaluated the full 2000–2020 dataset with state agricultural statistics, finding strong agreement with median planting dates for maize (R2 = 0.76, MAE = 4.4 days) and slightly lower agreement for soybeans (R2 = 0.65, MAE = 5.4 days) when aggregated to the state level. We then used this new dataset to analyze environmental determinants of planting dates at a finer-scale than previously possible, controlling for unobserved variation at the sub-state district level. We found that during 2000–2020, each standard deviation increase in early-season rainfall delayed planting by ∼2.5 days, and fields with higher soil productivity ratings tended to be planted earlier. We did not find meaningful trends over the last two decades in planting dates for maize or soybeans, in contrast to trends towards earlier planting dates late last century and predicted for this period in climate adaptation studies. We hypothesize increases in early season rainfall or persisting early-season frost risks may have inhibited these shifts towards earlier planting. Remotely sensed planting dates will be a useful tool for yield gap analyses, crop simulation modeling, and ongoing assessment of climate adaptation.

Dissecting the contribution of weather and management on water table dynamics under present and future climate scenarios in the US Corn Belt

Dissecting the contribution of weather and management on water table dynamics under present and future climate scenarios in the US Corn Belt

Leveraging Important Covariate Groups for Corn Yield Prediction

Accurate yield information empowers farmers to adapt, their governments to adopt timely agricultural and food policy interventions, and the markets they supply to prepare for production shifts. Unfortunately, the most representative yield data in the US, provided by the US Department of Agriculture, National Agricultural Statistics Service (USDA-NASS) Surveys, are spatiotemporally patchy and inconsistent. This paper builds a more complete data product by examining the spatiotemporal efficacy of random forests (RF) in predicting county-level yields of corn—the most widely cultivated crop in the US. To meet our objective, we compare RF cross-validated prediction accuracy using several combinations of explanatory variables. We also utilize variable importance measures and partial dependence plots to compare and contextualize how key variables interact with corn yield. Results suggest that RF predicts US corn yields well using a relatively small subset of climate variables along with year and geographical location (RMSE = 17.1 bushels/acre (1.2 tons/hectare)). Of note is the insensitivity of RF prediction accuracy when removing variables traditionally thought to be predictive of yield or variables flagged as important by RF variable importance measures. Understanding what variables are needed to accurately predict corn yields provides a template for applying machine learning approaches to estimate county-level yields for other US crops.

Using a machine learning approach and big data to augment WASDE forecasts: Empirical evidence from US corn yield

AbstractThis paper investigates the accuracy of corn yield forecasts using machine learning with satellite and weather data. In addition, the study examines the incremental value of these forecasts to augment the World Agricultural Supply and Demand Estimates (WASDE) forecast. To illustrate the potential of machine learning methods for agricultural forecasting, publicly available data are collected from 1984 to 2021 for national corn yield, state corn yield, satellite variables, and weather variables and used with the XGBoost algorithm. The results show that the XGBoost model performed about the same but did not outperform the WASDE corn yield forecasts over a 12‐year out‐of‐sample period. The incremental value analysis results suggest that the XGBoost and WASDE forecasts capture similar information, and no incremental information exits. Although the XGBoost model does not outperform the WASDE August forecast, it is near real‐time and can be produced using publicly available data. The results indicate that the XGBoost machine learning models can produce reasonably accurate crop yield forecasts.

Modeling days suitable for fieldwork using machine learning, process-based, and rule-based models

ContextPrediction of days suitable for fieldwork is important for understanding the potential effects of climate change and for selecting machinery systems to improve efficiency in field operations and avoid soil damage. Yet, we lack predictive models to inform decision-making at scale. ObjectiveWe filled this knowledge gap by developing and testing five new workability models. MethodsOne model follows soil moisture-based methods (APSIM), one uses simple rain and temperature thresholds, and three follow machine learning techniques (Random Forest, Decision Table, Neural Network). We parameterized the models using USDA survey data from Iowa, USA and evaluated their temporal and spatial prediction over twelve US Corn Belt States and different time periods using multiple statistical indexes and sensitivity analysis. The models operate at a 5-arcminute resolution. Results and conclusionsResults indicated that the simple rule model, the Decision Table model, and the process-based model predicted field workable days with an agreement index of 0.88, 0.86, and 0.84, respectively for the testing datasets (n = 22,671), and hence were deemed sufficient for future use. The selected models are better suited for large timespan evaluations of workability (monthly to annual, normalized root mean square error, nRMSE = 8 to 15%)) than weekly predictions (nRMSE = 21%). The machine learning models tended to cluster their predictions around a mean value and were about 50% less responsive to precipitation than the process-based or rule-based models. We concluded that simple approaches are more robust to be applied at scale than complex approaches with many data input requirements. SignificanceThe developed models enhance our capacity to predict climate change impacts on workability, a valuable indicator for decision-making and overall sustainability.

Contrasting impacts of dry versus humid heat on US corn and soybean yields

The impact of extreme heat on crop yields is an increasingly pressing issue given anthropogenic climate warming. However, some of the physical mechanisms involved in these impacts remain unclear, impeding adaptation-relevant insight and reliable projections of future climate impacts on crops. Here, using a multiple regression model based on observational data, we show that while extreme dry heat steeply reduced U.S. corn and soy yields, humid heat extremes had insignificant impacts and even boosted yields in some areas, despite having comparably high dry-bulb temperatures as their dry heat counterparts. This result suggests that conflating dry and humid heat extremes may lead to underestimated crop yield sensitivities to extreme dry heat. Rainfall tends to precede humid but not dry heat extremes, suggesting that multivariate weather sequences play a role in these crop responses. Our results provide evidence that extreme heat in recent years primarily affected yields by inducing moisture stress, and that the conflation of humid and dry heat extremes may lead to inaccuracy in projecting crop yield responses to warming and changing humidity.