Location: Plant Physiology and Genetics Research2020 Annual Report
Objective 1: Improve ecophysiological models for quantitative prediction of G x E x M. Sub-objective 1A: Models improved with respect to their ability to simulate phenology and canopy architecture from genetic information. Sub-objective 1B: Models improved with respect to simulating crop responses to energy and water balances. Sub-objective 1C: Models assessed to understand how improved responses to energy and water balances affect cropping system responses. Objective 2: Characterize the temperature response of agronomic crops using an exceptionally wide range of natural air temperatures. Sub-objective 2A: Responses of cereal crops to temperature characterized using an exceptionally wide range of natural air temperatures emphasizing near-lethal high temperatures under well-watered conditions. Sub-objective 2B: Responses of cereal crops to temperature characterized by combining deficit irrigation and an exceptionally wide range of natural air temperatures. Objective 3: Develop tools for proximal sensing and remote sensing for improved quantification of crop growth. Sub-objective 3A: Tools developed for assessing crop architecture and light interception through proximal and remote sensing. Sub-objective 3B: Tools developed for quantifying the conversion of intercepted radiation to biomass via proximal and remote sensing. Sub-objective 3C: Tools developed for assessing partitioning of vegetative and reproductive growth via proximal and remote sensing.
Objective 1: To improve ecophysiological models that quantify crop responses to G x E x M, the project will strengthen simulation of phenology, canopy architecture, and crop energy balances (CEB) and water balances (CWB). Targeting common bean, soybean and sorghum, research on phenology and architecture focuses on improving how genetic differences within a crop species are represented in existing models such as the Cropping Systems Model (CSM). The work exploits large phenotypic datasets from multiple environment trials, linked to data on daily weather conditions, crop management and crop genetics. Improved simulation of crop energy and water balances should benefit overall simulation of cropping systems. To improve calculation of the three-source (sunlit and shaded leaves, soil surface) CEB as implemented in the CSM, planned work will ensure that crop and soil temperatures estimated are correctly transferred to routines for other temperature-sensitive processes and calculation of the CEB is numerically stable. Improved calculation of the CWB builds on comparisons of over 30 maize models (including CSM), which we lead as part of the global Agricultural Model Intercomparison and Improvement Project. This work will identify approaches providing the best estimates of crop water use and indicate how model calibration affects CWB estimates. Objective 2: Through detailed monitoring of crop growth and development, field trials will be used to compare responses of cereal crops to thermal stress at near-lethal and lethal temperatures. This will provide a unique dataset to analyze how temperature affects multiple processes of crop growth and development. Four spring cereals (bread wheat, durum wheat, barley and triticale) will be sown on sequential dates that expose the crops to the exceptionally high mid-day air temperatures. In a second phase, a water deficit treatment will augment the range of temperatures experienced by the four crops as well as allow characterizing how temperature and water deficits interact to affect cereals at near-lethal temperatures. Objective 3: Crop models require high quality data on growth, novel sensor systems will be used to monitor growth at lower cost, higher accuracy and higher throughput than previously possible. This builds on advances in high throughput phenotyping, which is usually associated with genetic research but is applicable to many aspects of crop research. The focus is to analyze data from the Transportation Energy Resources from Renewable Agriculture Phenotyping Reference Platform (TERRA REF) field scanner to quantify growth of sorghum and wheat using a conceptual framework of light interception and radiation use efficiency. The first year, four seasons each of sorghum and durum diversity panels will have been variously scanned with stereo cameras, a thermal camera, a 3-D laser scanner and two hyperspectral cameras (covering 400 to 2,500 nanometers) and imaged with unmanned aerial vehicles. In collaboration with image analysts of TERRA REF, we will quantify crop growth and architecture, cross-validating results with light interception, crop height and biomass data from our manual assessments.
Under Objective 1, loss of personnel resulted in no progress under Sub-objective 1A. Under Sub-objective 1B, energy and water balances model routines were improved: (1) to determine how maize growth models can simulate evapotranspiration (ET), an inter-comparison study was conducted by a retired ARS collaborator, 4 ARS scientists and numerous international cooperators. The project was coordinated through the Agricultural Model Inter-Comparison and Improvement Project (AgMIP) using eddy covariance data from ARS’s Ames, Iowa, location. A “blind” phase provided weather, soils, phenology, and management information to the modelers. Estimates of seasonal ET varied by about ±250 mm from observed 450 mm, and even after receiving all information, ET estimates still varied by a factor of two or more. Another simulation of 40 maize models was initiated using more detailed eddy covariance data from Mead, Nebraska, and lysimeter data from ARS’s Bushland, Texas location. A “blind” phase provided weather, soils, phenology, and management information to the modelers. However, there was a wide range of daily ET estimates among models. Next, growth and ET data will be provided to see how well the models perform when fully calibrated and find avenues for model improvement; (2) a retired ARS collaborator, and scientists from Brazil and Universities of Florida, Washington, and Nebraska have implemented an energy balance routine. This model was presumed theoretically sound with energy and water balances conducted on sunlit leaves, shaded leaves and the soil surface. Model simulations were compared against eddy covariance data on soybean from Mead, Nebraska. Results were mixed, so several improvements were made to the code, and now it out-performs the FAO-56 method for estimating ET; (3) the ARS collaborator, and other ARS and University of Arizona scientists have assembled data observed during the 1998 and 1999 free-air CO2 enrichment (FACE) sorghum experiments at Maricopa, Arizona. Once published these data will be available to sorghum modelers to validate their models. Under Sub-objective 1C, progress was made indirectly via progress on 1B. For Sub-objective 2A, a novel and cost-effective means to elucidate thermal response for 4 cool-season cereal crops (bread wheat, durum wheat, barley, triticale) simultaneously was employed to utilize intra- and inter-annual variations in ambient air temperature in a semiarid desert region such as Maricopa, Arizona – especially, to emphasize near-lethal to lethal growth temperatures. In the Thermal Regime Agronomic Cereal experiment (TRACE) we determine [Genotype (G) by Environment (E) by Management (M) (GxExM)] interactions by staggering sowing dates, from the normal cropping season in November-December, to be in closer intervals during April-June. Use of day-neutral cultivars negated any vernalization requirement or photoperiod effect. These 4 cereal grain crops were planted in a 12-ha field in 4 replicates, over 5, 8, and 8 (total of 21) sowing dates (November to June) during 2016, 2017, and 2019, respectively, under a well-watered flood irrigation (100% ET) management regime. Across a wide range in ambient air temperature growth, phenological development, biomass and yield components databases were collected and are currently being analyzed. Preliminary results from the 3-year field trials (November-June) plantings under well-watered M regime, where air temperatures ranged from -2 to 44 °C, indicated that yields decreased as temperature increased. However, the response varied with crop and year, as did the lethal air temperature. To complete the well-water flood irrigation M trial, a 3-year database of inter-season and final biomass, grain yield, weather and season-long remotely sensed high throughput phenotyping (HTP) using tractor, cart and drone [tri-metric: spectral indices; sonic (height); infrared thermometry (temperature)] technologies continues to be developed - including a data pipeline to standardize the format (support Objective 3). Preliminary results for the 3-year experiment enhanced our understanding of the impact of global climate uncertainty on cereal grain producing regions of the globe. Furthermore, TRACE data have been distributed to colleagues at the Potsdam Institute of Climate Impact Research (PIK), Potsdam, Germany where a “stochastic” multivariate principal component regression analysis that formulates the time-dependent temperature effect on cereal grain crops is ongoing. Time-dependent temperature effects on final biomass and grain yields were proportional to the cumulative sum of total temperature exposure during a cropping season. Also, a proof of concept manuscript is underway, that compares results derived from prior experimental, “deterministic point-source,” and “gridded-modeling” AgMIP wheat-team modeled results for intercomparing and improvement of models. Once a proof of concept is developed, the 3-year TRACE field trial results for the well-watered flood irrigation M regime will serve as a trifecta test bed for a comparison of stochastic, deterministic point-source and gridded modeling approaches in AgMIP. An additional M component (100 and 70% ET) was added to complete the 3-way GxExM interaction with a wet-well pumping station, corresponding drip-tape irrigation system and real-time soil moisture monitoring system. This automated irrigation control drip-tape irrigation system, which will be monitored via a real-time automated soil volumetric water content probe system, combined with the already constructed flood irrigation system, will enable more accurate manipulation of soil moisture regimes. This will be particularly relevant later in the project plan for Sub-objective 2B, which includes a 2-year drought treatment. These databases from both well-watered and drought trials will be interfaced with Objective 1, thereby, enabling investigations and development of adaptation and mitigation strategies to ensure global food security. Support of Objective 3 was accomplished by the development of a cost-effective portable solution for high throughput phenotyping (HTP) to support genetic improvement and management. This new HTP system was designed as a portable and wireless solution with low power and light weight to deliver a plug-n-play module with quick and easy mounting to ground (Tractor, cart) or aerial (plane, drone) platforms. The portable system measures phenotypic metrics of the spectral signature, temperature, and height of the plant canopy from a multispectral camera, an IR thermometer, and a Mini LiDAR array, respectively. All sensors were interfaced to a microcontroller that features light weight and low power consumption. The system was configured for wireless connection to the cloud via public or hidden network for remote access and control of the HTP system regardless of indoors and outdoors. Comparisons of macro array (tractor, cart systems) and portable (mini array) are ongoing. Furthermore, this on-the-go HTP system was easily converted to a stationary plant monitoring system that measures height and a vertical temperature profile of plant canopy via solar powered wireless control. Software design for GUI-based user-friendly image analytics was developed to process images and point data and deliver plot-level metrics. Once developed and tested this work will provide overall support for Objective 2. To specifically support Sub-objective 3A, plant height was assessed in field year 2018 using three different sensor types. The results showed that simple ultrasonic transducers were enough for capturing accurate plant height. To assess crop architecture, particularly leaf or branch angles, a camera was mounted to a small row-bot. The row-bot traverses the field below the crop canopy in the inter-spatial rows (non-planted), which provided images looking up into the crop canopy. In field year 2019, images were captured for both cotton and sorghum throughout the growing season. The intent is to use these images to develop machine learning algorithms that will identify a leaf or branch from the main stem of the plant and report the angle of growth, thus providing useful information on the crop architecture. In support of Sub-objective 3B, a cart developed by the cotton breeding program that measures plant height and the normalized difference vegetative index (NDVI) was used to supply information to two biomass models for cotton and sorghum germplasm grown under well-watered conditions. These same plots were also imaged using a camera mounted to a done. Plant height and biomass from cotton and sorghum plots were collected on the same days as the flights and cart runs and used to determine the accuracy of the models. Initial results indicate the biomass models using canopy height and NDVI from the cart were not accurate for estimating cotton or sorghum biomass. Analysis is still ongoing using the drone derived images. Next steps will include applying the proximal information to different models before re-evaluating the overall approach. In support of Sub-objective 3C, a proximal imaging cart was developed to provide three-dimensional renderings of experimental plots. The cart was used to collect images on cotton germplasm grown under differing irrigation treatments during two important phenological time points, peak flower and fully open boll. Machine learning algorithms were developed from these images to identify and count the number of flowers or bolls present in the plot at the time of capture. The outputs from the algorithms were compared to manual counts from sample plots and found to be 80-90% accurate. In the future, this information will be used to determine the potential yield potential due to heat stress in cotton.
1. Development and testing of a portable High Throughput phenotyoing (HTP) system. High throughput phenotyping (HTP) research is currently accomplished through numerous technologies. An ARS researcher from Maricopa, Arizona, developed and tested a low-cost wireless HTP system using a microcontroller (Arduino) and a single-board computer (Raspberry Pi) powered by a solar rechargeable battery for plant phenotypic metrics of vegetation index, canopy temperature, and height from a multispectral camera, an infrared (IR) thermometer, and mini LiDAR sensors, respectively. Use of this portable HTP system is of interest to plant breeding and phenotyping research with innovative features such as a portable plug-n-play feature providing quick mounting to outdoor ground\aerial platforms; easy adoption to indoor platforms such as vertical farms, growth chambers, and greenhouses; a wireless interface that allows end-users to remotely monitor and control the HTP system from smartphones or computers; and a self-powered feature offering simple cable-free installation. The HTP system was adopted for automated irrigation control and plant phenotyping in growth chambers in Maricopa, Arizona, and enabled controlled water scheduling based on soil water condition and automated collection of phenotypic data months. This innovation will potentially enhance HTP research programs.
2. Fluctuations of CO2 in Free-Air CO2 Enrichment (FACE) depress plant photosynthesis, growth, and yield. Various techniques have been employed to provide a CO2 enrichment treatment to investigate plant responses. ARS researches from Gainesville, Florida, Maricopa, Arizona, Beltsville, Maryland, and Big Spring, Texas, and from the National Institute for Agro-Environmental Sciences, Tsukuba, Japan; and the University of Florida, Gainesville, Florida, completed a review of the effects of fluctuating concentrations of CO2, such as exist in free-air CO2 enrichment plots, on plant growth and yield responses. They concluded that it is likely that growth and yield responses in fluctuating CO2 are about 2/3 the size of the stimulations found using a corresponding steady CO2 enrichment. Therefore, even though FACE experiments are conducted under more natural open-field conditions than those using chambers, although conservative, the FACE results likely have a bias to be low. These results affect projections of the likely effects of the increasing atmospheric CO2 concentration on future agricultural productivity.
3. Carbon Dioxide (CO2), water supply, and nitrogen fertilizer aspects of Community Land Model tested using Maricopa, AZ FACE (free-air CO2 enrichment) data. The Community Land Model is widely used to predict the likely effects of Earth’s rising atmospheric CO2 concentration on future climate and carbon balance, but it had never been validated for the likely effects of CO2, drought, and N supply on the portions of the Earth devoted to wheat production. Scientists from the National Center for Atmospheric Research, Boulder, Colorado, together with a retired ARS collaborator at Maricopa, Arizona, used data from FACE (free-air CO2 enrichment) experiments conducted at Maricopa with ample and limiting levels of water and nitrogen fertilizer from 1993-1997 to test the model. The tests revealed that the model overestimates growth responses to CO2, as well as a few other aspects that need improvement in order to make more reliable forecasts of future climate. These data will be made available to sorghum modelers to validate their models.
Kim, J.Y. 2020. Roadmap to high throughput phenotyping for plant breeding. Journal of Biosystems Engineering. 45:43-55. https://doi.org/10.1007/s42853-020-00043-0.
Allen, L.H., Kimball, B.A., Bunce, J.A., Toshimoto, M., Harazono, Y., Baker, J.T., Boote, K.J., White, J.W. 2020. Fluctuations of CO2 in Free-Air CO2 Enrichment (FACE) depress plant photosynthesis, growth, and yield. Agricultural and Forest Meteorology. 284. https://doi.org/10.1016/j.agrformet.2020.107899.
Lu, Y., Kimball, B.A. 2020. Validation of spring wheat responses to elevated CO2, irrigation, and nitrogen fertilization in the Community Land Model 4.5. Earth and Space Science. 7(6). https://doi.org/10.1029/2020EA001088.