A rapidly expanding world population and extreme climate change have made food production a crucial challenge in the twentyirst century. Therefore, improving crop production through agricultural management could be an effective solution for this challenge. However, due to the associated cost and time to perform field works, researchers widely rely on agricultural system modeling to examine the impacts of different crop management scenarios. However, due to the complexity of agricultural system modeling, their applications in producing practical knowledge for producers are limited. Concurrently, deep learning techniques have been recognized as a preferred method when dealing with large datasets. This study was performed in three phases. First, A deep learning network was utilized and trained by incorporating a large number of datasets produced by the Decision Support System for Agrotechnology Transfer (DSSAT) model. To the best of our knowledge, no research has been done in the literature on modeling a cropping system by deep learning. An model accuracy level of around 98\% was obtained, and it was 770 times faster than classical crop models DSSAT in calculating 900,000 different crop growth scenarios. However, The second phase of the study examined the robustness of the deep learning model under a wider range of environmental factors (e.g., different irrigation and climatological conditions) while a deep learning structure was desired compare to the first study. To optimize the deep learning structure, three variable reduction methods were used (Bayesian, Spearman, and Principal Component Analysis). The result of this study showed that a deep learning structure could be developed that has a similar accuracy level as the original model while the structural size was reduced up to 80 times. In the third phase of the study, three techniques (L1/L2 regularization, and neurons dropout) were used to address the overfitting problem in some deep learning models. The L2 regularization was identified as the most effective method that increased model generalization and reduced overfitting. The overall results from this study demonstrated the effectiveness of the proposed deep learning technique in replicating the yield results from crop modeling under different climatological and management conditions.

Corn is one of American agriculture's greatest success stories, where we have witnessed incredible increases in yield potential over the last half-century. However, abiotic stress is still the primary limiting factor preventing plants from reaching their true yield potential. In addition, agriculture is not exempt from the deleterious effects of changing weather patterns and the altered climate our world will face as time progresses. Thus, increasing our understanding of how crops interact with their environment both above and below the soil will be crucial to increasing production on a global scale while maximizing profitability at a local level. Five studies were carried out to quantify the relationship between corn and multiple abiotic factors, including temperature, moisture, and nitrogen. In study one, Corn seed germination as a function of temperature was quantified, and the effects were compared between soybean and cotton, both major agronomic crops in Mississippi. Study two determined the effects of osmotic stress on corn seed germination, and commercial corn hybrids' variability was explored. In study three, functional relationships between temperature and early season growth and development were quantified, and the concept of a simple mathematical model for predicting growth as a function of temperature was explored. Study four exemplified the effects of increasing waterlogging durations from 0 to 14 days and determined critical limits for above and below-ground growth. Finally, in study five, growth, development, and physiology were determined as a function of nitrogen concentration. In addition, optimum concentrations to maximize vigor were estimated. Data generated from these studies exemplify how abiotic stress significantly affects corn during germination and early season growth and development. These datasets will be valuable foundations to build on as we explore how abiotic stress affects all growth stages of corn and other important agronomic crops. Functional relationships generated from these studies will be useful to update crop simulation models. Both simple and complex mathematical models have promising implications in emerging and developing precision and predictive agricultural technologies.

Le semis précoce du tournesol, d’un à deux mois par rapport à la période habituelle (Avril dans le sud ouest de la France), a été envisagé pour esquiver les périodes de sécheresses estivales. Cette stratégie conduit à un abaissement des températures de l’ordre de 5 à 10°C durant les premières phases de développement de la culture. L’objectif de ce travail est donc d’identifier des génotypes de tournesol adaptés à des conditions de basses températures en début de cycle, et de fournir des outils pour la sélection de ces nouveaux idéotypes. Dans un premier temps le modèle de culture SUNFLO, développé pour l’analyse des interactions génotype x environnement chez le tournesol, a été utilisé pour identifier par simulation un idéotype pour le semis précoce. Cette étude a révélé que le type variétal valorisant le mieux de semis précoce présente une levée précoce et un cycle tardif. Dans un deuxième temps, la variabilité génétique d’une population de lignées recombinantes de tournesol a été une analysée pour des traits agro-morphologiques et physiologiques caractérisant le développement (vitesse de germination, phénologie) et la croissance à basse température (élongation de l’hypocotyle, production de biomasse, et traits physiologiques impliqués dans la tolérance au froid). L’analyse génétique de ces caractères a permis d’identifier les régions chromosomiques impliquées dans la variation de ces caractères (QTLs) ainsi que les marqueurs moléculaires associés à ces QTLs qui représentent des marqueurs d’intérêts pour la sélection. L’analyse des processus impliqués dans la levée (germination et élongation de l’hypocotyle) montre que la température de base pour l’élongation de l’hypocotyle présente un gain génétique significatif à basse température. Ce trait est sous le contrôle génétique de deux QTLs majeurs dont l’un, qui explique 40% de la variabilité phénotypique observée, est lié au marqueur SSR ORS1128. Le temps thermique du semis à la floraison est un caractère contrôlé par des QTLs spécifiques en conditions de semis précoces, parmi lesquels deux sont colocalisés avec des QTLs identifiés pour des traits relatifs à la levée. L’étude des traits physiologiques impliqués dans la réponse aux basses températures a révélé que le tournesol a un potentiel de sélection pour la tolérance au froid, notamment pour le potentiel osmotique. Le maintien des membranes plasmiques stables à basses températures est également un trait jouant un rôle important dans la tolérance au froid. Un QTL à effet majeur lié au marqueur SSR ORS331_2 a été identifié pour ce trait et pourrait être utilisé pour aider à la sélection de génotypes de tournesol adaptés au froid. / Early sowing to escape the drought during summer was studied in sunflower. Sowing one or two months earlier leads to reduce about 5 to 10°C during first stages of development compared with traditional sowing (April in south parts of France) in this species. The aim of this study is to identify sunflower genotypes adapted for low temperature and to identify tools for selecting them. Firstly the crop model SUNFLO, Which is developed to analyze “genotype x environment” interactions in sunflower, was used to identify by simulation favorable ideotypes for early sowing. Results show that they have to present early emergence and a late development cycle. Then, several experiments were undertaken to study genetic variability for agro-morphologic and physiologic traits under early sowing in sunflower. A population of 95 recombinant inbred lines and their two parents were used at low temperature in all experiments. Germination rate, hypocotyl elongation, biomass production and some physiological traits for cold tolerance were studied. Genetic analyses were performed and genomic regions (QTLs) involved in the variation of these traits as well as SSR markers associated with them were identified. Analysis of physiological processes related to emergence (germination and hypocotyl elongation) show that the base temperature of hypocotyl elongation presents a significant genetic gain at low temperature. This trait is controlled by two major QTLs and one of them explains 40% of the phenotypic variance and contains the SSR marker ORS1128. The thermal time from sowing to flowering is controlled by specific QTLs in early sowing and two of them are collocated with QTLs detected for emergence related-traits. The study of physiological traits implied with response to low temperature showed that sunflower present a high potential for cold tolerance variability, especially for the osmotic potential. The cell membrane stability at low temperature is also an important trait for cold tolerance. A major QTL associated with the SSR marker ORS331_2 was identified for this trait and should be used to select sunflower cold tolerant genotypes.

The increasing demand for energy, depletion of fossil fuels, rising global warming, and greenhouse gas emissions have stimulated the need for widespread development and adoption of renewable energy sources (RES) worldwide. Among these sources, solar energy has emerged as a major contender to meet the growing demand. It offers adaptable applications and provides an alternative to traditional energy sources. A brand-new application of solar panels is agrivoltaics. Agrivoltaics consists in installing solar panels above farming lands such as crops. The combination of solar energy production and farming on the same lands increases the overall yield of the land and brings several other opportunities. However, agrivoltaics is also very challenging. An improper installation of solar panels above crops may result in a dramatic drop of the farming yield. Thus, it is of major importance to understand how to maximize the solar energy production without harming the plants or decrease the farming yield. This master’s thesis focuses on the impact of agrivoltaic systems on the micro-climate close to the crop. The goal is to link the modified physical phenomena within an agrivoltaic system and their impact on the crops. The methodology is based on Computational Fluid Dynamics (CFD). The idea is to realize high fidelity simulations of the different physical phenomena and their coupling, and compare them to experimental data. Flow simulations coupled with radiative models and a surface model are realized in this perspective. The master’s thesis is divided in three parts. 1. Based on experimental data collected during three years at the EDF lab les Renardières, determine which physical phenomena impact the most the crop and what are the key parameters to study the growth of the plants. 2. Validate with experimental data from the atmospheric laboratory the SIRTA (Site Instrumental de Recherche par Télédétection Atmosphérique) of the engineering school Polytechnique, the radiative models and the surface model of the CFD software. 3. Study the impact of an agrivoltaic system on the identified physical phenomena with a simple geometry composed of one pitch of solar panel. The data study shows clearly that the plant temperature, the groundwater, and the radiation play crucial roles in the growth of the plant. A lack of radiation or groundwater will limit the growth of the crops. In addition, extreme temperatures can harm the crops. Consequently, this research project will firstly focus on capturing the impact of the solar panels on these three key parameters. Simulations are using a coupling of a 1D radiative model which is computationally fast and that can therefore be applied on a very large domain to compute the absorption of the atmospheric layers and the clouds, and a 3D radiative model which is able to capture the impact of an obstacle such as a solar panel. This coupling is validated for the shortwave radiation and the longwave radiation. Finally, full U-RANS simulations with the radiative models, the surface model and the - turbulence model are realized. The impact of the panels on the radiation field, the soil temperature, the specific humidity and on other fields such as the wind speed is well captured.

The Mediterranean basin is widely recognized as a climate change hotspot, with climate models projecting increasingly warmer and drier conditions that will impact local ecosystems, communities, and economies. Agriculture will be among the most affected sectors, with harsher conditions for crops’ growth, greater water needs, and lower yields. One of the most resilient crops to limiting and stressful conditions is barley, which is often sown in areas where other crops and cereals would struggle. This work analyzed the impacts of climate change on rainfed barley using the province of Almeria as a case study. This is one of the most arid areas of the Mediterranean basin, where agriculture is among the main economic resources, and where barley is the main crop produced outside greenhouses. Barley growth was modeled using the AquaCrop model in its Python implementation, AquaCrop-OSPy. Setting the model up to avoid local re-calibration of the barley parameters and to capture multi-year trends in productivity change, rather than its interannual variability. The study focused on two 30-year time periods: mid-century (2041-2070), and end-century (2071-2100); and on Shared Socioeconomic Pathways scenarios SSP1-2.6, SSP2-4.5, and SSP5-8.5. For each time period and SSP scenario, the research also evaluated three sub-scenarios of soil water content at sowing: with the parameter set respectively at 10%, 20%, and 30% of the Total Available Water (the water present in the soil available for the crop to sustain its life). Having estimated climate change impact, the research analyzed different adaptation pathways (irrigation, the application of mulches, and the change in sowing date), to evaluate their performances for climate change adaptation in the area.  The results indicate the importance of soil water content for maintaining good yields, or reducing losses, and indicate the possible average yield change to be between +14% and -45% at mid-century, and between +12% and -55% at end-century. The greater variability in productivity is associated with the soil water content at sowing rather than on the SSP scenario, with SSP5-8.5 being the only one showing a marked difference compared to the others. Regarding irrigation, the results show how with a soil water content at sowing of 10% of the Total Available Water, irrigation up to 100 m³/ha might not be sufficient to avoid productivity losses. Also, the study indicates that an optimal threshold to trigger irrigation for adaptation purposes might be found between 0% and 20% of the Total Available Water. Overall, it indicates how adaptation through irrigation can be viable in the province. The work moreover suggests the effectiveness of mulches as an adaptation strategy to partially limit irrigation water needs in the future and improve the yield performance of the crop. However, the research does not indicate a clear benefit linked to changing the sowing date to earlier or later sowing dates but suggests the importance of correctly seizing the sowing window to reach optimum yield in the future. Lastly, the work shows that the approach used to carry out this research is suitable to assess trends in yield change at multi-year scale, if the analyzed time window is indicatively larger or equal to 10 years, and if an error of around 10% on the results is accepted. / Medelhavsområdet är allmänt erkänt som en hotspot för klimatförändringar, och klimatmodellerna förutspår allt varmare och torrare förhållanden som kommer att påverka lokala ekosystem, samhällen och ekonomier. Jordbruket kommer att vara en av de mest drabbade sektorerna, med tuffare förhållanden för grödornas tillväxt, större vattenbehov och lägre avkastning. En av de grödor som är mest motståndskraftiga mot begränsande och stressande förhållanden är korn, som ofta sås i områden där andra grödor och spannmål skulle ha svårt att klara sig. I det här arbetet analyserades klimatförändringarnas inverkan på regnkorn med provinsen Almeria som fallstudie. Detta är ett av de torraste områdena i Medelhavsområdet, där jordbruket är en av de viktigaste ekonomiska resurserna, och där korn är den viktigaste grödan som produceras utanför växthus. Kornets tillväxt modellerades med hjälp av AquaCrop-modellen i dess Python-implementering, AquaCrop-OSPy. Modellen ställdes in för att undvika lokal omkalibrering av kornparametrarna och för att fånga fleråriga trender i produktivitetsförändringar, snarare än den mellanårliga variationen. Studien fokuserade på två 30-årsperioder: mitten av århundradet (2041-2070) och slutet av århundradet (2071-2100), och på scenarierna SSP1-2,6, SSP2-4,5 och SSP5-8,5 för de gemensamma socioekonomiska vägarna. För varje tidsperiod och SSP-scenario utvärderade forskningen också tre underscenarier av markvatteninnehåll vid sådd: med parametern inställd på 10%, 20% respektive 30% av det totala tillgängliga vattnet (det vatten som finns i jorden som är tillgängligt för grödan för att upprätthålla sitt liv). Efter att ha uppskattat effekterna av klimatförändringarna analyserade forskningen olika anpassningsvägar (bevattning, applicering av mulcher och förändring av sådatum) för att utvärdera deras prestanda för anpassning till klimatförändringar i området.  Resultaten visar att markvattenhalten är viktig för att upprätthålla god avkastning eller minska förlusterna, och visar att den möjliga genomsnittliga avkastningsförändringen är mellan +14% och -45% vid mitten av århundradet och mellan +12% och -55% vid slutet av århundradet. Den större variationen i produktivitet är förknippad med markvatteninnehållet vid sådd, snarare än på SSP-scenariot, med SSP5-8.5 som det enda som visar en markant skillnad jämfört med de andra. När det gäller bevattning visar resultaten att med en markvattenhalt vid sådd på 10% av det totala tillgängliga vattnet, kan bevattning upp till 100 m³ / ha inte vara tillräcklig för att undvika produktivitetsförluster. Studien visar också att en optimal tröskel för att utlösa bevattning i anpassningssyfte kan hittas mellan 0% och 20% av det totala tillgängliga vattnet. Sammantaget visar studien hur anpassning genom bevattning kan vara genomförbar i provinsen. Arbetet tyder dessutom på att mulcher är effektiva som en anpassningsstrategi för att delvis begränsa bevattningsvattenbehovet i framtiden och förbättra grödans avkastning. Forskningen visar dock inte på någon tydlig fördel med att ändra sådatumet till tidigare eller senare sådatum, men antyder vikten av att korrekt utnyttja såfönstret för att nå optimal avkastning i framtiden. Dessutom visar arbetet att den metod som används för att genomföra denna forskning är lämplig för att bedöma trender i avkastningsförändringar på flerårig skala, om det analyserade tidsfönstret är större eller lika med 10 år, och om ett fel på cirka 10% på resultaten accepteras.

Crop improvement is necessary for food security as the global population is expected to exceed 9 billion by 2050. Limitations in water resources and more frequent droughts and floods will make it increasingly difficult to manage agricultural resources and increase yields. Therefore, we must improve our ability to monitor agronomic research plots and use the information they provide to predict impacts of moisture stress on crop growth and yield. Towards this end, agronomists have used reductions in leaf expansion rates as a visible ‘plant-based’ indicator of moisture stress. Also, modeling researchers have developed crop models such as AquaCrop to enable quantification of the severity of moisture stress and its impacts on crop growth and yield. Finally, breeders are using Unmanned Aircraft Systems (UAS) in field-based High-Throughput Phenotyping (HTP) to quickly screen large numbers of small agronomic research plots for traits indicative of drought and flood tolerance. Here we investigate whether soybean canopy cover and color time series from high-resolution UAS ortho-images can be collected with enough spatial and temporal resolution to accurately quantify and differentiate agronomic research plots, pinpoint the timing of the onset of moisture stress, and constrain crop models such as AquaCrop to more accurately simulate the timing and severity of moisture stress as well as its impacts on crop growth and yield. We find that canopy cover time series derived from multilayer UAS image ortho-mosaics can reliably differentiate agronomic research plots and pinpoint the timing of reductions in soybean canopy expansion rates to within a couple of days. This information can be used to constrain the timing of the onset of moisture stress in AquaCrop resulting in a more realistic simulation of moisture stress and a lower likelihood of underestimating moisture stress and overestimating yield. These capabilities will help agronomists, crop modelers, and breeders more quickly develop varieties tolerant to moisture stress and achieve food security.

In the last decade, Brazilian soybean and maize, cultivated in succession, accounted for 23.8 ± 1.9% and 6.9 ± 0.9% of world\'s production, respectively. More than 80% of soybean and maize production in Brazil is under rainfed conditions, which results in a high interannual yield variability and, consequently, increasing the risks for food supply, not only in the country but also around the world. Among the natural phenomena that cause climate and yield variability in Brazil, El Niño Southern Oscillation (ENSO) is the most important. The best way to minimize the impacts of ENSO, mainly those associated to water deficit in rainfed crops, is by defining the most favorable sowing dates, when the probability of crop failure is small. Based on that, this study aimed: to determine the best sowing dates for the soybean-maize production system, based on the economic profitability at national scale; to assess the influence of the ENSO phases (El Niño, La Niña and Neutral) on spatial and temporal soybean and maize off-season yield variabilities for different sowing dates; and to determine the magnitude of the current soybean- maize succession yield gap due to water deficit and crop management in different Brazilian producing regions. To achieve such goals, soybean and maize off-season simulations were performed using three previously calibrated and validated crop simulation models (FAO-AZM, DSSAT and APSIM), in a multi-model approach. Soybean and maize yields were simulated for 29 locations in 12 states, with soybean sowing dates ranging from 21st September to 1st January, for a period of 34 years (1980-2013). Maize sowings were simulated in the same day soybean was harvested. The optimal sowing dates for soybean-maize succession varied according to the Brazilian region, with water deficit, solar radiation and air temperature being the main weather variables that influenced this crop system. ENSO phases affected soybean and maize yields across the country, having, in general, opposite effects during the warm (El Niño) and cold (La Niña) phases, but also depending on the sowing date considered. The yield gap (YG) of soybean-maize succession varied among locations, sowing dates and growing seasons. However, the yield gaps caused by water deficit (YGw) were, on average, higher than those caused by sub-optimal crop management (YGm), which can be explained by the high inter-annual and spatial climate variability observed in the Brazilian territory. / Na última década, a soja e o milho safrinha, cultivados em sucessão no Brasil, contribuíram com 23.8 ± 1.9% e 6.9 ± 0.9% da produção mundial, respectivamente. Mais de 80% da soja e do milho brasileiro são produzidos em condições de sequeiro, o que resulta em uma alta variabilidade interanual da produtividade e, consequentemente, aumenta os riscos de falhas no abastecimento alimentar no Brasil e no mundo. Entre os fenômenos causadores da variabilidade climática e da produtividade agrícola no Brasil, o El Niño Oscilação Sul (ENOS) é o mais importante. A melhor maneira para minimizar os impactos do ENOS, principalmente os associados ao déficit hídrico em culturas de sequeiro, é definindo as datas de semeaduras mais favoráveis, onde as chances de grandes perdas são menores. Assim, os objetivos deste estudo foram: determinar a melhor data de semeadura para o sistema de produção em sucessão soja - milho safrinha, baseado na rentabilidade econômica em escala nacional; indicar a influência das fases do ENOS (El Niño, La Niña e Neutro) sobre a sucessão soja - milho safrinha em escala espacial e temporal, em diferentes datas de semeaduras; e determinar a magnitude da quebra de produtividade da sucessão soja - milho safrinha devido ao déficit hídrico e ao manejo sub ótimo do cultivo. Para atingir os objetivos, simulações de produtividade para soja e milho safrinha foram realizadas usando três modelos de simulação de cultura (FAO-AZM, DSSAT e APSIM), previamente calibrados, em uma abordagem multi-modelos. As produtividades das culturas da soja e do milho foram simuladas para 29 locais em 12 estados, com as datas de semeadura da soja variando de 21 de setembro a 1º de janeiro, para um período de 34 anos (1980-2013). A semeadura do milho ocorreu imediatamente após a colheita da soja. A data de semeadura ótima para a sucessão soja - milho safrinha variou de acordo com a região brasileira, tendo o déficit hídrico, radiação solar e a temperatura do ar como as principais variáveis que influenciam o sistema. As fases do ENOS afetaram a produtividade da soja e do milho safrinha no Brasil, tendo, efeitos opostos durante as fases quentes (El Niño) e frias (La Niña). Os impactos das fases do ENOS também variaram de acordo com as datas de semeadura. As quebras de produtividade da sucessão soja - milho safrinha variaram entre os locais, datas de semeadura e safras. Entretanto, as quebras de produtividade causadas pelo déficit hídrico foram, em média, superiores àquelas causadas pelo manejo subótimo das culturas, o que pode ser explicado pela alta variabilidade espacial e interanual das condições meteorológicas no território brasileiro.

No description available.