Canavalia brasiliensis is a tropical forage legume of high agronomic and ecological value, widely recognized for its ability to enhance the productivity and sustainability of livestock systems in tropical and subtropical regions. This species is characterized by high biomass production, good palatability, and adequate crude protein content, making it a strategic alternative for animal feeding, particularly during periods of forage scarcity.This dataset provides information on the response of Canavalia brasiliensis genotypes to water stress and different soil pH conditions. To generate the data, three genotypes of C. brasiliensis were established under water stress (drought) and two soil pH conditions (acidic and neutral). Each experimental unit (genotype × water stress × soil pH) was replicated six times.The dataset includes measurements taken every two days using MultispeQ and QualitySpec over a five-week period. Final harvest data includes leaf area, number of leaves, number of stems, soil moisture, photographic records of the root system, root biomass, and aboveground biomass.This dataset is useful for evaluating the physiological, morphological, and productive responses of C. brasiliensis genotypes to the interaction between water stress and soil pH, allowing for analysis of genetic variability and phenotypic plasticity under contrasting environmental conditions. Periodic measurements with MultispeQ and QualitySpec, combined with final biomass and root system data, facilitate the identification of early indicators of stress tolerance and support studies on genetic improvement and sustainable management of forage legumes in tropical systems.<br><br> Methodology:Plant material, experimental design, and study site: In this study, seeds from three Canavalia brasiliensis genotypes were used, obtained from the germplasm collection of the Genetic Resources Program of the Bioversity Alliance and CIAT. The seeds were established in a sandy loam Vertisol characterized by high fertility and a pH of 7.5 (neutral) and in a clay loam Ultisol with a pH of 4.5 (acidic).The experiment followed a completely randomized design with a 2 × 2 × 3 factorial arrangement, conducted in soil cylinders under semi-controlled conditions:• Factor 1: Soil acidity (neutral pH and acidic pH)• Factor 2: Water stress (control and drought)• Factor 3: C. brasiliensis accession (7318, 17009, and 17462)Each seed was grown in a PVC cylinder containing 8,250 kg of soil for acidic pH and 7,750 kg for neutral pH soil. Each combination (genotype × water stress × soil pH) was replicated six times.The trial was conducted in a greenhouse equipped with rails to control sunlight exposure during the day and allow coverage to prevent dew formation. The greenhouse is located on the CIAT campus in Palmira, Valle del Cauca, Colombia.Photosynthesis and energy exchange: Measurements were conducted using the MultispeQ, a portable sensor that allows for rapid, non-destructive assessment of the plant’s physiological status by measuring variables related to photosynthesis, energy exchange, and environmental stress. The device quantifies chlorophyll fluorescence parameters associated with the functioning of photosystems II and I. Measurements were taken on the last trifoliate of fully expanded leaves every two days, from the onset of water stress (26 days after seed germination) until the end of the experiment (61 days of age).Spectral reflectance: Measurements were conducted using the QualitySpec, a portable spectroradiometer that allows for the characterization of the plant’s physiological and structural status. Measurements were taken simultaneously with the MultispeQ over the same time period.For both devices, the order of the initial plots was rotated during data collection to minimize measurement error.At 58 days after the onset of treatments, a final destructive harvest was conducted. Leaf area was determined using an LI-3100 area meter (LI-COR, Lincoln, NE, USA). The number of leaves per plant and the stems were recorded. Soil moisture was measured at multiple points along each cylinder (point 1 = 10 cm from the top, and subsequently every 20 cm along the length) using a Bluelab Pulse Meter EC.Photographic records of root growth were taken on both sides of each cylinder to cover 100% of the root area. Subsequently, roots were separated from the soil and, along with the leaf samples, were oven-dried in paper bags at 60 °C for 72 hours.

This dataset contains orthomosaics and individual Regions of Interest (ROIs) of forage grasses in crop fields from experimental trials of CIAT’s tropical forages breeding program; and annotations in Common Objects in Context (COCO) format derived from that data. The ROIs were manually annotated on UAV imagery and exported in common objects in context (COCO) format compatible with different machine learning models and architectures. 9,554 ROIs in the geospatial data and 12,365 annotations of forage grasses in COCO format.<br><br> Methodology: The dataset was generated through a multi-step process beginning with data acquisition of forages crop fields via UAV flights (DJI Phantom 4 Multispectral drone) with RTK determining the geolocation. These images were processed in Agisoft Metashape to generate georeferenced orthomosaics as raster files. Manual annotation of forage grasses ROIs was performed in QGIS and the geospatial data for 8 different orthomosaics was later converted to COCO format using custom python scripting. To ensure compatibility witch COCO standards and optimize training efficiency, the large orthomosaics where clipped to the annotations’ extents with additional 1% spatial buffer and split into tiles with a maximum dimension close to 1024 pixels for the larger side and 25% overlap.

Megathyrsus maximus is a forage species widely used in livestock production systems due to its high biomass yield, good nutritional quality, and ease of propagation. Understanding the response of the species’ genetic diversity to growth under shade conditions is essential for assessing its potential use in silvopastoral systems. These systems represent a sustainable alternative to mitigate the negative impacts of monoculture forage production on soil health, animal welfare, and the environment.This dataset includes information on the response of M. maximus genotypes to shade conditions. To generate the data, vegetative propagules from 33 genotypes were established under two light environments in a glass greenhouse: full sunlight (no shade) and 75% shade. Each experimental unit (genotype × shade condition) was replicated three times. The dataset comprises weekly measurements collected over a six-week period, including SPAD values, plant height, stem height, gas exchange parameters, and chlorophyll fluorescence. In addition, reference measurements of red light, far-red light, and photosynthetically active radiation (PAR) were recorded under each shade condition. Final harvest data include leaf area, number of leaves, biomass partitioning, root-to-shoot ratio, and specific leaf area.This dataset provides valuable information for breeding programs aimed at selecting M. maximus genotypes with potential for use in silvopastoral systems. Moreover, it offers a useful resource for plant scientists seeking to better understand the physiological and morphological mechanisms underlying shade adaptation in this species.<br><br> Methodology:Plant material, experimental design, and study site: Vegetative propagules of 33 Megathyrsus maximus genotypes from the germplasm collection of the Alliance Bioversity and CIAT Tropical Forages Breeding Program were used in this study. Plants were established in a sandy loam Vertisol characterized by high fertility and a pH of 7.5. The experiment followed a split-plot design with a 33 × 2 factorial arrangement, where genotype (33 levels) and shade condition (75% shade and full sunlight) were the experimental factors. Shade treatment was assigned to the main plots, while individual plants constituted the experimental units. Each plant was grown in a 3-L pot, and each genotype × shade combination was replicated three times. The trial was conducted in a glass greenhouse located on the CIAT campus in Palmira, Valle del Cauca, Colombia. Red and far-red radiation were measured using a LightScout Red/Far Red Meter (Spectrum Technologies, USA), while photosynthetically active radiation (PAR) was measured using a SpectraPen Mini (Photon Systems Instruments, Czech Republic).Gas exchange and chlorophyll fluorescence: Stomatal conductance, transpiration rate, and the effective quantum yield of photosystem II were measured using a LI-600 porometer/fluorometer (LI-COR, Lincoln, NE, USA). Measurements were taken weekly on the third fully expanded leaf of each plant, from the initiation of shade treatments until 45 days after treatment establishment.Plant growth: From the onset of treatments until day 45, SPAD values were recorded weekly using a SPAD 502 Plus chlorophyll meter (Spectrum Technologies, Aurora, IL, USA), with measurements taken on the third fully expanded leaf. Plant height was measured from the stem base to the tip of the longest leaf, while stem height was measured from the base to the apical meristem of the most vigorous stem per plant. At 45 days after treatment initiation, a final destructive harvest was conducted. Leaf area was determined using an LI-3100 area meter (LI-COR, Lincoln, NE, USA). The number of leaves per plant was recorded, and stems (including inflorescences), leaves, and roots were separated and oven-dried in paper bags at 60 °C for 72 hours. The root-to-shoot ratio was subsequently calculated, and specific leaf area was estimated as the ratio of leaf area to leaf dry mass.

OverviewThis dataset contains 80 4k videos of 20 individual Megathyrsus maximus plants. Each plant was recorded at four distinct developmental stages for 4 weeks under greenhouse conditions. The dataset is intended as a primary source for 3D reconstruction using Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS).

OverviewThis dataset contains 80 4k videos of 20 individual Megathyrsus maximus plants. Each plant was recorded at four distinct developmental stages for 4 weeks under greenhouse conditions. The dataset is intended as a primary source for 3D reconstruction using Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS).

OverviewThis dataset contains a time-series collection of 135 high definition video sequences of tropical forage grasses (Megathyrsus maximus and Urochloa spp.) grown under open field conditions. The study follows various genotypes with 3 replications each, captured over a growing period.Week 1 (Pre-Pruning): This sequence represents the full biomass before the pruning event.Weeks 2–9 (Regrowth): This sequence represents the regrowth period.The dataset is intended as a primary source for 3D reconstruction using Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS).

OverviewThis dataset contains a time-series collection of 135 high definition video sequences of tropical forage grasses (Megathyrsus maximus and Urochloa spp.) grown under open field conditions. The study follows various genotypes with 3 replications each, captured over a growing period.Week 1 (Pre-Pruning): This sequence represents the full biomass before the pruning event.Weeks 2–9 (Regrowth): This sequence represents the regrowth period.The dataset is intended as a primary source for 3D reconstruction using Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS).

This is the second part of the dataset Dataset: Multi-view video sequences of Megathyrsus maximus and Urochloa spp. in field conditions for 3D reconstruction and temporal phenotypingOverviewThis dataset contains a time-series collection of 135 high definition video sequences of tropical forage grasses (Megathyrsus maximus and Urochloa spp.) grown under open field conditions. The study follows various genotypes with 3 replications each, captured over a growing period.Week 1 (Pre-Pruning): This sequence represents the full biomass before the pruning event.Weeks 2–9 (Regrowth): This sequence represents the regrowth period.The dataset is intended as a primary source for 3D reconstruction using Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS).

This is the second part of the dataset Dataset: Multi-view video sequences of Megathyrsus maximus and Urochloa spp. in field conditions for 3D reconstruction and temporal phenotypingOverviewThis dataset contains a time-series collection of 135 high definition video sequences of tropical forage grasses (Megathyrsus maximus and Urochloa spp.) grown under open field conditions. The study follows various genotypes with 3 replications each, captured over a growing period.Week 1 (Pre-Pruning): This sequence represents the full biomass before the pruning event.Weeks 2–9 (Regrowth): This sequence represents the regrowth period.The dataset is intended as a primary source for 3D reconstruction using Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS).

Soil organic carbon (SOC) sequestration through vigorous and deep root systems is a promising strategy to mitigate the accumulation of greenhouse gases. The presence of carbon compounds with low decomposition rates, such as suberin, is another root trait that can contribute to this process. Forage grasses exhibit a high capacity for SOC sequestration because they develop abundant and deep root systems and allocate a substantial proportion of their photoassimilates to belowground organs.Urochloa humidicola is an ideal species in the context of climate change adaptation due to its strong tolerance to abiotic stresses such as drought, waterlogging, low soil fertility, and acidic soils, in addition to its high biological nitrification inhibition activity. Megathyrsus maximus is characterized by high forage yield and nutritional quality, as well as extensive, abundant, and deep root systems.To generate information on the potential of selected genotypes of U. humidicola and M. maximus to sequester SOC, vegetative propagules of both species were established in a hydroponic system using two contrasting nutrient solutions: an acidic, low-fertility solution and a neutral, high-fertility solution. The dataset includes information on root morphological traits such as root length, diameter, surface area, volume, branching, and number of root tips, as well as data on the proportion of root suberization, determined through image analysis of histochemical cross-sections.The information provided in this dataset is valuable for plant breeding programs focused on sustainable production, as it enables the evaluation of key root traits associated with SOC sequestration and supports the identification of promising genotypes for this purpose.<br><br> Methodology:Root morphological and anatomical data were obtained from hydroponic experiments conducted at the Alliance Bioversity & CIAT campus in Palmira, Valle del Cauca, Colombia (3°50′38″ N, 76°35′36″ W). Three tropical forage grass genotypes with contrasting root vigor were evaluated for each species: Megathyrsus maximus (high: Pm21_3522; medium: CIAT_16055; low: CIAT_16379) and Urochloa humidicola (high: Bh16b_1618; medium: CIAT_6133; low: CIAT_26146). Plants were grown under two contrasting nutrient solutions: a neutral-pH solution with high nutrient availability and an acidic-pH solution with low nutrient availability, representing typical soil fertility conditions in Colombia. Each genotype was grown with three biological replicates.Root systems were scanned at 600 dpi using an Epson Expression 12000XL scanner, and morphological traits—including total root length, mean diameter, surface area, volume, number of tips, and branching parameters—were quantified using RhizoVision Explorer software.For anatomical and histochemical analyses, root samples were collected from basal, medial, and apical sections, fixed in an acetic acid–ethanol solution (3:1), and transversely sectioned using a Leica VT1000 S vibratome. Sections were stained with Sudan Red 7B to visualize suberization and imaged using a light microscope. Image preprocessing included background removal using the PlantCV and cv2 libraries in Python, followed by binarization. Colorimetric analyses were then conducted using the colordistance library in R. Suberin, identified by red tonalities, and was detected through the integrated use of PlantCV, cv2, matplotlib.pyplot, and NumPy. Ten chromatic clusters were defined per image, and the relative proportion of each cluster was converted into a colorimetric descriptor. This information, combined with the total image area (in pixels), was used to estimate the relative area occupied by suberin.