Phenology and health of Stenocereus Queretaroensis: A multimodal dataset combining multispectral imagery and spectrophotometry

This data article presents a multimodal, non-invasive dataset documenting the physiology and growth stages of Stenocereus queretaroensis (pitayo), a native species from the arid and semi-arid regions of Southern Zacatecas, Mexico. In particular, Stenocereus spp. are important cacti in the region due to its nutritional properties, role as an economic resource, and cultural significance.It is worth emphasising that these cacti traditionally grow wild (i.e., without deliberate cultivation); accordingly, controlled cultivation is uncommon and remains understudied. With the aim of producing a formal, comprehensive analysis and compendium, the data were collected across multiple phenological stages to provide a complete representation of the plant development cycle, from vegetative growth through to fruiting. To achieve this, the collection process combined high-resolution multispectral imaging with field spectrometry in the 400-700 nm range. Standardized acquisition protocols were applied in field conditions to capture consistent reflectance data, and environmental variables such as illumination, temperature, and geographic coordinates were recorded for each session to ensure reproducibility. The dataset integrates several components: (i) multispectral images that provide spatial information on canopy and structural characteristics, (ii) field spectral signatures with detailed reflectance values for each sampled plant, and (iii) metadata describing phenological stage, acquisition date and time, environmental conditions, and equipment settings. For subsequent analysis, data was preprocessed and normalized to enable reliable comparisons between growth stages and across acquisition sessions, resulting in a clean, structured resource ready for computational analysis. In this regard, this dataset has been organized to facilitate its direct application across multiple research and development contexts. Specifically, potential applications include the training and validation of machine learning and computer vision models for automated phenological stage classification, harvest time estimation, and development of species-specific vegetation indices. Moreover, owing to its standardized design, the resource can serve as a benchmark for comparing methods, validating algorithms, and supporting reproducible workflows in precision agriculture and remote sensing. Beyond Stenocereus queretaroensis, the documented acquisition and preprocessing methodology can be replicated or adapted to generate similar multimodal datasets for other climate-resilient crops, particularly those cultivated in arid and semi-arid regions. This could enable comparative analyses across species and provide a reference for extending multimodal sensing approaches to underrepresented plants of ecological and economic importance.