Rural aqueducts play a vital role in sustaining the communities they serve by managing the essential water resource. A common challenge for rural aqueducts is related to water loss and rudimentary monitoring methods. Internet of Things (IoT) enabled systems offer a promising solution to improve water management, yet their design complexity and costs often exceed technical and financial capacities in rural aqueducts. This study proposes a Model-Based Systems Engineering (MBSE) approach to develop a modular IoT architecture tailored to rural aqueducts. Using the ARCADIA methodology and the Capella modeling tool, the process translates informal user needs into structured requirements and logical system designs. The resulting architectures are modular, adaptable, and aligned with enduser conditions. The methodology is accessible enough to be implemented by undergraduate students, positioning academic institutions as effective intermediaries for technology transfer. By formalizing needs and systematically guiding architecture development, MBSE offers a structured pathway to create IoT solutions for water monitoring and loss prevention. This approach provides a scalable, explainable framework to support the modernization of rural aqueduct systems.
Vision-language models (VLMs) offer a promising approach to estimate floodwater depth from images, a critical metric for emergency response. While their accuracy has been benchmarked against human estimators, their reasoning has mostly been assessed qualitatively and post hoc. In this work, we hypothesize that bounding boxes provide a systematic framework to evaluate VLM reasoning and localization. To test this, we inferenced the GLM-4.5V VLM on a dataset of flooding images and compared its depth estimates and reference-object bounding boxes against the consensus of four human annotators under four perturbation scenarios. We employ the distance intersection over union (DIoU) metric to quantify the agreement between bounding boxes, and the mean absolute error (MAE) to characterize the accuracy of floodwater depth estimates. Results show that VLMs use bounding boxes with accuracy comparable to textual descriptions, though below human performance, and that numerical estimates depend on the visual content of the chosen boxes. This demonstrates that bounding boxes can capture VLM reasoning traces and systematically reveal failure cases previously identified only through qualitative analysis.