ASABE Technical Library - Abstract

Click on “Download PDF” for the PDF version or on the title for the HTML version.
If you are not an ASABE member or if your employer has not arranged for access to the full-text, view access and subscription options

Abstracting External Models for Decision-Ready Agricultural IoT: H

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.org

Citation:  2026 ASABE Annual International Meeting  2600959.(doi:10.13031/aim.202600959)
Authors:   Md. Samiul    Basir ,  , Fabio A. Castiblanco, Yaguang     Zhang ,  , Dennis   R.  Buckmaster ,  , James  V.  Krogmeier ,  
Keywords:   Abstraction, data integration, data plenitude, data completeness, forecast, IoT

Abstract. Agricultural IoT systems automate data acquisition to support decision-making, often integrating heterogeneous sources such as sensors, remote sensing platforms, vehicles, expert systems, machine learning models, and public or private repositories. These environments may also depend on external model-derived signals produced outside the IoT system. Data abstraction is a proven approach for achieving interoperability by exposing relevant information while hiding source-specific formats, representations, and processing details, thereby providing a stable, usage-independent interface to downstream systems. This work presents a framework for abstracting external model outputs into decision-ready semantics for automated agricultural IoT systems and demonstrates its applicability through a weather-driven hay-making case study. External weather forecast models provide location- and time-specific data that drive automated insight logic. The abstraction layer decouples decision systems from provider-specific data formats, spatial resolutions, and model representations. As an example, we use the NOAA/NWS National Digital Forecast Database, which provides 7-day forecasts of variables such as temperature, precipitation probability, and wind speed derived from the National Blend of Models. The abstraction service ingests forecasts in their native gridded binary formats and translates them into location- or area-specific weather time series using an application-defined JSON schema. To demonstrate utility, we compute sequential probabilities supporting hay-making decisions for mowing, raking, and baling, and present the results through an informative visualization. The decision logic consumes abstracted forecast signals and applies domain-specific probabilistic rules without direct dependence on raw forecast products. In this case, the primary decision variable is the cumulative probability of no precipitation over consecutive days, supplemented by anticipated mean and maximum temperatures. The abstraction framework can also maintain a persistent archive of daily forecasts, enabling forecast-informed decision support that leverages both current and historical data while supporting auditability and future decision analysis.

(Download PDF)    (Export to EndNotes)