Authors: Noah B. Beckage, Patrick J. Clemins, Scott Turnbull, Panagiotis D. Oikonomou, Muhammad Adil, Mirce Morales, and Asim Zia – University of Vermont
Title: Data Acquisition Framework Design for Model Input Data
Abstract: Thoughtful construction of data acquisition and preparation workflows for model input data allows for greater flexibility in future research directions, quicker and more efficient changes between input datasets, and enables code reuse. These properties are achieved by virtue of modular design principles such as code interchangeability and independence. Although developing software with these attributes require a greater investment of labour upfront, they ultimately yield a reduction in coding time and labour in the long term. This work implements these design principles through a modular data acquisition framework, capable of acquiring hydrological and meteorological data from several observation and forecast products for use in water resources forecasting applications. The framework consists of data acquisition modules developed in Python and based on two primary principles: that (1) each module should acquire data and return them in a readable, standardized data structure, and (2) each module should have the same generalized parameters. With these specific design principles, the framework organizes acquired data in a standardized data structure which can then be written out to model-specific input formats, making the framework generalizable to future modelling efforts while simultaneously improving workflow efficiency.