ALM Software Engineering

Abstract

From a software engineering perspective, a lot of effort has been placed on software structure analysis to improve software quality and computing performance. We think it is the time to develop new computational frameworks to analyze complicated environmental modeling system with emphasis on hypotheses, scientific workflow, and numerical methods inherited from existing model development. Herein, we present our new effort on scientific function test framework for modular environmental model development.  Evolved from the traditional concepts on software unit testing, our framework is designed to provide innovative and convenient (piece-by-piece) ways for process-based multiscale model verification and validation, covering both model structure and scientific workflow. It expedites model modification and enhancement; it also enables environmental model reconfiguration, reuse and reassembly. We have applied these methods to the ACME Land Model (ALM) with three typical scenarios: 1) benchmark case function validation, 2) observation-constraint function validation, and 3) a virtual root module generation for root function investigation and evaluation. In this poster, we also present that our experience/strategies/plans on ALM software refactoring, which can be beneficial to many other research programs that adapt integrated environmental modeling methodology.