Software Carpentry and the Hydrological Sciences

Scientists are spending an increasing amount of time building and<br>using hydrology software. However, most scientists are never taught<br>how to do this efficiently. As a result, many are unaware of tools and<br>practices that would allow them to write more reliable and<br>maintainable code with less effort. As hydrology models increase in<br>capability and enter use by a growing number of scientists and their<br>communities, it is important that the scientific software development<br>practices scale up to meet the challenges posed by increasing software<br>complexity, lengthening software lifecycles, a growing number of<br>stakeholders and contributers, and a broadened developer base that<br>extends from application domains to high performance computing<br>centers. Many of these challenges in complexity, lifecycles, and<br>developer base have been successfully met by the open source<br>community, and there are many lessons to be learned from their<br>experiences and practices. Additionally, there is much wisdom to be<br>found in the results of research studies conducted on software<br>engineering itself. Software Carpentry aims to bridge the gap between<br>the current state of software development and these known best<br>practices for scientific software development, with a focus on<br>hands-on exercises and practical advice. In 2014, Software Carpentry workshops targeting earth/environmental<br>sciences and hydrological modeling have been organized and run at the<br>Massachusetts Institute of Technology, the US Army Corps of Engineers,<br>the Community Surface Dynamics Modeling System Annual Meeting, and the<br>Earth Science Information Partners Summer Meeting. In this<br>presentation, we will share some of the successes in teaching this<br>material, as well as discuss and present instructional material<br>specific to hydrological modeling.