|'A Software Stack for Water Resources GIS Web Apps'
The purpose of this presentation is to review existing technologies that can be used in developing water resources web modeling applications. The mobile web revolution is making it possible to bring high performance computing to palm-sized mobile devices. This has exciting implications for water resources and engineering—especially in the field of hydrologic modeling. Moving common modeling scenario exploration analyses from a desktop environment to a web-based platform has several advantages. Some of these advantages include:
• a user interface built with familiar web form elements
• access to distributed cloud and super computing resources
• data sharing
• preventing loss of organizational knowledge
There are a number of hurdles hydrologic web modeling developers will face. One of these is how to work with the geospatial data inherent with this class of models. Supporting geospatial data in a website is beyond the capabilities of typical web frameworks because it requires the use of additional software. The functionality needed for a web modeling applications can be achieved with three elements:
1. a geospatially enabled database
2. a map server
3. a geoprocessing toolbox.
ESRI has developed the most popular set of tools that can be used to this end. The Python scripting module, ArcPy, allows developers to script to the extensive ESRI geoprocessing library. The data can be served using ArcServer and users can even build custom webpages using ArcGIS.com As a proof-of-concept, a well-permitting web application was developed for the Utah Division of Water Rights by researchers at Brigham Young University using some elements of the ESRI framework. The model uses an existing geodatabase MODFLOW model of aquifers in the state of Utah to predict the drawdown in the aquifer due to a new well. This system is now being implemented for all of the aquifers in the state.
While ArcGIS may provide a robust and rich set of tools they are proprietary and for many organizations the cost barrier is high. In response to this concern a number of open source alternatives have emerged and the most mature will be presented. We have conducted an extensive review of these alternatives in conjunction with the Cyber Infrastructure Water project. We recommend a software stack for geospatial web application development comprising: MapServer, PostGIS, and 52 North with Python as the scripting language.