The Interactive Catchment Explorer (ICE) is a dynamic visualization interface for exploring catchment characteristics and environmental model predictions.
ICE was created for resource managers and researchers to explore complex, multivariate environmental datasets and model results, to identify spatial patterns related to ecological conditions, and to prioritize locations for restoration or further study.
ICE is part of the Spatial Hydro-Ecological Decision System (SHEDS).
Alternative versions of ICE have been created and customized to datasets and model results for the following projects.
Management and restoration of hydro-ecological systems often relies on the use of large environmental datasets and model simulations. However, many of these datasets can be extremely large and difficult to process and understand using traditional methods. To increase the utility of these large environmental datasets and simulations, we need new ways of exploring the data so that we may ask questions and identify the trends and patterns needed to effectively manage and restore these valuable natural resources.
The Interactive Catchment Explorer is a data visualization tool that was developed to facilitate exploration of large environmental datasets representing various geospatial characteristics and model simulations. The purpose of ICE is to allow users with a range of scientific and management backgrounds to better understand the complex, multi-variate relationships of environmental systems through an intuitive map-based interface at multiple spatial scales.
ICE was originally developed to present the results from stream temperature and brook trout occupancy models of the northeastern U.S. that were developed by our research group at the USGS Conte Anadromous Fish Branch. From the beginning, however, ICE was intended to become a general platform that could be easily adapted for other environmental datasets and models. It has since been used to help other research groups across the country focus on their own datasets and model results in order to better understand the patterns and trends of their target systems.
ICE provides a highly interactive and map-based user interface for exploring large environmental datasets representing geospatial characteristics and model results. These datasets are all based on a delineation of catchment areas across a larger region. A catchment represents the drainage area of a single stretch of river or stream. For the primary SHEDS version of ICE, there are nearly 400,000 catchments across the northeastern U.S. that have an average area of 2 square kilometers. Due to the sheer number and small size of these catchments, traditional GIS-based methods for viewing the data were slow and difficult to interpret, especially at the regional scale.
Therefore, instead of displaying all catchments simultaneously across a large region, ICE performs spatial aggregation by grouping these catchments into larger watershed areas. Each watershed contains many individual catchments, and thus represents a larger drainage area comprised of multiple rivers and streams. The watersheds are based on the standard Hydrologic Unit Code (HUC) delineations for the U.S. There are multiple "levels" of HUC delineations such that each level represents a different resolution and average watershed size. Higher levels (e.g. HUC10) contain smaller watershed areas, and smaller levels (e.g. HUC6) represent larger areas. The spatial aggregation is performed by computing the area-weighted mean value of each watershed based on the values and areas of the catchments within it. The aggregated value associated with each watershed thus accurately reflects the average across its catchments but accounting for differences in catchment areas.
In addition to spatial aggregation, ICE also allows users to filter the catchments through the use of one or more criteria in order to focus only on areas that have certain characteristics. For example, the user could specify a filter for percent forest cover between 50 and 100% to focus only on highly forested areas. As a result, ICE will remove all catchments that do not meet that criteria, and then re-compute the spatial aggregation of each watershed. The map will then show the area-weighted mean value for each watershed based only on the catchments with forest cover of 50% or more. The user can specify more than one filter criteria, and interactively change each filter and see the map respond in near-real time. This interaction provides powerful insight into the spatial patterns and relationships between variables in a way that is not possible using tradition GIS interfaces.
In order to achieve the spatial aggregation and catchment filtering in a highly responsive way, ICE was developed as a client-side web application, which means all computations are performed within the user's web browser (as opposed to remotely on the web server). The application is comprised of two primary components:
The design of ICE was inspired by the Visual Information-Seeking Mantra described by Schneiderman (1996):
"Overview first, zoom and filter, then details on demand"
Specifically, ICE provides the following core features and functionalities:
ICE was developed using the following software and libraries:
ICE is currently supported on the latest versions of all major web browsers. Google Chrome is highly recommended for the best user experience. ICE is not intended for use on mobile devices.
Note that ICE is a memory-intensive application. Older computers may have difficulty rendering the interface resulting in sluggish performance. If you run into issues, we recommend closing all other programs and browser tabs to increase available memory.
Development of ICE is currently ongoing and future updates will include performance improvements, additional features, and greater generalization so that it may be more easily adapted to new datasets for other projects and in other regions. If you have any questions, discover any bugs, or are interested in applying ICE to your dataset, please contact us at firstname.lastname@example.org.
Modeling and database support was provided by Dan Hocking, PhD and Kyle O'Neil. Additional web development support was provided by Chris Jennison.