You are here:
Novel GIS approaches to watershed science and management: Description, prediction, and integration
Citation:
Flotemersch, J., M. McManus, AND M. Scown. Novel GIS approaches to watershed science and management: Description, prediction, and integration. Geography, Environment and GIS, Targoviste, ROMANIA, May 19 - 21, 2016.
Impact/Purpose:
The purpose of this presentation is to first, share recent EPA developments in the area of spatial analyses of watersheds. Secondly, to gain feedback and ideas from conference attendees on how to improve these approaches. Thirdly, to discuss potential collaboration and expansion of these tools globally.
Description:
Spatial data and geographic information systems (GIS) are playing an increasingly important role in watershed science and management, particularly in the face of increasing climate uncertainty and demand for water resources. Concomitantly, scientists and managers are presented with ever-growing databases of spatial information, some of which are ‘big data’, and require novel approaches to render them usable and relevant for watershed applications. We present an overview of three case studies applying descriptive, predictive, and integrative analyses to watershed research, and discuss the challenges faced in undertaking such tasks. The first case study provides an example of hydrogeomorphic characterisation of a river network using GIS. Thirteen hydrogeomorphic variables were extracted from a range of spatial datasets, including streamlines, a digital elevation model (DEM), rainfall grid, and lithology polygons, at regular intervals throughout the Little Miami River network, Ohio, USA. Multivariate clustering was then used to group the sample points into discrete ‘Functional Process Zones’ (FPZs) based on their hydrogeomorphic character. The emergent FPZs were then mapped throughout the river network, and various diversity metrics commonly applied in community ecology were used to determine the structural diversity of the river network at the watershed scale. The results revealed that potential threats at the watershed scale, such as loss or fragmentation of physical habitats, may differ among FPZs because of their spatial organisation. In addition, important structural features, such as floodplains, were found to be distributed throughout the entire watershed, rather than concentrated in the lower reaches. This hydrogeomorphic characterisation approach has management applications for identifying rare physical habitats, mapping ecosystem services, and environmental asset trading throughout watersheds. An example of ‘spatial stream network’ (SSN) modelling of in-stream nutrients is provided in the second case study. SSN modelling has recently gained support as an improvement on traditional regression models in river networks. SSN modelling enables both covariates and spatial autocorrelation in the response variable to be incorporated into the final model. Modelling of total phosphorus (TP) was conducted throughout the East Fork Little Miami River watershed, Ohio, USA, using traditional regression and SSN techniques. In addition, both models were initially fitted using nationally-available spatial covariate coverages, then fitted again with additional covariate coverages only available in the study watershed. The results revealed that SSN modelling improved upon traditional regression for predicting TP throughout the watershed, and that additional covariate coverages, namely waste water treatment plant release loads and the location of septic systems, should be included in national spatial databases in order to improve their utility for modelling in-stream nutrients. The third and final case study is an example of integrating social and environmental spatial data for interdisciplinary watershed applications. Relationships between an Index of Watershed Integrity (IWI) and components of a Human Well-Being Index (HWBI) were investigated in counties of the Ohio River Basin, USA. Challenges faced with integrating the two types of data included differences in the delineation of spatial units and different data resolutions. These challenges were overcome by adopting an image processing approach and multivariate analyses. The three case studies presented here are examples of novel approaches to using large spatial datasets for watershed applications and may prove useful to scientists and managers in other geographic settings. The case study findings may also be useful for guiding spatial database development both within and across political boundaries.