ELOHA extends the use of limited ecological data by assuming that ecosystems with similar streamflow attributes and geomorphic characteristics respond similarly to flow alteration.  Freshwater ecosystems are classified for many different purposes, which means there are many different classification systems.  For ELOHA, classification is based initially on ecologically-relevant streamflow characteristics.  For example, Kennard et al (2010, p. 179-184), describe the flow regimes of Australia's 12 river types.  Subclassification is based on other factors that influence how biota respond to hydrologic alteration.  Non-hydrologic factors could include water quality (e.g., water temperature), geomorphology (e.g., channel form and materials), or biologic conditions (e.g., communtity composition or endemic species).  

Classification of river types for ELOHA is distinguished from regional classification systems such ecoregions (Omernick, 1987) and hydrologic landscape regions (Wolock, 2003) in that ELOHA river types are not geographically contiguous, as illustrated below for river types in lower Michigan, USA (source: Paul Seelbach, Michigan Dept of Natural Resources).  For example, headwater tributaries and a mainstem river in the same basin would likely be classified as different river types.

 RESOURCES  

ELOHA Flow Chart in
   Spanish
   English

SIFN Stream Network Classification Project

Journal Articles and Reports

HYDROLOGIC CLASSIFICATION 

Apse et al (2008)  Pennsylvania, USA

Henriksen et al (2006) USGS HIP user's manual

Kennard et al (2010)

Kennen et al (2007) New Jersey, USA

Olden and Poff (2003) USA

Sanborn and Bledsoe (2005) Australia Northwestern USA

Thompson et al (2001) New South Wales, Australia

Elliot and Jacobson (2006) Missouri River, USA

OTHER CLASSIFICATIONS

Michingan: Michigan Groundwater Conservation Advisory Council (2007); Brendan et al (2008)

New Zealand:
Snelder et al (2005); NZ River Classification User Guide and Data

More Case Studies 

Australia

United Kingdom

Michigan, USA

Missouri, USA

New Jersey, USA

Texas, USA

Washington, USA

As indicated on the ELOHA flow chart, the basis for hydrologic classification in ELOHA is unaltered hydrology of rivers and streams.  Historical hydrologic information on unaltered streamflow may be available at some sites or may be inferred easily for some streamflow characteristics.  In other cases, little may be known about either past (unaltered) or current (altered) hydrology at a site.  Hydrologic-based classification can still be used if streamflow characteristics can be synthesized (e.g., by building a hydrologic foundation).  Otherwise, classification may have to be based on basin attributes that are related to streamflow characteristics and can be estimated at any site using a geographic information system (GIS).  For example, Snelder et al (2005) classified New Zealand's rivers according to topography and climate.  Data and instructions can be found here.  A primary components analysis indicates that the resulting classes have distinct flow regimes.

Examples

The Pennsylvania (USA) Instream Flow Advisory Committee recently completed a hydrologic classification of all rivers within the 117,348-km² State of Pennsylvania according to the U.S. Geological Survey's Hydroecological Integrity Assessment Process (HIP).  To define river types, first, 205 hydrologic metrics were derived from mean daily discharge data for 136 relatively unimpaired stream gauges.  Next, a principal components analysis (PCA) eliminated redundancy, thereby reducing the number of metrics from 205 to 151.  Three different clustering procedures were then tested for their ability to group Pennsylvania's streams according to flow regime and to identify the ecologically relevant flow metrics that best characterize them.  Ultimately, five river types were defined based on a subset of 11 metrics that describe streamflow magnitude, variability of high flows, and flood frequency (Apse et al. 2008).  This approach was similarly used to define river types in New Jersey, Missouri, Texas, and Massachusetts (Kennen et al. 2007, 2009; Cade 2008).

In Washington, USA, researchers have completed a statewide hydrologic classification based on 99 metrics describing ecologically relevant characteristics of the natural flow regime (J.D. Olden unpublished data).  Metrics were calculated from continuous time series (>15 years of record) of mean daily discharge data for 52 stream gauges, and classification was undertaken using a fuzzy partitional method - Bayesian mixture modeling.  This analysis has identified distinctive flow regime types that differ in their seasonal patterns of discharge, variation in low flow and flood magnitude and frequency, and other aspects of flow predictability and variability.  Factors related to catchment (watershed) topology, surficial geology, and climate were found to be strong discriminators of flow regime, and this information is being used in statistical models to predict flow regime type and flow metrics for streams and rivers across the state.  The spatial context provided by the hydrologic classification improves understanding of the interaction between hydrology and ecology in rivers of the Pacific Northwest United States, and provides a benchmark against which flow-ecology relationships can be assessed.

In Australia, two different steps were used to classify river types. The first step involved empirically derived streamflow discharge data from 830 gauges located throughout Australia. A total of 120 hydrologic metrics were calculated for each gauge and a Bayesian clustering technique was used to group gauges according to similarity in flow regime. The second step used data derived from Geographic Information Systems (GIS) to model streamflow for 1.2 million stream segments to derive a classification flow.  A specialized River Analysis Package was used to prepare data and metrics used in the classification. The Australian hydrologic classification contains 12 distinct classes of river types.(Pusey et al, 2009; Kennard et al, 2010).

Water temperature is also a key component of environmental flows and is strongly influenced by ground-and surface-water hydrology.  In Michigan, USA, 11 river types have been delineated, based on hydrology, temperature, and catchment size (Brenden et al 2008; Michigan Groundwater Conservation Advisory Council 2007).  

Geomorphology is an important mediator of biological response to flow alteration.  For example, in a homogenous stream reach, extensive dewatering could cause a stressful habitat bottleneck that induces a threshold-type reduction in fish populations; but if the river has deep pools, then these refuges could make possible a more gradual and continuous (linear) ecological response.  Therefore, it is useful to subgroup river types according to geomorphic setting.  For example, Elliot and Jacobson (2006) used a multiscale classification to identify a hierarchy of naturally occurring clusters of reach-scale geomorphic characteristics for segments of the Missouri National Recreational River in South Dakota and Nebraska, USA.  Thompson et al (2001) developed a procedure that evaluates and links instream habitat and geomorphology at four different scales, and applied the procedure to the Manning catchment in northern New South Wales, Australia.

Omernick, J.M. 1987. Ecoregions of the conterminous United States, Annals of the Association of American Geographers 77: 334-340. 

Wolock, D. 2003, Hydrologic landscape regions of the United States, US Geological Survey Open File Report 03-145, geospatial data available at http://water.usgs.gov/lookup/getspatial?hlrus

More articles and reports on characterizing river types

Apse et al (2008) explain options for classifying rivers in Pennsylvania, USA, and apply USGS HIP to all Pennsylvania rivers.

Brenden et al (2008) describes Michigan's river classification system for environmental flow standards.

Henriksen et al (2006) is the user's manual for USGS HIP.

Higgins 2003 shows how researchers in the Snohomish River basin of Washington State used channel slope, discharge, stream power, and bankfull width and depth to delineate geomorphic stream classes.  Rivers and streams in several parts of the United States have been similarly subgrouped according to macrohabitats with unique combinations of size, elevation, gradient, geology and connectivity. 

Kennen et al (2007) applied USGS HIP in New Jersey, USA.

MichiganGroundwater Conservation Advisory Council (2007)classified rivers according to temperature, hydrology, and size.

Olden and Poff (2003) provide a framework for identifying ecologically relevant hydrologic indices that adequately characterize flow regimes in a non-redundant manner.

Sanborn and Bledsoe (2005) present a methodology for stratifying streamflow regimes of gauged locations, classifying the regimes of ungauged streams, and developing models for predicting a suite of ecologically pertinent streamflow metrics for streams in Colorado, Washington, and Oregon, USA.

Photo credits (left to right): Photo © Harold E. Malde (vernal pools at Table Mountain); Photo © Cheryl Rose (cormorant in wetlands habitat).